88%(8)88% нашли этот документ полезным (8 голосов)

12K просмотров702 страницыExcellent Applied Calculus Text

Applied Calculus For Business Gordon .pdf

© Attribution Non-Commercial (BY-NC)

PDF, TXT или читайте онлайн в Scribd

Excellent Applied Calculus Text

Attribution Non-Commercial (BY-NC)

88%(8)88% нашли этот документ полезным (8 голосов)

12K просмотров702 страницыApplied Calculus For Business Gordon .pdf

Excellent Applied Calculus Text

Attribution Non-Commercial (BY-NC)

Вы находитесь на странице: 1из 702

Warren B. Gordon / Walter O. Wang / April Allen Materowski

Baruch College City University of New York

Applied Calculus for Business, Economics, and Finance, by Warren B. Gordon, Walter O. Wang, and April Allen Materowski. Published by Pearson Learning Solutions. Copyright 2007 by Pearson Education, Inc.

Cover designed by Joshua Gordon. Photograph of the TI-89 Platinum Graphing Calculator reproduced by permission of Texas Instruments.

Copyright 2007, 2006 by Pearson Custom Publishing All rights reserved. Permission in writing must be obtained from the publisher before any part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage or retrieval system. All trademarks, service marks, registered trademarks, and registered service marks are the property of their respective owners and are used herein for identi cation purposes only.

501 Boylston Street, Suite 900, Boston, MA 02116 A Pearson Education Company Applied Calculus for Business, Economics, and Finance, by Warren B. Gordon, Walter O. Wang, and April Allen Materowski. Published by Pearson Learning Solutions. Copyright 2007 by Pearson Education, Inc.

Preface

Applied Calculus for Business, Economics and Finance is a combination of the authors two texts Precalculus and Elements of Calculus and Applied Calculus. This single text may be used to cover the content of an applied calculus course for non-science majors in a variety of ways. This text continues the approach used in its precursor texts, that is, the integration of precalculus with the calculus as well as the integration of technology. Most sections of this text conclude with Calculator Tips, illustrating how the calculator may be used to enhance the understanding and appreciation of the topics considered in that section. While the text illustrates the TI 89 calculator, many of the calculator examples and exercises may be worked with most graphing calculators. As most students are now required to have a basic knowledge of spreadsheets, we utilize them several times in the text. They make the calculations involved in Newton s method painless, they allow us to easily illustrate the use of left, right and midpoint variations in computing Riemann sums, and even suggest the notion of speed of convergence. Many of the graphs and tables found in this text were created using MAPLE, the TI 89 calculator and Excel. The exponential and logarithmic functions are introduced in this text, assuming no prior knowledge. There are abundant examples illustrating the importance of these functions, perhaps more than time may allow and therefore selection is left to the instructor. In an introductory applied calculus text we believe it is inappropriate to give a detailed treatment of multivariable functions. We make no attempt to develop the methods used to sketch surfaces, but instead, examine functions of two variables using level curves. Our objective in this chapter is for students to understand the notion of a partial derivative, and the optimization of multivariable functions, including the method of Lagrange multipliers, an important tool in Economics and Finance. The chapter concludes with an examination of double integrals and their application to areas, volumes and probability. To learn mathematics, students must work out exercises and we have included many in this text. The text includes answers to both the odd and even exercises. For many students, applied calculus is a terminal course. Most of the exercises were written with the goal of reinforcing the concepts and skills introduced in the section. We also include, in most exercise sets, problems which will allow the stronger students to stretch their understanding of the calculus. These exercises can be assigned by the instructor as extra credit or honors problems. As there are a significant number of students majoring in quantitative disciplines who will continue with their study of mathematics, they may find some of these exercises useful in easing their transition to the next course. On the publisher s web page http://www.pearsoncustom.com/gordon_appliedcalcbus/ there is located a link to the text videos. These videos, presented by two of the authors, correspond to the text material, section by section. We have found that the videos are very useful to students who miss a class, or just need another detailed look at the material. The videos allow students to go over the section material at their own speed. There is more material in this text than can be included in a one semester course, and as a result, there is some flexibility in the inclusion of topics. Chapter 0 is provided for students who need an algebra refresher. It has been our experience that students should be referred to this material as needed (or to the corresponding videos). The content of this

iii

Applied Calculus for Business, Economics, and Finance, by Warren B. Gordon, Walter O. Wang, and April Allen Materowski. Published by Pearson Learning Solutions. Copyright 2007 by Pearson Education, Inc.

iv

* **

Preface

chapter is aimed directly at the algebraic skills needed for calculus. Appendix A, Matrices and Linear Systems, is generally not found in most applied calculus texts. We include it as many business schools require their students to be familiar with matrices and their applications. No other chapter in the text uses this material, and it may be included or omitted as required by the needs of the course. Trigonometric functions do not appear in this text. While we have been tempted to include them, we realize that periodic functions arise mostly in the study of physics, infrequently, if ever, in business applications. A supplement can be made available to those who wish to briefly examine the calculus of these functions. In the inside cover of this text are images of the TI 89 and TI 89 Titanium calculators. Note that in addition to the symbol on each key there are two additional symbols associated with each key; on the TI 89, either in orange, accessed by first pressing the orange 2nd key and then the key with the associated symbol, or in green, accessed by first pressing the green diamond key and then the key with the associated symbol. We shall indicate the green diamond by the symbol * in this text. The TI 89 Titanium model has the * key in light green and the 2nd key in blue. Both of these keys are found in the upper left portion of the keypad on each of the calculators. Also note the purple alpha key which is needed for the space and and insertions. We gratefully acknowledge the help of many of our colleagues at Baruch College who have provided suggestions and offered constructive criticism of this text, in particular, we thank Sherman Wong who is our local Maple guru. We thank Delia Uherec, Acquisitions Editor at Pearson Publication for her assistance with publication of this text, and Curt Hinrichs and Ann Day, Editors at Thompson Brooks/Cole Publishing, for allowing us to use material written by the first author from the text Succeeding in Applied Calculus: Algebra Essentials, and its website. Warren B. Gordon Walter O. Wang April Allen Materowski April, 2006

Applied Calculus for Business, Economics, and Finance, by Warren B. Gordon, Walter O. Wang, and April Allen Materowski. Published by Pearson Learning Solutions. Copyright 2007 by Pearson Education, Inc.

Contents

Preface iii

Chapter 0

2 3 5 5 6 9 10 12 14 15 17 18 21 22 24 25 26 27 28 29 32 32 38 39

Addition and Multiplication Properties Linear Equations with Fractions Linear Equations with Decimals Solving for a Particular Variable Applications of Linear Equations Calculator Tips Exercises Isolation of Squared Term Isolation of Squared Binomial Term Calculator Tips Exercises Completion of the Square Calculator Tips Exercises

Quadratic Formula Clearing Fractions Applications Equations Reducible to Quadratics Calculator Tips Exercises Sign Analysis Interval Notation Calculator Tips Exercises

Chapter 1

1.1 The Line

42 43 45 46 47 49 v

Two Dimensional Coordinate System Horizontal and Vertical Lines The Slope Intercept Form Graphing The Point-Slope Equation The Slope Formula

vi

* **

Contents

51 53 56 58 60 61 62 64 66 67 68 69 70 70 71 75 79 80 82 83 84 88 90 91 91 94 94 97 99 102 103 104 105 109 111 113 114 115 116 117 119 124 124 126

Definition of a Function Functional Notation Difference Quotient Domain and Range Independent and Dependent Variables Vertical Line Test Combining Functions Composition Decomposition Functions of Several Variables Calculator Tips Exercises Break-Even Analysis Depreciation Piecewise Linear Graphs Calculator Tips Exercise

Scaling Vertical Translation Axis of a Parabola Horizontal Translation Locating the Vertex Graphing a Parabola in the form y = ax2 + bx + c Applications to Optimization Calculator Tips Exercise Definition of a Circle Equation of a Circle Graphing a Circle Tangent Line The Ellipse Calculator Tips Exercises Supply Function Demand Function Market Equilibrium Revenue, Cost and Profit Functions Marginal Functions Calculator Tips Exercises

Contents

* **

vii

Using the Zeros Even Functions Symmetry About the y-axis Odd Functions Symmetry About the Origin Rational Functions Vertical Asymptotes Horizontal Asymptotes Translations Calculator Tips Exercises

127 130 131 132 135 136 144 147 149 151 153 155 156 156 161 162

1.8 Regression

Scatter Plot Line of Best Fit Linear Regression Correlation Coefficient Non-Linear Regression Calculator Tips Exercises Chapter Summary

Chapter 2

An Introduction to Calculus

169 170 171 171 174 175 176 177 179 180 181 182 184 185 186 188 190 191 193 196 196 198 199

Slope of a Tangent Line The Slope as a Limit Slope of a Curve Equation of a Tangent Line A Place Where No Tangent Exists The Derivative Calculator Tips Exercises Derivative of a Linear Function The Simple Power Rule The Constant Multiplier Rule The Sum Rule Calculator Tips Exercises The Limit Limits by Substitution One Sided Limits Jumps and Holes Continuity Removable Discontinuities Differentiability and Continuity Calculator Tips Exercises

viii

* **

Contents

Limits at Infinity Dominant Terms Horizontal asymptotes Infinite limits Vertical asymptotes Calculator Tips Exercises The Product Rule The Quotient Rule Calculator Tips Exercises

202 203 205 207 207 208 209 210 211 214 214 217 218 221 221 222 224 225 227 229 229 232 232 234 235 237 237 239 242 242 243 245 246 246 247 248 249 252 252 253

The Chain Rule The General Power Rule Calculator Tips Exercises Marginal Functions Average Cost Velocity Average and Instantaneous Rates of Change Calculator Tips Exercises Finding a Tangent Line Finding the Derivative Calculator Tips Exercises Vertical Angles Parallel Lines Similarity Congruence Midpoint Formula Exercises

A Geometric Example An Ecological Example An Economic Example Using Similarity Exercises Newton s Method Calculator Tips Exercises Chapter Review

Contents

* **

ix

Chapter 3

256 257 258 259 261 264 265 266 267 269 274 274 276 277 278 279 281 281 288 289 290 294 295 295 298 299 300 302 303 306 309 310 312 314 315 315 317 319

Continuity Maximum and Minimum Values Extreme Value Theorem Relative Maxima and Minima Critical Numbers and Critical Points Calculator Tips Exercises Increasing and Decreasing Functions The First Derivative Test Sign Diagrams Calculator Tips Exercises The Second Derivative Higher Order Derivatives Velocity and Acceleration Concavity The Second Derivative Test for Concavity The Second Derivative Test for Relative Extrema Implicit Differentiation and Curve Sketching Calculator Tips Exercises Area and Perimeter Optimization Procedure Volume Distance and Velocity Calculator Tips Exercises

Price, Demand and Revenue Cost and Average Cost Elasticity of Demand Exercises

Linearization Differentials The Differential Approximation Differentiable Functions Differential Formulas Exercises Chapter Review

* **

Contents

Chapter 4

322 322 323 325 329 331 332 332 335 336 339 340 342 343 343 346 349 350 353 353 354 355 357 359 359 362 363 364 364 367 368 370 371 373 375 376 377 378 380 382 383 384 384

One-to-One Function Horizontal Line Test Increasing and Decreasing Functions Inverse Function Composition Property Derivative of the Inverse Calculator Tips Exercises Exponential Expressions The Graph of y = f1x2 = bx Solving Special Exponential Equations Finding the Exponential Function Growth and Decay Rates Power Function Calculator Tips Exercises Continuous Compounding of Interest The Constant e Calculator Tips Exercises The Simple Exponential Rule The Generalized Exponential Rule Exponential Domination Calculator Tips Exercises

Definition of a Logarithm Base 10 and e pH of a Solution Graphing Logarithmic Functions The Simple Logarithmic Rule The Generalized Logarithmic Rule Calculator Tips Exercises Multiplicative and Division Properties Exponential Property Derivatives Using the Properties Logarithmic Equations Exponential Equations Change of Base Derivatives in Different Bases Logarithmic Differentiation Calculator Tips Exercises

Contents

* **

xi

Exponential Growth Population Growth Continuous Compounding Radioactive Decay Carbon Dating Logistic Growth Richter Scale Calculator Tips Exercises Chapter Review

387 387 388 388 389 390 392 393 394 395

Chapter 5

398 399 399 401 402 403 404 405 407 410 411 412 413 414 415 418 419 421 421 423 424 425 426 433 433 435 436 436 439 440 444 445

Antiderivative Integration Theorems Simple Power Rule Simple Logarithmic Rule Simple Exponential Rule Calculator Tips Exercises

Particular Solutions Equations of Motion Marginal Functions Separable Differential Equations Calculator Tips Exercises Reversing the Chain Rule Generalized Power Rule Generalized Logarithmic Rule Generalized Exponential Rule Calculator Tips Exercises Areas by Rectangles Left Endpoints Right Endpoints Midpoints Calculator Tips Exercises

Sigma Notation Linearity Property Summation Formulas Riemann Sums Areas by Riemann Sums Calculator Tips Exercises

xii

* **

Contents

Definite Integral Fundamental Theorem of Integral Calculus Basic Properties Calculator Tips Exercises Substitution Odd and Even Functions Average Value Derivative of an Integral Calculator Tips Exercises

447 448 450 456 459 462 464 466 468 470 471 472 479 481 483 484 485 486 489 492 494 494

Area Between Curves Consumer and Producer Surplus Continuous Income Flow Probability Calculator Tips Exercises Substitution Integration by Parts Tabular Integration Exercises Chapter Review

6.1 Functions of Several Variables

Functions of Several Variables Difference Quotients Three-Dimensional Coordinate System Surfaces Calculator Tips Exercises 498 499 499 500 501 503 504 507 509 511 513 516 518 519

Partial Derivative Visualization of the Partial Derivative Level Curves Contours Cobb-Douglas Production Function Utility Functions and Indifference Curves Higher Order Partial Derivatives Calculator Tips Exercises

T

his chapter reviews some of the basic concepts from elementary and intermediate algebra. Each section begins with a pretest which you may take to see if you remember the material. If you do, then the section may be skipped, if you cannot solve most of the problems in the pretest, then you should review the entire section.

* **

Section 0.1

0.1

Addition and Multiplication Properties Linear Equations with Fractions Linear Equations with Decimals Solving for a Particular Variable Applications of Linear Equations Calculator Tips

Each question is worth one point. Solve for the unknown: 1. x + 2 = 3 3. 3y - 2 = 7 5. 2z - 5 = 5z - 3 7. 9.

3 5x

3 4w

= 12 =

4 3

1 4

= 2 3w +

7 2

2x 3x - 2

In this section we review the method of solving a linear equation. The objective of this section is to remind you how to solve an equation of the form ax + b = c for x (we assume, of course, that a Z 0). The basic idea of solving an equation is to isolate the unknown on one side of the equation. The solution is then the number on the other side. Usually, two steps are required. First isolate the term containing the unknown, then isolate the unknown itself. This often takes the application of two properties: the first is the addition property which states that the same term may be added (or subtracted) to each side of an equation. For example if A + D = E then A + D - D = E - D or A = E - D. This property may be rephrased as follows: an expression may moved from one side of an equation to the other provided its sign is reversed. Notice we move D from the left side of the equation A + D = E to the right side by changing its sign and writing A = E - D. Sometimes the procedure is called transposition. Once the unknown term is isolated, we use the multiplication property to solve for it; this property states that two sides of an equation may be multiplied by the same (non-zero) expression. For example, if A = B then AC = BC or if AB = D, then if we multiply this equation by 1/B (or equivalently divide by B) we have that A = D/B. We illustrate how these properties are used in the following examples. Example 1 Solve the following equation for x: x + 2 = 5. Solution We isolate the x term by transposing the 2 to the right hand side of the equation and write x = 5 - 2, or x = 3.

Section 0.1

* **

It is important to check the result. This is performed by substituting our result for x wherever it appears in the original problem. In this example, ? 3 + 2 = 5 5 = 5 Thus, our solution checks the original equation

Example 2 Solve the following equation for x: 3x - 2 = 7 Solution We isolate the x term by transposing the - 2 to the right hand side of the equation and write 3x = 7 + 2 or 3x = 9 we next solve for x by multiplying both sides of the equation by the reciprocal of 3, 1/3 yielding, 11/323x = 11/329 or x = 3 We check our result. ? 3132 - 2 = 7 ? 9 - 2 = 7 7 = 7 which checks our solution. Note: in the last example, instead of multiplying each term on both sides of the equation by 1/3, we could have equivalently divided each term by 3. Example 3 2 Solve the following equation for x: x = 5 3 Solution Method I Since the x term is already isolated, we need only multiply both sides of the equation by the reciprocal of 2/3, which is 3/2. Thus, 2 x = 5 3 3 2 3 a b x = a b5 2 3 2 15 x = 2 Method II Recall that whenever an equation contains fraction, we may multiply every term on each side of the equation by the LCD (lowest common denominator) to clear the fractions. In

* **

Section 0.1

this example, we need only multiply each term on either side of the equation by the LCD which is 3 and then complete the solution as follows. 2 x = 5 3 2 3# x = 3#5 3 2x = 15 1 1 a b 2x = a b 15 2 2 15 x = 2 We leave the checking of the solution in this and the rest of the examples as exercises for you. As we see from the last example, there is often more than one way to solve an equation. Sometimes one method may be preferred to another as it results in less work in obtaining the solution. Example 4 Solve the equation 5y - 7 = 2y - 11 for y. Solution We need to isolate the y term. First we transpose the - 7 from the left to the right side of the equation by changing its sign, so we have 5y = 2y - 11 + 7 5y = 2y - 4 We next transpose the 2y term to the left hand side by changing its sign to obtain 5y - 2y = 4 or 3y = 4 and finally, we multiply both sides by 1/3 (or equivalently divide both sides by 3) to obtain y = 4/3

Example 5 2 3 1 5 Solve the equation w + = - w + for w. 4 6 2 3 Solution We multiply every term on both sides of the equation by the LCD which is 12 to obtain 2 3 5 1 12 # w + 12 # = 12 # a - w b + 12 # 3 4 6 2 8w + 9 = - 2w + 30 8w = - 2w + 30 - 9 8w + 2w = 21 10w = 21 w = 21/10

Section 0.1

* **

Example 6 Solve the equation 312m - 42 + 315 - 6m2 = 9 - 417 - 4m2 for m. Solution We first distribute, combine like terms, and transpose to isolate m and then solve. 6m - 12 + 15 - 18m = 9 - 28 + 16m - 12m + 3 = - 19 + 16m - 12m = - 19 + 16m - 3 - 12m = - 22 + 16m - 12m - 16m = - 22 - 28m = - 22 m = 22/28 = 11/14

Example 7 Solve the equation 0.5x + 0.213 - 2x2 = 0.04 Solution Just as with fractions, where they are cleared by multiplying by the LCD, we may do the same thing with decimals. Observe that the smallest decimal involves hundredths (0.04), so we multiply each term on each side of the equation by 100 to obtain 100 # 0.5x + 100 # 0.213 - 2x2 = 100 # 0.04 50x + 2013 - 2x2 = 4 50x + 60 - 40x = 4 10x = - 56 x = - 56 10 = - 5.6 (Note that we could have written the answer as - 28/5, but since the original problem involved decimals, we left the answer in decimal form.) Sometimes, more than one variable may appear in an equation and we are asked to solve for one of them in terms of the others. Our tactics are still the same, isolate the term containing the desired variable. Example 8 Solve the equation 3x + 4y = 24 for (a) x (b) y. Solution (a) As before we isolate the x term by transposing the y term to obtain 3x = 24 - 4y Now we divide each side of the equation by 3, to obtain x = or x = 8 4 y. 3 24 - 4y 3 Solving for a Particular Variable Linear Equations with Decimals

(b)We next solve the original equation for y. We isolate the y term by transposing the x term to the right hand side of the equation to obtain

* **

Section 0.1

4y = 24 - 3x Dividing each side of the equation by 4 gives y = often, this last equation is written as 3 y = - x + 6 4 3 24 - 3x = 6 - x 4 4

It is important that we be able to use mathematics as means by which we may solve applied problems. The next few examples illustrate how we translate an applied problem into a mathematical equation and then solve the equation. Application of Linear Equations Example 9 A 24 foot rope is cut into two pieces. The longer piece is 3 feet longer than twice the shorter piece. Find the length of the longer piece. Solution Let x = the length of the shorter piece. Let us translate the statement the longer piece is 3 feet longer than twice the shorter piece. The longer piece = 3 + 2x (Note that is translates into = , more than into + .) The sum of the shorter piece and the longer piece is the total length which is 24 feet (see Figure 1), therefore, we have e !!!! 24 !!!! e !!!! e ! !!!!! x 3 + 2x Figure 1 x + 13 + 2x2 = 24 3x + 3 = 24 3x = 24 - 3 3x = 21 x = 7 Thus, the shorter piece is 7 feet long and the longer piece is 3 + 2172 = 17 feet long.

Example 10 The sum of Susan and Mark s present ages is 27 years. Ten years from now, twice Susan s age (then) will exceed Mark s age (then) by 28 years. What are their present ages? Solution Let Mark s present age = m, since the sum of their present ages is 27, we have Susan s present age = 27 - m. In 10 years their ages will be 1m + 102 and 127 - m + 102 respectively. At that time, twice Susan s age (two times her age) will be 28 years more than Mark, so we must have 2127 - m + 102 = 1m + 102 + 28

twice Susan s age is 28 years more that Mark s age

Note that twice Susan s age exceeds Marks age by 28, so for them to be equal, we need to add 28 to Mark. Solving, we have

Section 0.1

* **

Exercise 11 Al earns $8,000 less than twice Bob s salary, together they earn $100,000. How much does each earn? Solution Let b = Bob s salary, Al s salary is $8,000 less than twice Bob s salary = 2b - 8000 Together they earn 100,000, so we have Al s salary + Bob s Salary = 100000 or 2b - 8000 + b = 100000 or 3b = 108000 or Bob s salary is b = 108000/3 = $36,000 and Al s salary is 2b - 8000 = 21360002 - 8000 = 72000 - 8000 = $64,000. (Alternately, we could subtract Bob s salary from $100,000 to compute Al s salary, why?)

Exercise 12 The Smiths want to add a rectangular porch to their home. Due to the remaining lot size, the length of the proposed porch must be 7 feet more than twice the width. If the perimeter of the room is to be 50 feet, determine the dimensions of the room. Solution We recall that the perimeter of a rectangle of length l and width w is P = 2l + 2w. We labeled the above rectangle, representing the porch, in Figure 2 with all the pertinent information. From the perimeter relationship, we have that 50 = 2l + 2w = 212w + 72 + 2w or 50 = 4w + 14 + 2w 36 = 6w 6 = w Thus, the width of the porch is to be 6 feet, and the length 2162 + 7 = 19 feet. w

l = 2w + 7

Figure 2: Porch

* **

Section 0.1

Recall that when objects (cars, planes, bicycles, people, etc.) are moving at a constant rate (speed), the relationship among the rate (r), the distance (d) and the time (t) is given by the formula rt = d (that is, rate times time is equal to distance.) If we solve this equation for r we have that r = d/t, or if solve this equation for t, we have that t = d/r. Depending on what needs to be solved for, we use one of these equivalent forms to solve motion problems. The next example illustrates a simple application of this result. Example 13 John and Isabel leave the parking lot and travel in opposite directions. If John averages 50 mph and Isabel averages 60 mph, how long does it take them to be 385 miles apart? Solution Let t be the time they need to travel to be 385 miles apart. Since John s average rate is 50 mph, the distance he travels in this time is 50t. Similarly, the distance Isabel travels in this time is 60t. Note that the sum of John and Isabel s distances must be 385 miles (see Figure 3), so we have the equation 50t + 60t = 385 or 110t = 385 t = 385/110 = 31*2 hours Thus, in 31*2 hours they will be 385 miles apart.

e !!!!!385 miles!!!!! e !!!!! e !!!!!!!! 50t 60t John s distance Isabel s distance Figure 3

When money is invested and earns simple interest, the relationship between the Principal (P - the money invested), the annual interest rate (r), and the time in years the money is invested (t) and the interest earned (I) is I = Prt Example 14 If Carl wishes to earn $500 in interest by investing $3000 for five years, what simple interest rate will be required? Solution We are given P = 3000, I = 500, t = 5, and need to find the annual interest r. Using I = Prt, we have 500 = 3000r152 or 500 = 15000r or r = 500/15000 = 0.033333 L 3.33%

We conclude this section with one more application which further reviews the concept of solving an equation using the method of cross multiplication. Recall that if two fractions are equal, that is, if

Section 0.1

* **

A C = B D Then by multiplying each side of the equation by the common denominator BD, it follows that AD = BC Example 15 The denominator of a fraction is one more than twice the numerator. If the numerator is doubled and the denominator is increased by two the resulting fraction is 2/3. Determine the original fraction. Solution Let x be the numerator of the original fraction, then the denominator is x 1 + 2x. Therefore the original fraction is represented by . If the numerator is dou2x + 1 2x bled and the denominator is increased by two, then the new fraction is . Since this 2x + 3 new fraction is 2/3, we have 2x 2 = 2x + 3 3 cross multiplying, we have 312x2 = 212x + 32 or 6x = 4x + 6 or 2x = 6 or x = 3 Therefore, the original fraction was 3 3 = 2132 + 1 7

The calculator may be used to solve most equations for an unknown; the solve command is used. You may obtain this command by pressing the Catalog key and then either scrolling down to s or pressing the letter s (the alpha key followed by s), scrolling down to solve and then press enter. For example, suppose we wish to solve the equation 2 7 x - = 2x + 4 for x. We illustrate in Figure 4. 3 5 Note the syntax, solve(eqn, variable), where eqn stands for the equation that is to be solved and variable represents the unknown you are solving for. Pressing enter produces the solution x = - 81/20, as seen in Figure 5.

Calculator Tips

10

* **

Section 0.1

Figure 5

In Exercise 1 21, solve for the unknown, and then check your solution.

1. x - 7 = 9 2. x + 11 = 5 3. 5x = 0 4. 2w - 7 = 5w - 7 5. 4w - 3 = w + 5 6. 7m - 4 = 3m + 12 7. 9 - 5r = 4r + 11 8. 314w - 52 = 719 - 3w2 9. 3p + 215p - 62 = 19 + 417 - 2p2 10. 2 + 315t - 92 = 7t - 319 - 3t2 11. 5n - 3 = 219n - 72 + 14 - 6n 12. 13 - 5x + 312x - 92 = 12x 13.

2 3x

3 5y = 9 4 5 2z = 3 3 1 4w + 2 = w - 2 2 1 1 3 4w - 3w + 4 = 6w 3 3 1 7 5 y - 4 = 4 y + 10

+ 3

19. 0.2x - 0.315 - 2x2 = 0.4 20. 0.35 + 0.24x = 0.215 - 3x2 21. 0.4r - 0.52140 - 2r2 = 100

22. For x: 5x - 2y = 19 23. For y: 5x - 2y = 19 24. For x: ax + by = c 25. For y: ax + by = c 26. For C: F = 9/5C + 32

= 8

Section 0.1

27. For h: V = pr2h 28. For b: A = *2h1a + b2 29. For S: 1/R = 1/S + 1/W 30. Today, Maria s age is 5 years less than twice Louie s age. In five years, the sum of their ages then will be 41. What are their present ages? 31. Mike is 5 years older then Joan. In ten years, the sum of their ages will be 57. What are their present ages? 32. Alphonse took four examinations and his average on these four exams was 83. If his grades on the first three exams were 82, 88 and 78, what grade did he get on the fourth exam? 33. The length of a rectangle is two feet less than three times its width. If the perimeter is 36 feet, what are the dimensions of the rectangle? 34. An isosceles triangle has a perimeter of 33 centimeters. If the equal sides are 6 centimeters less than twice the unequal side, what is the length of the two equal sides? 35. The sum of three consecutive integers is 96, find the three integers. 36. The sum of two consecutive odd integers is 15 more than the next odd integer. Find the three integers. 37. Boston and New York are approximately 280 miles apart. One train leaves Boston traveling towards New York at an average speed of 80 mph. Another train leaves New York at the same time traveling toward Boston at an average speed of 60 mph. (a) How long will it take them to meet. (b) How far has each train traveled when they meet? 38. Two cars leave a city traveling in the same direction. One travels at 60 mph and the other at 45 mph. How long before the cars are 60 miles apart?

1

* **

11

39. Two planes leave JFK airport at the same time, one flying north at 500 mph and the other south at 600 mph. In how many hours will they be 4400 miles apart? 40. Two angles are said to be complementary if the sum of the measures of their angles is 90 degrees. If one angle is 15 degrees more than twice the other, what is the measure of the smaller angle? 41. Two angles are said to be supplementary if the sum of the measures of their angles is 180 degrees. If one angle is 30 degrees less than six times the other, what is the measure of the larger angle? 42. James has a collection of nickels and dimes which total to $5.15. If he has 5 fewer nickels than dimes, how many dimes has he? 43. The denominator of a fraction is 1 less than three times the numerator. If the numerator is increased by 1 and the denominator is decreased by 2 the resulting fraction is 1*2. Determine the original fraction. 44. The numerator of a fraction is three less than the denominator. Tripling the numerator and quadrupling the denominator results in a fraction equal to 3/5. Determine the original fraction. 45. $6,000 is invested for two years earning annual simple interest at 4%; to what does it accumulate? 46. How long does it take $5,000 to earn a total interest of $600 if the annual simple interest rate is 4%? 47. What annual simple interest does a $10,000 investment earn if it accumulates to $11,200 after three years?

Each question is worth one point. Solve for the unknown: 1. 3 - x = 2 2. 5x - 2 = 8 3. 5 - 7y = 26 5. 312z - 52 - 412 - 3z2 = 519 + 2z2 7.

3x 2 - 5x

4. 6.

4 3x 1 6x

= 20 2 3

= 3 4x + 2

3 4

8. 8.213 + 2x2 - 1.115x - 42 = 2 10. A nine foot rope is cut into two pieces so that one piece is twice the other. How long is the larger piece?

9. Solve for y: 3x - 4y = 9

12

* **

Section 0.2

0.2

Solve each of the following quadratic equations for the unknown. 1. 4x2 = 5x 3. 9x2 - 49 = 0 5. 12x - 1213x + 22 = 24 7. 3 2 2 x = 0 4 3 2. 5x 2 - 20x = 0 4. 16x2 - 12 = 0 6. 4x 2 + 32 = 0 8. 1x - 522 - 18 = 0 10. 3 12x - 522 - 14 = 1 4

9. 31x - 522 + 81 = 0

You should recall how to solve quadratic equations by factoring, for example the equation 6x2 - x - 12 = 0 may be factored as 12x - 3213x + 42 = 0. We recall that if AB = 0, then either A = 0 or B = 0, so we have the two possibilities, 2x - 3 = 0 or 3x + 4 = 0, the first equation yields x = 3/2 and the second x = - 4/3, so the two solutions to the given quadratic are x = - 4/3 and x = 3/2. You may not be able to (easily) factor every quadratic equation, therefore, we shall examine other techniques for solving quadratic equations which will easily yield their solutions, also called their roots. Consider the equation x2 - 9 = 0. This problem is most easily solved by factoring. We write, 1x - 321x + 32 = 0 Isolation of Squared Term and we immediately find that x = - 3 or x = 3, or more succinctly x = ; 3. We can also solve this equation by isolating the x 2 term. Consider the following. Rewrite the equation with the x2 term isolated on one side of the equation. x2 = 9 We now take the square root of each side, remembering that x could also be negative, so we have x1 = 29 = 3 or x2 = - 29 = 3. or written more succinctly as x = ; 29 = ; 3 More generally, given the equation ax 2 - b = 0, we can write

Section 0.2

* **

13

ax2 = b b x2 = a b x = ; Aa Of course, we would have to simplify the radical expressions, and sometimes, the solutions could be complex numbers. We illustrate the various cases in the examples that follow. One general remark: for simplicity in carrying out the algebraic manipulations, the first thing we should do is to clear any fractions. That is, if fractions appear in the quadratic equation, first multiply every term on each side of the equation by the LCD. This results in an equivalent quadratic equation all of whose terms are integers. Example 1 Solve the quadratic equation 4x 2 - 25 = 0. Solution Isolating the x2 term, we have 4x2 = 25 25 x2 = 4 25 x = ; A4 5 x = ; 2 Example 2 Solve the equation x 2 - 8 = 0. Solution Isolating the x2 term, we have x2 = 8 x = ; 28 x = ; 2 22 Example 3 2 1 Solve the equation x 2 - = 0 4 3 Solution We begin by first clearing fractions. We multiply every term on each side of the equation by the LCD which is 12. 2 1 12 # x2 - 12 # = 12 # 0 3 4 2 8x - 3 = 0 8x2 = 3 3 x2 = 8 3 3#2 26 x = ; = ; = ; A8 A8 2 4

14

* **

Section 0.2

Example 4 Solve the equation x 2 + 8 = 0. Solution Isolating the x2 term, we have x2 = - 8 x = ; 2 - 8 = ; i 28 = ; i 24 22 = ; 2 22i

Note that in the last example, the radicand was negative resulting in imaginary solutions. In the study of calculus we will be mostly concerned with solutions which are real numbers. A slight variation on the above method of solution occurs when the left hand side is not x2, but some expression squared. In such a case, we isolate the squared expression and then take the square roots, as illustrated in the next two examples. Example 5 Solve the quadratic equation 1z - 522 - 4 = 0. Solution We first isolate the squared term by rewriting the expression as 1z - 522 = 4 We next take the square roots, z - 5 = ; 24 z - 5 = ;2 z = 5 ; 2 z1 = 5 - 2 or z2 = 5 + 2 z1 = 3 or z2 = 7

We remark that if we actually squared the expression and rewrote the quadratic as z2 - 10z + 21 = 0, we could have easily solved the problem by factoring. However, that could not be done so easily on the next example. Example 6 Solve the quadratic equation 14y + 622 - 12 = 0 Solution. 14y + 622 = 12 4y + 6 = ; 212 = ; 2 23 4y = - 6 ; 2 23 - 6 ; 2 23 y = 4 We can reduce this expression by either breaking it into two fractions or by factoring. If we break it into two fractions, we have y = 6 2 23 3 23 - 3 ; 23 - 6 ; 2 23 = - ; = - ; = 4 4 4 2 2 2

Section 0.2

* **

15

- 3 + 23 - 3 - 23 or y2 = . If we need a numer2 2 ical approximation, we can use a calculator to find y1 L - 0.634 and y2 L - 2.366.

Try solving this last example by multiplying out the expression and factoring. You will quickly see that it is not a simple matter to find the factors since they are irrational numbers. Note that the two solutions to the quadratic equation in the previous examples are conjugates of each other, that is of the form a - 2b and a + 2b. This will be true in general, as we shall see. We now make one very important observation. Any quadratic equation written in the form 1x + B22 - C = 0 may be solved by isolating the squared binomial term and then take the square roots. It is precisely this observation that will enable us to solve any quadratic equation. What we shall do is to learn to rewrite any given quadratic equation in the above form. To do so, we must first learn how to complete the square, the object of the next section. You may recall in the last section we introduced the solve command to solve a linear equation and quadratic. This command will work on most equations, linear or not, and will yield the real solution(s). Consider the problem

Calculator Tips

(Note that is used for exponentiation.) The calculator produces the two solutions. For approximate solutions you need only press (the green diamond and then Enter) Enter, see Figure 2.

16

* **

Section 0.2

Figure 2: Obtaining Approximate Answers with the Calculator Note that if the solutions are not real, as in Example 4, the solve command results in the answer false, meaning there are no real solutions, see Figure 3.

The complex solution may be obtained by using csolve in place of solve, see Figure 4.

Section 0.2

* **

17

Find the unknown in each of the following exercises.

1. x = 25 2. x2 = 36 3. y2 - 4 = 0 4. y2 - 49 = 0 5. z2 - 25 = 39 6. x + 7 = 16 7. x2 - 4 = 21 8. y - 40 = 24 9. 5x - 7 = 2x + 20 10. 17x + 20 = 11x + 74 11. x2 = 18 12. 3z - 24 = 0 13. 8x - 96 = 0 14. 2y = 48 15. x2 + 25 = 0 16. w = - 36 17. x + 4 = 0 18. y + 49 = 0 19. x2 + 16 = 7 20. r + 21 = 5 21. x + 20 = 0 22. 3x + 24 = 0 23. 6w2 + 96 = 0 24. 5y2 + 12 = 3y2 - 24 25. 6y2 + 9 = 4y2 - 45 26. 27. 28.

2 2 3x 3 2 4y 5 2 6r 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

29. 30.

31. 1x - 22 = 9 32. 1x - 321x + 22 = 14 33. 1x + 1212x - 32 = 12 34. 1x + 322 = 25 35. 1x - 422 + 3 = 39 36. 12y - 322 = 49 37. (3z - 4)2 = 64 38. 14x - 722 = 18 39. 15x + 422 = 20 40. 12r - 322 - 24 = 0 41. 16y - 222 + 9 = 41 42. 1x + 322 = - 9 43. 1x - 422 + 16 = 0 44. 1y - 522 + 12 = 0 45. 13x - 222 + 25 = 0 46. 14y + 722 + 36 = 0 47. 12y - 322 + 24 = 0 48. 13x - 522 + 18 = 0 49. 16x - 422 + 48 = 0 50. 51. 52. 53. 54. 55. 56. 57.

2 2 3 1x + 32 = 12 3 2 2 13y - 22 - 18 = 0 4 2 5 15w + 22 + 25 = 0 3 2 412z - 32 + 24 = 9 3 15 3 2 81x + 72 + 4 = 2 3 15 3 2 8 12r + 72 + 2 = 4 2 1 2 513x - 52 + 3 = 2 1 2 2 5 12z - 52 + 2 = 3

5 2 1 5 6w - 4 = 3 1 2 8x + 3 = 0 2

- 8 = 0 + 24 = 0 2 3

3 4

Solve each of the following quadratic equations for the unknown. 1. 7x2 = 3x 3. 16x2 - 81 = 0 5. 14x - 5214x + 52 = - 9 7. 4 2 3 x = 0 5 10 2. 4x 2 - 18x = 0 4. 20x2 - 45 = 0 6. 6x 2 + 120 = 0 8. 1x + 322 - 24 = 0 10. 5 15x + 422 - 19 = 1 3

9. 41x - 522 + 48 = 0

18

* **

Section 0.3

0.3

Each question is worth two points. 1. What number should be added to x2 - 8x so that it is a perfect square? 2. What number should be added to x 2 + 3x so that it is a perfect square? Complete the square to solve each of the following quadratic equations. Leave the answer in simplest form. 3. x2 - 10x - 3 = 0 4. x2 + 5x - 5 = 0 5. 3x 2 - 12x - 4 = 0

In the preceding section we observed that any quadratic equation written in the form 1x + B22 - C = 0 (1)

may be solved by isolating the squared term and then take square roots. The question that naturally arises is given any quadratic equation, how do we rewrite it so that the unknown appears within a square as in (1)? The procedure by which this is accomplished is called completion of the square, and in addition to the solution of quadratic equations is useful in numerous applications. Let us make some observations on several expressions which are identities: x 2 + 6x = 1x + 322 - 9 The expression may be rewritten as x2 + 6x = A x + 1*2162 B 2 - A 1*2162 B 2 x 2 - 8x = 1x - 422 - 16 The expression may be rewritten as x2 - 8x = A x + 1*21 - 82 B 2 - A 1*21 - 82 B 2 x 2 + 5x = 1x + 5/222 - 25/4 The expression may be rewritten as x 2 + 5x = A x + 1*2152 B 2 - A 1*2152 B 2 Do you see the pattern in each of the examples? We spell it out more generally. The sum x 2 + Bx may be written as a perfect square as follows: In words, take one-half the coefficient of the x term, (the sign is part of the coefficient) add it to x, square the sum, then subtract the square of one-half the coefficient of the x term from the previous squared sum. x 2 , Bx + a x ,

2 2 1 1 Bb * a Bb 2 2

(2)

Section 0.3

* **

19

Of course, you may verify that the identity is true by simply multiplying out the right hand side of (2), and comparing it to the left hand side. This identity is useful not only in solving quadratic equations, but in other applications as well, as you shall see later in this text as well as in calculus. Example 1 Rewrite the expression x2 + 12x in the perfect square form as given on the right-hand side of (2). Solution The coefficient of x is 12, therefore 1*2 1122 = 6, so we have x2 + 12x = 1x + 622 - 1622 or x 2 + 12x = 1x + 622 - 36

Example 2 Rewrite the expression x2 - 10x in the perfect square form as given on the right-hand side of (2). Solution The coefficient of the x term is - 10, one-half this number is - 5, so we have x2 - 10x = 1x - 522 - 1 - 522 = 1x - 522 - 25

Example 3 Rewrite the expression x2 - 3x in the perfect square form as given on the right-hand side of (2). Solution The coefficient of the x term is - 3, one-half this number is - 3/2, so we have x2 - 3x = 1x - 3/222 - 1 - 3/222 = 1x - 3/222 - 9/4

It is just one more step to see how rewriting a quadratic may be used to solve any quadratic equation. Consider the quadratic equation x2 - 10x + 10 = 0 We isolate the two terms involving x, x2 - 10x = - 10 We rewrite the left-hand side of this equation as in Example 2 as x 2 - 10x = 1x - 522 - 25. The quadratic equation may now be written as 1x - 522 - 25 = - 10 The problem is now in the form we studied in the previous section. We isolate the square term and solve. 1x - 522 = 15 x - 5 = ; 215 x = 5 ; 215

20

* **

Section 0.3

We illustrate how the method of completion of the square is used to solve quadratic equations in the following examples. Example 4 Solve the quadratic equation x2 - 8x + 4 = 0 Solution We rewrite the equation as x 2 - 8x = - 4 and next rewrite the left-hand side as x 2 - 8x = A x + 1*21 - 82 B 2 - A 1*21 - 82 B 2 = 1x - 422 - 16. Therefore, we have 1x - 422 - 16 = - 4 1x - 422 = 12 x - 4 = ; 212 = ; 2 23 x = 4 ; 2 23

Example 5 Solve the quadratic equation x2 - 7x - 6 = 0 Solution We rewrite the equation as x2 - 7x = 6 x2 - 7x = 1x - 7/222 - 1 - 7/222 = 1x - 7/222 - 49/4 therefore we have, 1x - 7/222 - 49/4 = 6 or 1x - 7/222 = 49/4 + 6 1x - 7/222 = 73/4 7 73 273 x - = ; = ; 2 A4 2 7 273 7 ; 273 x = ; = 2 2 2 In all the examples considered so far, the coefficient of the x 2 term has been 1. When this is not the case, multiplying each term on both sides of the equation by the reciprocal of the coefficient reduces the problem to an equivalent one with coefficient 1. We illustrate in the following examples. Example 6 Solve the quadratic equation 4x2 + 12x - 25 = 0. Solution Multiplying each term on both sides of the equation by the 1/4, the reciprocal of 4, gives x2 + 3x - 25/4 = 0 rewriting, we have x 2 + 3x = 25/4

Section 0.3

* **

21

completing the square and solving, we have 1x + 3/222 - 13/222 = 25/4 1x + 3/222 - 9/4 = 25/4 1x + 3/222 = 25/4 + 9/4 1x + 3/222 = 34/4 3 234 = ; 234/4 = ; 2 2 3 234 - 3 ; 234 x = - ; = 2 2 2 x + We remark, that if approximate answers are needed, we would use a calculator to approximate the square roots; for example, in the previous example, the roots to the nearest onethousandth are 1.416 and - 4.416. Example 7 Solve the quadratic equation 2x 2 - 8x + 25 = 0 Solution. We first multiply each term on each side of the equation by the reciprocal of the coefficient of the x2 term, 1*2, giving x2 - 4x + 25/2 = 0 or x2 - 4x = - 25/2 completing the square, we have

or 1x - 222 - 4 = - 25/2 1x - 222 = 4 - 25/2 1x - 222 = - 17/2 x - 2 = ; A 34 234 17 17 17 # 2 = ;i = ;i = ;i = ;i A4 2 2 A2 A2 2 x = 2 ; 234 4 ; 234i i = 2 2

The use of completion of the square in this section was to solve a quadratic equation. As we illustrated in the last section, we may use the solve function to solve equations having real solutions. For example, if we want to solve the quadratic in Example 6, we enter solve(4x2 + 12x - 25 = 0, x) the calculator will give us the two solutions. However, you should be warned, if you try to solve a problem with complex solutions with the solve function, it tells us there is no solution, that is, if you entered the last example, solve 12x 2 - 8x + 25 = 0, x2 the calculator responds with false, indicating there is no real solution. Instead, use the function csolve which will solve for complex (as well as real) solutions. Thus, to solve this problem for the complex solutions, you would enter csolve 12x 2 - 8x + 25 = 0, x2 the calculator then provides the solution obtained above. We illustrate these tips along with the screen shots on other examples at the end of the next section.

Calculator Tips

22

* **

Section 0.3

Complete the square in each of the following by putting in the form 1x + B22 - C.

1. x2 + 4x 2. x2 + 6x 3. x2 - 6x 4. x - 8x 5. x - 10x 6. x + 12x 7. x2 + 3x 8. x - 3x 9. x - 5x 10. x + 5x 11. x2 - 7x 12. x + 9x

2 2 2 2 2 2 2

28. 1x - 321x + 52 = 12 29. 1x + 321x + 22 = - 10 30. 12x - 3213x + 52 = - 20 31. x2 - 6x + 23 = 0 32. x2 - 12x = - 48 33. x2 + 10x + 50 = 0 34. x2 - 8x + 36 = 0 35. x2 - 4x = 8 36. x2 + 6x + 21 = 0 37. x2 - 3x + 3 = 0 38. x2 + 5x + 10 = 0 39. 2x2 + 8x + 1 = 0 40. 2x2 + x - 6 = 0 41. 6x2 + 5x - 6 = 0 42. 12x2 - 7x = 12 43. 6x2 + 23x = - 20 44. 2x2 + 5x + 10 = 0 45. 4x2 - 5x + 8 = 0 46. 1*2 x2 + 5x + 2 = 0 47. 48.

1 2 2x 3 2 5x

Solve each of the quadratics by first completing the square. When the roots are irrational, also give the solutions to the nearest onethousandth.

13. x2 + 2x - 10 = 0 14. x2 + 6x + 3 = 0 15. x2 - 2x - 10 = 0 16. x2 - 6x + 3 = 0 17. x2 + 6x = 3 18. x2 - 8x = 4 19. x2 - 12x - 4 = 0 20. x2 + 5x = 5 21. x2 - 5x = 5 22. x2 - 3x - 1 = 0 23. x2 + 3x - 1 = 0 24. x 2 + 2x + 10 = 0 25. 1x - 421x + 12 = 6 26. 1x - 321x + 52 = 20 27. 12x - 3213x + 12 = - 4

- 3 4x = - 3 2x =

5 6 7 10

49. 3 - 4x2 = 8x 50. 9 - 3x2 + 5x = 8 + 2x2 51. The area of a triangle is 20 square feet. If the height is three feet less than the base, find the length of the height and base of the triangle. 52. A rock is thrown down from the ledge of a mountain 200 feet above the ground with an initial velocity of 48 feet per second. If the height, h, is given by the equation h = - 16t2 - 48t + 200, where t is the time in seconds, how long does it take for the rock to hit the ground? Give your answer to the nearest one-hundredth of a second. 53. A farmer wants to set aside a rectangular plot of land to contain 100 square meters. If the width of the plot is 10 meters less than the length, find the dimensions of the plot. Give your answers rounded to the nearest tenth of a meter.

Each question is worth two points. 1. What number should be added to x2 - 12x so that it is a perfect square? 2. What number should be added to x 2 + 5x so that it is a perfect square? Complete the square to solve each of the following quadratic equations. Leave the answer in simplest form. 3. x2 - 12x - 4 = 0 4. x2 + 3x - 9 = 0 5. 4x 2 - 16x - 8 = 0

Section 0.4

* **

23

0.4

Quadratic Formula Clearing Fractions Applications Equations Reducible to Quadratics Calculator Tips

Solve each of the following using the quadratic formula. Leave answers in simplest radical form. Each question is worth two points. 1. 2x2 - 4x - 5 = 0 3.

3 2 4x

2. 4x - x 2 = - 1 4. 3x2 - 6x + 4 = 0

= 2 3x +

1 6

5. The vertical height h of a rocket measured in feet at time t measured in seconds is given by the equation h = - 16t2 + 3200t. To the nearest thousandth of a second, how long does it take the rocket to reach a height of 5,280 feet?

We saw in the last section that any quadratic equation could be solved by the method of completion of the square. Suppose we apply this method to the general quadratic equation ax2 + bx + c = 0 (1)

What should then happen is that our solution should depend on a, b and c. That means we will have a formula for the solutions to any quadratic equation. We proceed by solving (1) using completion of the square. We assume that (1) is indeed a quadratic, which means that a Z 0. We multiply every term on each side of (1) by 1/a, obtaining x2 + we rewrite this equation as x2 + we next complete the square and obtain ax + b 2 b 2 c b - a b = a 2a 2a b c x = a a b c x + = 0 a a

24

* **

Section 0.4

taking square roots, we have x + Solving for x, we find x = or writing as a single fraction we have * b_ 2b 2 * 4ac 2a b 2b2 - 4ac ; 2a 2a b2 - 4ac b ; 2b2 - 4ac = = ; 2 2a 2a B 4a

x +

(2)

Quadratic Formula

The solution given in (2) is the solution to the general quadratic equation and is known as the quadratic formula. We remind you that there are two solutions, as the ; symbol is a shorthand that tells us first use the + sign to write one solution, and then use the - sign to write the other solution. Sometimes, we call one solution x1 and the second solution x2. Sometimes, we call the solutions the roots. We illustrate the use of this formula on the following exercises. First, some suggestions that will make the use of the formula more convenient.

Given any quadratic equation, perform the following steps, if required: 1. Rewrite the equation so that the terms are in descending powers. 2. Rewrite the equation so that the coefficient of the squared term is positive. 3. Clear all fractions so the coefficients are all integers.

We remark that Step 2 is easily accomplished by multiplying each term in the equation by - 1, and Step 3 is accomplished by multiplying each term in the equation by the least common denominator. These remarks will be illustrated in the examples that follow. Example 1 Use the quadratic formula to solve the equation 6x2 + 5x = 6. Solution We first rewrite the equation as 6x2 + 5x - 6 = 0. We identify the coefficients; a = 6, b = 5 and c = - 6. We next substitute into the quadratic formula to obtain x = - 5 ; 21522 - 41621 - 62 - 5 ; 225 + 144 - 5 ; 2169 - 5 ; 13 = = = 2162 12 12 12

therefore, we have as our two solutions (roots), x1 = 1 - 5 + 132/12 = 8/12 = 2/3 and x2 = 1 - 5 - 132/12 = - 18/12 = - 3/2. Note that the preceding example could have easily been solved by factoring as 6x 2 + 5x - 6 = 13x - 2212x + 32 = 0 yields x = 2/3 and x = - 3/2.

Section 0.4

* **

25

Example 2 Use the quadratic formula to solve the equation 4x = - 3x2 + 8. Solution We first apply Step 1 and rewrite the equation as 3x 2 + 4x - 8 = 0. We now identify a = 3, b = 4, and c = - 8. Substituting into the quadratic equation, we have - 4 ; 21422 - 41321 - 82 - 4 ; 216 + 96 - 4 ; 2112 = = 2132 6 6

x =

We next simplify this expression, remembering what we learned about the simplification of radicals. We have, x = 2 A - 2 ; 2 27 B - 4 ; 4 27 - 4 ; 216 # 7 - 2 ; 2 27 = = = 6 6 6 3

thus, we have irrational conjugate expressions as the two roots (solutions) x1 = - 2 + 2 27 - 2 - 2 27 and x2 = 3 3

If we needed numerical solutions, we would compute them using a calculator to approximate the square root. Example 3 2 3 1 (a) Solve the quadratic equation x 2 - x + = 0. (b) Approximate the roots to the 4 6 3 nearest thousandth Solution (a)We first clear fractions by multiplying each term by the least common denominator 12. We obtain 8x2 - 9x + 2 = 0. We set a = 8, b = - 9, and c = 2. This gives x = - 1 - 92 ; 21 - 922 - 4182122 9 ; 281 - 64 9 ; 217 = = 2182 16 16

(b) Using 217 L 4.1231056 we obtain, to three decimal places, x1 = 0.305 or x2 = 0.820.

Example 4 1 5 1 Solve the quadratic equation z2 + = z. 8 2 4 Solution We clear fractions by multiplying each term by the LCD which is 8, giving 2z2 + 5 = 4z. We rewrite this equation as 2z2 - 4z + 5 = 0. We set a = 2, b = - 4 and c = 5. Substitution into the quadratic formula, gives

Clearing Fractions

26

* **

Section 0.4

z =

- 1 - 42 ; 21 - 422 - 4122152 4 ; 216 - 40 4 ; 2 - 24 = = = 2122 4 4 2 A 2 ; 26i B 4 ; i 224 4 ; i 24 26 4 ; 2 26i = = = 4 4 4 4 z = 2 ; 26i 2 2 - 26i 2 + 26i and z2 = . 2 2

Applications

In applications involving quadratic equations, we sometimes find that both solutions make sense and sometimes we find that one of the solutions needs to be rejected as it makes no physical sense. Consider the following examples. Example 5 The altitude h of a rocket fired vertically upward is given by the equation h = - 16t2 + 2500t, where t is the time from firing in seconds. (a) How long does it take the rocket to reach an altitude of 4700 feet? (b) How long does it take for the rocket to return to the ground? Solution (a) We are asked to find t when h = 4700 feet. Therefore we need to solve the equation 4700 = - 16t2 + 2500t We rewrite this equation as 16t2 - 2500t + 4700 = 0 Using the quadratic formula, we have t = 2500 ; 225002 - 4 # 16 # 4700 32 2500 ; 25949200 t = 32

Using a calculator, we find the two solutions are t1 L 154.347 seconds or t2 L 1.903 seconds. Why are there two solutions? When the rocket is going upward, it reaches a height of 4700 feet in approximately 1.903 seconds. However, this rocket will achieve its maximum altitude and then begin to come down. At approximately 154.347 seconds after launch it will again be at this position but coming down. (b) When the rocket returns to the ground its altitude h = 0, so we need to solve the equation 0 = - 16t2 + 2500t This equation may be solved by factoring 0 = - t116t - 25002 yielding t = 0 or t = 2500/16 = 156.25 seconds. When t = 0, the rocket is being launched, so the time it takes for the rocket to return to the ground is 156.25 seconds.

Section 0.4

* **

27

Exercise 6 One leg of a right triangle is 6 inches and the hypothenuse is 14 inches, find the length of the other leg of the triangle to the nearest thousandth of an inch. Solution Let x represent the length of the unknown leg, see Figure 1. Using Pythagoras theorem, we have x + 6 = 14 isolating the unknown, we find that x2 = 160, x = ; 2160 L = ; 12.649 Since x represents a length, it must be a positive number, therefore, we reject the negative solution. The other leg of the triangle is therefore approximately 12.649 inches.

2 2 2

x 6 Figure 1

14

Exercise 7 (a) If the profit P (in dollars) earned by selling x units of some item is given by the equation P = - 3x2 + 200x - 2000, find (to the nearest integer) the value of x which produces a profit of $1,000. (b) Find the quantity (to the nearest integer) which produces a break even situation 1Profit = 02. Solution (a) We need to solve the equation - 3x 2 + 200x - 2000 = 1000 or equivalently 3x2 - 200x + 3000 = 0 We may solve this equation using the quadratic formula. We find (verify) x = 10 A 10 ; 210 B 3

Or to the nearest integer, x = 23 or 43. (b) A break even situation arises when the profit is zero, so we solve - 3x2 + 200x - 2000 = 0 or 3x 2 - 200x + 2000 = 0 20 A 5 ; 210 B Solving, we find x = , or to the nearest integer, x = 12 or 54. 3 Sometimes, we need to perform some algebraic simplifications on an equation before we recognize it as a quadratic, as the next two examples illustrate. Exercise 8 Solve for x: (note that x Z - 8, 2. Why?) 24 20 + = 12. x + 8 x - 2 Equations Reducible to Quadratics

28

* **

Section 0.4

Solution We multiply each term on both sides of the equation by the LCD which is 1x + 821x - 22, yielding 241x - 22 + 201x + 82 = 121x + 821x - 22 Multiplying out and collecting similar terms, we obtain 12x 2 + 28x - 304 = 0 or 413x 2 + 7x - 762 = 41x - 4213x + 192 = 0 Therefore x1 = - 19/3 and x2 = 4. We leave it as an exercise for you to check the solutions.

Sometimes, a trinomial may be at first glance appear not to be a quadratic equation, but by a simple substitution it may be transformed into a trinomial which is a quadratic, as illustrated by the next example. Exercise 9 3 3 Solve the equation 8x 4 - 19x 8 - 27 = 0. Solution The key to this problem and problems similar to it is that the exponent of the highest power term in the trinomial is twice the exponential of the next highest power term. Thus, if we let the smallest power be called u, i.e., u = x8 then u2 = A x 8 B 2 = x 8 = x 4 (Recall that when raising a power to a power, we keep the base and multiply the exponents.) We may rewrite the original equation as 8u2 - 19u - 27 = 0 which is a quadratic in the variable u and factors as 18u - 2721u + 12 = 0 Thus u = 27/8 or u = - 1. 3 8 3 8 8 However, we need to find x. Since u = x8, we have 1u23 = A x 8 B 3 or x = u3. Therefore, when u = - 1, we have x = 1 - 128/3 = 1; when u = 27/8, we have x = 127/828/3 = 1127/821/328 = 13/228 = 6561/256, however, x = 1 does not check (it is extraneous that is, it does not check the original equation), thus, we have x = 6561/256.

3 6 3 3

Calculator Tips

The solve command will work well on all the problems considered in this section. Consider the last example, 8x 3/4 - 19x 3/8 - 27 = 0, we enter the equation into the calculator as illustrated in Figure 1 and press enter. You ll note that the equation is too big for the screen, so the ellipsis symbol (...) appears on the input line to indicate you need to scroll to see the rest of the equation. Similarly, the black arrow appears on the screen to indicate the entire image is not there and you need to scroll to see the remainder of the image.

Section 0.4

* **

29

In exercises 1 15 rewrite the equation in the form ax2 + bx + c = 0, with a 7 0, and then determine the coefficients a, b, and c. Do not solve the equation.

1. x + 4x - 11 = 0 2. 3x2 + 7x - 8 = 0 3. 5x2 - 8x - 12 = 0 4. 7x + 3x2 - 10 = 0 5. - 27x + 10 - 5x2 = 0 6. 4x = 9x2 + 4 7. 3 - 2x = 8x + 5x2 - 11 8. 2x2 = 10 9. 5x2 = - 11 10. 3x = 7x 2 11. 4x2 = 9x 12. 1x - 2213x + 42 = 5 13. 12x - 5214x + 32 = 7 14. 413x - 4215x - 62 = 25 15. 1x - 7212x - 32 = 3x12x - 42

2

22. x2 - 12x - 4 = 0 23. x2 + 5x = 5 24. x2 - 5x = 5 25. x2 - 3x - 1 = 0 26. x2 + 2x + 10 = 0 27. 1x - 421x + 12 = 6 28. 1x - 321x + 52 = 20 29. 12x - 3213x + 12 = - 4 30. 1x - 321x + 52 = 12 31. 1x + 321x + 22 = - 10 32. 12x - 3213x + 52 = - 20 33. x2 - 6x + 23 = 0 34. x2 - 12x = - 48 35. x2 + 10x + 50 = 0 36. x2 - 8x + 36 = 0 37. x2 - 4x = 8 38. x2 + 6x + 21 = 0 39. x2 - 3x + 3 = 0 40. x2 + 5x + 10 = 0 41. 2x2 + 8x + 1 = 0 42. 2x2 + x - 6 = 0 43. 6x2 + 5x - 6 = 0 44. 12x2 - 7x = 12 45. 6x2 + 23x = - 20 46. 2x2 + 5x + 10 = 0 47. 4x2 - 5x + 8 = 0 48. 1*2 x2 + 5x + 2 = 0 49.

1 2 2x

In each of the following, solve the given quadratic equation exactly using the quadratic formula. Write the solutions in its simplest form. Using a calculator, determine all irrational solutions to the nearest thousandth.

16. x2 + 2x - 10 = 0 17. x2 + 6x + 3 = 0 18. x2 - 2x - 10 = 0 19. x2 - 6x + 3 = 0 20. x2 + 6x = 3 21. x2 - 8x = 4

- 3 4x =

5 6

30

50.

* **

Section 0.4

7 10

62. John bikes a distance of 120 miles and then returns over the same route. On his return his average speed is 2 miles per hour more than when going. If the combined time for both trips was 22 hours, what was his speed each way? 63. Mary can build a computer in two hours less time than Tim. Working together, they can build a computer in 2 hours and 24 minutes. How long does it take Mary alone to build a computer? 64. The profit in dollars in producing x-items of some commodity is given by the equation P = - 2x2 + 400x - 15000. (a) How many items should be produced to break even? (b) How many items should be produced to maximize profit? (c) What is the maximum profit? 65. The profit in dollars in producing x-items of some commodity is given by the equation P = - 20x2 + 1300x - 15000. To the nearest integer, how many items should be produced to (a) yield a profit of $2,000? (b) break even?

3 2 5x

- 3 2x =

2

51. 3 - 4x = 8x 52. 9 - 3x2 + 5x = 8 + 2x2 53. The sum of the squares of three consecutive integers is 110, find the integers. 54. The sum of the squares of three consecutive odd integers is 515, find the integers. 55. If the legs of an isosceles right triangle are each 12 inches, how long, to the nearest one-thousandth of an inch, is the hypothenuse? 56. The hypothenuse and one leg of a right triangle are 18 and 12 inches respectively, how long, to the nearest one-thousandth of an inch is the other leg? 57. A ball is thrown vertically upward from the ledge of a building 75 feet above ground. The ball s height h in feet above the ground at time t in seconds is given by the equation h = - 16t2 + 80t + 75. (a) how long does it take the ball to reach a height of 90 feet? (b) How long before the ball is back to its original position (at 75 feet)? (c) How long before the ball hits the ground? Give each answer to the nearest one-thousandth of a second. 58. An object is dropped from a helicopter hovering at 250 feet above the ground. The objects height in feet, h is given by the equation h = - 16t2 + 250, where t is measured in seconds. How long before the object is (a) 100 feet above the ground? (b) 50 feet above the ground? (c) on the ground? Give each answer to the nearest one-thousandth of a second. 59. A rectangular swimming pool is 30 feet by 40 feet. If a rectangular strip of grass of uniform width is to go around the pool, and the area of this strip is 624 square feet, how wide is the strip? 60. Two cars leave an intersection at the same time, one goes north and the other goes east. Some time later they are 125 miles apart. If the car moving north traveled 12 miles more than the one going east, how many far (to the nearest mile) did each car travel? 61. Barbara wants to purchase an area rug for her dining room whose dimensions are 20 feet by 24 feet. If the rectangular rug she purchases has an area of 216 square feet and is placed an equal distance from each wall (a) how wide is the uniform strip of uncovered flooring? (b) what are the dimensions of the rug?

66. 67. 68. 18 9 12 + = x - 4 x + 4 2 18 8 = 1 2x + 3 x + 5 4x 16 - 3x + = 2 3x - 2 2x + 2

69. x4 + 5x2 - 36 = 0 70. 6x6 - 17x3 + 12 = 0 71. x3 - 35x 3 + 216 = 0 72. 2x4 - 4x2 - 3 = 0 73. Here is another proof of the quadratic formula. Begin with ax2 + bx + c = 0, and multiply each term of the equation by 4a. This gives 4a2x 2 + 4abx + 4ac = 0. Rewrite the equation as 4a2x 2 + 4abx = - 4ac and add b2 to each side giving 4a2x2 + 4abx + b2 = b2 - 4ac. Now factor the left-hand-side of this equation and complete the proof. 74. Given the quadratic equation ax2 + bx + c = 0 and cx2 + bx + a = 0, prove that the roots of one equation are the reciprocals of the roots of the other equation.

2 1

Solve each of the following using the quadratic formula. Leave answers in simplest radical form. Each question is worth two points. 1. 4x2 - 3x - 2 = 0 3.

2 2 5x

2. x - 2x 2 = - 4 4. 4x 2 - 5x + 3 = 0

1 10 x

1 2

5. A ball is thrown vertically upward from a ledge of a building 200 feet above the ground. Its height above the ground s, measured in feet, in terms of its time of flight t, measured in seconds, is given by the equation s = - 16t2 + 86t + 200. How long does it take the ball to hit the ground?

Section 0.5

* **

31

0.5

Each question is worth two points. Solve the given inequality 3 - 2x 1. 3x2 + 10x 8 2. 2x3 - 12x2 + 18x 6 0 3. x - 5 2x 31x 2 - 3x + 22 2x - 3 4. 5. 7 2 0 3x + 4 x2 - 16 0

Given the linear inequality, 3x - 2 7 0, we wish to find all values of x for which this inequality is a true statement. We recall that the rules for manipulating an inequality are the same as those for solving an equation with one exception, namely when an inequality is multiplied or divided by a negative number the inequality symbol reverses. Thus, to solve the inequality 3x - 2 7 0, we could write 3x 7 2 or x 7 2/3. This inequality is completely solved, that is, for any x-value greater than 2/3, the inequality is true. Let us now re-examine the same problem and suggest another approach which will work on any inequality which may be factored and is to be less than or greater than zero. We first solve 3x - 2 = 0. This is of course yields x = 2/3. We draw the number line and indicate this value on the line as indicated in Figure 1. Sign of 3x-2 I 0 II !!!!!! e !!!!!! value of x 2/3

Figure 1: Examining the Sign of 13x - 22 The only time the term 3x - 2 is zero is when x = 2/3. For any other value of x it is either positive (greater than 0) or negative (less than 0). Therefore, in each of the Regions I or II the sign 3x - 2 must always be the same. Therefore, we need only choose any test value for x in each of these regions and examine the sign of 3x - 2 at the test value. Consider Table 1 where we examine the sign at a test value in each region. Table 1: Sign of 13x - 22

Region I II Test Value 0 1 sign of 13x - 22 3102 - 2 = 3112 - 2 = +

Thus, we see that in region II the expression 13x - 22 is positive therefore our solution is x 7 2/3.

32

* **

Section 0.5

We now generalize the second method and consider the following non-linear inequality in factored form 1x - 521x + 22 6 0 Our problem is to find those value of x which satisfy the inequality, that is, find those values of x for which the product of the two factors is negative. We first observe that the right hand side of the inequality is the number zero. This is important as we shall see. Observe that the only time the product on the left-hand-side is zero is when either factor is zero, that is, when x = - 2 or x = 5. That means for any other value of x the product of these factors must always be positive or negative. A simple way to determine this is as follows: draw the number line and indicate these two values for x on the number line as indicated in Figure 2. sign of (x - 5)(x + 2) value of x I 0 -2 II 0 5 III x

Sign Analysis

Figure 2: Examining the Sign of 1x - 521x + 22 Note that at x = - 2 and x = 5, we put a zero in the diagram. That is to indicate that the product of the two factors is zero at each of those x values. Moreover, we included letters for the three regions they created. Region I corresponds to the region in which x 6 - 2, Region II corresponds to - 2 6 x 6 5 and Region III to x 7 5. Since the product is only zero at the indicated two x-values, it will be of the same sign, that is, only positive or only negative in each of the three indicated regions. All we need to do is to select any x-value in each of these regions and determine the sign of the product. We do this in Table 2. Table 2: Sign of 1x - 521x + 22

Region I II III Test value x = -3 x = 1 x = 6 Sign of Product 1x - 521x + 22 1 - 821 - 12 = + 1 - 42132 = 112182 = +

We can now redraw the sign diagram in Figure 2, indicating the sign of the product in each of the three regions. This is done in Figure 3. + 0 -2 0 5 + x

Figure 3: Sign Diagram for 1x - 521x + 22 This tells us that the solution to the inequality 1x - 521x + 22 6 0 is any x in the interval - 2 6 x 6 5, because in that region the product is negative. Sometimes the answer is written as 1 - 2, 52, this is shorthand for the interval - 2 6 x 6 5. More generally, (a, b) is shorthand for the interval a 6 x 6 b, [a, b] is shorthand for the interval a x b, and (a, b] is shorthand for a 6 x Note that the bracket is used to include the endpoint and b. the parenthesis is used to exclude it. (How would you indicate a in interval x 6 b notation?) Notice that the given inequality could have been written as x 2 - 3x - 10 6 0, or 2 x - 3x 6 10, or some other equivalent formulation could have been given. All that we

Interval Notation

Section 0.5

* **

33

need do is rewrite the expression so that one side of the inequality is the number zero, and then proceed as above. The above suggests a general procedure for the solution of an inequality.

Solving an Inequality

1. Rewrite the inequality so that the expression involving x is on one side of the inequality symbol and the number 0 is on the other side of the inequality symbol. 2. Solve for those x-value at which both the numerator = 0 and the denominator = 0. 3. Draw a number line indicating those x-values (in increasing order) found in Step 2. Put the number 0 in the sign diagram above each of the zeros of the numerator and ND for each zero of the denominator. 4. Choose any x-value in each of the regions separated by the values found in Step 2. Compute the sign of the expression at each of these values and enter them on the sign diagram. 5. Read the solution from the sign diagram.

We illustrate the procedure on the following examples. Example 1 Determine the values of x satisfying the inequality (a) 2x 3 - 8x 2 (b) 2x3 - 8x 2 - 8x. Solution (a) First, we rewrite the inequality so that zero is on one side. We have 2x3 - 8x 2 + 8x 0

- 8x

We want to determine when the expression on the left will be negative or zero. Factor the left hand side, to obtain 2x1x2 - 4x + 42 or 2x1x - 221x - 22 or finally, 2x1x - 222 0 0 0

The left hand side will be zero when 2x = 0, or x - 2 = 0, yielding x = 0 or x = 2. We indicate the number line in Figure 4. I 0 0 II 0 2 III

Figure 4: Examining the Sign of 2x1x2 - 4x + 42 We choose values in each of the regions and determine the sign of the expression at these test values. They are indicated in Table 3.

34

* **

Section 0.5

Region I II III Test value x = -1 x = 1 x = 3 Sign of Product 2x1x - 222 21 - 121 - 322 = 21121 - 122 = + 21321122 = +

The expression will be negative or zero when x 0 or x = 2. We allow x = 2 because the inequality is 2x3 - 8x 2 + 8x 0, allowing the expression to equal to zero. Using the interval notation we may write the solution as 1 - q , 02 or x = 2. (The symbol - q is read minus infinity. ) (b) this reduces to solving the inequality 2x 3 - 8x 2 + 8x 0. Its solution is read immediately from Figure 4, namely the interval x 0, or in interval notation, as [0, q 2. (Note that when x = 0 or 2, the expression is zero, elsewhere in the interval it is positive.)

Example 2 2x - 5 Find those values of x so that 0 x + 3 Solution The problem reduces to determining when the quotient on the left of the inequality is negative or zero. Since the rules of signs for quotients is the same as for products, (when the factors have the same sign the quotient will be positive, and when they are different it will be negative) we proceed in almost the identical way as above. The factors that determine the sign are the numerator and denominator. The numerator will be zero when x = 5/2 (verify!), and the denominator will be zero when x = - 3. Be careful, x is not allowed to be - 3, because then we would be dividing by zero. We indicate this on the number line by writing ND (not defined) above - 3. See Figure 6.

ND -3

II

0 5/2

III x

We next choose and test points in each of the regions. They are indicated in Table 4.

Section 0.5

- 5 Table 4: Sign of 2xx + 3

* **

35

Region I II III

Test value

- 5 Sign of 2xx + 3

x = -4 x = 0 x = 4

(Notice that all we really need is the sign not the actual value of the expression at the test values.) We now can complete the sign diagram as given in Figure 7. + ND -3 0 5/2 + x

We may now read our solution directly from the sign diagram, namely - 3 6 x 5/2 or in interval notation, 1 - 3, 5/2]. Note that x Z - 3, but x = 5/2 since the expression is not defined at - 3, and is zero at 5/2.

Example 3 3x - 8 Solve the inequality 7 1. 2x + 5 3x - 8 Solution We first rewrite the inequality so that one side is zero. We have - 1 7 0. 2x + 5 We combine into one fraction to obtain 3x - 8 - 12x + 52 3x - 8 2x + 5 x - 13 = = 7 0 2x + 5 2x + 5 2x + 5 2x + 5 The method of solution is now similar to the previous example. The numerator is zero when x = 13, and the denominator is zero (the fraction is not defined) at x = - 5/2. We begin the sign analysis with Figure 8. I ND -5/2 II 0 III 13 x

- 13 Table 5: Sign of x 2x + 5

Region I II III

Test value

- 13 Sign of x 2x + 5

x = -3 x = 0 x = 14

36

* **

Section 0.5

- 13 This expression x 2x + 5 is positive when x 6 - 5/2 or x 7 13 (which may also be written as - q 6 x 6 - 5/2 or 13 6 x 6 q . Using interval notation, we may write the solution as 1 - q , - 5/22 or 113, q 2.

We remark that some texts use the set theoretic symbol instead of writing the word or. Thus, in place of 1 - q , - 5/22 or 113, q 2 they would write 1 - q , - 5/22 113, q 2. Also, if an inequality has no solution we indicate this by writing , which represents the empty set. Our next example illustrates that as long as an expression is written in factored form, the procedure for finding the sign of an expression generalizes to any number of factors. Example 4 Solve the inequality 4x41x3 - 4x2 1x2 + 7x + 1221x 2 - 92

0.

Solution We must first factor the given expression. 4x41x 3 - 4x2 1x2 + 7x + 1221x 2 - 92 = 4x 51x + 221x - 22 4x 4x1x 2 - 42 = 1x + 321x + 421x - 321x + 32 1x + 421x + 3221x - 32 4x51x + 221x - 22 1x + 421x + 3221x - 32

The numerator will be zero when x = 0, - 2, or 2, and the denominator will be zero (the expression will not be defined) when x = - 4, - 3 or 3. We indicate these point on the number line in Figure 10. I ND II ND III 0 IV 0 V 0 VI ND VII -4 -3 -2 0 2 3 x

4x51x + 221x - 22 1x + 421x + 32 1x - 32

2

We need seven test points to determine the sign in each region. In determining the sign of the expression, rather than actually evaluating it for each test value, we will instead indicate the sign of each factor for each test value. All we need do is count the total number of negative signs, an odd number means the expression is negative, and even number means it is positive. Be careful when working with exponents. For example, 1 - 25 counts as five

Section 0.5

* **

37

negative signs, 1 - 22 counts as two. Table 6 summarizes the sign of the expression

4x51x + 221x - 22 1x + 421x + 32 1x - 32

2

Table 6: Sign of

Region I II III IV V VI VII

2

Test value

Sign of

2

-5

- 3.5 - 2.5 -1 1 2.5 4

41 - 251 - 21 - 2/1 - 21 - 221 - 2 = 41 - 251 - 21 - 2/1 + 21 - 221 - 2 = + 41 - 251 - 21 - 2/1 + 21 + 221 - 2 = + 41 - 251 + 21 - 2/1 + 21 + 221 - 2 = 41 + 251 + 21 - 2/1 + 21 + 221 - 2 = + 41 + 251 + 21 + 2/1 + 21 + 221 - 2 = 41 + 251 + 21 + 2/1 + 21 + 221 + 2 = +

The information obtained from Table 5 is now placed into our sign diagram in Figure 11. ND + ND -4 -3 0 -2 0 0 + 0 0 2

2

ND 3

+ x

We see that the expression will be zero or positive when - 4 6 x 6 - 3 or -3 6 x - 2 or 0 x 2 or x 7 3. Using the interval notation, we can write it as 1 - 4, - 32 or 1 - 3, - 2] or [0, 2] or 13, q 2.

In each of the examples we considered so far, the expression that needed to be analyzed was easily factorable. Even when this is not the case, we can sometimes perform the analysis needed to determine the sign of a given expression. The next example illustrates. Example 5 Solve the inequality x2 + 2x - 2 6 0. Solution The zeros of the quadratic x 2 + 2x - 2 are found by the quadratic formula to be x1 = - 1 - 23 L - 2.732, and x2 = - 1 + 23 L 0.732. We now proceed in the usual way. We begin our sign diagram in Figure 12. I 0 II 0 III x x1-*2.732 x2 * 0.732 Figure 12: Examining the Sign of x2 + 2x - 2

38

* **

Section 0.5

Region I II III Test value Sign of x 2 + 2x - 2

x = -3 x = 0 x = 1

9 - 6 - 2 = + -2 = 1 + 2 - 2 = +

We now complete the sign diagram as given in Figure 13. + 0 x1-*2.732 0 + x x2 * 0.732 Figure 13: Sign of x2 + 2x - 2 Therefore, we see that x2 + 2x - 2 6 0 when - 1 - 23 6 x 6 - 1 + 23.

Suppose we have a quadratic whose zeros are complex numbers. That means there is no real number at which the quadratic is zero, therefore the sign of the quadratic is always positive or always negative. We need only test any particular x-value to determine the sign of the quadratic. The next example illustrates this. Example 6 Solve the inequality x2 + 9 7 0. Solution Since x2 + 9 has no real zeros (verify!), we need only determine its sign at any convenient x-value. We choose x = 0. We see that for this value of x the quadratic is positive, therefore it is everywhere positive, so we have as our solution the interval - q 6 x 6 q .

Note for the inequality x2 + 9 6 0, the same analysis tell us that there is no solution, or the solution is , the empty set. The calculator cannot directly solve an inequality, it can be used to perform some of the analysis for us, in particular the determination of the sign at the test points. Consider the determination of the sign of the expression x2 + 2x - 2 at the test points x = - 3, 0

Calculator Tips

Section 0.5

* **

39

and 1. We proceed as follows: we go to the Y = screen by pressing * F1, and for y1(x) we enter x 2 + 2x - 2, see Figure 14. We then return to the home screen by pressing the HOME key. The expression x 2 + 2x - 2 examined in Example 5 is now in the calculator s memory and it is named y1(x). To compute the value of the expression at 1, we type in y1(1) and the calculator gives 1, which means in this Region III the expression is positive, similarly y11 - 32 = 1 which means in Region I the expression is also positive, and y1102 = - 2, which means in Region II the expression is negative.

In each of the following exercises, solve the given inequality.

1. 2x - 3 7 0 2. 2x - 3 0 3. 1x - 321x + 42 0 4. 1x - 321x + 42 6 0 5. 13x + 5214x - 72 7. 15x - 321x + 52 0 8. 15x - 321x + 52 6 0 9. 15 - 2x21x + 42 7 0 10. 13 - 4x212 - 3x2 0 11. x12x - 52 0 12. x12x - 52 6 0 13. 1x + 2221x - 32 0 14. 13x - 42 12x - 32 15. x1x - 221x + 32 0 0

4

1x + 221x - 12 x - 3 1x + 221x - 12 7 x - 3 12x - 5213x - 22 1x - 42 1x - 42 x12x - 32 1x - 42 1x - 42 10 x - 3 10 x - 3 7x x + 3 7x x + 3

3 2 2 2

0 0 7 0 0

6. 13x + 5214x - 72 7 0

12x - 5213x - 22

2

0 6 0

x12x - 32

3

6 2 2 4 6 4

16. 1x + 121x - 221x + 52 0 17. 12x + 3213x - 521x + 12 18. x2 + 7x + 12 0 19. x2 - 7x + 10 6 0 20. x + x - 42 7 0 21. 6x2 - x - 12

2 2

3 x x - 1 2 x 3 x - 1 6 2 1x - 121x + 22 x - 3 1x - 121x + 22 x - 3 x 2 - 16 2 x - 25

4

7 1 1

20 11

22. 6x - x - 12 7 0 23. 10x2 + 7x - 12 0 24. 10x + 7x - 12 6 0 25. 24x2 + 10x - 25 0 26. 24x2 + 10x - 25 6 0 27. x - 16x 0 28. x3 - 16x 6 0 29. 25x - x3

3 3 2

7 0

7

4

5 5 6 2 3

0 0 7 0

923

6

1x - 22 12x + 12

0 7

59. x2 - 8

30. 25x - x 7 0 31. x2 - 9 32. x 2 + 4 7 5 33. 12x + 321x - 32 11 34. 12x + 321x - 32 6 11 35. 13x - 2212x + 32 6 28 36. 13x - 2212x + 32 28

40

* **

Section 0.5

Each question is worth two points. Solve the given inequality 1. 10x2 + 11x 6 4. 2 - 7x 7 -1 4 + 5x 5. 2. 24x4 - 42x3 + 18x 2 6 0 12x 3 - 38x 2 + 20x 25 - 9x2 0 3. 2x - 5 8x + 12 0

This chapter examines the basic notions of a function. It begins with the linear function, and then the quadratic functions the parabola. The functional concept needed for the study of calculus are included as well as relevant applications to economics and statistics.

42

* **

Section 1.1

The Line

1.1

The Line

Two Dimensional Coordinate System Horizontal and Vertical Lines The Slope Intercept Form Graphing The Point-Slope Equation The Slope Formula The General Linear Equation An Economic Application Calculator Tips

Each question is worth one point. 1. Determine the equation of the line whose slope is 2 and whose y-intercept is (0,5). For questions 2 4, use the line 5x - 2y = 9. 2. 3. 4. 5. 6. 7. 8. 9. Determine the slope of the line. Determine its y-intercept. Determine its x- intercept. Find the equation of the line with slope 3 passing through the point 12, - 52. Find the equation of the line passing through (5, 2) and 15, - 72. Find the equation of the line passing through (5, 2) and 1 - 3, 22. Find the equation of the line passing through (4, 2) and (6, 7). Find the equation of the line parallel to the line 3x - 2y = 7 and passing through the point (6, 5).

You have learned about lines at several different times in your mathematical education. We give a complete development of the line and slope in this section along with some applications. We believe a detailed review of this topic is useful as some of the basic notions of calculus generalize ideas examined in this section. Once again, we provide a short pretest for those of you who believe you remember the topic well. The box above indicates the major points of discussion of this section. Coordinate geometry is one of the most useful tools in gaining a visual understanding of functions. With coordinate geometry, algebraic formulas may be translated into graphs. In many cases, having the graph is the end of the problem. As you know from everyday experience, a picture may be far more informative than a collection of data. In other cases, the picture may reveal the solution to a problem that might otherwise appear to be too difficult to attack. Although none of this should be new to you, let us review the ideas and thereby fix what will become our standard notation and terminology. We remark that our approach is probably not the one you first learned, but as a second time through the material, it will most quickly obtain the results needed to understand linear functions. We begin with a

Section 1.1

The Line

* **

43

two-dimensional universe consisting of all ordered pairs of real numbers, usually denoted by R2. Examples of points in this universe are (1,4), (0.5 , 6), A 0, 22 B , and 1 - p, 3.12. Because the two real numbers are ordered, (2,7) and (7,2) are different points. When coordinate axes are introduced in a plane, every ordered pair is associated with a point in the plane; and, conversely, every point in the plane has attached to it a unique ordered pair of coordinates. Let us briefly explain how this is done. First, as in Figure 1, a pair of number lines are drawn at right angles to one another, intersecting at the point zero on each line. The horizontal line is called the x-axis and the vertical line is called the y-axis. Construct a vertical line through any point in the plane. At the point where this line crosses the x-axis is a number called the x-coordinate (or abscissa) of the point. Now, construct a horizontal line through the point. At the point where this line crosses the y-axis is a number called the y-coordinate (or ordinate) of the point. If the point is called P, and its coordinates are (x, y), then we may refer to it as P(x, y). In Figure 1, the points P(2, 9), Q1 - 4, - 52, and R1 - 5, 42, S14, - 32 are shown. Notice that every point on the x-axis has ordinate 0 and every point on the y-axis has abscissa 0.

P(2,9)

R(-5,4)

(0,0)

S(4,-3)

Q(-4,-5)

The point where the axes intersect, called the origin, has coordinates (0,0). If we have a relation between two quantities x and y, then we may plot all the points whose coordinates satisfy the rule. The resulting picture is the graph of the relationship. Of course, there may be an infinite number of points, so that actually plotting them all is impossible. On the other hand, in most cases of practical interest, the graph assumes an easily observable pattern which we can visualize without seeing every point. It is also possible to go the other way. That is, we could have a verbal description of the geometric shape that we want as a graph and then try to find the algebraic relationship between the coordinates, that would produce it. For example, the vertical line crossing the x-axis at 3 would consist of all points with x-coordinate, 3. Thus, it would be described by the rule x = 3 and y = anything. Since the last restriction is no restriction at all, we shall simply refer to this graph as the line x = 3. In general, any vertical line will be represented by the equation

44

* **

Section 1.1

The Line

x = a, where a is some constant. If a is positive, then the line lies to the right of the y-axis. If a is negative, the line lies to the left of the y-axis. Of course, x = 0 is the y-axis. See Figure 2. We note that the graphs in Figure 2 cannot be graphs of functions of x, since there are many y values corresponding to the value x = 3. In a similar manner, the equation y = b has as its graph a horizontal straight line which crosses the y-axis at (0, b). If b is positive the line lies above the x-axis; if b is negative the line lies below the x-axis; and, naturally, y = 0 is the x-axis. See Figure 3 where we graph the lines y = 4 and y = - 3. The graph defined by y = b is a very simple one, and this relation is called the constant function. (We shall discuss functions in detail later in this chapter.)

y

x = -1 x=3

y=4

y=-3

Let us now see what kind of function gives rise to any other straight line graph. To begin, we consider a line passing through the origin and making an angle u with the positive x-axis (see Figure 4). Let P(x, y) be any point on the line, other than the origin. For any choice of (x, y), the triangles formed by the given line, the vertical line through P and the x-axis are similar. Therefore, from the figure, it is clear that the ratio y/x is the same for every point P. That is, y/x is a constant. We will denote the constant by m, and call this the slope of the line. That is, the coordinates of every point on this line except the origin must satisfy the condition that y/x = m, which is a constant. Multiplying by x, we obtain y = mx and in this form we can also allow 1x, y2 = 10, 02, so that y = mx

The Line

* **

45

is the equation of a line of slope m passing through the origin. Indeed, every non-vertical line passing through the origin is the graph of y = mx for some value of m. Notice that for m = 0, this equation reduces to just y = 0, which we saw above is an equation for the x-axis.

y = mx

(0, b)

y = mx + b

Figure 5

Now consider the line parallel to the line y = mx that cuts the y-axis at (0,b). If b is positive, then this line lies above the original line y = mx (as drawn in Figure 5) and for each x, the y-coordinate of every point on the new line is just b more than the value y = mx on the first line. Therefore, we conclude that every point on this line satisfies the equation y = mx + b. Of course, if b is negative, the line y = mx + b will lie below the original line, but in all other respects the analysis will be the same. Note that parallel lines have the same slope. Technically, the y-intercept is the point where the line crosses the y-axis and it has coordinates (0, b). In practice, the number b will be referred to as the y-intercept. Thus, we conclude that the equation of any non-vertical straight line may be written in the slope intercept form which follows.

46

* **

Section 1.1

The Line

An equation for any non-vertical line is

y * mx + b

where m and b are constants, m is called the slope of the line, and b is called the y-intercept.

Drawing the graph of a straight line is particularly simple, since we know that two points determine a line. Thus, we need locate only two points on the line and lay a straight edge across them. Let us look at some examples. Graphing Example 1. Plot the graphs of (a) y = 2x - 1 (b) y = - 1*2x + 4 (c) y = - 3 (d) x = 3 Solution. In (a) we have a line of slope 2 and y-intercept - 1. That is, the line crosses the y-axis at 10, - 12. In order to draw its graph we need only one additional point. So, we substitute any convenient value for x into the equation, say, x = 1. When x = 1, y = 2112 - 1 = 1. That is, the line passes through 10, - 12 and (1, 1). We plot these two points and draw the line as shown in Figure 6(a). In (b), the slope is - 1*2 and the y-intercept is 4. This line passes through (0, 4). Picking x = 2 for convenience, we get a second point y = - 1*2122 + 4 = 3. Thus, for a second point we have (2, 3). The graph is shown in Figure 6(b). We see no x term in equation (c). However, we recognize this as being of the form y = constant, which is a horizontal line consisting of all points with y-coordinate - 3. Its graph is Figure 6(c). We note that this line could be thought of as y = 0x + 1 - 32, that is, a line of zero slope, parallel to the x-axis, with y-intercept - 3. Finally, in (d), there is no y term and we recognize this line as a special case that cannot be put into slope intercept form. Its graph is the vertical line shown in Figure 6(d).

y = 2x -1

y = * *+ x + 4

Figure 6(a): y = 2x - 1

Section 1.1

The Line

* **

47

x=3

y = -3

Figure 6(c): y = - 3

Figure 6(d): x = 3

Example 2 Find an equation of a line having slope 3, and which passes through the point 1 - 1, 42. Solution. form Since this is not a vertical line, its equation can be written in slope intercept y = 3x + b, where b is to be determined. Since the coordinates of the point 1 - 1, 42 must satisfy the equation, we have 4 = 31 - 12 + b 4 = -3 + b Thus, b = 7, and the equation of the line is y = 3x + 7

In the many applications, we shall frequently encounter examples in which one knows the slope of a line and one point on the line as in Example 2. Therefore, it will be convenient to have a simple formula into which such data can be substituted to find the equation of the line directly. Let us suppose that we know the slope, m, and one point 1x1, y12 on a line. As in the previous example, we know that the equation must be y = mx + b, where b is to be determined. Substituting the known point 1x1, y12, yields y1 = mx1 + b We solve for b, obtaining y1 - mx1 = b We substitute this for b into y = mx + b, to obtain y = mx + 1y1 - mx12 = mx + y1 - mx1 Now we subtract y1 from both sides of the equation, giving y - y1 = mx - mx1.

48

* **

Section 1.1

The Line

Finally, we factor the m from both terms on the right hand side of the equation, to get y - y1 = m1x - x12. This is the so-called point-slope equation of a straight line.

If you are given the slope of a line m, and know the coordinates of one point 1x1, y12 on the line then you may determine an equation for the line by simply substituting the information into the point-slope equation

y + y1 * m1x + x 12

Of the various forms of the equation of the straight line, the Point-Slope form is the most useful for calculus. When simplifying the expression, we either write the final expression in the form y = mx + b or in the form Ax + By = C where A, B, and C are integers. Example 3 Using the point-slope equation, rework Example 2. Solution. In Example 2, we were asked to find an equation for a line of slope 3, that passes through 1 - 1, 42. The given information exactly suits the point-slope formula that was just derived. Therefore, we substitute directly m = 3, x1 = - 1, y1 = 4 to obtain y - 4 = 31x - 1 - 122 y - 4 = 31x + 12 y - 4 = 3x + 3, and adding 4 to both sides, we obtain y = 3x + 7 the same form as before.

Here is another frequently encountered type of problem. Example 4 (a) Find an equation for the line that passes through (2, 9) and 1 - 5, - 62. (b) Identify its slope and y-intercept. Solution. (a) Although we know more than one point, we cannot use the point-slope formula yet, since we do not know the slope. However, since both points satisfy the same equation, y = mx + b, we substitute each set of coordinates to get two equations 9 = 2m + b - 6 = - 5m + b

Section 1.1

The Line

* **

49

If we subtract the lower equation from the upper, we will eliminate b, thus, 9 - 1 - 62 = 2m - 1 - 5m2 15 = 7m m = 15/7 Now we may use the point-slope equation using either of the given points for 1x1, y12. Let us use the first point (no reason to burden ourselves with extra negative signs): y - 9 = 115/721x - 22 Multiply through by 7 to simplify, yielding 7y - 63 = 151x - 22 7y - 63 = 15x - 30 7y = 15x + 33 (b) Notice that the equation we just obtained looks nice but it is not in the form y = mx + b (or, if you prefer, y = f1x2). This is not unusual. Since we are asked to identify the slope and y-intercept of this line we must divide the equation by 7 to get y = 15 33 x + 7 7

It is not uncommon to encounter cases in which two points are known, and an equation of the line determined by them is needed. What would help is a simple way to find the slope; then you can use the point-slope formula as in the last example. Therefore, let us suppose that we know the coordinates of two points on a non-vertical line. Call them P1x1, y12 and Q1x2, y22. Proceeding as in the preceding example, we realize that both points satisfy the equation y = mx + b So y2 = mx2 + b and y1 = mx1 + b subtracting the lower equation from the upper, we have y2 - y1 = mx2 - mx1 = m1x2 - x12. Since the line is not vertical, no two points on it have the same x-coordinate. Therefore, x1 Z x2 and x2 - x1 Z 0. Thus, dividing by 1x2 - x12, we obtain the slope formula.

Let 1x1, y12 and 1x2, y22 be any two points on a line for which x1 Z x2, then the slope of the line is given by

m =

y2 - y1 x2 - x1

50

* **

Section 1.1

The Line

This result is known as the slope formula. Do we have to worry about the possibility that x2 - x1 = 0? Not really. If the difference is zero, this means that x2 = x1. If the two points are different points and x2 = x1 then our line passes through two points with the same x-coordinate. That is, the line is vertical. The slope of a vertical line is undefined in any event, so we can safely use the above formula, and if the denominator is zero, we will know that the desired line is vertical. Note that it is correct to write m = but m Z y2 - y1 y1 - y2 and m Z x 1 - x2 x2 - x 1 1Why?2 y1 - y2 x 1 - x2

Example 5 Using the slope formula, rework Example 4. Solution. We know the two points, 1x1, y12 = 12, 92 and 1x2, y22 = 1 - 5, - 62. Substituting into the slope formula, m = -6 - 9 - 15 15 = = -5 - 2 -7 7

of course, now that we have m, we may finish the problem exactly as in Example 4.

The slope formula also gives us a good indication of what the slope really represents. Call the two given points, P1x1, y12 and Q1x2, y22 We can always think of Q as lying to the right of P (see Figure 7). That is, always identify x2 and x1, so that x2 7 x1; then x2 - x1 is always positive. Thus, in the slope formula, the denominator is always positive. If Q is higher than P, the numerator is also positive; hence, m is positive. So, if the line is rising as you go from left to right, the slope is positive; otherwise it is negative. A zero slope indicates that the numerator is zero, which means that y2 = y1 and the line is horizontal.

P(x1, y1) Q(x2, y2) P(x1, y1) Q(x2, y2)

The numerator is the change in y as you go from the first point to the second, and the denominator is the change in x. Thus, m is sometimes referred to colloquially as the rise over the run; that is, the vertical change divided by the horizontal change. It now becomes clear that if

Section 1.1

The Line

* **

51

the magnitude of the slope is large, then the change in y is large for a relatively small change in x. That is, steep lines have large slopes. Here, large negative m means m is large. (Remember, m means the absolute value or magnitude of m. So, for example, 6 = 6 and - 6 = 6.) Incidentally, the division between large and small in this context is 1. A line of slope 1 makes a 45* angle with the x-axis ( - 1 means the line makes an angle of 45* but measured from the negative portion of the x-axis.) Figure 8 shows several examples to give you an idea of how steep, lines of different slopes are.

m3 m2 m1

Figure 8: Lines with Slopes m1 6 m2 6 m3 In summary, we have the following cases summarized in Table 1. Table 1: The Relationship Between a Line and its Slope

Slope Property y decreases (gets smaller) as we move from left to right, i.e. the line slopes downhill. y increases (gets larger) as we move from left to right, i.e. the line slopes uphill. line is horizontal (parallel to x-axis). line is vertical.

m 6 0 m 7 0 m = 0

undefined

We shall see, when we study the calculus, how these observations generalize to nonlinear graphs. In short, we have seen equations of a straight line may take any of three forms: 1. x = a, a vertical straight line. 2. y = b, a horizontal straight line. 3. y = mx + b, any other straight line. Of course, the second form of a line is only a special case of Form 3 with m = 0. All this can be summarized in the following:

Theorem:

Every equation of the form Ax + By * C, A, B, and C constants, A and B not both zero, is an equation of a straight line. Accordingly, every such equation is called a linear equation.

52

* **

Section 1.1

The Line

This may be verified as follows. Since A and B cannot both be zero, first consider the case where B = 0, and A is nonzero. The equation is now Ax = C, and we can divide by A yielding x = C/A. That is, x = constant, which is a vertical line. On the other hand, if B Z 0, then we can solve for B, By = - Ax + C Now divide by B, y = - 1A/B2x + 1C/B2 which is of the form y = mx + b. Note that the slope of the line Ax + By = C is m = - A/B.

Example 6 Find the slope and y-intercept of the line whose equation is 3x + 4y = 8. Also find the x-intercept (the point where the line crosses the x-axis) and plot the line. Solution. The equation is 3x + 4y = 8, which we solve for y in order to get it into the usual form: 4y = - 3x + 8 Dividing by 4, 3 y = - x + 2 4 Now, by inspection, we see that the slope is - 3/4 and the y-intercept is (0, 2). To find the x-intercept, we are really asking to find the value of x for which y = 0. In other words, we substitute y = 0 into the given equation and solve for x. 3x + 4102 = 8 3x = 8 x = 8/3 The x-intercept is (8/3, 0). In general, unless a line passes through the origin, (or is horizontal or vertical) the easiest way to draw its graph is to plot the intercepts. Thus, in Figure 9 we show the intercepts and the line.

3x + 4y = 8

Section 1.1

The Line

* **

53

Example 7 Determine an equation for the line parallel to the line 3x - 2y = 8 and passing through the point 11, - 22. Solution. As above, we put the given line in the usual form by solving for y. We find that y = 3/2x - 4. Since the required line is parallel to the given line, its slope must also be 3/2. Thus, the required line has slope m = 3/2, passes through the point 11, - 22, and by the point-slope formula its equation is y - 1 - 22 = 3 2 1x - 12 Simplifying, we find that the slope-intercept equation of the required line is y = 3/2x - 7/2. We can also write an equation for the line in the form 3x - 2y = 7. An Economic Application

The slope is defined as the change in y-values divided by a corresponding change in x-values. Thus, slope is essentially the average rate of change of y with respect to x. It is precisely this interpretation of slope that is essential to our understanding many applications, especially in Economics and Finance. The next example illustrates such an application. Example 8 When a wholesaler sold CD players at $60 per player, weekly sales averaged 150 players. For each $5 drop in the wholesale price the average number of players sold increased by 15. (a) Describe the relationship between the wholesale price and average weekly sales and (b) what is the average weekly sales if the wholesaler charges $42 per unit? Solution. We let x represent the average weekly sales, and y the wholesale price. (a) We first plot some points to see if we notice a pattern. When y = 60, x = 150, that is, (150, 60) is our first point. If we decrease the price, y by 5, then weekly sales, x increases by 15, so we have the point (165, 55), if we drop y again by 5, then x increases by another 15, and we have as our third point, (180, 50). These points are plotted in Figure 9. Note that as we move the sales right 15 units to the right (increase), the price falls 5 dollars. Thus, we go from the point (150, 60) to the point (165, 55). Once again, as we move from the point (165, 55) and move right 15 units, we then fall another 5 units to the point (180, 50) and this trend continues. This ratio of the price to the weekly sales, or equivalently, the average rate of change of price with respect to weekly sales is the slope of the line connecting the points. Thus, we see a linear relationship between the variables y and x, the price and sales. The rate of change of price with respect to sales is 5/15 = 1/3. Since sales are increasing as the price falls, we have a negative rate of change, that is, the slope is - 1/3. We can now write the equation that represents this relationship. We have m = - 1/3 and choosing the point (150, 60) we have y - 60 = - 1/31x - 1502 or simplifying, we have 3y + x = 330

54

* **

Section 1.1

The Line

Since it makes sense to think of the price as driving the demand for the CDs, we solve for x as a function of y. This gives x = - 3y + 330 (b) When the wholesale price y = 42, we have x = - 31422 + 330 = 204. Thus, at a price of $42 per unit, the average sales are 204.

15 5 15 5

Figure 10: Price Plotted versus Average Weekly Sales We have already seen that parallel lines have the same slope, and conversely that two lines with the same slope are parallel. We now investigate the relationship between the slope of perpendicular lines, that is, lines that intersect at an angle of 90 degrees right angles. Let us first dispense with the case where one line is vertical, then any line perpendicular to it must be horizontal. Thus, we assume in what follows that neither line is vertical. We remind you that the formula for the distance d between the points 1x1, y12 and 1x2, y22 is given by d = 21x2 - x122 + 1y2 - y122 Without loss of generality, assume the two lines intersect at the origin. (Or equivalently, think of the origin as being at the intersection of the two lines.) Let m1 and m2 be the slope of these two lines, then their equations are y = m1x and y = m2x (why?). Choose points A and B on each of these lines with x-coordinate 1, then the corresponding y-coordinates are m1 and m2. Consider Figure 11. Let us first assume the lines intersect at right angles at the origin. The triangle shown with sides a, b, and c is therefore a right triangle, and Pythagoras theorem is applicable. By the distance formula, we have a = 211 - 022 + 1m1 - 022 = 21 + m1 2 b = 211 - 022 + 1m2 - 022 = 21 + m2 2 and the vertical line connecting A to B has distance c = m1 - m2

Section 1.1

The Line

* **

55

y = m1x m1 a c b

m2 y = m2x

by Pythagoras theorem a 2 + b2 = c 2 we have 1 + m1 2 + 1 + m2 2 = 1m1 - m222 or 2 + m1 2 + m2 2 = m1 2 - 2m1m2 + m2 2 or 2 = - 2m1m2 or m1m2 = - 1 Thus, we have shown that if two lines are perpendicular, the product of their slopes is - 1, or equivalently one is the negative reciprocal of the other, that is, m2 = - 1/m1 We also observe that each step in the above argument is reversible, that is, beginning with two lines whose slope product m1m2 = - 1, by working backwards, we obtain the first statement which is Pythagoras theorem, implying the triangle is a right triangle, therefore the lines intersect at right angles.

Example 9 Determine the slope of the line perpendicular to the line (a) y = 3x (b) y = - 2/5x + 7 (c) 5x + 2y = 11 (d) x = 2 (e) y = 1. Solution. (a) m1 = 3 therefore the slope of the perpendicular line is its negative reciprocal, m2 = - 1/3. (b) m1 = - 2/5 therefore its negative reciprocal is m2 = 5/2.

56

* **

Section 1.1

The Line

(c) rewriting the line in the form y = - 5/2x + 11/2, we see its slope m1 = - 5/2 therefore the slope of the line perpendicular to it is 2/5. (d) x = 2 is a vertical line, therefore any line perpendicular to it is horizontal. Horizontal lines have 0 slope. (e) The line y = 1 is a horizontal line, any line perpendicular to it is vertical and vertical lines have no slope.

Example 10 Determine the equation of the line passing through the point 12, - 32 and perpendicular to the line 3x - 2y = 12. Solution. We first find the slope of the line 3x - 2y = 12, by rewriting it as y = 3/2x - 6, we see immediately that its slope m1 = 3/2. The slope of the line perpendicular to it is the negative reciprocal so we have m2 = - 2/3. By the point slope formula, the equation of the required line is y - 1 - 32 = - 2/31x - 22 or y = - 2/3x - 5/3 or 2x + 3y = - 5 Calculator Tips The calculator may draw graphs for us in regions called windows. If we do not set the window, the calculator chooses the default window which may or may not be what we desire. We illustrate with lines. We write the equation in the form y = mx + b. You will notice that on top of the F1 button of the calculator, in green is Y = . To get the screen associated with Y = , you press the green diamond and then the F1 key1 F12. You now see the screen indicated in Figure 12. It has on separate lines, y1, y2, y3, and so on. For a single line, we only need y1. Suppose you want to draw the line y = - 2x + 7, then we enter - 2x + 7. Be careful, the negative sign in front of the 2 is input by pressing the 1 - 2 key to the left of the Enter key on the bottom of the calculator. (The calculator distinguishes between the operations subtraction and negation.)

Section 1.1

The Line

* **

57

Figure 13: y1 = - 2x + 7

Stored in the calculator s memory is the equation y = - 2x + 7. The calculator calls it y1, in the event that you store other equations in its memory, it gives each one a separate name. If you now press the graph button 1*F32 it draws the line in the default window, see Figure 14.

Figure 14: y = - 2x + 7

You can see what the default window is by pressing the window key 1* F22, see Figure 15.

58

* **

Section 1.1

The Line

xmin, the lowest x-value is - 10, the highest is xmax which is 10. xscl = 1 means each x tick mark is 1. Similarly, the lowest y-value (ymin) in the window is - 10, the largest (ymax) is 10, each y tick mark is 1. The last item, xres is the resolution, we usually leave this alone. If you change any of the setting indicated above, you change the window. Experiment, change some of them and then press the graph key to redraw the line. Note the zoom button is F2; this button allows you to change the window more quickly by zooming in and out on the graph. Once again, experiment. One note of caution. Horizontal lines all have zero slope, it is sometimes useful to enter the equation of the horizontal line say y = a as y1 = 0x + a. Thus, the line y = 7 would be entered as y1 = 0x + 7.

In Exercises 1 5, determine the equation of the line whose slope and y-intercept are given. 1. 2, (0, 4) 2. - 3, (0, 0) 3. 0, 10, - 22 4. 1*2, (0, 1/4); 5. 1/4, (0, 3) In exercises 6 10, find the slope, x-intercept and y-intercept of the given line. 6. 3x - 4y = 12 7. 2x + 4y - 9 = 0 8. 4y = 5x + 12 9. 1.3x + 4.7y + 11.2 = 0 10. 16 - 4y = 35 In exercises 11 15 give an equation of the line with the given slope and passing through the given point. 11. - 3, (2, 5) 12. 1*2, 1 - 2, 02 13. 1/6, (0, 0) 14. 14, 1 - 1, 0.32 15. 4.1, (3, 0) In exercises 16 20 find equation for the given line. 16. A vertical line passing through 1 - 1, 82. 17. A horizontal line passing through 12, - 62. 18. A line with intercepts (0,4) and 1 - 2, 02. 19. A vertical line passing through (12,2). 20. A horizontal line passing through 1 - 3, 82. In exercises 21 40 plot the lines found in 21. Exercise 1 22. Exercise 2 23. Exercise 3 24. Exercise 4 25. Exercise 5 26. Exercise 6 27. Exercise 7 28. Exercise 8 29. Exercise 9 30. Exercise 10 31. Exercise 11 32. Exercise 12 33. Exercise 13 34. Exercise 14 35. Exercise 15 36. Exercise 16 37. Exercise 17 38. Exercise 18 39. Exercise 19 40. Exercise 20 41. (a) Find the slope of the line whose equation is 2x - 5y = 6. Find the x and y-intercepts of the line. Plot the line. (b) Find the equation of the line parallel to the line given in (a) and passing through 1 - 1, 72. Plot the line on the same set of axes. 42. (a) Find the slope of the line whose equation is 3x + 7y + 42 = 0. Find the x and y-intercepts of the line. Plot the line. (b) Find the equation of the line parallel to the line given in (a) and four units above it. Plot the line on the same set of axes. 43. (a) Find an equation for the horizontal line passing through the y-intercept of the line in 41(a). (b) Find an equation for the vertical line passing through the x-intercept of the line in 41(b). 44. (a) Find an equation for the horizontal line passing through the y-intercept of the line in 42(a). (b) Find an equation for the vertical line passing through the x-intercept of the line in 42(b). 45. Find the equations of two lines parallel to the line y = - 3, and 4 units from it.

Section 1.1

46. Find the equations of two lines parallel to x = 2, and 6 units from it. In Exercises 47 58 find the slope of the line passing through each pair of points 47. 11, - 22 and 11, - 1.42. 48. 12, - 92 and (12, 5). 49. 1 - 1/4, 2/52 and (0, 0). 50. (1, 4) and (2, 4). 51. (12, 16) and 112, - 732. 52. (0, 3) and 1 - 6, 02. 53. A 1*2, - 2 B and 11/4, - 1/42. 54. 10, - 92 and (1/2,3). 55. 1 - 1/3, 2/32 and (0,0). 56. 11, - 52 and 12, - 52. 57. A 1*2, 16 B and A 1*2, - 73 B . 58. (0, 4) and 1 - 7, 02. In Exercise 59 70 find the equations of the lines and plot the lines from 59. Exercise 47. 60. Exercise 48. 61. Exercise 49. 62. Exercise 50. 63. Exercise 51. 64. Exercise 52 65. Exercise 53 66. Exercise 54. 67. Exercise 55. 68. Exercise 56. 69. Exercise 57. 70. Exercise 58. 71. Determine an equation for the line parallel to y = 3x - 7 and passing through the point 11, - 52. 72. Determine an equation for the line (a) parallel (b) perpendicular to 2x - 5y = 9 and passing through the point 1 - 2, - 42. 73. Determine an equation for the line (a) parallel (b) perpendicular to 3x + 7y = 11 and passing through the point 11, - 32. 74. (a) Plot the line 4x + 6y + 12 = 0. Find the area of the triangle formed by the line, the x-axis and the y-axis. (b) Repeat for Ax + By + C = 0 for A, B, C positive.

The Line

* **

59

75. Find an equation of a line whose y-intercept is 4 and such that the area of the triangle formed by the line and the two axes is 20 square units.(Two possible answers.) 76. The area of a triangle formed by a line and the two axes is 40 and the slope of the line is - 5. Find an equation for the line. (Two possible answers.) 77. (a) Find the length of the portion of the line 5x + 12y = 84 that is cut off by the two axes. (b) Repeat for Ax + By + C = 0 for A, B, C positive. 78. (a) Plot the points 1 - 1, - 72, (4, 2) and (8,4). (b) Do they lie on the same line? (c) How can you tell without plotting? 79. In 1990 the Massachusetts Non-Resident State Income Tax calls for a tax of 5% on earned income and 10% on unearned income. Suppose a person has total income of $40,000 of which amount x is earned. Find her tax, t, as a function of x. 80. When the price for a color television is $240, the average monthly sales for this item at a department store is 450. For each $10 increase in price, the average monthly sales fall by 20 units. What is the average monthly sales if the price is $400 per color television? 81. When the price is $50 per radio, a producer will supply 100 radios each month for sale. For each $2 increase in price the producer will supply an additional 6 radios. How many radios are supplied if their per unit price is $72? 82. Plot each of the following lines on the same set of axes. (a) y = 2x (b) y = 21x - 32 (c) y = 21x + 32 (d) How are these lines related? 83. Plot each of the following lines on the same set of axes. (a) y = 2x (b) y - 4 = 2x (c) y + 4 = 2x (d) How are these lines related? 84. Plot each of the following lines on the same set of axes. (a) y = 2x (b) y - 4 = 21x - 32 (c) y + 4 = 21x - 32 (d) y - 4 = 21x + 32 (e) y + 4 = 21x + 32 (f) How are these lines related? 85. In general, how are the lines y = mx + b and y - k = m1x - h2 + b related? (m, b, h, and k are constants.) 86. Show if ab Z 0, then the line with intercepts (a, 0) and (0, b) has the equay x tion + = 1. a b 87. Using the previous exercise, determine the equation of the line with intercepts (a) (3, 0), (0, 6) (b) (2, 0), 10, - 42 (c) A 1*2, 0 B , (0, 2/3). 88. Find the point on the line y = 2x + 3 that is equidistant from the points 1 - 5, 62 and (0, 0). 89. Given the two parallel lines y = mx + b and y = mx + B, determine the perpendicular distance between these two lines. 90. Let A1x1, y12 and B1x2, y22 be in any two points in the plane. (a) Plot these points, (b) Obtain the right triangle formed by drawing a horizontal line from A and a vertical line through B. What are the coordinates of the point at which these two lines intersect? (c) Using the theorem of Pythagoras, derive the distance formula.

60

* **

Section 1.2

Each question is worth one point. 1. Determine the equation of the line whose slope is 3 and whose y-intercept is 10, - 32. For questions 2 4, use the line 2x + 7y = 12. 2. Determine the slope of the line. 3. Determine its y-intercept. 4. Determine its x- intercept. 5. Find the equation of the line with 6. Find the equation of the line slope 3/2 passing through the point passing through 1 - 2, 62 and (3, 6). 1 - 8, 72. 8. Find the equation of the line 7. Find the equation of the line passing passing through (3, 4) and 1 - 8, 52. through (5, 3) and 15, - 82. 9. Find the equation of the line parallel to 2x + 5y = 9 and passing through the point 110, - 32.

Definition of a Function Functional Notation Difference Quotient Domain and Range Independent and Dependent Variables Vertical Line Test Combining Functions Composition Decomposition Functions of Several Variables Calculator Tips

Definition of a Function

The concept of a function comes up regularly in everyday usage; for example, your tax rate is a function of your total earnings, a child s height is usually a function of age, taxicab fare is a function of the distance traveled. Let us look at a few of these examples to understand what we mean by a function. The more you earn, the more you (usually) pay in taxes, two people with the same income (and with all other deductions being equal) should pay the same taxes; taxicab fare as a function of distance means to each trip we associate a fare, trips of the same distance should cost the same, and longer trips should cost more than shorter ones. With these examples in mind, we may now formulate the definition of a function. A function is a set of ordered pairs obtained by some rule so that to each first element in the ordered pair there corresponds a unique second element in the ordered pair. Observe the definition indicates that a function is a set. For example, f = 511, 32, 12, 52, 19, 1126 is an example of a function (not a very interesting one) that we named f (we often use the letters f, g, or h, to represent functions, but any letter may be chosen). f

Section 1.2

* **

61

contains three members, (1,3), (2, 5), (9, 11), instead of saying 3 corresponds to 1, 5 corresponds 2 and 11 corresponds to 9, we prefer to write this correspondence as f112 = 3, f122 = 5 f192 = 11. Notice that each first element (in this example 1, 2, 9) corresponds to a unique second element. We could not have (1,3) and (1,7) belonging to the function since we would have two different second elements (3 and 7) corresponding to the same first element (1). In general, when we write y = f1x2, we mean that y is the value associated with the number x. This is sometimes called functional notation. Often the rule by which we obtain the correspondence between x and y is given by an equation of this form. The equation can be a simple one, or it maybe more complex, as we shall see. For example, we might say a function is defined by the equation f1x2 = 2x + 1. This means if x is any first element in an ordered pair, then the second element is found by multiplication of the first element by 2 then adding 1. For example if x = 5 then the corresponding y-value is 2152 + 1 = 11, or we would write f152 = 11. Note that in this example x could be any real number. That means that the function has an infinite number of elements. Since it is impossible to list the elements of the function, we could instead describe this function by its graph. We know that y = 2x + 1 (remember y is the same as f(x), that is, y * f(x)) is the equation of a line, so each point on the line is an element of f. Thus, we indicate the graph of f in Figure 1.

Function Notation

f(x) = 2x + 1

Figure 1: The Graph of f1x2 = 2x + 1 Note that the caption in Figure1 states the graph of f1x2 = 2x + 1, it would be wrong to say the graph of the function f1x2 = 2x + 1; y = f1x2 is not a function, it is the equation which defines the functional relationship between x and its corresponding y-value. The function is the set of all ordered pairs obtained through the use of this equation. Note that any non-vertical line defines a function. (Why did we say non-vertical?) Example 1 Given the function defined by the equation y = f1x2 = x 2 - 3x + 4. Determine (a) f(0) (b) f(1) (c) f1 - 12 (d) f(2) (e) f(2x).

62

* **

Section 1.2

Solution. We have f1x2 = x2 - 3x + 4, therefore, (a) f102 = 1022 - 3102 + 4 = 4 (b) f112 = 1122 - 3112 + 4 = 2 (c) f1 - 12 = 1 - 122 - 31 - 12 + 4 = 8 (d) f122 = 1222 - 3122 + 4 = 2 (e) f12x2 = 12x22 - 312x2 + 4 = 4x 2 - 6x + 4

Difference Quotient

Note that to determine the y-value at any x-value, we replace x everywhere it appears with the given x-value. Thus, to find f(2x) in the preceding example, we replaced x by 2x wherever x appeared. One of the most important uses of functional notation in the study of the calculus is the evaluation of the difference quotient. It is precisely this quotient that generalizes the concept of slope which is fundamental to the development of the calculus. Given a function defined by the equation y = f1x2, and a non-zero number h, then the difference quotient is defined as

f1x + h2 - f1x2 h

(1)

We remark that for a linear function, that is, a function defined by an equation whose graph is a non-vertical line, say f1x2 = mx + b, the difference quotient will turn out to be precisely the slope of the line. You will verify this remark in the exercises (Exercise 23). We illustrate the computation of the difference quotient in the next example. Example 2 Compute the difference quotient for the function defined by the equation f1x2 = x2 - 3x + 4. Solution. We first compute f1x + h2. This means we replace x by x + h, everywhere it appears in the equation. We have, f1x + h2 = 1x + h22 - 31x + h2 + 4 = x 2 + 2xh + h2 - 3x - 3h + 4 and f1x + h2 - f1x2 = x2 + 2xh + h2 - 3x - 3h + 4 - 1x2 - 3x + 42 = 2xh + h2 - 3h therefore, f1x + h2 - f1x2 h12x + h - 32 2xh + h2 - 3h = = 2x + h - 3 = h h h

We momentarily digress and remind you about the rationalization of denominators and numerators that you learned about in algebra. The basic idea was to eliminate the square roots appearing in a binomial expression. The trick was to multiply the binomial to be rationalized by its conjugate expression. Recall, given the expression a 2b + c 2 d, its conjugate is a 2b - c 2 d. The product of two conjugates

A a 2b - c 2 d B A a 2b + c 2 d B = a2b - c2d

contains no radicals.

Section 1.2

* **

63

Consider the expression (for h Z 0) h 2 - 24 + h and let us rationalize the denominator. We have h 2 - 24 + h h A 2 + 24 + h B 4 - 14 + h2 = h

A 2 + 24 + h B

A 2 - 24 + h B A 2 + 24 + h B

h A 2 + 24 + h B h

= -

= - 2 - 24 + h

This last expression is algebraically equivalent to the original expression when h is not zero. Note the cancellation of h in both the numerator and denominator of the last expression. Sometimes, in problems similar to the last one, we have to rationalize the numerator in order to perform the necessary evaluation, as the next example illustrates Example 3 29 + h - 3 Rationalize the numerator of the expression for h Z 0. h Solution.

A 29 + h - 3 B A 29 + h + 3 B 29 + h - 3 9 + h - 9 = = h h A 29 + h + 3 B h A 29 + h + 3 B

= h h A 29 + h + 3 B = 1

A 29 + h + 3 B

We shall see, when we study the calculus, that after performing rationalizations like these, we would like to determine what the rationalized expression equals if we allow h to be 0. Thus, in this last example, if we evaluate the expression with the numerator rationalized with h = 0, we obtain the result 1/6. You will note that evaluation before the rationalization produces 0/0 an expression which is called an indeterminate form. However, after rationalization, h cancelled in both the numerator and denominator. Example 4 Determine the difference quotient for f1x2 = 22x + 1, the result should not contain any radicals in the numerator. After the rationalization process is completed, determine the value of the expression for h = 0. Solution. f1x + h2 = 221x + h2 + 1 = 22x + 2h + 1 subtracting f(x) from this expression yields f1x + h2 - f1x2 = 22x + 2h + 1 - 22x + 1 therefore, the difference quotient is

64

* **

Section 1.2

f1x + h2 - f1x2 22x + 2h + 1 - 22x + 1 = h h We rationalize the numerator by multiplying numerator and denominator by the conjugate of the numerator. f1x + h2 - f1x2 22x + 2h + 1 - 22x + 1 # 22x + 2h + 1 + 22x + 1 = B R B R= h h 22x + 2h + 1 + 22x + 1 12x + 2h + 12 - 12x + 12 h A 22x + 2h + 1 + 22x + 1 B = 2h h A 22x + 2h + 1 + 22x + 1 B 2 =

A 22x + 2h + 1 + 22x + 1 B

Notice the cancellation of the h in the numerator and denominator. If we now set h = 0, we have 2 22x + 1 + 22x + 1 = 2 2 22x + 1 = 1 22x + 1

Some texts use the symbol x in place of h. In those texts, the difference quotient is written as Domain and Range f1x + x2 - f1x2 x In does not matter which symbol is used, its interpretation is the same. When using a calculator to compute the difference quotient, the form using h is easier to enter. The set of all first elements in the ordered pairs contained within a function is called the domain of the function, and the set of all second elements is called the range of the function. In the example in which f = 511, 32, 12, 52, 19, 1126, this function has three first elements 1, 2, and 9 and three second elements 3, 5, 11. Thus this function has as its domain D = 51, 2, 96 and its range R = 53, 5, 116. When the correspondence which determines the function is given by an equation, the domain is determined by examination of the permissible values of the (independent) variable. We also assume that only real numbers are allowed in the domain. The range is often obtained from the sketch of the graph. We illustrate in the next few examples. Example 5 Determine the domain and range of the function defined by the equation y = f1x2 = - 2x + 5 if 1 x 4. Solution. Note that the domain is specified in this example, namely the set of all x such that 1 x 4. Therefore, we need only determine the range. We sketch the graph of this linear function in Figure 2. Note that the graph of this function is a line segment connecting the two endpoints (1, 3) and 14, - 32. y can assume all values between - 3 and 3, therefore the range is the set of all y-values with - 3 y 3.

Section 1.2

* **

65

(4, -3)

Example 6 Determine the domain and range of the function defined by the equation y = f1x2 = 22x - 3. Solution. To determine the range recall that we must deal only with real numbers. Therefore, we must require that 2x - 3 0 (if it were not, then the radicand would be negative resulting in imaginary values for y). Thus, solving the inequality 2x - 3 0 for x, we obtain that the domain is x 3/2. Sometimes this is written in interval notation as [3/2, q 2. To determine the range we will sketch a graph of the function. We do this by plotting some points as indicated in Table 1. Of course, we begin with x = 3/2. Table 1: Points to Plot the Graph of y = f1x2 = 22x - 3

x 3/2 2 3 4 5 6

y = f1x2 = 22x - 3

0 1

3

Using these points, we draw the graph in Figure 3. Note that the lowest y-value on the graph is at the point (3/2, 0) and the graph rises and will assume all positive y-values. Thus, we have as the range, y 0 or in interval notation [0, q 2.

The next example illustrates another way the domain may be restricted.

66

* **

Section 1.2

Figure 3: The Graph of the Function Defined by y = f1x2 = 22x - 3 Example 7 Determine the domain of the function defined by the equation f1x2 =

x . x 2 - 7x + 12 Solution. All x-values are allowable with the exception of those which make the denominator zero. The denominator will be zero when x 2 - 7x + 12 = 0, or 1x - 321x - 42 = 0, or when x = 3 or 4. Thus, the domain consists of all x-values except 3 or 4.

Note that we did not find the range of the function defined in the previous example. This function is an example of a rational function and will be considered in detail later in this chapter. At that time, you will learn how to easily sketch its graph and (sometimes) obtain its range from examination of its graph. Usually, we call the elements in the domain the independent variables and the elements in the range the dependent variables. In the illustrations above, the length of the ride determined the taxicab fare, the fare (the dependent variable) is dependent on the ride s length (the independent variable); the amount paid in taxes (the dependent variable) depends upon the income earned (the independent variable). The terminology certainly makes sense. Often we use the letter x to indicate the independent variable and f(x) or y to indicate the dependent variable, but any symbol may be used. For example velocity as a function of time is often written as v = f1t2, where v represents velocity and t time. Sometimes, it is convenient to visualize a functions as a machine as indicated in Figure 4. This machine accepts an input, x and produces an output, y.

x+

+ y = f(x)

Section 1.2

* **

67

The machine performs an operation on the input x producing the output y. (This gives the correspondence between x and y.) If a machine is reliable, you would expect that its output should be replicated by using the same input (independent variable), that is, to each input, there corresponds a unique output (the dependent variable). The set of all possible inputs accepted by the machine is its domain, and the set of all outputs produced by the machine is its range. Note in Figure 1 that if we were to draw a vertical line anywhere on the graph it would intersect the graph (which represents the function) exactly once. This is true for all graphs which represent functions. Remember, to each x-value (first element) there corresponds and unique y-value (second element). A graph represents a function if and only if any vertical line intersects the graph in at most one point. Consider the graph in Figure 5. Note that a vertical line passing through this graph will intersect the graph in more than one point, as shown, when x = 10, there are two different y values, - 1 and 5 (violating the condition that each x corresponds to a unique y), therefore this graph cannot represent a function.

x =10 (10, 5)

(10, -1)

Figure 5: A Graph That Does Not Represent a Function Example 8 Consider the graphs in Figure 6, which of these represent a function?

Figure 6a

Figure 6b

Figure 6c

68

* **

Section 1.2

Solution. In Figure 6a or Figure 6b, no matter where you draw a vertical line, it intersects the graph in exactly one point, therefore, each of these graphs represents a function. If we draw a vertical line through any point (except x = 2) in Figure 6c, we see in Figure 7 that it intersects the graph it two different points, this violates the vertical line test and therefore, this cannot be the graph of a function.

Often functions arise in many different formats. The next example illustrates one which you may have some familiarity with. Figure 7: A Vertical Line Intersecting a Graph at Two Points Example 9 The fare for riding in a cab in a small town is as follows: $1.50 for the first mile or part and $2.50 for each additional mile or part traveled. (a) Determine an equation that describes this function, (b) What is the fare for a ride that is 5 miles long? (c) 8.3 miles long? Solution. (a)Let d represent the distance traveled, rounded up to the next integer when d is not an integer (so 1.4 miles is round up to 2 miles). The cost C as a function of distance may be determined as follows: C1d2 = 1.50 if d 1

if d 7 1 the cost will be $1.50 plus $2.50 times each additional mile traveled; d - 1 gives the number of additional miles traveled after the first, so we have C1d2 = 1.50 + 2.501d - 12 = 2.50d - 1 if d 7 1. Note that we have two different formulas for the cost, depending on whether the distance is less or more than one mile. We may combine these two pieces as follows: c1d2 = e 1.50 2.50d - 1 if if 0 6 d 1 d 7 1

A function written like this is sometimes called a piecewise function since its definition depends upon which piece you are on. (b) To determine C(5), we use the description C1d2 = 2.50d - 1, (since d = 5 which is greater than 1) which gives C152 = 2.50152 - 1 = $11.50. (c) Since we round up, C18.32 = C192 = 2.5192 - 1 = $21.50.

Combining Functions

Functions defined over a common domain may be added, subtracted, multiplied and even divided so long as we do not divide by zero. These operations are easily performed, as illustrated in the next example. Example 10 Let f1x2 = 2x + 3 and g1x2 = x2 + 4x + 4 (b) f1x2 - g1x2 (c) f(x)g(x) (d) f(x)/g(x) determine (a) f1x2 + g1x2

Solution. Since the two functions have a common domain (what is it?). These operations are defined. (a) f1x2 + g1x2 = 12x + 32 + 1x2 + 4x + 42 = x 2 + 6x + 7. (b) f1x2 - g1x2 = 12x + 32 - 1x2 + 4x + 42 = - x 2 - 2x - 1. (c) f1x2g1x2 = 12x + 321x2 + 4x + 42 = 2x 3 + 11x 2 + 20x + 12. (d) f1x2/g1x2 = 12x + 32/1x 2 + 4x + 42, for x Z - 2.

Section 1.2

* **

69

Functions have one additional way of being combined, called the composition of functions. Some care needs to be taken with regard to the domain. Suppose we are given two functions defined by the equations y = f1x2 and y = g1x2. We wish to consider under what conditions can we define the new composite functions y = f1g1x22 and y = g1f1x22 When these functions are defined, the first is the composition of f with g and the second is the composition of g with f. We examine the conditions under which we may define y = f1g1x22, the conditions for y = g1f1x22 are similar and will be left as an exercise. First, x must be in the domain of g otherwise, we could not obtain g(x). Second we are using the value g(x) to compute f(g(x)), therefore the number g(x) must be in the domain of f. Thus, the domain of the composite function defined by the equation y = f1g1x22 is the set of all x such that x is in the domain of g and g(x) is in the domain of f. For any x-value in this domain, the composite function determined by the equation y = f1g1x22 is defined. We illustrate this pictorially in Figure 8. Note, for the g machine to operate, x must be in its domain, and for the f machine to operate, g(x) must be in the domain of f.

Composition

x+

g(x)+

f (g(x))+

Example 11 Let f1x2 = 2x + 3 and g1x2 = x 2 + 1, determine (a) f(g(3)), (b) g(f(3)), (c) f(g(x)) and (d) g(f(x)) Solution. (We note that the domain of each function - q 6 x 6 q , so there are no concerns about the domain.) (a) g132 = 32 + 1 = 10, therefore f1g1322 = f1102 = 21102 + 3 = 23. (b) f132 = 2132 + 3 = 9, therefore g1f1322 = g192 = 92 + 1 = 82. (c) f1g1x22 = 21g1x22 + 3 = 21x2 + 12 + 3 = 2x 2 + 5. (d) g1f1x22 = 1f1x222 + 1 = 12x + 322 + 1 = 4x 2 + 12x + 10.

(Verify that setting x = 3 in (c) and (d) yields the same results found in (a) and (b).) The next example illustrates our concern about the domain of the composite function.

70

* **

Section 1.2

Example 12 Let f1x2 = 1/x and g1x2 = Solution. x + 1 , determine the domain of f(g(x)). x + 2

however when does this make sense? First, we observe that x Z - 2 as it is the domain of g (that is, g1 - 22 is not defined). Also note that g(x) must not be 0 as 0 is not in the domain of f. To determine when g1x2 = 0, we need only solve the equation x + 1 = 0 x + 2 whose solution is x = - 1 (verify!). Thus, the domain of the composite function f1g1x22 = 1x + 22/1x + 12 is all x except - 2 and - 1.

Decomposition

Example 13 Consider the function defined by the equation y = 12x 2 - 3210, find f(x) and g(x) so that f1g1x22 = 12x2 - 3210. Solution. Suppose we let u = g1x2, then we may write f1g1x22 = f1u2. There are many possible choices for the two functions, we will make the choice that is most useful when such problems arise in the study of the calculus. We choose u to be the inner function, that is, if we view f1 2 = 1 210, then u is to be the expression inside the parenthesis, thus, we choose u = g1x2 = 2x 2 - 3. A simple calculation yields f1g1x22 = 12x2 - 3210. Note that another possibility is to choose f1x2 = x 2, and u = g1x2 = 12x2 - 325, verify this is correct.

Breaking apart a composite function, as in the previous example, is called decomposition. Note that some texts write 1f g21x2 to mean f(g(x)). We prefer the latter notation. Productivity may depend on the amount of available capital as well as the number of units of labor available, that is, productivity is a function of the two variables, capital and labor. The cost of an automobile may depend on the labor, cost of steel, fiberglass and rubber, that is, the cost is a function of four variables. It follows that in realistic applications, we may often need to deal with functions of more than a single variable. The definition of function extends in a very natural way. Consider first the case of a function of two variables. To each pair, (x, y), we associate a unique value z, written z = f1x, y2. The definition extends is a natural way to any number of variables. Computations with functions of two or more variables is handled the same way as with a function of a single variable. You substitute for each variable as it appears in the equation, as the next two examples illustrates. Example 14 1 3 Suppose the productivity z is given by the equation z = f1x, y2 = 100x4y4, where x is the number of units of capital available, and y is the number of units of labor available. Determine f(81, 16).

Section 1.2

* **

71

Solution.

3

Example 15 Given the function defined by the equation w = f1x, y, z2 = xy - xz + 3yz, determine f1 - 1, 3, 22. Solution. f1 - 1, 3, 22 = 1 - 12132 - 1 - 12122 + 3132122 = 17.

The idea of the difference quotient also can be generalized, as the next example illustrates. Example 16 Given z = f1x, y2 = x2 + 2xy + y 2, assuming neither h nor k is zero, determine f1x + h, y2 - f1x, y2 f1x, y + k2 - f1x, y2 , (b) . (a) h k Solution. (a) f1x + h, y2 = 1x + h22 + 21x + h2y + y 2 = x 2 + 2xh + h2 + 2xy + 2hy + y 2 f1x, y2 = x2 + 2xy + y 2 f1x + h, y2 - f1x, y2 = h 1x2 + 2xh + h2 + 2xy + 2hy + y 22 - 1x 2 + 2xy + y 22 = h h12x + h + 2y2 2xh + h2 + 2hy = = 2x + 2y + h h h (b) f1x, y + k2 = x 2 + 2x1y + k2 + 1y + k22 = x 2 + 2xy + 2yk + y 2 + 2yk + k2 f1x, y2 = x2 + 2xy + y 2 f1x, y + k2 - f1x, y2 1x 2 + 2xy + 2xk + y 2 + 2yk + k22 - 1x 2 + 2xy + y 22 = k k 2 k 1 2 x + 2 y + k 2 2xk + 2yk + k = = 2x + 2y + k k k

We remark that a function of two variables defines a three dimensional graph. Suppose you want to draw the graph of the function whose equation is y = f1x2 = 22x - 3. Of course, we realize the domain is x 3/2. We proceed as follows: We go to the Y = window and define y1. However, we also want to include the domain. There is a key that allow us to insert the domain, the vertical line key to the left of the 7 key. We define the function in Figure 9. (To get the symbol press . (Period key) similarly for the , 0.)

Calculator Tips

72

* **

Section 1.2

We can improve this graph by changing the window, see Figure 11.

Section 1.2

* **

73

Figure 12: Redrawn Graph of y1 = 22x - 3 It should be noted that calculations are easily done on the calculator with functions. As y1 = 22x - 3, if your are in the HOME screen (press the HOME button to get to this screen) and want to evaluate this function when x = 3, you need only enter y1(3) and press Enter. The calculator gives the answer as 23. If you want a numerical approximaion, you press * Enter to obtain 1.73205. (You can determine how many decimal places will be shown using the MODE button and then modify Display Digits). What happens if you try to compute y1(1)? Why? The composition of two (or more) functions is easy to do as well, just enter one as y1(x) and the other as y2(x) in the Y = screen 1*F12. Consider the functions defined in Example 12, we enter these as illustrated in Figure 13.

Figure 13 Returning to the HOME screen will allow us to do computations with these two functions. Compute, for example, their product, y11x2 # y21x2, their compositions y1(y2(x)) and y2(y1(x)). What happens if you try to compute y11y21 - 122? Piecewise functions can be drawn several ways. Perhaps the easiest way is to define each piece separately. Consider the following example. y = f1x2 = c 2x - 3 -1 2 x - 17 if if if x 1 -2 1 6 x 6 4 x 4

74

* **

Section 1.2

Figure 14: Piecewise Defined Functions Note that the space bar key is located above the 1 - 2 key on the bottom of the calculator. The Alpha key is needed to type in the and. We next press the graph key and obtain Figure 15.

Figure 15: Graph of the Piecewise Function The graph looks like it has a jump at x = 4, it does not, but that is how it is displayed by the calculator. The calculator cannot always accurately portray the continuity of the graph, because of the way it displays the graph by means of pixels being on or off. We shall examine the concept of continuity more fully later in the next chapter. We remark that the calculator can easily perform most functional notation operations including the computation of the difference quotient, see below. Computations with functions of two variables can also be performed on your calculator. You must first press the MODE key and change FUNCTION to 3D (three dimensional) by scrolling down. Once this is done then the Y = screen becomes a Z = screen, see Figure 16.

Figure 16

Section 1.2

* **

75

Suppose z = f1x, y2 = 100x 1/4y 3/4, and we want to evaluate f(81, 16), we illustrate in Figure 17, where we use z1 in place of z.

Figure 17 We next return to the home screen and enter z1(81, 16), see Figure 18.

Figure 18 Remember to press MODE and change GRAPH back to FUNCTION when you are done. The calculator computes the difference quotient using the expression avgRC, found in the Catalog. This is illustrated at the end of Section 3.7.

In exercises 1 8 determine if the given set is a function. 1. f = 511, 32, 12, 52, 112, 232, 1 - 1, 926 2. g = 512, - 52, 13, 42, 17, 92, 15, 1126 3. h = 513, - 52, 14, - 52, 17, 112, 111, 1926 4. r = 512, 72, 15, 32, 1 - 3, 72, 119, 526 5. s = 513, 52, 17, 22, 13, - 52, 19, 1126 6. T = 5112, - 12, 13, - 12, 15, - 12, 16, - 126 7. V = 514, 22, 15, 32, 14, 72, 19, - 2226 8. W = 511, 52, 12, 52, 13, 52, 14, 526 In exercise 9 16 determine the domain and range of the set given in 9. Exercise 1 10. Exercise 2 11. Exercise 3 12. Exercise 4 13. Exercise 5 14. Exercise 6 15. Exercise 7 16. Exercise 8 17. Given the function defined by the equation f1x2 = 3x - 5, determine (a) f(0), (b) f1 - 12, (c) f(2) (d) f(2x), (e) f1x + h2 18. Given the function defined by the equation f1x2 = 2x 2 - 1 determine (a) f(0), (b) f1 - 22, (c) f(1), (d) f(3x), (e) f1x + h2 x + 1 19. Given the function defined by the equation f1x2 = determine 2x - 3 (a) f(0), (b) f1 - 12, (c) f(3), (d) f(2x), (e) f1x + h2 20. Given the function defined by the equation g1x2 = 2x 2x + 1 determine (a) g(0), (b) g1 - 12, (c) g(3), (d) g(8), (e) g(2x)

76

* **

Section 1.2

f1x + h2 - f1x2 47.

, h Z 0. h Whenever possible, simplify the expression so that the resulting expression is defined when h = 0. 21. f1x2 = 3x - 2 22. f1x2 = 5x + 7 23. f1x2 = mx + b 24. f1x2 = x2 - 5x + 3 25. f1x2 = x2 + 3x + 9 26. f1x2 = 2x2 + 3x - 7 27. f1x2 = 5x2 - 4x + 11 28. f1x2 = x3 29. f1x2 = 2/x 30. f1x2 = x + 1 x - 3

48.

31. f1x2 = 2x + 3 Hint: refer to Example 19 Section 5. 32. f1x2 = 22x - 1 Hint: refer to Example 19 Section 5. In exercises 33 41 (a) determine the domain, (b) sketch the graph and (c) determine the range of the function defined by the given equation. 33. f1x2 = 3x - 5 34. f1x2 = 4x + 1 35. g1x2 = x + 5 36. g1x2 = - 2x2 + 4 37. h1x2 = 1x 38. r1x2 = 22x + 1 39. w1x2 = 23x - 4 40. f1x2 = x 41. g1x2 = 2x - 3 In exercises 42 45, determine the domain of the function defined by the given equation. 42. f1x2 = 43. g1x2 = 44. h1x2 = 45. r1x2 = 3x + 2 2x - 3 x + 1 2x + 5 5x + 3 6x2 - x - 12 2x + 1 50.

2

49.

x3 - 9x In exercises 46 54 use the vertical line test to determine if the given graph may represent a function. 46. 51.

Section 1.2

52.

* **

77

In Exercises 63 69 determine (a) f(g(x)) and the domain of the composite function, (b) g(f(x)) and the domain of the composite function. 63. f1x2 = 2x + 3 64. f1x2 = x + 3 x + 2 65. f1x2 = x - 3

2

53.

g1x2 = x2 - 1 g1x2 =

70. Determine the domain of the composite function defined by y = g1f1x22. 71. Let f1x2 = ax + b and g1x2 = cx + d determine the conditions on a, b, c and d if f1g1x22 = g1f1x22. 72. The absolute value function is defined as follows:

y = x = e 54. Sketch the graph of y = x . 73. Sketch the graph of y = x - 2 . 74. Sketch the graph of y = x + 2 . 75. Sketch the graph of y = 2x - 3 . The Floor and CEILING functions are defined as follows: -x x x 6 0 x 0

55. The fare for riding a cab is as follows: $1.75 for the first mile or part, $0.50 for each additional mile or part. (a) Determine C(d) an equation representing the cost C of the ride in terms of the distance d traveled. (b) How much does it cost for a 10 mile ride? 56. The Elite Limo company charges by the hour for use of its cars. The charge for the first hour or part is $80. The charge for each additional hour or part up to 6 hours is $55, and for each hour or part above 6 hours the hourly charge is $48. (a) Determine C(h) an equation represent the cost C of the ride in terms of the hours h. What is the charge for (b) 4 hours, (c) 7 hours. 57. Height is a function of age. Suppose at birth the average male is 16 inches tall and grows 3 inches per year until age 12. From age 12 to age 19 growth is 2 inches per year, assume there is no growth after age 19. (a) Determine an equation which give the height h as a function of age a. What is the height at age (b) 9? (c) 15? In exercises 58 62, determine (a) f1x2 + g1x2, (b) f1x2 - g1x2, (c) f(x)g(x), (d) f(x)/g(x) when defined. 58. f1x2 = 3x - 2 59. f1x2 = 2x - 3 60. f1x2 = 3x + 7 61. f1x2 = x2 + 5 62. f1x2 = 3 x + 2 g1x2 = 4x + 7 g1x2 = x 2 g1x2 = 2x - 3 x + 1 g1x2 = 2 4x + 2 g1x2 = 2 x - 4

2

y = flr1x2 = e y = ceil1x2 = e

Use these functions to solve exercises 76 81. 76. (a) flr(5), (b) flr1 - 52, (c) flr(5.1), (d) flr1 - 5.12 77. (a) ceil(5), (b) ceil1 - 52, (c) ceil(5.1), (d) ceil1 - 5.12 78. Sketch the graph of y = flr1x2 for - 3 79. Sketch the graph of y = ceil1x2 for - 3 x x 3. 3. x x 3. 3.

80. Determine a relationship between flr(x) and ceil(x). 81. Sketch the graph of y = flr1x2 + x for - 3 82. Sketch the graph of y = ceil1x2 + x for - 3

In exercises 83 86 determine two simple functions who composition is f(g(x)). 83. f1g1x22 = 13x2 - 2x + 23218 84. f1g1x22 = 12x8 - 11x5 - 2x + 19254 85. f1g1x22 = 2x2 - 3 86. f1g1x22 = 2 3 x 4 - 7x2 + 3x - 2 87. Show how to build, by composition, the following function f1x2 = 4x + 3x + 2x + 1x.

78

* **

Section 1.3

neither h nor k is zero, determine (c) (d) (f) f1x, y + k2 - f1x, y2 k f1x, 1 + k2 - f1x, 12 , (e) f1x + h, y2 - f1x, y2 ,

88. A power station is on one side of a straight river which is five miles wide, and a factory is on the other side, 20 miles up-river. A power line is to be run from the power station, under the river to some point A, which is x miles upriver from P, and then over land to the factory (see Fig. 27). If it costs $80 per mile to run the line under water and $50 per mile to run it overland, estimate x, the point on the other side of the river where the power line comes out of the river, if the total cost of the power line is to be as small as possible (a sketch of the grap or a table may be useful)

h f12 + h, y2 - f12, y2 h

. (g) In (c) and (e) what happens if you allow h to k equal 0 at the end of your calculations? (h) Same question for (d) and (f) if you allow k to equal 0 at the end of your calculations.

92. Given f1x, y2 = 8x3y2 determine (a) f (3, 2), (b) f(2, 5). If neither h nor k if f1x, y + k2 - f1x, y2 , (d) , h k f12 + h, y2 - f12, y2 f1x, 3 + k2 - f1x, 32 (e) , (f) . (g) In (c) and h k (e) what happens if you allow h to equal 0 at the end of your calculations? (h) Same question for (d) and (f) is you allow k to equal 0 at the end of your calculations. zero, determine (c) f1x + h, y2 - f1x, y2 93. Given f1x, y, z2 = x 2 - 2xy 2 - 3y 3z2 + z2, determine (a) f11, - 2, 32 (b) f10, 1, - 22 if neither h, k, l is zero, determine

(This problem and the next will be re-examined when we learn more aboutcalculus.) 89. Suppose the power station in the previous exercise is moved one mile inland find the point A at which the power line enters the river. 90. Referring to the previous two problems, discuss what effect (if any) the overland and under water costs have on the determination of x. 91. Given f1x, y2 = x2 - 2xy2 - 3y3, determine (a) f12, - 12, (b) f1 - 1, 22, if

(c) (e)

. (d)

f1x, y + k, z2 - f1x, y, z2 k

Break-Even Analysis Depreciation Piecewise Linear Graphs Calculator Tips

In the previous section we defined and examined the notion of a function, in this section we consider linear functions in some detail. Many real processes are modeled by linear functions. We will look at some examples from business. To start with, the simplest description of production costs is a linear one. Suppose that a manufacturer has fixed overhead costs of D dollars per month. Suppose further that the cost of producing each of a particular item is m dollars. If the assembly line produces x items per month, then the total monthly cost, C(x), is given by the rule Cost = Overhead + 1Cost per item2 # 1Number of items2. In symbols, this is C1x2 = D + mx. Thus, C(x) is a linear function of x. The slope of the line y = C1x2 is the cost per item m (also known as the marginal cost), and the y-intercept is the overhead, D.

Section 1.3

* **

79

Now, suppose that the manufacturer sells each item produced for p dollars per item. The total revenue to the producer, R(x), is given by Revenue = 1Price per item2 # 1Number of items2. That is, R1x2 = px Thus, the revenue, R(x), is also a linear function of x. In this case, the slope is the selling price per item p, and the y-intercept is zero. ( You sell nothing, you get nothing. ) The selling price per item p should be higher than the cost of producing it. In such a case, the revenue line would have a higher slope than the cost line but starts at the origin for zero sales. As x increases, revenue will eventually match and finally exceed cost. The point at which revenue equals cost is called the break-even point. Its x-coordinate is the number of sales for which revenue would equal cost. See Figure 1(a). Alternatively, you could simply consider the profit, which is revenue minus cost. We shall use P(x) to denote profit (note upper case P represents profit, lower case p represents price), and write P1x2 = R1x2 - C1x2 = px - 1D + mx2 = px - D - mx = 1p - m2x - D. The profit function is also linear. This time, the slope is p - m and the y-intercept is - D. Notice that the slope, p - m should be positive; otherwise, the manufacturer is selling things for less than it costs to make them and will surely never make a profit! The x-value for which P1x2 = 0, that is, the x-intercept of the line, D/1p - m2, is the break-even point. To the left of this value, P(x) is negative (i.e., a loss), and to the right, P(x) is positive (i.e., a profit). See Figure 1(b).

Break-Even Analysis

y = R or C R(x) = px

y=P

break even point

x

break even point

Figure 1(a): Linear Revenue and Cost Here is a simple numerical example.

Example 1. BBZ Wineries has a fixed monthly overhead of $2000 and a cost of $20 per bottle to make and bottle wine. The wine sells for $30 per bottle. (a) Find the cost function, revenue function, and break-even point. (b) Find the profit function. (c) How large must sales be in order to achieve a profit of $1500 per month?

80

* **

Section 1.3

Solution. (a)Let x = the number of bottles produced per month. The cost equation is C1x2 = 2000 + 20x The revenue equation is R1x2 = 30x (In the general notation, D = 2000, m = 20, and p = 30.) The two lines intersect when R1x2 = C1x2, that is, when 30x = 2000 + 20x 10x = 2000 x = 200. Thus, when sales are 200 bottles per month, the winery breaks even (See Figure 2(a)). That is, R12002 = C12002 = $6000.

Figure 2(a): Linear Revenue and Cost (b) The profit function is

P1x2 = 30x - 12000 + 20x2 = 30x - 2000 - 20x = 10x - 2000 (c) (The break-even point can be determined (again) by setting P(x) equal to zero. This gives x = 2000/10 = 200, as above.) We want P1x2 = 1500, so we solve 1500 = 10x - 2000 3500 = 10x x = 350 That is, on sales of 350 bottles, the winery shows a monthly profit of $1500. The graph of this situation is shown in Figure 2(b).

Depreciation

Another common business situation where a linear function arises is in modeling the process of depreciation of assets. Suppose that a company or individual invests in a piece of equipment, building, or other tangible asset that does not have an unlimited life expectancy. Let the value of that item when new be V. It is not reasonable to carry that asset on your books at value V year after year, only to have it drop to value zero in the year in

Section 1.3

* **

81

which it finally wears out. Instead, you normally write off the value of the item by some systematic plan over the item s expected lifetime, T. There are several ways this can be done, of which only a few are acceptable to the business community and the Internal Revenue Service. The simplest way is to assume that the asset depreciates by a fixed amount each year, called linear or straight-line depreciation. For example, suppose that a taxicab driver values his cab at $40,000 when new. Suppose also that he anticipates a usable life of 4 years, at which time he will junk the cab. In other words, the value of the cab will drop to zero in T = 4 years. If the cab is assumed to depreciate a constant fixed amount each year and drop to zero after 4 years, it must lose 1/4 of its original value each year. That is, it must depreciate $10,000 (that is 1 4 of $40,000) each year. Letting t be time measured in years, and A(t) be the value of the cab at time t, we have Table 1. Table 1: Depreciation of a $40,000 Taxi over Four Years

t

A(t) 0 40,000 1 30,000 2 20,000 3 10,000 4 0

It is not difficult to see that these points (t, A(t)) lie along the straight line with slope - 10000 and y-intercept 40,000. Notice, we are using linear depreciation. Therefore, the rate at which the taxicab is depreciating to zero is the total change in value divided by the length of time, or - 40,000/4. Thus, we may interpret the slope as the rate of depreciation. Notice also that the slope is negative since the value is decreasing. The equation is A1t2 = - 10,000t + 40,000. In general, for this method, the value of the item plotted versus time will have a constant negative slope - V/T and y-intercept V: A1t2 = 1 - V/T2t + V Of course, it is not necessary that an asset be depreciated to zero. It is perfectly natural for the cab driver to assume that there will be some salvage value to the car at the time it is disposed of. Consider the following example. Example 2. Ms. Winslow buys a new car for her business for $13,000. She expects to use the car for five years and then sell it to a used car dealer for $2,000. Find the function A(t) that describes the value of the car after t years 10 t 52 if she uses straight line depreciation. Solution. Since at time 0, the car is worth 13,000 and at time t = 5, the salvage value of the car will be 2,000, we know that the line passes through the two points (0,13000) and (5,2000). The slope of the line must be m 2000 - 13000 - 11000 = = - 2200 5 - 0 5

and since we know the y-intercept is 13,000, A1t2 = - 2200t + 13000. Notice again that the slope is the rate of change of value and it is negative, because the asset (car) loses value over time.

82

* **

Section 1.3

Actually, the straight-line depreciation relationships are true only for integer values of x and the graphs really consist of isolated dots rather than continuous straight lines. For simplicity, however, we shall continue to let x be any real number. If our answer for a break-even point had not turned out to be a whole number, then we could make the following natural interpretation: that is, the next larger integer value would give the first profit. In addition, there are cases in which fractions make perfect sense. For instance when the x-values are not expressed in units but in hundreds or thousands of units as is common in most real business problems. Remember the vertical line criterion for functions. The graph of a function can be distinguished by the fact that every vertical line intersects the graph in at most one point. Given such a graph, if we are lucky, we can figure out an algebraic equation represented by the graph and express f(x) that way. For given straight lines, that is easy. However, consider the graph in Figure 3. Graphs such as this are called piecewise linear . By inspection, it is certainly the graph of a function (by the vertical line criterion). We express the function in algebraic form. We see that to the left of x = 1, the graph is a straight line that passes through (0,0) and (1,2). Therefore, its slope is 2 and its y-intercept is 0. So the equation of this line is just y = 2x. Thus, for x 6 1, f1x2 = 2x. Next, we see that between x = 1 and x = 3, the graph is a horizontal line with height y = 2. So, for 1 6 x 6 3, f1x2 = 2. Finally, if x is to the right of 3, the graph is a straight line passing through (3,2) and (5,0). It is easy to calculate that the slope is - 1, and using the point-slope formula, we get y = - x + 5 as its equation. What about the transition points x = 1 and x = 3? At x = 1? In this case, it does not matter whether you use the rule y = 2x or y = 2, since both give the same y-value, 2. So we can pick either.

y

Figure 3: A Piecewise Linear Graph The transition point x = 3 is handled similarly. (If the two rules gave different values at such a point, we would have been forced to decide precisely which was intended, since every x must define exactly one y.) At any rate, this graph can be described as: f1x2 = c 2x 2 -x + 5 1 if x if 1 6 x if x 7 3 3

Section 1.3

* **

83

Example 3. The 2003 Schedule X is used by single taxpayers to determine their tax, Table 2 is a portion of this schedule. Table 2: Portion of Schedule X

Schedule X If Taxable Income Is THEN Is Over Single $0 $7,000 $28,400 $68,800 But Not Over $7,000 $28,400 $68,800 $143,500 This Amount Plus This % Of the Excess Over $0 $700 $3,910 $14,010 10% 15% 25% 28% $0 $7,000 $28,400 $68,800 The Tax is

Letting x be your taxable income and t be your tax, find t = T1x2 for the domain 0 6 x 143,500. Solution. If x 7,000, then t = 0.10x. If x is between 7,000 and 28,400, then the tax is 700 plus 15% of 1x - 7,0002. Thus, t = 0.151x - 7,0002 + 700. Note that at the splitting point, the rule gives the same value for t. Proceeding similarly, for x between 28,400 and 68,800, we have t = 0.251x - 28,4002 + 3,910, and for x between 68,800 and 143,500, we have t = 0.281x - 68,8002 + 14,010. In short The graph of T(x) is shown in Figure 4. 0.10x 0.151x - 70002 + 700 T1x2 = d 0.251x - 284002 + 3910 0.281x - 688002 + 14010 if 0 6 x 7000 if 7000 6 x 28400 if 28400 6 x 68800 if 68800 6 x 143500

We show how to use your calculator to sketch the graph of this function below.

The graph shown for T(x) in the previous example may be drawn with your calculator by entering each of the pieces of the function separately on the Y = screen. We illustrate in Figures 4(a) and 4(b). Some remarks: first, the space bar is located above the 1 - 2 key and you must press Alpha and then this key. The condition (in this case the interval of definition) is inserted using the vertical bar key to the left of the number 7 key. To get the symbol you press and then the 0 key, similarly, to get the symbol you press and then the # (period) key.

Calculator Tips

84

* **

Section 1.3

Notice the black arrow to the right of y2, y3 and y4, that means there is not enough room to show the rest of the line on the screen and you need to scroll (use the arrow keys) to see it. Now we know we are working with larger numbers here, so we need to set an appropriate window. The default window will not be accurate. Since this graph has such a large range of values, the shifts from one piece of the graph to another is not easily seen, to see them more clearly would require you to zoom in near each transition point. One possible window is given in Figure 4(b).

Figure 4(b): A Window to View T(x) Using this window, we have the sketch in Figure 4(c). (Note that the first switch in pieces occurs when x = 7000, the second at x = 28,400 and the third at x = 68800. The calculator shows a slight bend at these x-values, the first bend is difficult to see in the present window as it has a relatively smaller y-value. Again, zooming in at these points will better show the transition (bend) from one piece of the graph to the next. Alternately, in this window we can press F2 (Zoom) and we obtain different zoom options. In most cases, you choose the portion of the graph you want to zoom in on by moving the cursor to the left and pressing enter and then move to the right and press enter. We leave it to you to experiment with these options.

1. Rework Example 1 for fixed overhead of $2,500, cost per bottle of $30, and selling price of $45. 2. Suppose that the manufacturer in Exercise 1 can manage to reduce the overhead to $1,200. How does this affect the break-even point? 3. Suppose that the manufacturer in Exercise 1 is forced to cut the selling price to $35 in order to clear inventories. If his overhead remains $2,500 and cost per bottle remains $30, what is the new break-even point?

Section 1.3

4. If a manufacturer has fixed costs of $700, a cost per item for production of $20, and expects to sell at least 100 items, how should he set the selling price to guarantee breaking even? 5. If the manufacturer of Exercise 4 wishes to guarantee profits of at least $600, what should he set the selling price to be? 6. A small bottling company finds that it costs $6,000 to prepare 10,000 sixpacks of cola, and $8,000 to prepare 15,000 six-packs. If each six-pack sells for $1.20, find and plot the graphs of the total cost function and the revenue function. Plot the profit function. Find the break-even point. 7. A manufacturer of small glass figurines discovers that it costs $1,800 for a production run of 1000 and $2200 for a production run of 1500. Assuming that cost is a linear function of number of items, find the overhead and the marginal cost of a figurine. Plot the cost graph. 8. At a selling price of $50, a lamp company breaks even on total sales of $3000. If the company s overhead is $1200, find the marginal production cost per lamp and find the profit on sales of 130 lamps. 9. The Amalgamated Flashlight Company shows a profit of $4,500 on a production of 6000 flashlights and a profit of $10,500 when the production is 10,000. (a) If flashlights sell for $2.25 apiece, find the cost equation for flashlight production. (b) Plot cost and revenue on the same axes and locate the break-even point. (c) Suppose that the capacity of the flashlight assembly line is 30,000 flashlights. How must the price be set to make possible a $90,000 profit? 10. A manufacturer of vases discovers that it sells 1800 vases at $12 per vase but only 1300 at $15 per vase. Assuming that demand is linearly related to price, find and graph the relation. 11. A machine worth $10,000 new and having a scrap value of $500 is to be depreciated over a ten-year life. Find the function that describes straight line depreciation for this situation. At what time will the machine be worth $5000 according to this model? 12. A house and lot valued at $100,000 is being depreciated over 25 years by the straight line method. After 10 years the book value of the asset is $64,000. Since only the value of the house depreciates, find the assumed value of the land. (Ignore inflation.) 13. Figure 5(a) shows the graph of a piecewise linear function f. Find the algebraic formulation of f(x). Find f1 - 22, f(0), f(2), and f(5).

* **

85

(-1,1)

+

(2, 1)

(1, -1)

Figure 5 (c)

15. Figure 5(c) shows the graph of a piecewise linear function f. Find the algebraic formulation of f(x). Find f1 - 22, f11*22, and f(5). 16. Consider the function given by f1x2 = e x + 3 -x + 3 if x 0 if 0 6 x

Find f1 - 52, f102, and f(3). Plot the graph of y 17. Consider the function given by 2 f1x2 = c x + 2 6

= f1x2.

if x 0 if 0 6 x if 4 6 x

Find f1 - 22, f102, f112, f142, and f(7). Plot the graph of y = f1x2 18. Consider the function given by - 2x + 1 2x - 3 -x + 6 if x 1 if 1 6 x if 3 6 x

f1x2 = c

Find f(0), f(1), f(2), f(3) and f(5). Plot the graph of y = f1x2. 19. (a) Sketch the graph of the line y = 2x - 3 on the same axes as the graph of y = f1x2 from Exercise 16. How many points of intersection do you see? What are their coordinates? (b) Sketch the graph of the line x - 2y + 2 = 0, on the same axes as the graph of y = f1x2 from Exercise 17. How many points of intersection do you see? What are their coordinates? 20. Sketch the graph of the line x - 2y + 1 = 0 on the same axes as the graph of y = f1x2 from Exercise 18. How many points of intersection do you see? What are their coordinates? 21. Wong Industries is depreciating the value of its machinery over a thirty-year life. Its original value was $1,000,000 and it is assumed to have a scrap value of $100,000. After 20 years, it is decided that the machinery still has another twenty years of life left but it will then be worth nothing. Find a piecewise linear function that describes the broken-straight-line depreciation that the company is using. 22. Schedule Y-1, to be used by married tax payers in 2003 is given in Table 2.

+

(3, 2)

+ (4,

2)

(1, 1)

(2, 1)

(3,- 4) +

(1,-5)

(0, -1)

Figure 5 (a)

Figure 5 (b)

14. Figure 5(b) shows the graph of a piecewise linear function f. Find the algebraic formulation of f(x). Find f1 - 12, f(1.5), and f(4)

86

* **

Section 1.4

Table 2: Schedule Y1

Schedule Y1 If Taxable Income Is THEN Is Over Married Filing Jointly or Qualifying Widow(er) $0 $14,000 $56,800 $114,650 $174,700 $311,950 But Not Over $14,000 $56,800 $114,650 $174,700 $311,950 This Amount $0 $1,400.00 $7,820.00 $22,282.50 $39,096.50 $84,389.00 Plus This % 10% 15% 25% 28% 33% 35% Of the Excess Over $0 $14,000 $56,800 $114,650 $174,700 $311,950 The Tax is

Letting x be your taxable income and t be your tax, find t = T1x2 for the domain 0 6 x 311,950 and plot its graph. 23. A local photocopying store advertises as follows. We charge 8 per copy for 100 copies or less, 6 per copy for each copy over 100 but not over 1000, and 4 per copy for all over 1000. Let x be the number of copies ordered and C(x) be the cost of the job. Write C(x) algebraically and plot its graph. 24. Metro City s Non-Resident Earnings tax is $6.50 per thousand dollars of earnings. However, if you earn under $10,000, you may subtract $3000 from

your earnings; if between 10,000 and 20,000, you may subtract 2000; and if between 20,000 and 30,000, you may subtract 1000. So, for example, on 26,000 earnings, you may subtract 1000 and pay tax on 25,000. Your tax is 125216.52 = $162.50. Express the tax, T(x), as a function of x in thousands. Plot the graph. Would you rather earn $19,999 1x = 19.9992 or $20,001 1x = 20.0012?

Scaling Vertical Translation Axis of a Parabola Horizontal Translation Locating the Vertex Graphing a Parabola in the form y * ax 2 + bx + c Applications to Optimization Calculator Tips

After studying the linear equation ax + b = 0, the next step was to examine the linear function defined by y = f1x2 = ax + b, and then its applications. Similarly, given the quadratic equation ax2 + bx + c = 0, it is natural to next consider the quadratic function defined by y = f1x2 = ax2 + bx + c, where a Z 0. The graph of this function is called a parabola. It turns out that once we understand how two special cases appear, then we almost know how any other parabola will appear. In particular, we examine the special case y = x2. We observe that when x = 0, the y-value is 0, and is positive for any other x-value. Moreover, equal and opposite x-values have the same y-value, therefore this graph is symmetric with respect to the y-axis, that is, the y-axis behaves like a mirror, the portion of the graph on each side of the axis is a mirror image of the other. In Table 1 we choose x-values and compute the corresponding y-value, and then we plot these points to obtain the graph of y = f1x2 = x 2.

Section 1.4

* **

87

x-value -3 y = f1x2 = x 2 y y y y y y y = = = = = = = 1 - 32 = 9 1 - 222 = 4 1 - 122 = 1 1022 = 0 1122 = 1 1222 = 4 1322 = 9

2

-2

-1 0 1 2 3

The graph is plotted using these point in Figure 1. Observe that the parabola is a U shaped graph, in this case opening upward. The graph is symmetric with respect to the y-axis, that is, the line x = 0. The lowest point on the graph, its turning point or vertex is the origin, (0, 0). What happens to the shape of the graph if we change the sign of the coefficient of the x2 term? In the previous example its coefficient was positive, resulting in the graph opening upward. Changing the sign will, as we shall see change the way the graph opens. As our next example, we consider the parabola y = - x 2. Table 2 gives the required points.

y = x2

x-value -3 -2 -1 0 1 2 3 y = - x2 y y y y y y y = = = = = = = - 1 - 32 = - 9 - 1 - 222 = - 4 - 1 - 122 = - 1 - 1022 = 0 - 1122 = - 1 - 1222 = - 4 - 1322 = - 9

2

88

* **

Section 1.4

The graph is given in Figure 2. Notice that the graph of the equation y = - x2 is also a U shaped graph opening downward, symmetric with respect to the y-axis - the line x = 0. Its turning point, or vertex, is the highest point on the graph. In fact the graphs of y = - x 2 and y = x2 are identical; if we turn one of them upside down we obtain the other. The reason is that for the same x-value, the y-value on one curve is the negative of the other.

y = -x2

Figure 2: The Graph y = - x2 We shall see soon that all parabolas will have one of the above shapes. That is, y = f1x2 = ax2 + bx + c will have an upward shaped appearance U when a 7 0, and a downward shaped one when a 6 0. Consider the graph y = 2x2. How does this graph differ from y = x 2? The answer is simple, for the same x-value, its y-value is doubled. Table 3, gives the points needed to plot this curve. Table 3: Points used to Plot y = 2x 2

x-value -3 -2 y = 2x 2 y y y y y y y = = = = = = = 21 - 32 = 18 21 - 222 = 8 21 - 122 = 2 21022 = 0 21122 = 2 21222 = 8 21322 = 18

2

-1

0 1 2 3

The graph is given in Figure 3. Note that this graph looks very much like the graph of y = x 2 except it rises faster. In fact, to better see that, we plot them both on the same coordinate system in Figure 4. Notice that the graph of y = 2x 2 is narrower, that is, for the same x-value its has twice the y-value. Thus, we see multiplying the coefficient of x2 by a constant does nothing more than scale the y-value. We have seen the effect of multiplying x 2 term by a constant. The effect of adding a constant to y = x 2 has the effect as moving the curve vertically, a vertical translation.

Scaling

Section 1.4

* **

89

y = 2x2

y = 2 x2

y = x2

Consider the graph of y = x2 + 1. We indicate in Table 5 the points used to sketch its graph which is given in Figure 5. Table 4: Points used to Plot y = x 2 + 1

x-value -3 -2 -1 0 1 2 3 y = x2 + 1 y y y y y y y = = = = = = = 1 - 32 + 1 = 10 1 - 222 + 1 = 5 1 - 122 + 1 = 5 1022 + 1 = 1 1122 + 1 = 2 1222 + 1 = 5 1322 + 1 = 10

2

90

* **

Section 1.4

y = x2 + 1

Figure 5: y = x2 + 1 Note that in comparing the graphs of y = x2 + 1 and y = x2, we see that for the same x-values, the y-value of the first graph is one unit above the y-value of the second graph. This is most easily seen in Figure 6, where the graphs are plotted together. Thus, we see, in general, that the c term in the quadratic function y = ax2 + bx + c moves the graph vertically upward or downward depending on c being positive or negative, without any change in the parabola s shape, that is y = ax 2 + bx and y = ax2 + bx + c are identically shaped but only differ in their vertical position by c units. The only remaining term we need consider is the x-term in the quadratic function. Consider the quadratic function defined by y = x 2 - 2x + 1 = 1x - 122. As indicated

y = x2 + 1 y = x2

Vertical Translation

Figure 6: The Graphs of y = x2 + 1 and y = x2 above, this graph of this function only differs from the graph of y = x 2 - 2x by one unit, that is, they are identical, but the first one is one unit above the other. Thus, by studying either, we can see what effect the x-term has on the quadratic. It is convenient to examine y = 1x - 122 since that graph s y values are most easily obtained by the substitution of its

Section 1.4

* **

91

x-values. In fact, the most convenient form of the parabola for plotting purposes is the form y = a1x - h22 + k. However, this form is not always the one presented, therefore, we shall consider other means, as considered below. Table 5 contains the points we need to draw its graph. Table 5: Points used to Plot y = 1x - 122

x-value -3 -2 -1 0 1 2 3 4 y = 1x - 122 y y y y y y y y = = = = = = = = 1 - 3 - 12 = 16 1 - 2 - 122 = 9 1 - 1 - 122 = 4 10 - 122 = 1 11 - 122 = 0 12 - 122 = 1 13 - 122 = 4 14 - 122 = 9

2

y = (x - 1)2

x=1

Observe that this graph is still symmetric with respect to the vertical line drawn through the vertex, except now the vertex is at the point (1, 0) so the graph is symmetric with respect to the line x = 1. The vertical line passing through the vertex is often called the axis of the parabola or its axis of symmetry. Notice that this graph is identical to the graph y = x 2 except it is shifted one unit to the right. This is most easily observed if both graphs are plotted together as in Figure 8. If we move the graph of y = x 2 one unit to the right, it coincides with the graph of y = 1x - 122. Thus, the x term in the quadratic function is the term that moves the graph horizontally, that is, causes a horizontal translation. Summarizing, we have that quadratic function defined by y = ax 2 + bx + c, with a Z 0 is always a U shaped graph, it opens upward if a 7 0 and downward if a 6 0. The b term effects the horizontal position of the vertex and the c term its vertical position.

Axis of a Parabola

Horizontal Translation

92

* **

Section 1.4

y = x2

y = (x - 1)2

Figure 8: The Graphs of y = 1x - 122 and y = x2 We observed that when the parabola is given in the form y = a1x - h22 + k, it is a simple matter to choose the appropriate points needed to plot the graph. The vertex occurs when x = h, since at this x-value the parenthesis term is zero, if a 7 0 then all other yvalues will be above this one or if a 6 0, all other y-values will be beneath it. Thus, in either case the point at which x = h is the vertex. We begin with this x-value and choose several values to its left and right, compute their corresponding y-values, plot the points, and then draw the graph, as was done above. Example 1 Sketch the graph of the parabola y = - 21x - 322 + 4. Solution. The x-coordinate of the vertex is x = 3. We choose several x-values to its left and right and compute their y-values. This is done in Table 6. Table 6: Points used to Plot y = - 21x - 322 + 4

x-value 0 1 2 3 4 5 6 y = - 21x - 322 + 4 y y y y y y y = = = = = = = - 210 - 211 - 212 - 213 - 214 - 215 - 216 32 322 322 322 322 322 322

2

point on graph = = = = = = = - 14 -4 2 4 2 -4 - 14

+ + + + + + +

4 4 4 4 4 4 4

(2, 2) (3, 4) (4, 2)

The question that remains is, how can we most easily draw the graph of any given parabola? A good representation of the graph should indicate its U shape, which means that we should first locate the vertex of the parabola, and then choose a few points to its left and right to obtain enough points to reasonably plot the graph. We observed that the vertex is always the lowest point on a parabola opening upward, or the highest point on a parabola opening downward. Consider the parabola

Section 1.4

* **

93

x=3

y = -2(x - 3)2 + 4

y = x2 + 4x + 5. We know even before plotting its graph that it opens upward (why?). Therefore, we shall search for its vertex which is its lowest point. The procedure by which we locate the vertex is nothing more than a variation of the method of completion of the square. We first isolate the x-terms and write y - 5 = x2 + 4x We next complete the square on the right hand side, remembering that whatever we do to one side of an equation, we must do to the other. 1 2 142 = 2 and 22 = 4, so we have y - 5 + 4 = x 2 + 4x + 4 or y - 1 = 1x + 222 or y = 1x + 222 + 1 Now observe that the lowest possible value for y will occur when the term in the parenthesis is zero. This will happen when x = - 2, and at this value, y = 1. Thus, the vertex of the parabola is at the point V1 - 2, 12. We remarked above that it is an easy matter to plot the parabola when it is written in the form y = 1x + 222 + 1. However, our objective is to plot the graph using the original form of the equation. This means, that if we knew the x-coordinate of the vertex, we could construct a table of values including the x-value - 2, and choose a few values to the left and right. Thus, if we were to plot this parabola we could begin our table as indicated in Table 7. Table 7: Constructing a Table of Values

x-value -5 -4 -3 x-coordinate at vertex : - 2 -1 0 1 y = x 2 + 4x + 5 point on graph

94

* **

Section 1.4

Remember, we want to locate the vertex and use the original form of the equation. Can we determine a formula that will easily give the x-coordinate of the vertex? If we do what we did above to the general quadratic function defined by y = ax 2 + bx + c, where a Z 0, let s see what happens. First, we divide by a so the coefficient of the x 2 term is 1. This gives y/a = x2 + b/ax + c/a We next isolate the x-terms, giving y/a - c/a = x2 + b/ax 1b b b 2 b2 We now complete the square. = , and a b = . 2a 2a 2a 4a 2 Therefore, y c b2 b b2 + = x2 + x + 2 a a a 4a 4a2 or y c b2 b 2 + b = ax + 2 a a 2a 4a b2 b 2 = aax + b 4a 2a b 2 b2 b + c 4a 2a

We can combine fractions and write this last equation as y = aax + b 2 4ac - b2 b + 4a 2a (1)

Graphing a Parabola

We see that the term in the parenthesis will be zero when x = - 2b a . Therefore, we conlude that given the parabola y = ax 2 + bx + c, the x-coordinate of its vertex is x = - 2ba. Thus, to quickly draw a parabola we need only determine the x-coordinate of the vertex and then construct the table of values needed to get enough points to accurately obtain the graph. This is conveniently done by choosing three successive integer values to the right and left of the x-coordinate of the vertex. The y-values are obtained by substituting the x-values into the equation of the parabola. Graphing a Parabola in the form y * ax 2 + bx + c (1) Locate the x-coordinate of the vertex which is x = - 2b a. (2) Make a table of x-values, including the vertex as well as three x-values larger and smaller than this x-value. (3) Complete the table by computing the y-values corresponding to each of the above x-values. (4) Plot each of the seven points determined from the above steps. (5) Draw a smooth graph through these seven points. (6) Draw and label the axis of the parabola - the vertical line through the vertex. You really do not need seven points to graph a parabola (in fact, as we shall see in the exercises, three points uniquely determine it), however, doing so may prevent errors in drawing its graph.

Section 1.4

* **

95

Example 2 Sketch the graph of the parabola y = - 2x2 - 8x + 12. Solution. The x-coordinate of the vertex is x = - 1 - 82/121 - 222 = - 2. We will choose three x-coordinates above and below this value and compute their y-coordinates, as given in Table 8. Table 8: Points used to Plot y = - 2x 2 - 8x + 12

x-value -5 -4 -3 x-coordinate at vertex : - 2 -1 0 1 y = - 2x 2 - 8x + 12 y y y y y y y = = = = = = = - 21 - 52 - 81 - 52 + - 21 - 422 - 81 - 42 + - 21 - 322 - 81 - 32 + - 21 - 222 - 81 - 22 + - 21 - 122 - 81 - 12 + - 21022 - 8102 + 12 - 21122 - 8112 + 12

2

We use the points from Table 8 to plot the graph given in Figure 10. (Note, that even before we begin the graph, we know that the graph opens down, because the coefficient of the x 2 term is negative.)

x = -2

y = -2x2 - 8x + 12

Figure 10: The Graph of y = - 2x2 - 8x + 12 Note that the above example, the y-intercept is indicated on the graph. In fact, this is always an easy point to plot since we need only set x = 0 and solve for y to find it. The x-intercepts are usually not convenient points to plot, as to find them we need to solve a quadratic equation. In the above case, the equation - 2x 2 - 8x + 12 = 0 However, we can read off the approximate solutions to this equation from our graph. In the above case we have x1 L - 5.2 and x2 L 1.2. The exact answers, found by the quadratic formula, are x1 = - 2 - 210, x2 = - 2 + 210. Notice, in the tables used above, corresponding points on either side of the vertex had the same y-coordinates. Of course, since the parabola is always symmetric about its

96

* **

Section 1.4

axis, this will always be the case for symmetrically chosen x-values. However, it may not always be convenient to plot the symmetric points. This is true when the x-coordinate at the vertex is not an integer. In such a case, we will choose integers values to the left and right of this x-value to minimize the arithmetic in computing the y-coordinates. The next example illustrates this remark. Example 3 Sketch the graph of the parabola y = 3x2 - 8x - 10. Solution. The x-coordinate at the vertex is x = - 1 - 82/121322 = 4/3. We shall choose integer values to the right and left of this x-value. Note the integer to the right of 4/3 is 2 and the one to its left is 1. The details are indicated in Table 9. Note that the only fractional arithmetic needed is to compute the y-value at the vertex, since we chose integer values for x at the other points. The graph is given in Figure 11.

x-value -2 -1 0 1 x-coordinate at vertex : 4/3 2 3 4 y = 3x 2 - 8x - 10 y y y y y y y y = = = = = = = = 31 - 22 - 81 - 22 31 - 122 - 81 - 12 31022 - 8102 - 10 31122 - 8112 - 10 31 - 222 - 81 - 22 31222 - 8122 - 10 31322 - 8132 - 10 31422 - 8142 - 10

2

V14/3, - 46/32

12, - 142 13, - 72 (4, 6)

x = 4/3

y = 3x2 - 8x - 10

V(4/3, -46/3)

Figure 11: The Graph of y = 3x2 - 8x - 10 Note that in the above example, the y-intercept is 10, - 102 and is easily found.

Section 1.4

* **

97

Example 4 Determine the x-intercepts in the graph given above. The exact intercepts are found by solving the equation 3x2 - 8x - 10 = 0. 4 - 1 246 4 + 246 Doing so, we find them to be x1 = , x2 = . From our graph, we 3 3 can see that their approximate values are - 0.9 and 3.6 respectively. Solution.

A typical problem that arises in calculus is to determine the maximum or minimum value assumed by a given function. If the function is quadratic, then it is clear from its graph that its optimal (maximum or minimum) value will occur at the vertex. Consider the following examples. Example 5 A ball is thrown vertically upward from the ledge of a building 160 feet above the ground with an initial velocity of 48 feet per second. It can be shown that its height h measured in feet above ground is given by the equation h = - 16t2 + 48t + 160, where t is the time measured in seconds. Determine (a) how long it takes the ball to reach its maximum height, and what is this height? (b) how long does it take for the ball to hit the ground? Solution. Since the height, as a function of time is a parabola, we plot its graph. Note that instead of y being a function of x, we now have h as a function of t. The t coordinate of the vertex is given by t = - 48/121 - 1622 = 3/2. The integer t-values to the left and right of this value are 1 and 2. Since t represents time, it makes no sense to use negative t-values, h represents the height of the ball above the ground level, so it too must be non-negative. We construct Table 10. Using these points, we plot the graph given in Figure 12. (a) We see that at the vertex, the ball is at its maximum height, therefore it takes 3/2 = 1.5 seconds to reach the maximum height which is 196 feet above the ground. (b) The ball hits the ground when its height h = 0. That means its t-intercept is the point at which this occurs. We solve the quadratic - 16t2 + 48t + 160 = 0 and find that t = - 2 or 5. Since we must have a positive value for t, we have that it takes 5 seconds for the ball to hit the ground.

Applications to Optimization

t-value 0 1 t-coordinate at vertex : 3/2 2 3 4 h = - 16t 2 + 48t + 160 h = - 16102 + 48102 + 160 = 160 h = - 161122 + 48112 + 160 = 192 h h h h = = = = - 1613/222 + 4813/22 + 160 = 196 - 161222 + 48122 + 160 = 192 - 161322 + 48132 + 160 = 160 - 161422 + 48142 + 160 = 160

2

point on graph (0, 160) (1, 192) (3/2, 196) (2, 192) (3, 160) (4, 96)

Once we recognize the function to be optimized is parabolic, it is no longer necessary to plot its graph as illustrated in the next two examples.

98

* **

Section 1.4

V(3/2, 196)

Figure 12: h = - 16t2 + 48t + 160 Example 6 A farmer will construct a rectangular enclosure from 1200 feet of fencing. He plans to subdivide the enclosure into three parts, as illustrated in Figure 13. Determine the dimensions of the enclosure that maximize the enclosed area.

y

x y

Figure 13: A Rectangular Enclosure of Width x and Length y Solution. We let x and y represent the dimensions of the enclosure as indicated in Figure 13. The problem is to maximize the total area A = xy. This is somewhat unusual, as A is a function of two variables, x and y. However, we can eliminate one of these variables, since the total amount of fencing used in constructing the enclosure is 4x + 2y = 1200. Solving for y we have y = - 2x + 600 therefore, A = xy = x1 - 2x + 6002 or A = - 2x2 + 600x Once again, we have an equation of a quadratic function whose graph is a parabola. Since x represents a dimension, we must have that x 7 0.

Section 1.4

* **

99

The x-coordinate of the vertex is x = - 600/121 - 222 = 150. Since this is a parabola opening down, (why?) the vertex is the highest point on the graph, so at the vertex the area is maximized. Thus, for the area to be maximized, the width x = 150 feet. From the equation y = - 2x + 600, we find that the corresponding length is y = - 211502 + 600 = 300. Thus, when the area is maximized, the width of the enclosure is to be 150 feet and its length 300 feet. The maximum area is 1150213002 = 45,000 square feet.

Typical business problems arising in applications are to maximize profit, minimize costs or maximize revenue. In the event that these functions are quadratic, the methods of this section are applicable. The next example illustrates. Example 7 The cost in dollars of producing x bicycles is given by the equation C = 2x2 - 2000x + 600000. Determine the number of bicycles that should be produced to minimize the producer s cost. Solution. The cost function is a parabola opening upward. Therefore, the lowest point on the graph is the vertex. The x-coordinate of the vertex is therefore the number of bicycles that should be produced to minimize the cost (which is the y-coordinate of the vertex). Thus, we have, x = - 1 - 20002/121222 = 500 The producer should produce 500 bicycles to minimize his cost. (The cost in producing the 500 bicycles is $100,000 and the average cost per bicycle is 100000/500 = $200.)

In the previous example, you might have asked yourself why the cost of producing zero bicycles is $600,000. This number is sometimes called the startup or overhead cost needed to begin production; for example, the cost of rent, machinery, initial labor costs, etc. In the previous section we indicated how to plot the graph of a function using the calculator. This procedure may be easily used to plot the graph of any parabola of the form y = ax2 + bx + c. To make sure you obtain the U-shaped portion of the graph in your window, you need only first calculate the x-coordinate of the vertex and make sure this is included in the window selected. Consider, for example, y = 2x 2 - 60x + 475. We go to the Y = window and define y1, see Figure 14.

Calculator Tips

100

* **

Section 1.4

Before we press the graph button, we compute the x-coordinate at the vertex, x = - 1 - 602/121222 = 15. Note, from the home screen we calculate y11152 = 25. Thus, we see that the vertex is (15, 25). Therefore, we want to include the vertex in our window. We can generate a table of values using the TABLE key 1*F52, we obtain Figure 15.

Figure 15: Using the TABLE Key The table generated does not include the vertex, so we use F2 (Setup), we obtain Figure 16.

Figure 16: Setting Up the Table tblStart is the x-value at which we want the table to begin and tbl is the incremental value, that is, the difference in succeeding x-values listed in the table. We will choose tblStart = 15 and (the x-value at the vertex) and tbl = 1, See Figure 17.

Section 1.4

* **

101

Next press Enter twice to get the table, see Figure 18.

Figure 18: The Table Starting at x = 15 Scrolling up and down will give different points on the graph. Thus, from the table, it is clear that we should could use a window containing x-values between 12 and 18 and the corresponding y-values would be between 25 and 43. If we also want to include the origin (we don t have to) then we could choose the window given in Figure 19 (there are many other reasonable choices for a window.).

Figure 19: Choosing an Appropriate Window The graph is now obtained by pressing the Graph key 1* F32, see Figure 20.

102

* **

Section 1.4

Once again, we stress that there are many other window choices that could have been made that would illustrate the U-portion of the parabola. You should experiment with various choices.

(a) In Exercises 1 21 draw the graph of the parabola. (b) From your graph, estimate the x-intercepts. (c) Check your estimate by finding the x-intercepts exactly. 1. y = 1x + 122 2. y = 1x - 222 3. y = 1x - 322 4. y = 1x + 322 5. y = 41x - 122 6. y = - 31x + 122 7. y = - 21x - 122 8. y = 31x + 122 - 3 9. y = 31x + 12 + 4 10. y = - 21x - 322 + 5 11. y = - 21x + 422 + 7 12. y = x2 - 4x - 5 13. y = x2 + 4x - 6 14. y = 2x2 + 4x + 5 15. y = - 2x2 - 4x + 9 16. y = 3x2 + 12x - 15 17. y = - 2x2 + 4x + 8 18. y = 2x2 + 5x - 9 19. y = - 2x2 + 7x + 12 20. y = - 3x + 5x - 8 21. y = - 4x2 - 5x + 12 22. y = - 4x2 - 5x - 12 23. y = 2x2 + 5x + 9 24. y = - 2x2 + 4x - 12 25. A ball is thrown vertically upward from the ground. If its height h measured in feet is given in terms of time t, measured in seconds by the equation h = - 16t2 + 96t, determine (a) how high the ball will go and how long does it takes to reach that height, (b) how long does it take for the ball to come back to the ground? 26. A gun is fired upward from the ledge of a 176 foot cliff. The bullet leaves the gun with a velocity of 80 feet per second. If its height h measured in feet is given in terms of time t, measured in seconds by the equation h = - 16t2 + 80t + 176, determine (a) how high the bullet will go and how long it takes to reach that height, (b) how long does it take for the bullet to come back to the ground? 27. A barnyard is to be fenced is as indicated in Figure 21. If 3600 feet of fencing is to be used in its construction determine the dimension x and y that will maximize its total area?

2 2

y x x x y x x

Figure 21

28. A barnyard is to be fenced is as indicated in Figure 22. If 6000 feet of fencing is to be used in its construction determine the dimension x and y that will maximize its total area?

2y x y 2x y y y x

2x

2x

Figure 22

29. A barnyard is to be fenced is as indicated in Figure 23. If 4200 feet of fencing is to be used in its construction determine the dimension x and y that will maximize its total area? s2 Hint: The area of an equilateral triangle with side s is 23. 4

x ,*+ x y y

,*+ x y

,*+ x

,*+ x

Figure 23

Section 1.5

30. The cost C, in dollars, in producing x large screen televison sets is given by the equation C = 4x2 - 2000x + 600000. (a) How many television sets should be produced to minimize the cost? (b) What is the minimum cost? (c) At this level of production, what is the average cost per set? 31. The revenue R, in dollars, derived from selling x computers is given by R = - 3x2 + 600x + 1000. (a) How many items should be sold to maximize the revenue? (b) What is the revenue? 32. Student Government at a University is chartering a plane for Spring Break. The plane can seat 150 passengers. The airline will charge $120 per passenger and added to this a surcharge of $15 per passenger for each unsold seat. Let x represent the number of unsold seats. (a) Show that the airline s revenue R = 1150 - x21120 + 15x2. (b) How many seats should be unused to maximize the airline s revenue? (c) What price would each passenger pay if the airline maximized its revenue? (d) Is this a good deal for the Student Government? 33. Given the parabola y = ax 2 + bx + c (a) determine the relationship between a, b and c if the parabola is to have no real x-intercepts. (b) When will its graph lie above the x-axis? (When will it lie below the x-axis? 34. Determine the equation of the parabola passing through the points (0, 10) (1, 9) and (4, 30). 35. Determine the equation of the parabola passing through the points (0, 8), (2, 6) and 1 - 1, 02. 36. A parabola may be defined as the set of all points equidistant from a fixed point called the focus and a fixed line called the directrix. (a) Suppose the

The Circle

* **

103

focus is located at (0, p) and the directrix has the equation y = - p, determine the equation of this parabola. (b) Suppose the focus is located at 10, - p2 and the directrix has the equation y = p, determine the equation of this parabola. 37. Find the equation of the parabola with (a) focus at 1h, k + p2 and directrix y = k - p. (b) focus at 1h, k - p2 and directrix y = k + p. 38. Find the equation of the parabola with focus on the line y = x at (h, h) and directrix with equation y = - x - 2h. (The axis of the parabola is the line y = x.) 39. This example illustrates another way of finding the vertex of a parabola (and deriving the quadratic formula). Given y = ax2 + bx + c multiply both sides of the equation by 4a, obtaining, 4ay = 4a 2x2 + 4abx + 4ac Now add b2 both sides of the equation yielding b2 + 4ay = 4a 2x2 + 4abx + b2 + 4ac Note the first three terms on the right hand side are a prefect square, so we may factor to obtain, b2 + 4ay = 12ax + b22 + 4ac Solve for y and then show that it follows that the vertex occurs when x = - b/2a. Also show that setting y = 0 results in the quadratic formula.

Definition of a Circle Equation of a Circle Graphing a Circle Tangent Line The Ellipse Calculator Tips

Suppose someone asks you to draw, as accurately as you can, a circle on a piece of cardboard. How would you go about doing it? One approach is to attach a pencil to one end of a taut string and hold the other end fixed. Keeping the string taut move the pencil. As you are doing so, you are drawing an arc of a circle. Continuing once around, the circle is drawn. The above method for sketching a circle may be used to define it. The circle is defined to be the set of all points equidistant from a given point. The given point is called the center of the circle and the distance each point is from the center is called the radius. This definition may be used to determine the equation of circle if we are given the coordinates of the center and the length of the radius. We illustrate in the following example. Example 1 Find the equation of the circle with center at the point C (1, 2) and radius 3. Solution. Let P (x, y) be any point on the circle, our objective is to determine an equation relating x to y. We know that every point on the circle is 3 units from the center, therefore, we have the distance from the center C (1, 2) to the point P(x, y) using the distance formula is

Definition of a Circle

104

* **

Section 1.5

The Circle

21x - 122 + 1y - 222 = 3 If we square both sides of the equation we obtain 1x - 122 + 1y - 222 = 9 This is one way we can leave the equation, or if we like, we could multiply out and combine like terms and write the equation in the form x2 + y 2 - 2x + 4y - 4 = 0

Equation of a Circle

What we did in Example 1, we may do in general to obtain the equation of the circle centered at the point C (h, k) with radius r. If the point P(x, y) is any point on the circle, then it is a distance r from the center. Therefore, by the distance formula, we have 21x - h22 + 1y - k22 = r or squaring both sides we have

1x - h22 + 1y - k22 = r2

(1)

Equation (1) is called the standard form and the most convenient form of the equation of a circle. Inspection immediately yields the center and radius. However, if equation (1) is expanded, we obtain x2 + y 2 - 2hx - 2ky + h2 + k2 - r 2 = 0 If we let a = - 2h, b = - 2k and c = h2 + k2 - r 2, then we may write the equation in the form

x2 + y2 + ax + by + c = 0

(2)

Equation (2) is called the general form, and is not the most useful form for the equation of the circle, as it is not immediately obvious that it is indeed a circle (as we shall see), and if it is, it is not immediately apparent where its center is located nor what is its radius. When the circle is centered at the origin, that is, h = k = 0, equation (1) reduces to

x2 + y2 = r2

(3)

The only difference between the circle whose equation is given by (1) and the one given by (3) is their location. Each point on the circle given by equation (1) is translated horizontally by h units and vertically by k units. We illustrate this remark in Example 5 below. Example 2 Given the equation of the circle 1x - 222 + 1y - 422 = 16, determine is center and its radius. Solution. Comparing with (1) we have immediately h = 2, k = 4, and r 2 = 16. Thus, the center of the circle is C (2, 4) and the radius is r = 216 = 4.

Section 1.5

The Circle

* **

105

Looking at (1), we see that h is the value of x that makes the first parenthesis zero and k is the value of y that makes the second parenthesis vanish, thus, to find the x-coordinate of the center we may set x - h = 0 and to find the y-coordinate we may set y - k = 0. Example 3 Given the equation of the circle 1x - 322 + 1y + 522 = 25, determine is center and its radius. Solution. Using the observation made above, set x - 3 = 0 and y + 5 = 0, yielding x = 3 and y = - 5. Thus, the center of the circle is C13, - 52. The radius is r 2 = 25 or r = 5.

Given the equation of a circle we can sketch its graph fairly easily. Actually, three points uniquely determine a circle, but we suggest the following approach which uses four points in addition to the center. Graphing a Circle 1. Plot the center 2. Move horizontally to the right from the center a distance equal to the radius and plot this point. 3. Move horizontally to the left from the center a distance equal to the radius and plot this point. 4. Move vertically above the center a distance equal to the radius and plot this point. 5. Move vertically below the center a distance equal to the radius and plot this point. 6. Connect these four points with a smooth curve. We illustrate this approach with the following example. Example 3 Sketch the circle whose equation is given in Example 2. Solution. The circle is centered at C (2, 4) and has radius 4. We choose the four points suggested above. Moving to the right 4 units from (2, 4) yields (6, 4). Moving to the left 4 units yields 1 - 2, 42. Moving up 4 units yields (2, 8) and moving below 4 units yields (2, 0). The points are plotted and the circle is drawn in Figure 1.

Graphing a Circle

When the general form is given for the equation of the circle we must first transform it into the standard form so we may easily locate its center and radius. In order to do this, we must complete the square, but in this case, twice; once for the x terms and once for the y terms. We must also remember that whatever we do to one side of an equation we must do to the other. We illustrate the procedure in the next example. Example 4 Given the circle whose equation is x2 + y 2 - 6x + 10y + 9 = 0. (a) Determine its center and radius. (b) Sketch its graph. Solution. (a) We rewrite the equation as x 2 - 6x + y 2 + 10y = - 9

106

* **

Section 1.5

The Circle

Figure 1: 1x - 222 + 1y - 422 = 16 We first complete the square for the x terms: 1 2 1 - 62 = - 3, and 1 - 322 = 9, so we have x2 - 6x + 9 + y 2 + 10y = - 9 + 9 or 1x - 322 + y 2 + 10y = 0 We next complete the square for the y terms: 1*21102 = 5 and 1522 = 25, so we have 1x - 322 + y 2 + 10y + 25 = 0 + 25 or 1x - 322 + 1y + 522 = 25 This is the circle given in Example 3. We see that the circle is centered at 13, - 52 and has radius 5.

Section 1.5

The Circle

* **

107

(b) we plot its graph by using the four points described above. Moving to the right from the center 5 units we have the point 18, - 52, moving to the left from the center we have 1 - 2, - 52. Moving up 5 units from the center we have (3, 0) and moving 5 units down from the center we have 13, - 102. We plot these points and the graph in Figure 2.

Example 5 Sketch the circles 1x - 122 + 1y - 222 = 9 and x2 + y 2 = 9 on the same coordinate system. Solution. The sketch of the two circles is given in Figure 3, and the points used to plot the graphs are indicated.

(x - 1)2 + (y - 2)2 = 9

x2 + y2 = 9

Figure 3: The graphs of 1x - 122 + 1y - 222 = 9 and x2 + y2 = 9 If we moved the circle centered at the origin so that its center would be at (1, 2) then the two graphs would be coincident. That is, if any point on the circle centered at (0,0) is moved one unit to the right and 2 units up then we get the corresponding point on the other circle. For example, take the point (3, 0), 13 + 1, 0 + 22 yields the corresponding point (4, 2). Not every equation of the form x2 + y 2 + ax + by + c = 0 is a circle. Consider the next example. Example 6 Classify each of the following: (a) x 2 + y 2 + 4x - 6y + 13 = 0, (b) x2 + y 2 + 4x - 6y + 15 = 0 Solution. We complete the square to understand what is happening here. (a) x2 + y 2 + 4x - 6y + 13 = 0 is rewritten as

108

* **

Section 1.5

The Circle

x 2 + 4x + y 2 - 6y = - 13

1

or 1x + 222 + y 2 - 6y = - 9

1

or 1x + 222 + 1y - 322 = 0 The only way a sum of squares can be zero is if each square term is zero, therefore we have x = - 2 and y = 3. That is, the equation reduces to a single point, namely the point 1 - 2, 32, (or the circle of radius zero centered at 1 - 2, 322. (b) x2 + y 2 + 4x - 6y + 15 = 0 is rewritten as x 2 + 4x + y 2 - 6y = - 15

1

or 1x + 222 + y 2 - 6y = - 11

1

or 1x + 222 + 1y - 322 = - 2 It is impossible for the sum of squares to total a negative number. The sum must be either zero or a positive number. Therefore the equation given defines neither a curve nor a point, it results in a contradiction, The last example indicates that the equation x 2 + y 2 + ax + by + c = 0 can either be a circle, a point or a contradiction. Note that a circle does not describe a function. This may be seen at once from the vertical line test. Drawing a vertical line through the circle intersects the circle at two points with different y-values. Consider the circle centered at the origin with radius 3, x2 + y 2 = 9 solving for y we have y2 = 9 - x2 or y = ; 29 - x 2 (Note that to each x-value there corresponds two y-values.)

Section 1.5

The Circle

* **

109

However, suppose we consider y = 29 - x2, this is the upper half of the circle and to each x-value their corresponds one y-value. This half of the circle does indeed define a function. Its graph is given in Figure 4.

Figure 4: The Graph of y = 29 - x2 The graph of y = - 29 - x 2 also defines a function. Its graph is given in Figure 5. Thus, a circle does not define a function, but its upper or lower halves taken separately do.

Figure 5: The Graph of y = - 29 - x2 The tangent line to a circle at a point P on its circumference is the line perpendicular to the radius of the circle at P. In Figure 1, The line T is the tangent line which is perpendicular to the radius of the circle at the point P. While the tangent line to a circle has the property that it is perpendicular to the radius at the point of tangency, it is not this property which generalizes to other curves. We shall make an observation about the tangent line to the circle which is carried over to other curves, and may be used as its defining property. Let us look at a specific example. Let the equation of the circle in Figure 6 be x2 + y 2 = 25. We may easily determine the equation of the tangent line to this circle at

Tangent Line

110

* **

Section 1.5

The Circle

the point P(3,4). First, we observe that the radius is a segment of the line passing through the origin (0, 0) and P(3,4), and its equation is y = 4 3 x (why?). Since the tangent line is perpendicular to this line, its slope is - 3/4 and passes through P(3, 4), using the pointslope formula, its equation is found to be y = -3 4x +

25 4

Let us compute y-values on both the tangent line and the circle for x-values near the point P(3, 4). Note that near P, we can solve for the y-value on the upper half of the circle which is found to be y = 225 - x 2 When x = 3.01, we find the y-value on the tangent line is y = - 3/413.012 + 25/4 = Figure 6: A Tangent Line to a Circle 3.9925, while the corresponding value on the circle is y = 225 - 13.0122 L 3.99248. In Table 1, we indicate other corresponding values as we vary x near P. Note the closer we get to the point P, the better the y-values agree, and near the point of tangency, the y-values on the circle and line are nearly the same. This observation motivates an alternative definition of a tangent line to a smooth curve at a point P.

DEFINITION The tangent line to a smooth curve at a point P is the best linear approximation to the curve at that point. That is, among all different lines that touch the curve at the point P, the tangent line is the one whose y-values best approximate the y-values on the curve near the point of tangency.

Notice the definition says nothing about the tangent line being perpendicular to other lines at the point of tangency, as in the case of the circle where it is perpendicular to the radius. The notion of perpendicularity is unique to the circle. (The term smooth used in the definition will be clarified later in this text when you study calculus, where you will discover that some curves do not have tangent lines at points which are sharp. ) Table 1: Comparing the y-values near P(3, 4) on y = - 3 4x +

x 2.90 2.91 2.92 2.93 2.94 2.95 2.96 2.97 2.98 2.99 3 3.01 3.02 3.03 3.04 3.05 3.06 3.07 3.08 3.09 3.10 y = -3 4x + 4.075 4.0675 4.06 4.0525 4.045 4.0375 4.03 4.0225 4.015 4.0075 4 3.9925 3.985 3.9775 3.97 3.9625 3.955 3.9475 3.94 3.9325 3.925

25 4

25 4

and x 2 + y 2 = 25

y = 225 - x 2 4.073082 4.065944 4.058768 4.051555 4.044305 4.037016 4.029690 4.022325 4.014922 4.007481 4 3.992480 3.984922 3.977323 3.969685 3.962007 3.954289 3.946530 3.938731 3.930890 3.923009

Section 1.5

The Circle

* **

111

As an illustration, we indicate in Figure 7, the tangent line to the graph of y = x 3 at the point (2,3). Observe how well the tangent line approximates the curve near the point, that is, the y-values are almost identical. The determination of the equation of the tangent line is studied later in this text.

Tangent Line

y = x3

The observation that a tangent line approximates the curve near the point of tangency is an extremely useful tool, one that is used over and over again in mathematics and is sometimes called linearization or linear approximation. A circle centered at the origin has the equation x2 + y 2 = r2. What happens to the graph if we change the coefficients of the squared terms so that they are not the same? For example, let us consider the graph of the equation 4x 2 + 9y 2 = 36. Such a graph is called an ellipse. Its graph is given in Figure 8.

The Ellipse

4x2 + 9y2 = 36

112

* **

Section 1.5

The Circle

The graph is obtained by finding the four intercepts. By setting y = 0, we obtain 4x2 = 36, and solving we have x = - 3 or 3, thus the points 1 - 3, 02 and (3, 0). Setting x = 0 we obtain 9y2 = 36 or y = - 2 or 2, yielding the points 10, - 22 and (0, 2). More generally, any equation of the form x2 + 2 = 1 is the equation of an ellipse a b centered at the origin. If the fractions are cleared, this equation has the form Ax2 + By2 = C. It is most easily drawn by finding the x and y intercepts and plotting them. The equation 4x2 + 9y 2 = 36 may be rewritten in the standard form by dividing by 36 to obtain x 9 + Example 7 Sketch the graph of the ellipse

2 2

y2

y2 4

= 1. y2 x2 + = 1. 25 9

Solution. Note that multiplying by the LCD which is 225, we could rewrite this equation as 9x2 + 25y 2 = 225. To find the x-intercepts, we set y = 0 and obtain 9x 2 = 225, or x2 = 25, solving, x = - 5 or 5, thus the x-intercepts are 1 - 5, 02 and (5, 0). To find the y-intercepts, we set x = 0 and obtain 25y2 = 225, or y 2 = 9, solving, y = - 3 or 3, thus the y-intercepts are 10, - 32 and (0, 3). The graph is plotted in Figure 9.

y2 x2 + = 1 25 9

When we examined the graphs of the two circles x 2 + y 2 = r 2 and 1x - h22 + 1y - k22 = r2, we saw that they were identical except the first one is centered at the origin and2 the second is centered at (h, k). Now consider the graphs of x2 y2 1x - 12 1y - 222 + = 1 and + = 1, they are also identical except that the second 25 9 25 9 one has its center 1x - h22 1y - k22 at the point (1, 2), see Figure 10. Thus, in general, + = 1 is identical to 2 2

x2 2 a

y2 b

2

= 1 except its center (and all its other points) are moved h units in the x-direction

and k units is the y-direction. We shall leave the analysis of these translated ellipses to the exercises.

Section 1.5

The Circle

* **

113

Consider the problem of comparing the y-values on the tangent line near the point of tangency with the y-values on the circle. One way of automating this calculation is to

use the Table feature of the calculator. What we do is define y1 to be the equation of the tangent line and y2 to be the equation of the circle. For the example considered above, 25 2 y1 = - 3 4 x + 4 and y2 = 225 - x . We must first go to the Y = Window 1*F12 and define these functions. See Figure 11. We next set up the table (*F5 or Apps then 5). A table is displayed, but we need to determine the proper x-values. Since we want to examine these two functions near x = 3, we will choose out starting point to 2.8 with an increment 0.1. We do this by pressing F2 Setup and then entering 2.8 in tblstart and 0.1 for tbl and then press Enter. See Figure 12. Alternately, a more extensive table could be generated using a spreadsheet like Excel. That is how the larger table given in Table 1 above was generated. To sketch the graph and the tangent line together, is a simple matter on the calculator as they are already entered in the Y = screen. Choose an appropriate window and they are both drawn. You can trace either, to see their value appear on the screen. To trace, press F3, the q p cursor arrows move you from one curve to the other, while the : ; move the cursor along the selected curve.

Calculator Tips

Figure 11

114

* **

Section 1.5

The Circle

In exercises 1 15 determine the center and radius of the given circle and sketch its graph. 1. 1x - 322 + 1y - 222 = 4 2. 1x - 422 + 1y - 222 = 9 3. 1x + 422 + 1y - 322 = 16 4. 1x - 322 + 1y + 222 = 25 5. 1x + 12 + 1y + 42 = 4 6. 1x + 322 + 1y + 222 = 9 7. x2 + y 2 = 4 8. x2 + y 2 = 25 9. x2 + y 2 + 6x - 10y + 9 = 0 10. x2 + y 2 - 2x + 4y + 1 = 0 11. x2 + y 2 + 8x - 6y + 9 = 0 12. 2x2 + 2y 2 - 6x - 10y + 9 = 0 13. 9x2 + 9y 2 - 12x + 24y - 101 = 0 14. 16x2 + 16y 2 - 48x + 8y - 27 = 0 15. 36x2 + 36y 2 - 48x + 60y - 283 = 0 In exercises 16 25 determine whether the given equations is a circle, a point, or a contradiction (no real graph). 16. x + y + 1 = 0 17. 1x - 12 + 1y + 22 + 4 = 0 18. 1x + 422 + 1y + 322 + 9 = 0 19. 1x + 42 + 1y - 32 - 9 = 0 20. 1x + 322 + 1y - 222 = 0 21. 1x - 322 + 1y + 222 = 0 22. x2 + y 2 - 2x + 4y + 12 = 0 23. x2 + y 2 + 4x - 6y + 13 = 0 24. x2 + y 2 - 10x + 4y + 29 = 0 25. x2 + y 2 + 8x + 6y + 16 = 0 26. Under what conditions on a, b, c, and d is ax2 + ay 2 + bx + cy + d = 0 (a) a circle, (b) a point, (c) a line. (d) a contradiction? 27. Let y = f1x2 describe the upper half of the circle x2 + y 2 = 16. Determine (a) f1 - 22, (b) f(0), (c) f(2). 28. Let y = g1x2 describe the lower half of the circle x2 + y2 = 16. Determine (a) g1 - 22, (b) g(0), (c) g(2). 29. Let y = f1x2 describe the upper half of the circle x2 + y 2 + 8x - 6y + 9 = 0. Determine (a) f1 - 32, (b) f1 - 62, (c) f1 - 12.

2 2 2 2 2 2 2 2

30. Let y = g1x2 describe the upper half of the circle x2 + y2 + 8x - 6y + 9 = 0. Determine (a) g1 - 32, (b) g1 - 62, (c) g1 - 12. 31. Let y = f1x2 describe the upper half of the circle x2 + y 2 + 6x - 10y + 9 = 0, and y = g1x2 the lower half of the circle. Determine (a) f1 - 42, (b) g1 - 42, (c) f(1), (d) g(1). 32. On the same coordinate axes, sketch the graph of: (a) x2 + y2 = 4 (b) 1x - 222 + 1y + 322 = 4, (c) 1x + 122 + 1y - 422 = 4 How are the graphs related. 33. (a) Determine the equation of the tangent line to the circle x2 + y2 = 169 at the point (5, 12). (b) Compare the y-values on the tangent line with those on the circle near x = 5. 34. (a) Determine the equation of the tangent line to the circle x2 + y2 = 25 at the point 13, - 42. (b) Compare the y-values on the tangent line with those on the circle near x = 3. 35. (a) Determine the equation of the tangent line to the circle x2 + y2 = 169 at the point 112, - 52. (b) Compare the y-values on the tangent line with those on the circle near x = 12. In exercises 36 41 sketch the graph of the given ellipse, labeling all intercepts. 36. 37. 38. 39. y2 x2 + = 1 9 25 y2 x2 + = 1 25 9 y2 x2 + = 1 4 16 y2 x2 + = 1 16 4

40. 9x2 + 4y 2 = 36 41. 4x2 + 9y 2 = 36 42. Let y = f1x2 denote the upper half of the ellipse y2 x2 + = 1, and 25 9 y = g1x2 its lower half. Determine (a) f(2), (b) g(2), (c) f1 - 12, (d) g1 - 12.

- 122 4 1x + 422 16

+ +

1y + 222 9 1y - 322 4

= 1. = 1

Section 1.6

Economic Functions

* **

115

Supply Function Demand Function Market Equilibrium Revenue, Cost and Profit Functions Marginal Functions Calculator Tips

In elementary microeconomic theory there is the fundamental assumption that, all other things being equal, the supply of a given commodity is a function of the price of the commodity. In particular, it is assumed that as the price increases, the producers will be willing to provide more of the item and hence the supply will rise. We would say that supply is an increasing function of price. In mathematical notation, let x be the number of units of the commodity supplied and let p be the price per unit, then x = S1p2 says that the amount supplied is a function of price. Now as p increases, so should x. That is, if p2 7 p1, then S1p22 7 S1p12. Graphically, this would mean that plotting x (horizontally) versus p (vertically), as we move from left to right the graph goes upward (see Figure 1).

p

Supply Function

p2 p1

x x1 x2

In the simplest case, S(p) is a linear function with positive slope. For example, suppose that the relation between price and quantity is p - 3x = 42 p = 3x + 42 Which is a straight line with positive slope 3. That is, we could think of x as an increasing function of p or we could think of p as an increasing function of x. There are applications in which one form might be preferred to the other. However, it is traditional to plot this graph with x on the horizontal axis and p on the vertical axis. We must also require that the values of both p and x remain non-negative (Why?). That is, the graph of the supply function is restricted to the first quadrant. So, in our example, the relationship is p - 3x = 42, x 0, p 0

116

* **

Section 1.6

Economic Functions

p - 3x = 42

Figure 2: Linear Supply We see that when x = 0, p = 42. Of course, you should read this the other way: when the price is 42 (or less) there will be no supply. In other words, at price less than 42, no supplier will bother to produce the commodity in question. This is very typical of a linear supply function, which must be of the form p = mx + p0 where p0 is the lowest price for which there will be any supply. Interestingly, we note that you could just as easily have solved for p in terms of x. Thus, Solving for x in terms of p yields 3x = p - 42 x = 11/32p - 14 This is a linear supply function with positive slope 1/3. In this rudimentary theory, there is also a relationship between the amount that can be sold and the unit price. This relationship is called the demand equation. When solved for x in terms of p, that is, in the form x = P1p2, it is called the Price function; When solved for p in terms of x and written in the form p = D1x2, it is called the demand function. There is a basic assumption that for this relationship, as p increases, x should decrease. That is, x is a decreasing function of p (see Figure 3). In symbols, we would write it p1 7 p2 then P1p12 6 P1p22. Again, not only is x a decreasing function of p, but if viewed from the other perspective, p is a decreasing function of x. As above, normally one plots p as a function of x. A typical example of a linear demand equation might be 2x + p = 72, Solving for x, x = 36 - 11/22p

x x1 x2

Demand Function

p1 p2

x 0, p 0

p = 72 - 2x

Section 1.6

Economic Functions

* **

117

that is, D1x2 = 72 - 2x. In either form, it is easy to see that the graph is a straight line with negative slope (see Figure 4)

p = -2x + 72

For this simple case of linear demand, there appear to be two significant values. When x = 0, p = 72. The price, p = 72, would appear to be the highest price possible. That is, at a price greater than or equal to 72, no one would be interested in buying this commodity. Similarly, when p = 0, x = 36. This amount, x = 36, would seem to be the maximum demand. That is, even if free, the most you could give away of this commodity is 36 units. Most economists feel that such upper and lower bounds on price, supply, and demand do not provide a realistic model. Therefore, although linear models are used to approximate more complex ones or to simplify computations, they should not be given much credence near their end points at the axes. It is easy to identify the supply or demand equation when each is linear. The supply equation must have positive slope since it is an increasing function, while the demand equation has negative slope because it is a decreasing function. We shall see that for nonlinear supply and demand equations this identification may be generalized. When supply and demand equations are plotted on the same set of axes, their intersection is called the point of market equilibrium. It is characterized by the fact that at the equilibrium price, supply and demand will be equal. If the price goes above this value, the supply will increase and the demand decrease, causing an oversupply and forcing the producers to lower their price. Similarly, if the price is below the equilibrium price, there will be a shortage followed by an upward adjustment in the market. Of course, real markets are always changing so things are rarely stable at an equilibrium point. However, in theory, the market is always trying to reach its equilibrium position. In order to find this position, it is necessary to solve the two equations simultaneously. If the equations are not linear, this can be quite difficult. For linear equations such as the ones given above, it is easy, as will now be illustrated. Example 1. Suppose the market for humidifiers is governed by the two equations

Market Equilibrium

118

* **

Section 1.6

Economic Functions

Supply: p - 3x = 42 Demand: p + 2x = 72 where p is in dollars and demand is in thousands of humidifiers. Find the point of market equilibrium. Solution. We solve simultaneously. Subtracting the demand equation from the supply equation, we get - 5x = - 30 so x = 6. Then, substituting into the supply (first) equation we get p - 3162 = 42 p - 18 = 42 p = 60 (Substitution into the demand equation would also yield p = 60, as can be verified.) Thus, the market is in equilibrium when the price is $60 per humidifier, at which time we will sell 6,000 humidifiers 1x = 62. The graph is shown in Figure 5.

p 3x = 42

p + 2x = 72

Figure 5: Finding Market Equilibrium It might happen that neither the supply nor the demand function is linear, as the next example illustrates. Example 2 Suppose the demand and supply functions are as follows: Demand: p + x2 = 12 Supply: p - x2 = 4 Determine Market Equilibrium. Solution. and We can solve each equation for p, yielding p = 12 - x 2 p = 4 + x2

Section 1.6

Economic Functions

* **

119

by substitution, we have 12 - x2 = 4 + x 2 or 2x2 = 8 x2 = 4 or x = 2 1remember, p 0 and x 02 substituting gives, p = 4 + 1222 = 8 thus, Market Equilibrium occurs at (2, 8), see Figure 6.

In the above example, did we need to specify which was the demand and supply curve?

p = 4 + x2

p = 12 - x2

Figure 6: Market Equilibrium for Non-Linear Demand and Supply We next examine the other common functions arising in Economics: revenue, cost and profit. The revenue is the producer s income derived from the sale of his product; namely, you multiply the number of items sold by the price per item. The price p paid for the product by the consumer is given from the demand equation p = D1x2. Thus, the total revenue function R is

R = xp

or

(1a)

R = xD1x2

(1b)

For example, suppose the demand equation is linear, as in Example 1, p = 72 - 2x, then the Revenue function

120

* **

Section 1.6

Economic Functions

R = xp = x172 - 2x2 = 72x - 2x 2 So, for example, when x = 10, p = 52 and the revenue is 520. (If x represented the actual number of units and p was the price in dollars, then the revenue is $520.) A sketch of the revenue function is given in Figure 7.

Figure 7: R1x2 = 72 - 2x2 A useful question to ask is when is the revenue maximized and what is the price and revenue when it is maximized. We have already examined this particular problem in Section 1.4. We recognize the graph of the revenue function to be a parabola, we know its maximum occurs at the vertex which is x = - 1722/21 - 22 = 18 at this value for x we have p = 72 - 21182 = 36 and R = 181362 = 648. Consider the problem when demand is non-linear as in the next example. Example 3 Suppose the demand equation is given as p = 28 - x, determine the revenue function, sketch its graph and estimate the point at which the revenue is maximized. Solution. We observe the domain is determined by the condition that 8 - x 0, or x 8. The revenue function is R = xp = x 28 - x This is a nonlinear function which we plot by choosing x values between 0 and 8 to calculate the corresponding R-values and then plot the points. The graph is given in Figure 8. Since this function is not parabolic, we have no simple formula to determine its turning point. Looking at the graph, the best we can do now is estimate that the revenue is maximized when x is between 5 and 6. Later on in the text, we will learn how to use the calculus to determine this point exactly.

In Section 1.3 we examined a linear cost function. Cost need not be linear, it may take many different forms. Let us assume that we are either given data by which we may determine the cost function or it is determined for us. For example, suppose the cost function in dollars is given by the equation C1x2 = 2x 2 - 200x + 100,000 when x 0. This is an example of quadratic cost. Once again observe that even if the producer produces no items his cost is $100,000, that is C102 = 100,000. This makes perfectly good sense as there

Section 1.6

Economic Functions

* **

121

Figure 8: R = xp = x 28 - x are expenses arising even before production begins, for example purchase of machinery and labor. This cost is sometimes called the overhead or fixed cost. A typical problem is to minimize the producer s cost. This particular example is quadratic so we could easily determine the minimum cost. However, the best we could do, at this time, when the cost function is not quadratic, is to estimate the minimum cost, as we illustrate in the next example. Example 4 Suppose the cost function is given by the equation C1x2 = 80x + 400/x + 500, x 7 0. Sketch its graph and estimate the point at which the cost is minimized. Solution. By selection a reasonable number of points (or by using the graphing capabilities of your calculator) we sketch the graph given in Figure 9.

Figure 9: C1x2 = 80x + 400/x + 500 From the sketch, it appears that the minimum occurs at an x-value between 2 and 2.5.

We saw that the overhead figures substantially into the cost function. Economists define a related function which averages the overhead cost among all the items produced; it is called the average cost, denoted by C. The definition is

122

* **

Section 1.6

Economic Functions

C1x2 =

C1x2 x

(2)

Thus, the average cost is the total cost divided by the number of items produced. Example 5 Determine the total and average cost of producing 50 items if the total cost is given by the equation C1x2 = 2x2 - 200x + 100,000. Solution. We have, C1502 = 215022 - 2001502 + 100,000 = 95,000 while C1502 = 95000/50 = 1900.

Which of the two costs do you think is more useful to use? Why? Profit is the difference between the money you take in (revenue) and the money you expend (cost) therefore, if we represent profit by P, we have

(3)

Usually, the producer s objective is to maximize the profit. Once again, except for very simple profit functions, we are unable, at this time, to determine the exact level of production, that is, the x-value, that results in the maximization of profit. We shall leave that determination to later on in this text. For now, we shall be content with determining the profit function and estimating its maximum as illustrated in the next example. Example 6 The demand for some commodity is given by the equation p = D1x2 = 28000 - 2x, while the production cost is given by the equation C1x2 = x2 - 1000x + 20000. Determine the Profit function and determine the profit and price charged when the level of production is 800 items. (Assume the price and cost are given in dollars) Solution. The profit function is P1x2 = x 28000 - 2x - 1x2 - 1000x + 200002 or P1x2 = x 28000 - 2x - x2 + 1000x - 20000 Therefore, when x = 800, P18002 = 800 28000 - 218002 - 180022 + 100018002 - 20000 = $204,000 The price charged p = 28000 - 218002 = 26400 = $80 A sketch of the profit function is given in Figure 10,

Section 1.6

Economic Functions

* **

123

P(x)=

- x2 + 1000x - 1000

Figure 10: P1x2 = x 28000 - 2x - x2 + 1000x - 1000 Note that our graph suggests the profit is maximized when the level of production, x, is about 550. The exact value of x may be found using calculus, as we shall see later on in this text.

Example 7 BBC Rentals Inc. owns a large apartment complex containing 400 identical apartments. If the monthly rent for each apartment is $800 all the apartments are rented. However, for each $60 monthly increase in rent 3 apartments become vacant. Find an expression which may be used to represent their monthly income and determine the monthly rent to be charged to maximize their income. Solution. We give two different approaches to the solution to this problem. Observe their monthly income is the revenue derived from the rent, therefore, the monthly income is the product of the monthly rent and the number of rented apartments.

METHOD 1. Table 1 indicates what is happening with regard to rent and the number of rented apartments if there are 1, 2, 3 increases of $60. We are looking for a pattern for n monthly increases in the rent.

Table 1

Monthly Rent 800 No. of Rented Apartments 400

+ + + +

= = = =

= = = =

Thus, we see the monthly revenue (income) is the product of the monthly rent and number of rented apartments, the R = 1800 + 60n21400 - 3n2 = - 180n2 + 21600n + 320000 Fortunately for us, the graph of R is a parabola, opening downward, so we know the maximum for R occurs at its vertex which is at n = - 21600/21 - 1802 = 60

124

* **

Section 1.6

Economic Functions

Thus the monthly rent to be charged is 800 + 601602 = $4,400. Note, the number of rented apartments is then 400 - 31602 = 220, that is, there are 180 vacant apartments when the monthly income is maximized.

METHOD 2 For every $60 increase in rent, 3 apartments become vacant. This is a linear relations between the rent p and the number of rented apartments x. so in particular when we have the following two points on the line (400, 800) the number of rented apartments when there are no increases, and (397, 860) when there is one $60 increase. Thus, the slope is 1860 - 8002/1397 - 4002 = 60/1 - 32 = - 20. Using the point slope formula, we have

p - 800 = - 201x - 4002 or p = - 20x + 8800 The monthly revenue is R = xp = x1 - 20x + 88002 = - 20x 2 + 8800x This is a parabola opening downward so the maximum value for R occurs at its vertex which is at 8800 x = = 220 21202 thus 220 apartments are rented when the revenue is maximized and the monthly rent is p = - 2012202 + 8800 = $4400.

In the above example, how is x related to n? Note that we were able to solve this problem exactly because it was a quadratic function. If it were not, then the best we could presently do would be to find an approximate solution using the methods described below. In Economics, a function based on another function is also defined, namely the marginal function. The marginal revenue, cost and profit are usually considered. For example, the marginal cost is defined by

MC = C1x + 12 - C1x2

(4a)

Marginal Functions

Its interpretation is clear, C1x + 12 is the cost of producing x + 1 items, C(x) is the cost of producing x items, so the marginal cost, their difference is the cost of producing the x + 1 item. Similarly the marginal revenue would be the revenue derived by producing the last item. More generally, let E(x) represent the equation of any realistic economic function then its marginal value is

E1x + 12 - E1x2

(4b)

Calculator Tips

Later in this text, when we study applications of the calculus, we shall give another definition of a marginal function that is virtually identical with (4). The calculator can be used for determining the intersection of two or more graphs (equations). There are two approaches, one is to use the solve command. Consider Example 1 where we needed to find the simultaneous solution to y - 3x = 42 and y + 2x = 72 (note we replaced p by y since y requires fewer keystrokes on the calculator). Entering, on the Home screen

Section 1.6

Economic Functions

* **

125

solve 1y - 3x = 42 and y + 2x = 72, 5x, y62 (Recall the space bar is Alpha 1 - 2). Note that braces enclose x and y. Similarly, if we want to solve the pair of equations in Example 2, we enter solve 1y + x2 = 12 and y - x2 = 4, 5x, y62 You will observe the calculator also gives the negative solution for x. To indicate that x 0 and y 0, you could enter instead solve 1y + x2 = 12 and y - x 2 = 4, 5x, y6 x 0 and y 02 Recall is obtained by pressing * # , the key is to the left of the 7 key. An alternate approach is to have the calculator graph the functions and then deter- Figure 11: Defining the Functions mine the point of intersection. To do Example 2, we solve for y and in the Y = screen we let y11x2 = 12 - x2 x 0, and y21x2 = 4 + x2 x 0. We inserted the condition x 0, since we recognize these graphs as parabolas and they are also defined for negative values of x. See Figure 11. Next we check our Window 1* F22 to make sure it produces a useful sketch in the first quadrant, see Figure 12. We can now press Graph 1* F32 to get the sketch of the two graphs, see Figure 13. We can now use the calculator to determine the point of intersection. Pressing the F5 key gives a screen of options, see Figure 14. We are interested in the fifth one, the Intersection. Scroll down and press Enter. The calculator indicates a point on the first curve, press Enter, then it indicates a point on the second curve, press Enter again. Now it asks for a Lower Bound, move the cursor to any point on the curve to the left of the intersection point (as you move it indicates the point on the curve on the bottom of the screen as Figure 12: Setting the Window xc and yc, press Enter when you are to the left of the intersection. It then asks for an Upper Bound; move the cursor to any point on the graph to the right of the intersection point and press Enter. The calculator produces the intersection point on the bottom of the screen, in this case, xc = 2 and yc = 8.

NOTES: (1) The reason the calculator asks you to confirm the curves, is that you may have more than two curves on the screen and the calculator needs to confirm which curve s intersection(s) you want to determine.

(2) It asks for a Lower Bound, a point to the left of the intersection, and an Upper Bound, a point to the right of the intersection so as to sandwich in the correct intersection. It might be that there are several points of intersection, this produces the desired one. In this example, the calculator produced the exact point of intersection, often, it will give a numerical approximation. (This is also true of the solve command.) The calculator can be very helpful in estimating the points at which functions achieve their maximum or minimum values. Later on, as you study the calculus, you will learn the exact method for locating such points. There are several ways we may obtain the approximate values, either by creating a table of values for the given function or by sketching the graph of the function. We illustrate by considering the problem of estimating the minimum cost for the function defined by the equation C1x2 = 80x + 400/x + 500 with x 0. We enter the equation in the Y = screen, see Figure 15. We then have the calculator sketch the graph, if the window is not a good one, ZOOM (press F2) and select ZOOMFIT. You should now have a picture of the graph, it may not be a very good one, but by tracing the graph (F3) you should see the lowest point on the graph is between x = 2 and 3. You could then refine your window and select values that give a better window and improve upon your estimate. Try this and see how well you estimate the x-value at he minimum point.

Figure 14

126

* **

Section 1.6

Economic Functions

Alternately press APPS and select TABLE. Press SETUP (F2) set tblStart = 2 and tbl = .1. and press Enter twice. Scrolling down the table the y1 value is minimized around x = 2.1. If you want to improve upon this estimate go back and redefine your table, starting at 2 and take the change tbl to 0.01 or smaller. The third method: with the graph displayed on the Y = screen, press F5 and then scroll down and select Minimum. As before, use the lower and upper bound to sandwich the minimum point on the graph and then press Enter to obtain the coordinates of the minimum. Figure 15

In Exercises 1 -7 (a) Determine if the given equation is a supply or demand equation. For a supply equation, find the minimum price for which there will be any supply. For a demand equation, find the maximum possible demand and the maximum price that can be charged. (b) Plot the graph of the given equation. 1. x + 4p = 16 2. 2p - 3x = 16 3. 2p + x - 12 = 0 4. 500p + 700x - 11000 = 0 5. 400p = 1000x + 16000 6. 2p - 8x2 = 20 7. 3p + 6x2 = 12 In exercises 8 12, you are given a pair of equations, one representing a supply curve and the other a demand curve, where p is the per unit price for x items. In each exercise (a) sketch their graphs on the same set of axes, (b) identify which is the supply and demand curve and the appropriate domain, (c) determine the coordinates of market equilibrium, (d) determine the revenue equation and (e) determine the revenue at market equilibrium. 8. 2p + x - 300 = 0, 9. - 5p + 2x + 70 = 0, 10. - 7p + 2x + 7 = 0, 11. p = 2x + 14, 12. x + 6x - 384 + 12p = 0

2

17. The maximum possible demand for a certain commodity is 20,000 tons. The highest price for which there is any demand is $40 per ton. If the demand equation is linear, find the demand function and the price function. 18. When the price of a certain farm product is 40 cents per bushel, the demand will be for 6 million bushels. If the price increases to 60 cents per bushel, the demand will be for 3 million bushels. If the demand equation is linear, find the demand function and the price function. Express x in millions of bushels. 19. When the price of a New York City bus ride was 35 cents, the number of paying riders per month was about 50 million. When the price was increased to 50 cents, ridership dropped 20%. Find the demand equation assuming that it is linear. Plot the graph and determine the maximum monthly ridership if bus rides were free. Does a linear model seem appropriate near p = 0? In Exercises 20 25 you are given a pair of supply and demand equations; identify which is supply and which demand. Plot both on the same set of axes and find the point of market equilibrium. 20. 2p - 3x = 17 2p + x = 30 21. p + x = 100 2p - 7x = 60 22. p - 2x2 = 1 p + 4x2 = 25 23. p - 3x - 200 = 0 2p + 5x - 950 = 0 24. 3p + 15x - 120 = 0 p = x + 6 25. p - 5x2 = 3 p + 3x2 = 75 26. The supply equation for a certain commodity is x = 4p - 3. The demand equation is 2x + 4p = 15. The commodity suddenly increases in popularity which is reflected in an upward adjustment in the demand curve. This adjustment creates a new demand curve, parallel to the old one and such that at every price the demand is 3 units higher than previously. Find the new demand equation. How does this change effect the point of market equilibrium? 27. Suppose, in Example 7, it costs the BBC realty company, on the average, $800 each month to maintain each vacant apartment. What monthly rent should they charge if their profit is to be maximized? 28. Suppose the cost of producing x items is given by the equation C1x2 = x2 - 3x + 2300, where x represents the number of items in hundreds, and C the cost in dollars. Determine (a) the overhead cost, (b) the cost

2

13. In a small town it has been discovered that the relationship between the number of radios demanded varies linearly with its price. When the price per radio was $15, 1000 radios were demanded; when the price was $30, 700 radios were demanded. (a) Determine the price per radio when 500 radios were demanded, (b) above what price will the demand be zero? (c) Why isn t the demand infinite when the price is zero? 14. The supply of wood to a mill is found to vary linearly with the price per cord. When the cord price is $50, ten cords are supplied; when the cord price is $100, thirty cords are supplied. (a) How many cords are supplied when the price per cord is $175. (b) What is the lowest price above which cords will be supplied? 15. Find the linear supply function satisfying the following conditions: the minimum price for which there is any supply is $3. For every dollar increase in price, the supply will go up 2.5 units. 16. Find the linear supply function satisfying the following conditions: when the price is $4, there will be no supply. When the price is $11, there will be 600 units supplied.

Section 1.7

of producing 100 items, (c) the cost of producing the 100th item (the marginal cost at x = 100), (d) the average cost function, (e) the average cost of producing 100 items, (f) the level of production which minimizes the cost and the minimum cost. 29. Suppose the cost, in dollars, of producing x items is C1x2 = 100x + 2000/x. Determine (a) the cost of producing 50 items, (b) the cost of producing the 50th item, the marginal cost when x = 50, (c) the average cost function, (d) the average cost of producing 50 items, (e) estimate the level of production which minimizes the cost and estimate the minimum cost. 30. Suppose the demand equation is 200x + 10p = 10000, where x is the number of items sold when the per unit price is p dollars. Determine (a) the revenue function, (b) the domain of the revenue function, (c) the revenue derived from the sale of the 20th item the marginal revenue when x = 20, (d) the level of production that maximizes the revenue, and the maximum revenue. 31. Suppose the demand equation is p = 22000 - 5x, where x is the number of items sold when the per unit price is p dollars. Determine (a) the revenue function, (b) the domain revenue function, (c) the revenue derived from the

More on Functions

* **

127

sale of the 100th item the marginal revenue when x = 100, (d) estimate the level of production that maximizes the revenue, and the maximum revenue. 32. When the price for a color television is $240, the monthly sales for this item at a department store is 450. For each $5 increase in price, the monthly sales fall by 5 units. How many color televisions should be sold to maximize monthly sales and what price should be changed to maximize sales? 33. Suppose it has been determined that the demand (in thousands of dollars) for 20 a certain item is given by the equation p = , and the cost of producing x 1x hundred items is given by C1x2 = 5x + 10. Determine (a) the revenue function, (b) the Profit function, (c) estimate the level of production x, at which profit will be maximized, (d) from your estimate in (c), what price should be charged to maximize profit? 34. In Example 7, find the relationship between n and x in the two methods used to solve the problem.

Using the Zeros Even Functions Symmetry about the y-axis

Odd Functions Symmetry about the Origin Rational Functions Vertical Asymptotes Horizontal Asymptotes Translations Calculator Tips

One of the most effective techniques for sketching the graph of a function is the calculus. At this point in the text we have not yet developed the machinery to allow us to use this technique. However, we can use some simple techniques like factoring and sign analysis of the function (reviewed in Section 0.5) to get some information about the function. These techniques will usually not yield all the information we require but we will resolve these deficiencies later. Consider the quadratic function defined by the equation f1x2 = 6x 2 - 5x - 6. Of course we know this is a parabola, and can sketch its graph very well using the methods of Section 1.4. Let us, for the moment, pretend not to recognize this function. Observe, that this particular function may be easily factored, and we have f1x2 = 12x - 3213x + 22 This factored form immediately provides the x-intercepts (also called the zeros) of the function, namely, x = 3/2 and x = - 2/3. That means that the only time the value of this function is zero is at these two x-values, everywhere else it must be either positive or negative. Using the sign analysis of Section 0.5, we have Figure 1.

128

* **

Section 1.7

More on Functions

0 + 2/3

0 + 3/2 x

Figure 1: The sign of 12x - 3213x + 22 This tell us the y-values are positive when x is less than - 2/3 or greater that 3/2 and negative elsewhere. Therefore, a graph illustrating these properties is given in Figure 2. However, this information does not yield the coordinates of the turning point which of course, is the vertex. Nor would we know, if we did not recognize this graph that it did not

Figure 2: A Representation of f1x2 = 12x - 3213x + 22 Using Sign Analysis have additional turning points. The best we can do at this point is to graph a possible representation of a function, with the understanding that, as your mathematical development increases, more information about functions will become available. This idea may be extended to higher order polynomial functions provided we can determine all the zeros of the function. Once the zeros are located (meaning we can factor the polynomial), we analyze the sign of the function and draw a possible graphical representation. Once again, note the use of the word possible. There is still a lot we do not know about the function, for example its turning points or how it increases or decreases, does it do it quickly or slowly? These questions are most effectively answered using the calculus. We consider a cubic equation in the next example. Example 1 Sketch the graph of f1x2 = x3 - 9x. Solution. This cubic may be easily factored. We have f1x2 = x1x 2 - 92 = x1x - 321x + 32 Therefore, the zeros of the function are at x = - 3, 0 and 3. Using sign analysis we have Figure 3.

0 -3 + 0 0 0 3 + x

Section 1.7

More on Functions

* **

129

Therefore, the y-values are negative when x 6 - 3 or when 0 6 x 6 3 and positive when - 3 6 x 6 0 or x 7 3. A sketch of a graph satisfying these conditions is given in Figure 4.

Once again, it is clear that we cannot yet determine the exact coordinates of the turning points of this graph which we shall see occur when x = ; 23. What do we do when we cannot factor the polynomial? At this point, not very much; however, the calculator can be very helpful in such cases, as we shall see later on this section. Sometimes, it may happen that a function has no zeros, what does this mean? If it has no zeros then the sign of the function must either always be positive or always negative (why?). Therefore its graph must either be above the x-axis or below it. Consider the function defined by f1x2 = x 4 + 4x 2 + 4, which may be factored and written as f1x2 = 1x2 + 222.. This function is always positive it has no real zeros therefore, it cannot cross the x-axis. A sketch is given in Figure 5.

While this graph may look like a parabola, it is not. It is flatter near the origin and rises more quickly than a parabola. Because of the symmetry of this graph (an idea we examine next) it follows that the lowest point on this graph is at the y-intercept which is (0, 4).

130

* **

Section 1.7

More on Functions

When we plotted parabolic functions, we considered the axis of symmetry, the vertical line about which, for equally spaced x-values to its left and right, it had the same y-values on the graph. That is, the axis of symmetry was like a mirror, the portion of the graph to its left was a reflection of the graph to its right. We are interested in determining under what conditions functions may have two other kinds of symmetries. Knowing a function has a symmetry is often useful in sketching its graph, as we shall see. The first symmetry we examine is symmetry with respect to the y-axis; functions which have this symmetry are called even functions. If a function whose equation is given by y = f1x2 is to be symmetric with respect to the y-axis, that means for equal and opposite x-values on its graph, the y-values must be the same. This translates into if f1 + x2 * f1x2 then f is an even function (1)

Sometimes, the equation representing the function is not given in this form. We can rephrase the condition as follows: The graph of an equation is symmetric with respect to the y-axis if replacing x by - x leaves the equation unchanged. Example 2 (a) Show the function whose equation is y = f1x2 = x 4 - 3x 2 + 12 is even. (b) Show the graph of the equation y3 - 3x2y - 23 = 0 is symmetric about the y-axis. Solution. (a) f1 - x2 = 1 - x24 - 31 - x22 + 12 = x 4 - 3x2 + 12 = f1x2 satisfying (1). (b) If we replace x by - x we have y 3 - 31 - x22y - 23 = 0 which becomes y3 - 3x 2y - 23 = 0, the same equation we started with. Therefore, the graph of the equation is symmetric with respect to the y-axis.

Do you see why we call functions which have symmetry with respect to the y-axis even functions? For polynomial functions to have this type of symmetry, x must appear only to even powers. However, there are functions which are not polynomials that also exhibit this symmetry, as we illustrate below. If we know a function has even symmetry, that means we need only worry about sketching its graph for x 0, the rest of its graph is just its reflection about the y-axis. Example 3 Assume the function whose graph for x 0 is given in Figure 6 is even, complete its sketch for x 6 0.

Figure 6

Section 1.7

More on Functions

* **

131

Solution. The complete graph is given in Figure 7. Note the portion drawn for x 6 0 is just the mirror image about the y-axis.

Figure 7: Completion of the Graph by Symmetry Non-polynomial functions can also exhibit even symmetry. Consider the absolute value function defined by the equation y = f1x2 = x , it follows that f1 - x2 = f1x2 (why?), therefore this function is even, that is, it is symmetric with respect to the y-axis. You should verify this conclusion by sketching the graph of this function. Another example is 2 the function defined by the equation f1x2 = x3. We also leave as an exercise for you to verify that the product or quotient of even functions is also an even function. Observe that a continuous even function (by continuous, we mean the graph has no holes or jumps ) must have a turning point at its y-intercept (why?) The other symmetry we wish to consider is symmetry with respect to the origin; functions which have this symmetry are called odd functions. If a graph is symmetric with respect to the origin, that means for example, if the point (2, 8) is on the graph, then its mirror image with respect to the origin is also on the graph, in this case 1 - 2, - 82. More generally, if (x, y) is a point on the graph so too is 1 - x, - y2. Now if we define y by y = f1x2, then the y-coordinate at - x is f1 - x2. Thus - y = f1 - x2 or - f1x2 = f1 - x2, multiplying by - 1, we write the condition as if f1 - x2 = - f1x2 then f is an odd function (2)

Sometimes, the equation representing the function is not given in this format. We can rephrase the condition as follows: The graph of an equation is symmetric with respect to the origin if replacing both x by - x and y by - y leaves the equation unchanged. Example 3 (a) Show the function defined by the equation y = f1x2 = 4x3 - 2x is odd. (b) Show the graph of the equation xy - 2x5y 3 = 0 is symmetric with respect to the origin. Solution. (a) f1 - x2 = 41 - x23 - 21 - x2 = - 4x 3 + 2x = - 14x 3 - 2x2 = - f1x2. (b) replacing x by - x and y by - y we have 1 - x21 - y2 - 21 - x251 - y23 = 0 or xy - 2x5y 3 = 0, the same equation we began with. Therefore, it has the required symmetry.

132

* **

Section 1.7

More on Functions

Example 4 The portion of the graph for some odd function is given for 0 8. Complete the sketch giving the graph for - 2 x 6 0.

y

2 is given in Figure

Figure 8 Solution. We need only reflect this graph about the origin to obtain the complete sketch. Note that the reflection of (0.5, 1), (1, 0), 11.5, - 12 and (2, 0) are respectively 1 - 0.5, - 12, 1 - 1, 02, 1 - 1.5, 12 and 1 - 2, 02. See Figure 9.

Figure 9: Completion of the Graph by Symmetry Note that continuous odd functions pass through the origin (why?). Also note that the shape of the curve changes at this point as well. We shall talk about this point more when we consider the calculus. The shape is sometimes called concavity. You might note we did not consider symmetry about the x-axis. Functions cannot exhibit this property (why?). However, the graph of an equation which has this kind of symmetry has the property that replacing y by - y leaves the equation unchanged, and conversely the graph of any equation which satisfies this property is symmetric with respect to the x-axis. We shall examine this more fully in the exercises. p1x2 A function defined by an equation of the form r1x2 = where p(x) and q(x) are q1x2 polynomials is called a rational function. Sketches of such functions illustrate their interesting features. There are two questions that we shall examine: (1) Division by zero is undefined. Suppose there are x-values at which the denominator vanishes (while the

Rational Functions

Section 1.7

More on Functions

* **

133

numerator does not). What happens to the behavior of the functional values (the y-values) at x-values near where the denominator approaches zero? (2) What happens to the functional values as x becomes a very large positive number or a very small negative number (note, - 1,000,000 is a very small negative number compared to - 100)? For example, x suppose r1x2 = 2 , it is clear that this function is defined for all real numbers except x - 4 x = - 2 and 2. How does the function behave near these two points? What happens to the functional values as x becomes very large, say 1,000,000 or 10,000,000 or even larger. What about negative values, say - 1,000,000 or - 10,000,000? We shall soon consider these questions, but first observe for this particular example that first examining its symmetry reveals useful information. Example 5 Discuss the symmetry of the function defined by the equation r1x2 = Solution. We have that r1 - x2 = 1 - x2

2

-x = 2 = - r1x2 therefore the func4 x - 4 1 - x2 tion is symmetric with respect to the origin, that is, it is an odd function.

x . x2 - 4

Note that r(x) is the ratio of two polynomials, its numerator is an odd function and its denominator is an even function, and the ratio of an odd and even function is always odd. We ask you to examine this and other symmetry combinations in the exercises. Let us now examine the first question. Suppose there are x-values at which the denominator of a rational function is zero while the numerator is not zero. In the previous example, when x is either - 2 or 2, the denominator vanishes. The domain of this function consists of all real numbers except these two values. Suppose, however, we substituted either of these values, the result would have the form N 0 Where N is a non-zero number. Such a quotient is said to be either positively or negatively infinite, to see why, suppose, in the above example we examine what happens if we select x-values near 2 and compute the corresponding y-values (we round the y-value to the nearest integer). See Table 1. Table 1: The behavior of r1x2 =

x 1.99 1.999 1.9999 1.99999 1.999999 2 2.000001 2.00001 2.0001 2.001 r1x2 =

x for x near 2 x - 4

2

x x2 - 4

134

* **

Section 1.7

More on Functions

Observe that if we take x-values to the left of 2 and approach 2, the y-values become more and more negative and their absolute value is very large. As we approach 2 from the right the y-values become larger and larger, ultimately becoming infinite as we get closer to x = 2. The function behaves similarly near x = - 2, see Table 2. Table 2: The behavior of r1x2 =

x - 1.99 - 1.999 - 1.9999 - 1.99999 - 1.999999 -2 - 2.000001 - 2.00001 - 2.0001 - 2.001 r1x2 =

x for x near - 2 x - 4

2

x x2 - 4

Thus, near x = - 2 and x = 2, that is the closer we get to these x-values, the larger their y-values are in absolute value. Sometimes we say that y-values become infinite at these points. To illustrate graphically, consider the graph of this function near x = 2, see Figure 10.

x=2

x x - 4

2

Note that we included the vertical line x = 2 in the sketch to indicate what happens as the curve approaches it, as the x-values get close to this vertical line, the y-values become very large (in absolute value). We have a similar behavior of the function near x = - 2 (which does not surprise us because the function is odd, its graph is symmetric with respect to the origin). Figure 11 is a sketch of the graph of the function; note we include the vertical lines x = - 2 and x = 2 to show the functional values become infinite near these lines. These lines are called vertical asymptotes.

Section 1.7

More on Functions

* **

135

x= 2

x=2

x x - 4

2

Note that the above rational function is one kind of an example of a discontinuous function, this function has separate pieces, going from one piece to another on either side of the asymptote results in an infinite jump. More generally, given any function, suppose when x = a the denominator vanishes and the numerator does not, then the line x = a is called a vertical asymptote of the function. (Alternately, the line x = a is a vertical asymptote if the functional values become infinite near x = a.) The domain of a rational function consists of all x-values with the exception of those at which the denominator is zero. Example 6 Determine the vertical asymptotes of the function defined by f1x2 = 2x - 6 . 2x - x - 3

2

Vertical Asymptotes

Solution. The denominator vanishes when 2x2 - x - 3 = 0, or 12x - 321x + 12 = 0. This yields x = - 1 and x = 3/2. Note at either of these values the numerator is not zero, therefore the two vertical asymptotes of the given function are x = - 1 and x = 3/2. The domain of this function is all real numbers except x = - 1 and 3/2.

In the previous example, the numerator is zero when x = 3, that is, when x = 3, the y-value is 0. The zeros of the numerator provide the x-intercepts or zeros of the function. Thus, the zeros of the denominator (which are not zeros of the numerator) are the vertical asymptotes of the function. Consider the rational function defined by the equation 2x + 2 f1x2 = , observe x = - 1 is both a zero of the numerator and denominator. 2x 2 - x - 3 This point is not in the domain of the function since the denominator is undefined there. Nor does the function have a vertical asymptote at this point. If we factor, we have, f1x2 = 21x + 12 12x - 3211x + 12

136

* **

Section 1.7

More on Functions

While you might be tempted to cancel the common factor 1x + 12 from the numerator and denominator, you may do so providing you indicate that after the cancellation x may not equal - 1 (because this value is not in the original function s domain), thus you could write f1x2 = 2 for all x = - 1 2x - 3

Note that if we did not exclude this x-value you might be tempted to write f1 - 12 = - 2/5. In fact the graph of this function has a hole or discontinuity at the point 1 - 1, - 2/52. (We discuss this notion more fully, later on in this text.) We indicated that near a vertical asymptote the y-values become infinite. The above example illustrates that a cancelling zero of the numerator and denominator cancels the infinite behavior of the function near this x-value. Many different types of functions, in addition to rational functions, have vertical asymptotes, the next example illustrates one such case. Example 7 Determine the vertical asymptotes for the function defined by f1x2 = 1x . x - 3 Solution. The domain of this function is all x 0 and x = 3. The denominator is zero when x = 3 and the numerator is not, therefore x = 3 is a vertical asymptote. A sketch of the function is given in Figure 12.

Note that the function defined in the previous example is not a rational function as the numerator is not a polynomial.

x=3

1x x - 3

Horizontal Asymptotes

We now examine the question of how a rational function behaves when the x-values are large, in absolute value. We shall see there are two possibilities; the y-values on the curve level off and approach a horizontal line or they do not. If they do, then the horizontal line they approach is called a horizontal asymptote. The line y = b is said to be a horizontal asymptote of a function if the functional values are near b as the x values get very large, in absolute value. In a sense, we may think of the y-values stabilizing as the x-values get large, sort of a steady state value of the function. It will be helpful to first make some observations. Consider k/x p, where k is a constant and p is any positive number. As x gets very large in absolute value what happens to this quotient? As the denominator gets very large, no matter what the constant value of k

Section 1.7

More on Functions

* **

137

is, the quotient gets very close to zero. Consider, for example 100/x, as x gets larger and larger, say 1000, 10,000, 1,000,000, . . . the quotient becomes, 0.1, 0.01, 0.00001 and so on, similarly if x is - 1000, - 10,000 the quotient also approaches 0. What happens to x p as x becomes a very large positive number? It should be clear p that x is a very large positive number as well. Sometimes we say it approaches positive infinity, written q . When x is a negative number whose absolute value is very large, then the result is infinite as well, but its sign depends on whether or not p is an odd or even integer. For example x 10 will be a large positive number for a negative x-value, on the other hand x11 will be negative for negative values of x. Thus we have the following: If x is a large number in absolute value, p 7 0 and k are constants, then k/xp : 0 and xp may either : + q or - q (The arrow is to be read approaches. ) From now on, instead of stating as x gets very large, we shall write x : q , similarly, instead of writing as x is a negative number which gets very large in absolute value we shall write x : - q . Consider the function defined by the equation f1x2 = 3x4 - 2x 3 - 3x + 11 6x4 + 3x 3 + 2x - 11

Let us examine the behavior of this function as x : q (a very large positive number sometimes we will say as x becomes infinite). With your calculator, you can see what is happening. Try large values of x like 1000, 10,000, 1,000,000 (of course these are finite values, but they give a picture of what is happening to the y-values as x gets large). Table 3 indicates the behavior of the function as x : q . It seems clear from Table 3 that as x : q , the y-values get arbitrarily close to 1*2. There is a simple algebraic procedure by which we may come to this same conclusion. We illustrate two related methods. 3x 4 - 2x 3 - 3x + 11 as x gets large 6x 4 + 3x 3 + 2x - 11

3x 4 - 2x 3 - 3x + 11 6x 4 + 3x 3 + 2x - 11

x 1000 10,000 1,000,000

f1x2 =

METHOD I Factor out the dominant (highest) power from both the numerator and denominator and perform any cancellation that may occur.

3x4 - 2x 3 - 3x + 11 f1x2 = = 6x4 + 3x 3 + 2x - 11 x4 a 3 2 3 11 2 3 11 - 3 + 4b a3 - 3 + 4b x x x x x x = 3 2 3 2 11 11 x4 a 6 + + 3 - 4b a6 + + 3 - 4b x x x x x x

138

* **

Section 1.7

More on Functions

as x : q , we already observed above that every term of the form k/xp approaches zero, that means the only terms left in the above expression are the 3 in the numerator and the 6 in the denominator. That is as x : q , f1x2 : 3/6 = 1*2, thus the horizontal asymptote is y = 1*2.

METHOD 2. As x : q , the dominant term in the numerator is the one with the highest power, all other terms are insignificant with respect to it for very large values of x. Similarly, the dominant term in the denominator, is the term with the highest power, all other terms are insignificant with respect to it, so we have, for very large x-values,

f1x2 = 3x4 3x 4 - 2x 3 - 3x + 11 1 = : 4 3 4 2 6x + 3x + 2x - 11 6x

giving y = 1*2 as the horizontal asymptote. This second method indicates that we need only keep the dominant terms in the numerator and denominator to determine if there are horizontal asymptotes. Note that x is a negative number, very large in absolute value, the above analysis is identical and the curve approaches y = 1*2, that is y = 1*2, is also a horizontal asymptote when x : - q . Example 8 12x 4 - 3x 3 + 17x + 21 Examine the behavior of the function f1x2 = as (a) x : q 5x2 - 2 + 3x 4 and (b) x : - q . Solution.

METHOD I

f1x2 = 12x4 - 3x 3 + 17x + 21 5x2 - 2 + 3x 4 x4 a 12 = 3 17 21 3 17 21 + 3 + 4b 12 + 3 + 4 x x x x x x = 5 2 5 2 - 4 + 3 x4 a 2 - 4 + 3 b x2 x x x 1 17 21 + 3 + 4 x x x : 12/3 = 4 5 2 - 4 + 3 x2 x

METHOD II Keeping the dominant terms in the numerator and denominator, we see that

f1x2 = 12x 4 - 3x 3 + 17x + 21 12x4 : = 4 5x2 - 2 + 3x 4 3x4

The next example illustrates that it is possible for a function to behave differently as x : q and as x : - q .

Section 1.7

More on Functions

* **

139

Example 9 Determine the horizontal asymptotes for the function defined by f1x2 = Solution. Using Method II, we have as x : q or as x : - q f1x2 = recall that 2x 2 = e thus, when x : q (x is positive) f1x2 = 4x 2x + 1

2

4x 2x 2 + 1

4x 2x + 1

2

4x 2x 2

x if x 0 - x if x 6 0

4x 2x

2

4x = 4 x

2

4x 2x

2

4x = -4 -x

Therefore, this function has two different asymptotes, as x : q , y = 4 is a horizontal asymptote and as x : - q , y = - 4 is a horizontal asymptote. It is straight-forward to sketch this function (which is odd). Note it passes through (0, 0) and since the denominator is always positive, the sign of the function will be determined by the sign of the numerator, which is positive for x 7 0 and negative for x 6 0. A sketch of this function is given in Figure 13.

y=4

y= 4

4x 2 x2 + 1

Note the two different horizontal asymptotes should not be surprising as the function in the previous example is odd. It may happen that a rational function may get infinite as x : q or as x : - q , as the next example illustrates.

140

* **

Section 1.7

More on Functions

Example 10. - 3x 5 - 2x 2 + 3x + 11 Investigate the behavior of the function defined by f1x2 = as 2x2 + 5x + 11 x : q and as x : - q . Solution. We have as x : q or as x : - q f1x2 = - 3x 5 - 2x2 + 3x + 11 - 3x 5 -3 3 = x : 2 2 2 2x + 5x + 11 2x

as x : q , we see that a large number to a power is large and multiplying it by a negative number makes it negative, and large in absolute value, that is, x : q , f1x2 : - q . On the other hand, as x : - q we have f1x2 : q . A sketch illustrating the behavior of the function is given in Figure 14. Note how the function increases without bound (becomes infinite) as you move to the left, and decreases as you move to the right. There is no leveling off of the y-values, that is, no horizontal asymptotes.

Figure 14: - 3x5 - 2x2 + 3x + 11 f1x2 = 2x2 + 5x + 11 Example 11 Determine the behavior of the function defined by f1x2 = and as x : - q . Solution. We have, as x : q or as x : - q , f1x2 = 3x2 - 2x + 9 3 3/5 3x2 = = 5 : 7 7 4 5 5x 5x - 11x + 21 5x x

3x 2 - 2x + 9 as x : q 5x7 - 11x 4 + 21

as x : q or as x : - q , k/xp : 0, therefore f1x2 : 0, therefore, in each case, y = 0 is a horizontal asymptote, that is, x : q or as x : - q the graph approaches the x- axis. See Figure 15.

Did you observe in the last two examples, when there is no horizontal asymptote or when the x-axis is a horizontal asymptote, what happens with regard to the relative sizes of the dominant terms in the numerator and denominator? (See Exercise 65.)

Section 1.7

More on Functions

* **

141

Solution. We have a vertical asymptote at x = - 3 (why?), and as x : q or as x : - q , y = 2 is the horizontal asymptote (verify this!). The y-intercept is 10, - 4/32. The zero of the function occurs when the numerator is zero, 2x - 4 = 0, or x = 2. The sign of the function is determined by the sign of the numerator and denominator, following the procedure described in Section 0.5, we have Figure 16 where ND means not defined (a vertical asymptote.)

ND + -3

0 + 2 x

2x - 4 . x + 3

We see to the left of the vertical asymptote x = - 3, the y-values are approaching q (since the sign of the function is positive), to its right they are approaching - q since the sign is negative. We begin our sketch in Figure 17 showing these basic facts, and complete the sketch, connecting the various segments of the graph.

y=2 x = -3

(2,0) (0,-4/3)

Figure 17: Segments of the 2x - 4 Graph of f1x2 = . x + 3 We now connect these segments, remembering that y = 2 is a horizontal asymptote. The graph is given in Figure 18. Note that the graph does not cross the horizontal asymptote, we can verify this by trying to find those points at which f1x2 = 2 (see Exercise 71), there are no solutions to this equation, meaning the graph does not intersect the asymptote.

142

* **

Section 1.7

More on Functions

Figure 18: The Graph of 2x - 4 f1x2 = . x + 3 Examining the sign of the function, we can sketch rational functions, however, the sketch will in general, not provide the location of the points where the graph turns. The calculus will provide those details. The next example illustrates the procedure. Example 13 Sketch the graph of the function defined by the equation f1x2 = 3x . x2 + 9 Solution. We first observe that this is an odd function, so its graph is symmetric with respect to the origin. There are no vertical asymptotes because the denominator is never zero (in fact it is always positive), verify these statements. To find the horizontal asymptote, observe that f1x2 = 3x 3x 3 : 2 = : 0 as x : q or as x : - q x x + 9 x

2

Thus, y = 0 is the horizontal asymptote. We also know the y-intercept is (0, 0). Since the denominator is always positive, the sign of the function will be determined by the numerator which is positive for x 7 0 and negative when x 6 0. We begin our sketch in Figure 19.

Figure 19: Beginning a 3x Sketch of f1x2 = 2 x + 9 Notice the piece of the graph on the right illustrating the fact that y is positive for x 7 0 and as x gets large it approaches the asymptote y = 0. Similarly, the piece on the left illustrates that y is negative when x 6 0, and also has y = 0 as x : - q . The graph

Section 1.7

More on Functions

* **

143

passes through the origin, to the left it is negative, to the right it is positive, giving the middle piece. We need only connect these pieces, which results in the two turning points on the graph. See Figure 20. (At this point, we do not know the coordinates of the turning points. We will have to wait until we study the calculus to find them.)

Observe from Figure 21 that the graph crosses its horizontal asymptote at the origin, in fact, there are functions for which the graph may cross its horizontal asymptotes one or more times; but a graph can never cross its vertical asymptote (why?). See Figure 21.

Figure 21: A Graph Intersecting Its Horizontal Asymptote In general, any function that can be expressed in terms of sums differences, products, quotients, or rational powers of polynomial functions is called an algebraic function. A function which is not algebraic is called transcendental, examples of which are exponential, logarithmic and trigonometric functions. We shall see that the calculus is an essential tool in examining the behavior of such functions and sketching their graphs. The calculator is also a useful tool in obtaining sketches of portions of their graphs.

144

* **

Section 1.7

More on Functions

Consider the graph of the function defined by the equation y = f1x2, how is it related to the function defined by the equation y = f1x - h2? We examined this question previously for the parabola and the circle. For example, consider f1x2 = x 2 - 2x + 3 and f1x - 12 = 1x - 122 - 21x - 12 + 3. If we construct a table of values for each what do we observe (Table 4a and Table 4b)? Table 4a: f1x2 = x 2 - 2x + 3

x -2 -1 0 1 2 3 4 f1x2 = x 2 - 2x + 3 11 6 3 2 3 6 11

x -1 0 1 2 3 4 5 f1x - 12 = 1x - 122 - 21x - 12 + 3 11 6 3 2 3 6 11

Translations

We observe the y-value for each x-value in Table 4b corresponds to an x value which is one less in Table 4a. In other words, the graph representing f1x - 12 = 1x - 122 - 21x - 12 + 3 is identical to the graph of f1x2 = x 2 - 2x + 3 except it is moved one unit to the right. See Figure 22.

Section 1.7

More on Functions

* **

145

Observe the graphs are identical, except the graph on the right is positioned one unit to the right of the graph on the left, that is, each point on the graph on the left is translated one unit to the right. More generally, the graph given by the equation y = f1x - h2 is identical to the graph of the equation y = f1x2 except each point on the first graph is translated h units to the right if h is positive and h units to the left if h is negative. See Figure 23 where we have y = f1x2, y = f1x - h2 and y = f1x + h2 for h 7 0. Observe that the curve on the left at x = a + h, the middle curve at x = a and the curve on the right at x = a - h each have the same y-value.

+* (a+h,b)

+* (a, b)

+* (a -h,b)

y = f(x+h)

y =f(x)

y = f(x - h )

Figure 23: Horizontal Translation y = f1x + h2, y = f1x2 and y = f1x - h2, h 7 0 Next, we ask the question how do the graphs y = f1x2 and y = f1x2 + k differ? It is apparent that once again the graphs are identical with the y-value on the second graph k units above the first if k 7 0 and k units below the first if k 6 0, see Figure 24.

(a, b +k)

+*

y = f(x) + k

(a, b)

+*

y = f(x) y = f(x) k

+*

(a, b k)

Figure 24: Vertical Translation y = f1x2 - k, y = f1x2 and y = f1x2 + k for k 7 0 For example, consider f1x2 = x2 - 2x + 3 and g1x2 = x 2 - 2x + 5 (observe, g1x2 = f1x2 + 2), their graphs are illustrated in Figure 25. Note how the graph of f is translated 2 units upward to obtain the graph of g. We can now combine the two translations, that is, if we are given the graph of y = f1x2 how does the graph of y = f1x - h2 + k differ from it? It should now be ap-

146

* **

Section 1.7

More on Functions

parent that each point on the second curve is translated h units horizontally (to the right if h is positive, to the left it is negative) and translated k units vertically (up if k 7 0, down if k 6 0).

f(x) = x2 * 2x + 3 g(x) = x2 * 2x + 5

We illustrate with an example. Example 13 Given Figure 26 which gives the graph y = f1x2, for some function, sketch the graph of y = f1x + 22 + 3.

(*2, 2)

(*3, 0) (*1, 0)

(2, 0)

(1, *2)

Solution. To sketch the required graph the graph in Figure 27 needs to be move 2 units to the left and 3 units upward. That is, take every point on the original graph and subtract 2 from its x-coordinate and add 3 to its y-coordinate. See Figure 27.

Section 1.7

More on Functions

* **

147

Thus, 1 - 3, 02 : 1 - 5, 32; 1 - 2, 22 : 1 - 4, 52 and so on for the other indicated points. See Figure 27.

Polynomials whose zeros are not easily found are hard to sketch without the calculus, even with a calculator. We first need to determine a reasonable window. Sometimes this can be difficult. However, there is a theorem (whose proof requires the more advanced mathematics studied in Complex Variable Theory) that is useful in determining a reasonable window in studying graphs with a calculator. We state the following theorem. Given the polynomial function defined by the equation f1x2 = xn + a n - 1xn - 1 + a n - 2x n - 2 + + a 2x 2 + a1x + a 0 Let z be any real zero of this function and A the maximum of 5 a n - 1 , a n - 2 , , a2 , a 1 , a 0 6 (that is, compute the absolute value of each coefficient and call the maximum value A) then z 6 1 + A The requirement that the coefficient of the highest term is one presents no problem as dividing a function by a constant does not change the position of its zeros. Thus, for exam3 2 ple f1x2 = 2x 3 - 3x2 - 2x + 3 and g1x2 = x 3 - 3 2 x - x + 2 have the same zeros; g was obtained from f by dividing by 2. Thus, we see that use of this theorem gives a window for the zeros. For example, in the above example, the window for the zeros of f are obtained from g, we have z 6 1 + max5 - 3/2 , - 1 , 3/2 6 = 1 + 3/2 = 2.5 z 6 2.5 means that - 2.5 6 z 6 2.5 This means all the zeros lie within this interval. (Note the zeros are actually x = - 1, 1 and 3/2.) Clearly, when we cannot easily determine the exact zeros, this theorem provides an initial window to use with our calculator. Thus, if we wanted the calculator to sketch f1x2 = 2x 3 - 3x 2 - 2x + 3, we choose a window such that xmin = - 2.5 and xmax = 2.5.

Calculator Tips

148

* **

Section 1.7

More on Functions

Example 14 Illustrate the above theorem for determining a window containing all the zeros fo the function defined by f1x2 = 3x5 - 2x 4 + 12x 3 - 5x 2 - 9x - 6. Use your calculator to sketch the graph. Solution. Since the theorem requires the coefficient of the highest term be one, we divide by 3 to obtain g1x2 = x 5 - 2/3x 4 + 4x 3 - 5/3x 2 - 3x - 2 The coefficient whose absolute value is largest is 4, so we have z 6 1 + 4 = 5. Therefore all zeros of g and therefore f are in the interval - 5 6 z 6 5. Thus we choose as our window as indicated in Figure 28. Note that we kept the default values for y, if we need to change that part of the window, we can do so afterwards.

Figure 28 We next have the calculator graph the function, see Figure 30

Figure 29 (We note that the usual default window would have produced a reasonable graph, but that is not always the case.) We can also sketch rational functions with the calculator, but you need to be careful near the vertical asymptotes, this can be dealt with by choosing a reasonable window in which to view the graph.

Section 1.7

More on Functions

* **

149

In Exercises 1 11, determine the zeros and then, using the sign of the function, draw its sketch. 1. f1x2 = x2 - x - 12 2. f1x2 = 2x + 3x - 9 3. f1x2 = 2x2 - x + 3 4. f1x2 = x4 - 4x2 5. f1x2 = 9x3 - x5 6. f1x2 = 4x - x3 7. f1x2 = 1x3 - 3x2214x2 - 92 8. f1x2 = 12x - x 212x - x + 32 9. f1x2 = 2x + 3 10. f1x2 = 29 - x2 11. f1x2 = x 21 - x In Exercises 12 23 determine if the function defined by the given equation is odd, even or neither. 12. f1x2 = 3 13. f1x2 = 2x2 + 1 14. f1x2 = 2x2 - x + 3 15. f1x2 = 2x3 - 3x 16. f1x2 = 17. f1x2 = 18. f1x2 = x x + 1 x x + 1 3x

2 4 5 2 2

1 x - 5 x + 5 x - 3 x x2 + 5x + 6 2

x2 - 4 x2 - 2x - 3 38. f1x2 = x - 3 In Exercises 39 39. f1x2 = 40. f1x2 = 41. f1x2 = 46 determine the horizontal asymptotes, if they exists.

2x x - 5 2 x - 5 2x2 + 5

4x2 + 9 19. f1x2 = x314x6 + 11210 20. f1x2 = x5 21. f1x2 = x x 22. f1x2 = x2 x x 23. f1x2 = ,x Z 0 x In Exercises 24 29, determine if the graph of the given equation is symmetric with respect to the y-axis, origin or neither. 24. xy = 1 25. x2 + x2y 2 + y 2 = 4 26. x2 + 4 + 3y = 0 27. x2 + y 2 = 5 28. x3 - y 3 = 0 29. x 2 + 4xy + xy 3 = 8 30. An equation is symmetric with respect to the x-axis if replacing y by - y leaves the equation unchanged. (a) Why can t such an equation represent a function? (b) In Exercise 24 29, which of the equations has this symmetry?

3x3 + 7 3x + 2 42. f1x2 = 2x - 3 43. f1x2 = 44. f1x2 = 45. f1x2 = 46. f1x2 = 2x2 + 7 3x + 4 21x - 2221x + 12 x3 + 3x + 2 12x - 1221x - 52 7x4 + 2 2x 9x2 + 1

In Exercises 47 55 use the zeros, vertical and horizontal asymptotes along with the sign of the function to sketch its graph. 47. f1x2 = 48. f1x2 = 49. f1x2 = 1 x - 2 2x x + 1 x + 1 4x2 - 9

150

* **

Section 1.7

x x - 9 3x 2 x2 - 16 4x2 x2 - 7x + 12 6x2 - 6x - 36

2

More on Functions

59. Given the graph of the function y = f1x2 in Figure 32, draw the graph with the corresponding points clearly labeled of (a) f1x - 1/42, (b) f A x + 1*2 B , (c) f1x2 + 3, (d) f1x2 - 1, (e) f1x - 1/42 + 1 and (f) f A x + 1*2 B - 2

225x2 - 1 57. Given the graph of the function y = f1x2 in Figure 30, draw the graph with the corresponding points clearly labeled of (a) f1x - 22, (b) f1x + 12, (c) f1x2 + 3, (d) f1x2 - 1, (e) f1x - 22 + 1 and (f) f1x + 12 - 2

60. Given the graph of the function y = f1x2 in Figure 33, draw the graph with the corresponding points clearly labeled of (a) f1x - 22, (b) f1x + 12, (c) f1x2 + 3, (d) f1x2 - 1, (e) f1x - 22 + 1 and (f) f1x + 12 - 2

58. Given the graph of the function y = f1x2 in Figure 31, draw the graph with the corresponding points clearly labeled of (a) f1x - 22, (b) f1x + 12, (c) f1x2 + 3, (d) f1x2 - 1, (e) f1x - 22 + 1 and (f) f1x + 12 - 2

61. Consider the function defined by the equation f1x2 = 1

2x2 - x - 3 (a) Determine its domain, and (b) Sketch its graph clearly showing all asymptotes. (c) Find its range by solving for x as a function of y. (d) Using (c), locate the coordinates of the turning point of the function. 2x - 4 does not cross its x + 3

62. Show the graph of the function defined by f1x2 = horizontal asymptote.

63. Show the product (or quotient) of an odd function with and odd function is even, an odd with an even is odd and an even with an even is even.

64. Does there exist a function which is both odd and even?

Section 1.8

65. Do there exist graphs of equations which are symmetric with respect to the y-axis and origin, and if there are, do they exhibit any other kinds of symmetry? 66. Suppose f is not a constant function with f1x + y2 = f1x2f1y2, show f1x + h2 - f1x2 f1h2 - 1 = f1x2 a b. (a) f102 = 1, (b) h h 67. Given the function defined by the equation

Regression

* **

151

y = 2x + 4

f1x2 =

a nx n + a n - 1xn - 1 + + a 1x + a 0 ; bmx m + bm - 1x m - 1 + + b1x + b0 2x2 + 6x + 8 x + 1 and its Slanted Asymptote Figure 34: r1x2 = y = 2x + 4

In Exercises 69 70, find all the asymptotes and plot the graph of the given function. 69. f1x2 = 70. g1x2 = 4x2 - 12x + 5 2x + 5 4x3 - 2x2 + 3x - 5

Determine the horizontal asymptotes (if they exist) if (a) n 6 m, (b) n = m, (c) n 7 m. 68. In the previous exercise, do m and n have to be integers, that is, must f be a rational function? Suppose the rational function r1x2 = q1x2 + l1x2 where l(x) approaches 0 as x approaches + q or - q , then r1x2 : q1x2 as x approaches + q or - q , or we say r(x) is asymptotic to q(x). For example,

r1x2 =

4 2x + 6x + 8 = 2x + 4 + x + 1 x + 1

as x approaches + q or - q , 4/1x + 12 : 0, so r1x2 : 2x + 4. Therefore y = 2x + 4 is an asymptote for r(x). A sketch is given is Figure 34. Note how the graph approaches the line y = 2x + 4 as x gets large. When l(x) is linear it is called a slant or oblique asymptote.

x2 - 9 71. Using your calculator, determine the vertical asymptote of the function defined in Example 11. 72. Show that the graph of the function defined in Example 12 does not cross its horizontal asymptote.

1.8 Regression

Scatter Plot Line of Best Fit Linear Regression

In applications, one is given a number of data points 1x1, y12, 1x2, y22, 1x3, y32, , 1xn, yn2 and seeks to determine a relationship between their x and y coordinates. Sometimes the points, when plotted, called a scatter plot, almost look linear, other times quadratic or perhaps some other shape. We first consider the case when the data appears linear, other cases will be considered at the end of this section. In the real world, if these points were taken from measurements, due to experimental or measurement errors, it would be almost miraculous for a set of many pairs of points to lie exactly on the same line even when the underlying model is exactly linear. When presented with a real set of data, one is faced with the problem of finding a line that best fits the data. However, when the model is linear or when the data appears to lie close to

Scatter Plot

152

* **

Section 1.8

Regression

some straight line, there is a standard technique to find the so-called line of best fit or the regression equation. The proof of the technique is considered later in the study of calculus, but let us look at the idea behind it. Suppose that we have n data points 1x1, y12, 1x2, y22, 1x3, y32, , 1xn, yn2. We plot these points and see that they lie roughly along a non-vertical straight line. Let us write y = ax + b for the equation of the line. Now for each x-value, xi, there is the observed y-value, yi, and the value of y predicted by the equation, that is y = axi + b (see Figure 1). This difference yi - 1axi + b2 is known as the error at x = xi.

(xn, yn)

yi - (axi + b) = Error at xi

What we would like to do is to pick the line so as to minimize the total error over all points. However, positive and negative errors would then cancel out and the fit might be very bad. In order to avoid this, one looks at the sum of the squares of the errors. That is, the squared error at a given xi is 1yi - 1axi + b222. One looks at the sum of these numbers for all the points and tries to determine the values of a and b that minimizes this sum. Although this may sound almost impossible, it is actually very simple. It does require us to use some new notation. We define the terms needed in this section. x1 + x2 + + xn 1average of the x values2 n y1 + y2 + + yn 1average of the y-values2 y = n sx = x1 2 + x2 2 + + xn 2 1sum of the squares of the x-values2 sx 1average of the sum of the squares of the x-values2 sx = n sy = y1 2 + y2 2 + + yn 2 1sum of the squares of the y-values2 sy 1average of the sum of the squares of the y-values2 sy = n = x1y1 + x2y2 + x3y3 + + xnyn 1sum of the products of the x and y values2 x =

sxy

Section 1.8

Regression

* **

153

(Note, the notation is easy to remember as follows: a bar above any symbol means it is an average, i.e., divide by n; the letter s is to remind you that you are summing squares of the variable in the subscript, and sxy sums the products of the x and y.) Using calculus, it can be shown that the line of best fit, the regression line has the equation y = ax + b where a = sxy x#y sx - x 2 and b = y - ax (1) (2) Line of Best Fit Linear Regression

It is clear that the calculations can be tedious and a calculator is useful. We shall show how the calculator finds the regression line very quickly at the end of this section Example 1 Median family income per year, in dollars, in the United States for the six year period from 1982 to 1987 is given in Table 1. Plot this data and find and plot the line of best fit. Use the line to predict the median family income for 1991. Table 1: Median Annual Income 1982 1987

Year Income 1982 23433 1983 24580 1984 25948 1985 27144 1986 28236 1987 29744

Solution. For simplicity, let 1982 be year 1, measure x in years and y in dollars. The years are now 1,2,3,4,5,6, so n = 6. Therefore, the sum of the x values is 1 + 2 + 3 + 4 + 5 + 6 = 21 and the average x, is given by x = 21/6. The sum of the y-values is 23433 + 24580 + 25948 + 27144 + 28236 + 29744 = 159085 and their average is y = 159085/6. Next we need the sum of the squares of the x-values, sx = 12 + 22 + 32 + 42 + 2 5 + 62 = 91 and the average, sx = 91/6 = 15.17. The sum of the products sxy = 1121234332 + 1221245802 + 1321259482 + 1421271442 + 1521282362 + 1621297442 = 578657 and the average is sxy = 578657/6. Substituting into the formula we have a = [578657/6 - 121/621159085/62]/[91/6 - 121/622] = 1249.11 b = 159085/6 - 11249.112121/62 = 22142.3 Rounding these numbers to the nearest integer, we have y = 1249x + 22142 as the line of best fit. Therefore, for 1991, we set x = 10 (remember we are counting years from 1982 as 1) and we forecast that the median family income at that time will be y = 1124921102 + 22142 = 34629, that is, $34,629 per year. The points and the best fit line (regression line) are shown in Figure 2.

154

* **

Section 1.8

Regression

y = 1249x + 22142

In the preceding example, we saw the process of inflation in the American economy. The median family income going up is a reflection of the general increase in the costs of goods and services, which, of course, must be balanced by an increase in salaries and other forms of earnings. What might be surprising is that the increase was so close to being perfectly linear during the mid-1980 s. On the other hand, some trends are linear but it is difficult to find a simple explanation for the observed phenomena. Consider the following example.

Example 2. Table 2 shows the number of deaths by falls in the United States over a five year period from 1978 through 1982.

Year No. of Deaths 1978 13690 1979 13216 1980 13294 1981 12628 1982 12077

Solution. Again, we measure x in years and let the first year (in this case 1978) be x = 1. We let y be the number of deaths and find that x = 3, y = 12981, sx = 11, sxy = 38180.2, and substituting, we find a = - 381.4 and b = 14125.2 Therefore, rounding off, we have y = - 381x + 14125. The results are graphed in Figure 3. Notice that this result predicts that the number of deaths by falls seems to be declining at a rate of about 381 per year, even though the population of the United States is growing. One would expect mortality numbers to increase with increasing population. Why do you think falls should be an exception?

Section 1.8

Regression

* **

155

y = *381x +14125

Figure 3: The Best Fit Line for Example 2 Given a scatter plot we need a measure to determine if the correlation between the variables is indeed linear. Linearity is unaffected by the position of the coordinate axis, so without loss of generality, assume the coordinate axes are translated so the origin corresponds to the point 1x, y2 in the original coordinate system. In this new X-Y coordinate system, if the points do appear to be linear then they will be mostly in quadrants I and III or in quadrants II and IV (why?). Note that in this translated coordinate system, the sum of the products will be positive if most of the points are in quadrants I and III, and negative if they are in quadrants II and IV. It is therefore reasonable to conclude that the larger sXY is in absolute value, the closer the points are to a straight line. However, x and y have units in which they are measured, and their magnitudes will depend on these units, it is useful in making comparisons to have a measure which is independent of units. To free the measure from units we divide the sum of products by another product of the same units. We divide s XY by 2sXsY, this product has the same units as sXY, so the quotient is dimensionless. We, therefore, define the correlation coefficient with respect to these translated variables as r = sXY 2sXsY (3a)

Correlation Coefficient

however, if we now return to variables in the original coordinate system (where X = x - x, and Y = y - y, we can show that) r = sxy - nx y 21sx - nx221sy - ny22 (3b)

Either of these expressions may be used to compute the correlation coefficient, r. It is clear that both involve some elementary but tedious computations, especially if the data set is large. Fortunately, we can automate the calculations using a spreadsheet like Excel, or as we shall see at the end of the section with our calculator. It can be shown that if r = - 1 or r = 1, the points lie on a straight line. If r is near zero then there is no linear relationship between x and y. Thus, if r is near - 1 or 1 then the

156

* **

Section 1.8

Regression

correlation between the points is linear, and if it near zero the relationship is not linear. (Note that from our discussion above r has the same sign as a, the slope of the regression line.) Example 3 Determine the correlation coefficient in Example 1. Solution. We may use either (3a) or (3b) to compute the correlation coefficient. Either way, the computations are straight-forward. If we use (3a) we need to use the translated variables X and Y, the computations are summarized in Table 3. (We actually generated the Table 3 using Excel.) Using these values, we have sXY = 21859.5, sX = 17.5, and = sY 27348356.8333333, giving r = sXY 2sXsY = 217.5 # 27348356.83333 218595.5 L 0.999207

x 1 2 3 4 5 6 21 3.5 y 23433 24580 25948 27144 28236 29744 159085 26514.16667 X = x - x - 2.5 - 1.5 - 0.5 0.5 1.5 2.5 Y = y - y - 3081.167 - 1934.167 - 566.167 629.8333 1721.8333 3229.8333 X2 6.2.5 2.25 0.25 0.25 2.25 6.25 17.5 Y2 9493588.02777779 3741000.69444445 320544.694444446 396690.027777776 2964710.02777777 10431823.3611111 27348356.8333333 XY 7702.91666666667 2901.25 283.083333333334 314.916666666666 2582.75 8074.58333333333 21859.5

Sum Mean

Alternately, we may use (3b) which uses the original variables, x and y. In Example 1, we found sxy = 578657, sx = 91 and sy = 234332 + 245802 + 259482 + 271442 + 282362 + 297442 = 424534561, n = 6, x = 21/6 = 3.5, y = 1599085/6 = 26514.1666667. and L 0.999207 159085 2 2 - 6 A 21 6 B B 14245354561 - 6 A 6 B Note that r is close to one, suggesting a strong linear correlation between x and y. r =

159085 578657 - 6 A 21 6 6 B

2A 6

91

Non-Linear Regression

Calculator Tips

We shall see at the end of the section that these calculations are done almost automatically for us with the calculator. Suppose the data now appears parabolic. Our goal should be to find the parabola that best fits the data. That is, suppose we want to find a, b, and c so that the parabola y = ax2 + bx + c best fits the data points. In the linear case, a and b had to be determined. Similarly, in the quadratic case we need to find a, b and c. What if the scatter plot looks like a cubic equation? Then we would expect a regression curve of the form y = ax 3 + bx 2 + cx + d, and we need to find a, b, c, and d. The formulas are messier than the linear case (that is to be expected). Instead of writing down the formulas for these coefficients, we shall show below, how to obtain the required equation and curve using the calculator, and leave the determination of the explicit formulas to you until after you study the calculus. The first thing we must do when determining a regression equation with the calculator is input the data. We shall use the data from Example 1 to illustrate. Basically, we

Section 1.8

Regression

* **

157

input a table for the given data sets and then will have the calculator compute the equation of the regression line and the correlation coefficient. We proceed as follows: 1. Press the MODE button on your calculator and check that the first line indicates FUNCTION as the Graph. Press the APPS (applications) button, and select Data/Matrix Editor, select New. Make sure Type is set to Data, Scroll down to Variable and give it a name and press Enter twice. We will name it xx , see Figure 4.

Figure 4: Naming the Data You now see a new screen (the Data/Matrix Screen) that looks like a table. Enter in C1 (column 1) the x-values, and in C2 the y-values, see Figure 5. (Note that Figure 5 shows only four of the entries, all six have been entered.)

Figure 5: Entering the Data After you enter all the data, Press F5 (Calculate), change Calculation Type to LinReg (scroll down), name x C1 and y C2. Also, enter y1 for RegEq; this tells the calculator to name the first column in the data set x, the second column y, and it stores the regression equation in memory as y1. see Figure 6. Press Enter twice.

158

* **

Section 1.8

Regression

The information required for the regression line now appears; as do the values for a and b as well as the correlation coefficient and its square. See Figure 7.

Figure 7: The Regression Equation and Correlation Coefficient Note we have the slope of the regression line, a = 1249.114286, its y-intercept, b = 22142.266667 and the correlation coefficient, r = 0.999207 (the last value is the square of the correlation coefficient).. Moreover, the calculator can also provide the scatter plot along with the regression line. Close the screen by pressing Enter. Press F5 (Calculate) and make sure all the settings are as before including Store RegEq which should be set to y1(x). Press F2 to display the Plot Set up Screen, press F1 to define Plot 1. Set Plot Type to Scatter, Mark to Box, x = C1 and y = C2. Save by pressing Enter twice. See Figure 8.

Figure 8: Defining the Scatter Plot Display the Y = Editor for y1(x), 1press * F12 and then select Style (2nd F1) set the Display Style to DOT and press Enter. Scroll up to highlight Plot 1. Next, press F2(Zoom) and select ZoomData.and then we obtain Figure 9, the scatter plot and regression line. Note that these directions can be found in the TI 89/92 Plus Manual beginning on page 83. However, the only difference between what is given here and the manual is that we did not select MedMed as the Calculation Type we selected LinReg and we stored the equation as y1(x) not y2(x). The information we entered is still stored in the calculator s memory, you may delete it by selecting VarLink (2nd-) scroll down to xx press F1(Manage), press Delete and then Enter. There is an alternative way of obtaining the same results with your calculator. You may proceed as follows:

Section 1.8

Regression

* **

159

(1) Store the x-coordinates and y-coordinates as lists named t1 and t2, as follows: 51, 2, 3, 4, 5, 66 STO t1 press ENTER 523433, 24580, 25948, 27144, 28236, 297446 STO t2 press ENTER Make sure you are using curly braces to enclose these two lists (2) Next enter on the entry line of the HOME screen (use the Catalog or type) LinReg t1, t2 and press ENTER then ShowStat and press ENTER (This produces Figure 7) (3) To draw the regression line, enter on the entry line in the HOME screen, Regeq(x) STO y1(x) and press Enter (This stores the regression equation as y1(x)) then enter on the entry line NewPlot 1,1, t1, t2 and press ENTER Choose WINDOW 1*F22, then press Zoom (F2), and scroll down and select ZOOMDATA (or press 9). (The regression line is now drawn. In some cases, you may need to press *F3) Often, data may resemble non-linear curves. In principle, the mathematics used to find the regression curve is the same as the linear case, that is, to minimize the square of the error between the data points and the best fit curve. Since the curve is non-linear, there will be more equations to solve to determine the coefficients (which will involve more complicated algebraic expressions). Nevertheless, the calculator automates the determination of the regression curve. Consider the following: Table 4 indicates the number of students who worked full-time while attending classes at a small urban university. Table 4: Students Working Full Time, 1991 1997

No. of Students Year 1023 1991 475 1992 430 1993 304 1994 411 1995 531 1996 982 1997

Find the parabola (quadratic regression curve) that best fits this data and estimate the number of students who will work full-time in 2004. The calculator steps will be almost identical to the linear example considered above with one exception, namely, we will choose QuadReg instead of LinReg. As before, let 1991 correspond to year 1. We enter this data into the calculator, see Figure 10, where we name the variable zz1

Figure 10: Naming the data As before, we enter the data and then when we Calculate: we choose QuadReg (for quadratic regression), C1 and C2 for x and y respectively and store RegEQ to y1(x), see Figure 11.

160

* **

Section 1.8

Regression

Figure 11: Defining the Columns Pressing Enter twice gives the required information, see Figure 12.

Figure 12: The Quadratic Regression Curve Rounding to the nearest integers, the parabola of best fit is y = 75x2 - 600x + 1496. In particular, our estimate for the number of students who work in 2004 (year 14) is y11142 = 7,767. As before, we can have the calculator draw the scatter plot and the regression curve, see Figure 13.

Figure 13: Quadratic Regression Curve and Scatter Plot As with linear regression, the alternative method may be used to produce the above, with QuadReg replacing LinReg. Likewise, we have the commands CubicReg and QuartReg for cubic and quartic regression curves.

Section 1.8

Regression

* **

161

Other types of regression equations may be calculated the same way and will leave their examination to the exercises. You may be asking yourself the question that if there are several possible regression curves for the same scatter plot, which one do we use? The answer is not always obvious. If you know something else about the data, and it suggests that quadratic regression best fits the model, then it is clear that this is the regression curve to use. Sometimes, other considerations come into play.

(Note, in the answers, we shall assume that the first year for which data is given is labeled as year 1.) 1. Table 5 shows the gross national product (GNP) of the United States in billions of dollars for the years 1978 through 1983. (a) Find the line of best fit. (b) Plot the line. (c) From the equation of the line, predict the gross national product in 1987. (The actual figure was 4488.5.)

Table 7

Year 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 Rate 3.5 3.7 4.1 4.4 4.6 4.9 5.0 5.0 5.2 5.4

Table 5

Year GNP 1978 2128 1979 2414 1980 2626 1981 2926 1082 3073 1983 3311

(d) What would this line predict about the divorce rate in 1987? See Exercise 11 for a continuation. 6. Baseball experts believe that there is a strong linear correlation between the number of games a team wins in a seasons and the teams batting average. Table 8 contains the data for the sixteen National League teams in the 2002 season. What is your conclusion?

2. Table 6 shows the median incomes, in dollars, of men and women in the United States for the years 1974 through 1980. (a) Find the line of best fit for women s incomes versus time and (b) the line of best fit for men s income versus time. (c) Plot both lines on the same set of axes. (d) Is the gap widening or shrinking? (e) What would this data predict about the earnings gap in 1986? (The actual gap in 1986 was 25,256 - 16,232 = 9024).

Table 8

Team Arizona Atlanta Chicago Cincinnati Colorado Florida Houston Los Angeles Milwaukee Montreal New York Philadelphia Pittsburgh San Diego San Francisco St. Louis No. of Wins 98 101 67 78 73 79 64 92 56 83 75 80 72 66 95 97 Team Batting Average 267 260 246 253 274 261 262 264 253 261 258 259 244 253 267 268

Table 6

Median Income Year 1974 1975 1976 1977 1978 1979 1980 Women 6970 7504 8099 8618 9350 10169 11197 Men 11889 12758 13455 14626 15730 17045 18612

3. (a) Use the data from Exercise 2 to find the line of best fit for men s income as a function of women s income. (b) What is the slope of this line? (c) Does it predict that the earnings gap is widening or shrinking? 4. The numbers of foreign investors in American enterprises for the years 1980, 1981, 1982, and 1983 were 1833, 1521, 1218, and 632 respectively. (a)Find the line of best fit for this data, and (b) the regression coefficient. (c) Does this indicate anything about the attractiveness of American investments to citizens of other countries? 5. The divorce rate (number of divorces per thousand couples) in the United States for the decade of the 1970 s is given Table 7. (a) Find and (b) plot the line of best fit and (c) determine the regression coefficient.

7. The following (x, y) pairs give the advertising dollars (in millions) spent on direct mail and newspaper advertisements respectively for the same years. (a) Find the line of best fit and the regression coefficient. (b) Roughly speaking, how do the two expenditures relate to one another? (c) Do you think that advertising people believe that one of these forms of advertising can substitute for the other. Data: (10.3,17.7), (8.9,16.5), (7.6,14.8), (4.1,8.2), (2.8,5.7), (1.8,3.7). 8. The amount of popcorn consumed in the United States (in millions of pounds) was: 353 in 1970, 393 in 1975, 568 in 1980, 670 in 1985, 700 in 1986, 741 in 1987, 807 in 1988, and 872 in 1989. (a) Find a line of best fit for this data and (b) use it to project the popcorn consumption in 1999. (c) In what year would you expect consumption to first exceed one billion pounds?

162

* **

Chapter Review

(a) Find the line of best fit for this data (label 1980 year 11 so as to continue with Exercise 5). (b) Combine this with the result of Exercise 5 to obtain a piecewise linear function that describes the American divorce rate over the eighteen year period. (c) Does this indicate any significant change in people s behavior? 12. Over the 9 year period, 1977 through 1985, the number of legal abortions (in thousands) performed in Bulgaria are given in the Table 12. (a) Fit a piecewise linear graph to this data. (b) Can you conclude anything?

Sometimes when one plots a set of data, it appears to be piecewise linear, causing one to ponder an explanation for the change in pattern at the corner(s). Consider Exercises 9 12. 9. The numbers of federal officials convicted of corruption in the period 1977 to 1986 is given in Table 9. The graph is easily visualized as two straight lines: one from 1977 to 1982, the other from 1983 through 1987. Find the lines of best fit for the data as so divided.

Table 9

Year 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 91 115 131 159 147 424 429 470 523 Number 94

Table 12

Year 1977 1978 1979 1980 1981 1982 1983 1984 1985 Number 141.7 147.7 147.9 155.9 152.4 147.8 134.2 131.1 132.0

13. The revenue (in thousands of dollars) resulting from the demand for a given item is indicated in Table 13. Find the equation of the best fit quartic for this data.

10. For the United States, the numbers of cyclists (not motorcyclists) killed in accidents with automobiles is given in Table 10. (a) Is there any obvious place where the data breaks into two straight lines? (b) Fit a piecewise linear function to the data. (c) Does this information say anything about the safety of riding a bicycle?

Table 10

Year 1960 1965 1970 1975 1980 1982 1984 1985 680 780 1000 1200 1100 1100 1100 Number 460

Table 13

Demand Revenue 100 0.8 200 3.8 300 5.7 400 7.2 500 5.6 600 4.1 700 1.1

11. Since 1979 the divorce rate statistics (see Exercise 5) are given in Table 11.

14. Given the data set (1, 0.9), (2, 3.8), (3, 6.2), (4, 4.1), (5, 2.2), (6, 5.8), (7, 9.3). Find the best fit regression (a) line, (b) quadratic, (c) cubic, (d) quartic. Which best fits the data? 15. Given the data set 10.0, - 12.82, 12.1, - 16.92, (3.01, 15.1), 13.92, - 0.932, (5.01, 30.5), find the best fit cubic. 16. Given the data set 1 - 1.1, - 9.12, 10.05, - 9.022, 11.05, - 10.92, (1.95, 14.9), (2.9, 146), find the quartic of best fit.

Table 11

Year Rate 1980 1981 1982 1983 1984 1985 1986 1987 5.2 5.3 5.1 5.0 4.9 5.0 4.8 4.8

CHAPTER REVIEW

Key Ideas

Two Dimensional Coordinate System Horizontal and Vertical Lines The Slope Intercept Form Graphing The Point-Slope Equation The Slope Formula Economic Applications The General Linear Equation Definition of a Function Functional Notation Difference Quotient Domain and Range Dependent and Independent Variables Vertical Line Test Combining Functions Composition Functions of Several Variables Break-Even Analysis Depreciation Piecewise Linear Functions Scaling Vertical Translation Axis of a Parabola Horizontal Translation Locating the Vertex Graphing a Parabola in the form y = ax 2 + bx + c Applications to Optimization Definition of a Circle Equation of a Circle Graphing a Circle Tangent Line The Ellipse Supply Function Demand Function Market Equilibrium Revenue, Cost and Profit Functions Marginal Functions Using the Zeros Even Functions Symmetry about the y-axis Odd Functions Symmetry about the origin Rational Functions Vertical Asymptotes Horizontal Asymptotes Translations Regression

Chapter Review

1. Plot the points 11, - 22, (2, 5), 1 - 3, 42 and 1 - 5, - 32. 2. Sketch the line (a) x = - 2, (b) 3x = 12, (c) y = 4, (d) 3y + 9 = 6. 3. Determine the equation of the line passing through 1 - 2, 42 with slope 2/3. 4. Determine the equation of the line with slope - *2 and passing through the point 10, - 32 5. (a) Determine the x and y- intercepts of the line 2x - 3y = 7 and (b) sketch its graph. 6. Determine the equation of the line passing through the points 12, - 32 and 1 - 6, 72. 7. Determine the equation of the line with x-intercept 3and y-intercept - 2. 8. Determine the equation of the line passing through 13, - 22 and (a) parallel, (b) perpendicular to the line 3x - 4y = 15. 9. Does the equation 4x - y 2 = 6 define a function? Explain. 10. Given f1x2 = (c) f1g1 - 122. 11. For the functions defined in the previous exercise, determine (a) f(x)g(x), (b) f(x)/g(x), (c) for what x values is (b) not defined? (d) for which x-values in g(x)/f(x) not defined? f13 + h2 - f132 12. Given f1x2 = 2x2 - 3x + 9, determine (a) , h f12 + h2 - f122 f1x + h2 - f1x2 (b) , (c) . (d) Substitute x = 2 and then h h x = 3 into (c) any observations? 13. (a) Determine the domain of the function defined by f1x2 = 23 - 4x, (b) Sketch its graph and (c) determine its range. 3x - 12 . 6x2 - 13x + 6 25 + 2x 15. Determine the domain of the function defined by f1x2 = 2 . x - 7x + 12 2 + x 16. Given f1x2 = , and g1x2 = 3 - 2x, determine (a) f(g(x)), for which 3 - 2x x-values is it not defined, (b) g(f(x)), for which x-values is it not defined? 14. Determine the domain of the function defined by f1x2 = 17. Given f1x, y2 = x3/4y 1/4, determine f(16, 81). 18. Given f1x, y, z2 = 2xyz + 3x2y3 + 9y2z2, determine f(2, 1, 3). 19. (a) Given p = 2x - 2 and p = 28 - x, identify which of these could represent a demand function and which a supply function. (b) Over what domain are they defined? (c) Determine the coordinates of market equilibrium. 20. A $1200 computer depreciates linearly of five years, at which time it is has a scrap value of $50. What is its value at the end of (a) two years, (b) four years? 21. Determine the domain and sketch the graph of the function defined by 3x - 4 x2 - 2 f1x2 = d 7 15 - 2x if if if if x 1 1 6 x 6 3 3 6 x 4 x 7 4 2 + x , and g1x2 = 4 + 3x, determine (a) f(3), (b) g(1), 3 - 2x

1

* **

163

22. Given a function defined by Table 1, determine, in tabular form, each of the following: (a) y = f1x - 22, (b) y = f1x + 22, (c) y = f1x2 + 2, (d) y = f1x2 - 4, (e) y = f1x - 22 + 3, (f) y = f1x + 22 + 4.

Table 1

x

f(x)

-2

3

-1

4

0

-2

1

-3

2

1

3

4

4

7

5

9

23. (a) Determine the coordinate of the vertex of the parabola y = - 2x2 - 4x + 8 and sketch its graph. (b) From your sketch, determine the graph s approximate intercepts, find them exactly and compare. (c) From your sketch, determine the approximate solution of - 2x2 - 4x + 8 = 6, and compare your approximate solutions with the exact solution. 24. The cost of producing x-items is given by the equation C1x2 = 2x2 - 200x + 6000, determine the number of items to be produced to minimize the cost as well as giving the minimum cost. 25. (a) The cost of producing x-items is given by the equation C1x2 = 2x2 - 200x + 6000, determine the cost of producing the 99th item. (b) Determine the average cost function, and then determine the average cost of producing the 99th item. 26. Determine the equation of the circle with center at 12, - 32 and radius 4. Give its sketch. 27. Determine the coordinates of the center of the circle and determine its radius if its equation is x2 + y2 - 4x + 6y - 12 = 0. Sketch this circle. 28. Determine the coordinates of the center of the circle and determine its radius if its equation is x2 + y2 - 4x + 6y + 13 = 0. Give a better description of this circle. 29. (a) Determine the equation of the tangent line to the circle of radius 13 centered at the point (*5, 12). (b) Compare the y-values on the circle with the yvalues on the tangent line when x = - 5.01. 30. Determine the equation of the ellipse centered at the origin with one x-intercept at (4, 0) and one y-intercept at 10, - 52. 1x - 222 1y + 422 31. Sketch the graph of the ellipse with equation + = 1. 4 3 32. Sketch the graph of the ellipse whose equation is 4x2 + 3y 2 + 8x - 12y + 4 = 0. 33. (a) Give a qualitative sketch of the graph of the polynomial whose equation is f1x2 = x1x - 2221x + 32312x + 52, (b) estimate where the turning points are on the graph. 34. Sketch the graph of f1x2 = 2x2 + 5x - 12 x2 - 1 , indicate all asymptotes and zeros.

35. Determine the symmetry of the function defined by (a) f1x2 = 5x3 - 7x, x2 (b) f1x2 = 2 , (c) f1x2 = x3 , (d) f1x2 = - x x3 , (e) f1x2 = x2 x 3 . x + 9 36. Given the data set (1, 2), (3, 5), (4, 8) and (6, 12), determine the (a) line of best fit, (b) quadratic of best fit, (c) cubic of best fit.

An Introduction to Calculus

In this chapter, we begin the study of the differential calculus of one variable. Starting with the notion of the slope of a straight line, we shall define the slope of a general curve as a function of x. We describe the fundamental method of finding this so-called derivative function from the definition and then develop the simplified rules for rapid calculation of derivatives that have been discovered by earlier generations of mathematicians. Next, we show that the derivative can also be interpreted as a rate of change and look at some applications of that interpretation, and then learn how to find the derivative of functions defined implicitly. The last section shows how the tangent line can be used to approximate the zeros of a function Newton s Method.

166

* **

Section 2.1

Slope of a Curve

2.1

Slope of a Curve

Slope of a Tangent Line The Slope as a Limit Slope of a Curve Equation of a Tangent Line A Place Where No Tangent Exists The Derivative Calculator Tips

We have seen that when x and y are linearly related, it is the slope of the line which gives us most of the information about the behavior of the function. A positive slope indicates that as x increases so does y. When the slope is negative the y-values decrease as x increases. A zero slope tells us that the line is horizontal and y is constant. When the slope is undefined, the line is vertical, and x is constant. In that case, of course, y is not a function of x. The slope itself can be thought of as the rise of a line divided by its run, or the change in y with respect to a change in x. Moreover, slope has different interpretations which depend upon the context. These include tax rate, rate of depreciation, and marginal cost, among others. Our object here is to generalize the notion of slope. We shall define the slope of an arbitrary curve and see how this concept can be used. Of course, for a linear function, this definition must produce the slope of the line. Let us see how we should define the slope of a curve. Consider the function whose graph is sketched in Figure 1. By the slope of the curve at the point P(x, f(x)), we mean the slope of the tangent line to the curve at this point. What do we mean by a tangent line? For a circle, that is easy. In classical geometry a tangent line to a circle is a line that touches the circle in only one point. However, for arbitrary curves this definition will not suffice. It is possible for a line to meet the intuitive sense of tangency, touching the curve in only one point, locally but, cross it again elsewhere. It is even possible, as you will see, for a tangent line to cross the curve at the point of tangency. Therefore, we resort to a physical notion. Think of the curve in Figure 1 as a road and imagine yourself driving along the road. Even though the road may be curving, at any instant you are looking in some direction that you think of as straight ahead. It is that straight line that we would call the tangent to the curve. That is, the tangent line at any point is the line that approximates the direction of the curve at that point. Take a ruler and try to place it at the point P(x, f(x)) so that it approximates the shape of the curve at this point. This line is the tangent line. No other line touching the curve at this point, will parallel the curve as well as the tangent line. In fact, we could define the tangent line as the best linear approximation to the curve near the point of tangency, that is as the line among all possible lines, whose y-values are closest to the y-values along the curve near the point P. Figure 2 shows the curve with several lines drawn at the point P. It is easy to pick out the tangent line.

Section 2.1

Slope of a Curve

* **

167

Tangent line at P

y = f(x)

+* P(x, f(x))

y = f(x)

P(x, f(x)) +*

For a linear function the slope is constant. However, you would not expect the slope of a curve to be constant. In Figure 3, we have drawn tangent lines to the curve at each of the points P, Q, R, and S. Notice that the slope of the curve is different at each of these points. To indicate that the slope of the curve depends upon the point P(x, f(x)) we write it as mtan1x2. Now remember, we said that we would define the slope of a curve at a point to be the slope of its tangent line. So, how do we find the slope of a tangent line? Before getting to the general definition, let us first look at a specific example. Consider the function defined by the equation f1x2 = x 2 + 1. You should recall that the graph of this function is a parabola. We would like to find the slope of this curve (the slope of the tangent line to

168

* **

Section 2.1

Slope of a Curve

Q P R

Figure 3: The Slope of the Tangent Line at Different Points along the Curve this curve) at the point P(2,5). Using a ruler, we could try to draw the tangent line at this point, find the coordinates of two points on the line and then compute the slope. Unfortunately, unless the graphs of the curve and tangent line are drawn exactly, our measurement of the slope will be at best an approximation. Nevertheless, as an exercise, you should try this approach and compare your answer to the one we obtain below (see Exercises 1 3).

Figure 4: f1x2 = x2 + 1 and the Tangent Line at P (2,5) In Figure 4, we have a sketch of the curve and its tangent line at (2,5). To find the slope of the line exactly, we shall proceed as follows. Choose a point on the graph either to the left or to the right of the point P, and call it Q. Make sure that this point is very close to the point P. Let us choose Q to be just to the right of P. In fact, let the x-coordinate of Q be 2 + 0.001 = 2.001. Since f1x2 = x 2 + 1, the y-coordinate of Q is f12.0012 = 12.00122 + 1 = 5.004001, and the point Q may be labeled Q(2.001,5.004001). Actually, Q is so close to P that you really could not tell them apart unless the graph was greatly magnified. Imagine that we have magnified our sketch and that Figure 5 reflects this magnification.

Section 2.1

Slope of a Curve

* **

169

We next draw a line connecting Q to P. This is not the tangent line but, it is expected to be close to the tangent line. In fact, in the unmagnified state, we probably could not tell them apart. It is due to the magnification that we are able to see any difference at all. Let us call the line connecting P to Q a secant line. We can compute the slope of this secant line which we shall denote by msec. It is just the difference of the y-values divided by the difference of their x-values. Thus, msec = 5.004001 - 5 = 4.001 2.001 - 2

Q(2.001, 5.004001)

msec

This slope is nearly equal to the slope of the tangent line. Can we do better? Why not take a point even closer to P, say with an x-coordinate 2 + 0.0001? We find the corresponding y-coordinate to be 12.000122 + 1 = 5.00040001, and the slope of the secant line connecting this point to P will turn out to be 4.0001. Similarly, suppose we took a point just to the left of P, with x-coordinate 2 + 1 - 0.00012 = 1.9999. Its y-coordinate computes to 4.99960001, and the slope of the secant line connecting this point to P turns out to be 3.9999. The basic idea is that as we take the points closer and closer to P, we should get better and better approximations to the slope of the tangent line. In fact, you have probably guessed the slope of the tangent line at the point P(2,5). Using points just to the right of P, we obtained secant lines with slopes just a little more than 4. Points to the left produced slopes just a little less than 4. As we let the points get closer and closer to P, the slope is clearly approaching the value 4. It is this so-called limiting value that we shall define to be the slope of the tangent line. Summarizing, we do the following: 1. Choose a point Q, which is very close to P(2, 5). Call the x-coordinate of Q, 2 + h. (When Q is just to the right of P, h is a small positive number. When Q is to the left of P, h is a small negative number.) The y-coordinate of Q is f12 + h2 = 12 + h22 + 1 = 4 + 4h + h2 + 1 = 5 + 4h + h2 2. Compute the slope of the secant line through the two points P and Q. msec = f12 + h2 - f122 4h + h2 5 + 4h + h2 - 5 = = 12 + h2 - 2 h h

P(2, 5)

Note that the difference in the x-coordinates is h. Now we can simplify by dividing by h, thus, msec = 4h + h2 = 4 + h h

3. Now let Q get even closer to P by letting h approach 0. Letting h approach zero, we have the slope of the tangent line at P, denoted by mtan122 Mathematicians have a special way of writing the instruction given in Step 3. They write mtan122 = lim msec = lim 14 + h2

h :0 h :0

The symbol lim is read the limit as h approaches zero of . . . . Thus, the slope of the h :0 tangent line is the limit as h approaches zero of the slope of the secant line. That is, the limit as h approaches zero of 14 + h2. Letting h approach zero 14 + h2 becomes 4. We could, using this notation, combine the instructions into one by writing, mtan122 = lim msec = lim

h :0

f12 + h2 - f122 h :0 h

170

Section 2.1

Slope of a Curve

In fact, let us generalize this with the following definition: The Slope as a Limit

DEFINITION 1 The slope of the tangent line (when it exists) to the function defined by the equation y = f1x2 at the point P(x, f(x)), is given by

mtan1x2 = lim

(1)

In trying to apply this definition, we must be very careful when we let h approach zero. If we do it immediately, we will obtain an expression of the form 0/0, which is indeterminate. When the calculations were done in our first example, we did not let h approach zero until after there was a cancellation of the common factor h in the numerator and the denominator. In general, only after the possibility of getting 0/0 eliminated do we let h approach zero. Here are some examples of limits taken as h approaches zero. We give a more detailed examination of limits in Section 3.3. Example 1. Compute (a) lim 110 + 4h2

h :0 h :0 h :0

Solution. (a) As h approaches zero, 4h (which is simply 4 times h) must also approach zero. Since 10 is a constant, it does not change, and the limit is just 10. That is, lim 110 + h2 = 10. (b) As in part (a), if we let h approach zero, 4h must approach zero. In fact, any constant multiple of h will approach zero. 3x is a variable but it does not depend upon h. Therefore, neither 3x nor - 6 changes as h approaches zero and the limit is 3x - 6, that is, lim 13x + 4h - 62 = 3x - 6. (c) Both 6xh and 3h are multiples of h. Although x is a variable, 6x does not change as h changes, so we can treat it as a constant. Therefore, both these terms go to zero and the limit is just x2 + 2. Thus,

h :0 h :0 h :0

lim 1x 2 - 6xh + 3h + 22 = x2 + 2.

(d) If we try to let h approach zero as this expression now stands, all the terms in the numerator are multiples of h and will go to zero. (Do you see that the last term, h2, also must go to zero? If h is a number like 0.001, what is h2?) The denominator is h, so that also goes to zero and we have the form 0/0, which is an indeterminate form and always requires more investigation. However, in this exercise, we can factor an h from each term in the numerator to get h1x2 + 4x + 3 + h2. Now we can cancel the common factor h in numerator and denominator. Thus, h1x 2 + 4x + 3 + h2 x 2h + 4xh + 3h + h2 = lim h:0 h:0 h h = lim 1x 2 + 4x + 3 + h2 = x 2 + 4x + 3 lim

h:0

Section 2.1

Slope of a Curve

171

We remark that this method may not always yield a real value for the slope. For example, a function may have points at which the tangent line is vertical. We know that the slope of a vertical line is undefined, so at such points our method is doomed to failure. In addition, some functions (such as piecewise linear functions, at the points where the pieces join) may have other peculiarities that will make it impossible to define a tangent line at one or more points in their domain. Example 2 Determine the slope of the curve defined by the equation f1x2 = x2 - 2x + 1 when (a) x = 3; (b) x = 0; (c) x = - 1; (d) x = 1 Solution. We use Definition 1 to compute mtan1x2 as a function of x and then substitute the various values for x. (We could compute each problem separately, but since the computation for each one is the same, it makes sense to compute the slope of the curve at an arbitrary point x and then substitute for x at the end.) To avoid any algebraic errors, we compute the slope in four steps. 1. We see that there are two terms in the numerator of (1) and we must compute their difference. We compute the two terms in the numerator.

f1x + h2 = 1x + h22 - 21x + h2 + 1 = x 2 + 2xh + h2 - 2x - 2h + 1 f1x2 = x2 - 2x + 1

Slope of a Curve

Notice that we line up similar terms, anticipating the next step, which is to simplify algebraically. 2. Take the difference to complete the numerator.

f1x + h2 - f1x2 = 2xh + h2 - 2h

3. Now form the quotient, factor out h in the numerator, and cancel it with the one in the denominator.

f1x + h2 - f1x2 h = h12x + h - 22 2xh + h2 - 2h = = 2x + h - 2 h h

mtan1x2 = lim

h:0

f1x + h2 - f1x2

lim

h:0 h h12x + h - 22

= lim

2xh + h2 - 2h = h:0 h

= lim 12x + h - 22 = 2x - 2

h:0

Check that last step. Just as we saw above, when we let h approach zero, 2x + h - 2 became simply 2x - 2. We can now answer the questions. (a) mtan132 = 2132 - 2 = 4, (c) mtan1 - 12 = 21 - 12 - 2 = - 4. (b) mtan102 = 2102 - 2 = - 2, (d) mtan112 = 2112 - 2 = 0

Example 3 Determine the equation of the tangent line to the curve y = x 2 - 2x + 1 at each of the points indicated in Example 2.

172

Section 2.1

Slope of a Curve

Solution. The equation of any line may be found using the point slope formula which was derived in Chapter 1. It is y - y1 = m1x - x12. (a) When x = 3, y = f132 = 1322 - 2132 + 1 = 4. We found in Example 2 that mtan122 = 4. Therefore, the equation of the tangent line is y - 4 = 41x - 32 or y = 4x - 8. (b) When x = 0, y = f102 = 1022 - 2102 + 1 = 1. We found in Example 2 that mtan102 = - 2. Therefore, the equation of the tangent line is y - 1 = - 21x - 02 or y = - 2x + 1. (c) When x = - 1, y = f1 - 12 = 1 - 122 - 21 - 12 + 1 = 4. We found in Example 2 that mtan1 - 12 = - 4. Therefore, the equation of the tangent line is y - 4 = - 41x - 1 - 122 or y = - 4x. (d) When x = 1, y = f112 = 1122 - 2112 + 1 = 0. We found in Example 2 that mtan112 = 0. Therefore, the equation of the tangent line is y - 0 = 01x - 02 or y = 0. (What kind of line is this?)

We emphasize that whenever we try to find mtan1x2, before we allowed h to approach zero, we had to factor and reduce the fraction in order to insure that we did not get the form 0/0. But what happens if we cannot factor and reduce the fraction? Let us take a look at such a situation. Example 4 Determine the equation of the tangent line to the curve defined by the equation f1x2 = 2x + 1 when x = 3. Solution. We first note that f132 = 23 + 1 = 2. Thus the equation of the tangent line is y - 2 = mtan1321x - 32. What remains is to determine the slope of the curve at x = 3, and substitute into the above equation. Since we want the slope at a particular point, we shall replace x by 3 in equation (1). We proceed as follows: 1. f13 + h2 = 213 + h2 + 1 = 24 + h f132 = 2 2. f13 + h2 - f132 = 24 + h - 2 3. f13 + h2 - f132 24 + h - 2 = h h

Note that now there is no obvious cancellation. If we allowed h to approach zero, we would have the form 0/0. However, let us see if a calculator will help. We cannot let h = 0, but we can let h be close to zero. So we try calculating the value of the fraction when h is 0.1, 0.01, and 0.001. If we try h = - 0.1, - 0.01 and - 0.001 we obtain 0.25158, 0.25016, and 0.25001 respectively. Again the same limit as h approaches zero from the left (through negative values). See Table 1. It certainly appears that the limit as h approaches zero will be 0.25 = 1/4. We feel safe in asserting that mtan132 = 1/4 and the equation of the tangent line is

Section 2.1

Slope of a Curve

173

h - 0.1 - 0.01 - 0.001 0.001 0.01 0.1 f13 + h2 - f132 h = 24 + h - 2 h

Although this is correct, let us see that we could actually have found the limit by algebra without resorting to the calculator. The algebraic device used in such problems is rationalization. Recall that when the two following binomial expressions are multiplied, we have

A a 2b + c 2d B A A a 2b - c 2d B = a2b - c2d

Note that the binomial expressions are identical with the exception of the connecting sign and are called conjugates of each other. The multiplication of these conjugates results in the clearing of all radicals, producing a rational expression. It is precisely this observation which allows us to perform the cancellation needed in the previous example. Example 5 Use rationalization to find the equation of the tangent line to the curve defined by the equation f1x2 = 2x + 1 when x = 3. Solution. In Example 4 we had f13 + h2 - f132 24 + h - 2 = h h There is no obvious cancellation of h. If we allow h to approach zero, we would have the form 0/0. However, we can get the appropriate cancellation by means of multiplying the numerator and denominator of the fraction by the conjugate of the numerator, that is, rationalizing the numerator. Thus,

A 24 + h - 2 B A 24 + h + 2 B

h

A 24 + h + 2 B

14 + h2 - 4 h A 24 + h + 2 B

h h A 24 + h + 2 B

1 24 + h + 2

Therefore, we have, Step 4. f13 + h2 - f132 1 1 1 = lim = = h :0 h : 0 24 + h + 2 h 4 24 + 2 lim Thus, mtan122 = 1*4, yielding y - 2 = 1*41x - 32, which is equivalent to the equation found above.

174

* **

Section 2.1

Slope of a Curve

Sometimes, the algebraic steps needed to determine the slope of the curve using (1) are not immediately obvious. As we have seen, it may be possible to get around this difficulty by using a calculator. However, the result may only be as accurate as the accuracy of the calculator used. Nevertheless, sometimes guessing the answer this way will help you to find the right trick. If there is a simple algebraic technique, we shall usually choose it over a numerical one. We indicated above that the slope of a curve might not exist at every point on the curve. For example, it might be that the tangent line to the curve is vertical. However, there are other possibilities. Generally speaking, if a curve is smooth then the tangent line will exist at each point. But what does smooth mean? In words, a graph is smooth if it has no corners or sharp points . At such points the graph will not have a tangent line. Some curves with corners and sharp points are sketched in Figure 5.

y y P

P x x

Figure 5: Curves Which Are Not Smooth at P What makes a point a corner? Remember how we proceed when trying to find the slope of the tangent line at a point P; we choose a point Q either to the left or right of P and draw the secant line connecting P to Q. When we pick Q, it must not matter which side of P it is on. As we let h approach zero, that is, as Q approaches P, the limiting value for the slope of the secant line is the slope of the tangent line. In Figure 5, you can see that if we approach P from the left, the limiting value for the slope will not be the same as when we approach it from the right. Thus, the slope of the tangent line at this point is not well-defined. Remember, if there is a tangent line at a point on the curve it must be unique. At a sharp point this is not the case. It is important to observe that the equation of the tangent line to each point along the graph of the linear function f1x2 = mx + b is, in fact, y = mx + b. Why is this so? Choose any point on the line, and draw its tangent line. The tangent line is precisely the line itself. The linear function is the only function which has this property. Thus, given any linear function in the form f1x2 = mx + b, mtan1x2 = m. That is, the slope of the tangent line is the same at each point, and equals the slope of the line. This observation gives us a whole class of examples of curves with sharp points. Consider any of the piecewise linear functions discussed in Chapter 1. The points where the slope changes are sharp points. As another example, consider the absolute value function. Example 6 Show that the absolute value function f1x2 = x = e has no tangent line at x = 0. x if x 0 - x if x 6 0

Section 2.1

Slope of a Curve

175

Solution. Referring to Figure 6, we see that if we approach P(0, 0) from the left, the tangent line has slope - 1, but if we approach P from the right, the slope of the tangent line is + 1. Since at (0,0), the slope coming from the right is not the same as the slope coming from the left, the slope of the tangent line does not exist at (0,0).

y = *x

y=x

P(0, 0)

Figure 6: f1x2 = x

f1x + h2 - f1x2 represents the slope of the tangent line at x. We h shall soon see that this expression may also have other meanings which depend upon the interpretation of the function. For example, if f represents position and x time, we shall see that the expression is to be interpreted as a velocity. Or, if f represents total cost of producing x items, then the expression will represent the marginal cost of producing one more item. The interpretation depends upon what the function is modeling. Therefore, a generic name is assigned to this expression. It is called the derivative. The derivative will be defined precisely in Definition 2. Each of the different interpretations of the derivative may be thought of as a particular brand name for the generic item. Thus, whether you are asked to find the slope of a curve, the velocity of a particle, or the marginal cost, you would do f1x + h2 - f1x2 the same thing that is, compute lim . h :0 h There are various symbols that are used to represent the derivative. Below, we list the most commonly used ones. We shall use them interchangeably. The expression lim

h :0

The Derivative

NOTATIONS: If y = f1x2 then each of the following may be used to represent the derivative at x (the full name is the derivative of f with respect to x.).

f 1 x2 dy dx y 1 x2 d [f1x2] dx 1read fprime of x2 1read dydx not dy over dx2 1read y prime of x2 1read the derivative of f1x22

176

Section 2.1

Slope of a Curve

Note that the various alternatives allow f(x) and y to be used interchangeably. If we want to indicate the derivative at x = a, we may write any of the following: dy d ` , [f 1x2] x = a. Thus, to indicate the slope of the tangent line at dx x = a dx x = 3, we could write f 132. Sometimes, instead of saying find the derivative of the function , we say differentiate the function. Similarly, if a function has a derivative at a point, we say that it is differentiable at that point. f 1a2, y 1a2, We may now rewrite Definition1 as follows:

DEFINITION 2 The derivative (when it exists) of the function defined by the equation y = f1x2 at the point P(x, f(x)) is given by

f 1x2 = lim

(2)

Notice that although we used the f 1x2 notation in (2), any of the alternatives would have been equally acceptable. Example 7 (a) Differentiate the function f1x2 = x 3. (b) Determine the slope of the curve defined by f1x2 = x 3 at x = 2. d 3 1x 2 (c) Find dx Solution. (a) We must find the derivative of the given function. We use (2). 1. f1x + h2 = 1x + h23 = x 3 + 3x 2h + 3xh2 + h3 f1x2 = x3

2. f1x + h2 - f1x2 = 3x2h + 3xh2 + h3 3. f1x + h2 - f1x2 h13x 2 + 3xh + h22 3x2h + 3xh2 + h3 = = 3x 2 + 3xh + h2 = h h h f1x + h2 - f1x2 = lim 13x 2 + 3xh + h22 = 3x2 h:0 h:0 h

4. f 1x2 = lim

Thus, we have as the derivative, f 1x2 = 3x2. (b) The slope of the tangent line at x = 2 is f 122 = 31222 = 12. (c) This is just another way of phrasing (a). So without additional work, d 3 1x 2 = 3x 2. dx

Calculator Tips

The calculator can be used to visually illustrate the concept of a tangent line. Suppose we have entered on the Y = screen, y11x2 = x 2 + 1 and want to draw the tangent line at x = 2. We proceed as follows: let the calculator draw the graph in the standard window. Next press the math key (F5) and scroll down to the Tangent option and press Enter. Next

Section 2.1

Slope of a Curve

* **

177

enter 2 (for x = 2) and press Enter, and the tangent line is drawn and its equation is given, see Figure 7 If you want to show additional tangent lines at other points, just repeat the process, that is, press F5, scroll down to the Tangent option press Enter, enter the x-value of the point where you want the tangent line, and press Enter.

It is also a simple matter to compute the derivative, via its definition, using the calculator. In the Calculator Tips Section 2.7 we illustrate that avgRC(f(x), x, h) gives the difference quotient, then the derivative is nothing more than the limit of this quotient as h approaches zero, that is limit(avgRC(f(x), x, h), h, 0).

Exercises 1 3 refer to Figures 8, 9, and 10. In each case, choose another point on the tangent line to determine the slope of the curve at P. 1. 2.

P(4, 6) P(8, 5)

Figure 8: Ex 1 Figure 9: Ex 2

178

3.

* **

Section 2.1

Slope of a Curve

23. f1x2 = 3/x 24. f1x2 = - 2/x 2 (compare with exercises 19 and 24). x In Exercises 26 and 27, short segments of the tangent lines are given at various points along a curve. Use this information to sketch the curve. 25. f1x2 = 1x 26. See Figure 11.

P(8, 7)

Figure 10: Ex 3

In Exercises 4 8, (a) sketch the graph of the given function, and then draw the tangent line at the point P. (b) Using your sketch, approximate the slope of the curve at P, (c) Use (1) to determine the exact value of the slope at P. 4. 6. 7. 8. f1x2 = x2 + 3 f1x2 = 2x + 1 f1x2 = x

3

5. f1x2 = - 2x2 + 3x + 3

f1x2 = - x2 + 2x - 1

9. Given the curve whose equation is f1x2 = x2 + 3. Let P be the point (1,4). (a) Determine the slope of the secant line joining P to Q, if Q has as its x-coordinate: (i) 1.01 (ii) 1.001 (iii) 1.0001 (iv) 0.99 (v) 0.999 (vi) 0.9999. (b) What limiting value does the slope of the secant line appear to be approaching as Q approaches P? 10. Given the curve whose equation is f1x2 = 2x + 4. Let P be the point (5,3). (a) Determine the slope of the secant line joining P to Q, if Q has as its x-coordinate: (i) 5.01 (ii) 5.001 (iii) 5.0001 (iv) 4.99 (v) 4.999 (vi) 4.9999. (b) What limiting value does the slope of the secant line appear to be approaching as Q approaches P? 11. Given the curve whose equation is f1x2 = x 0.3. Let P be the point (1,1). (a) Determine the slope of the secant line joining P to Q, if Q has as its x-coordinate: (i) 1.001 (ii) 1.00001 (iii) 0.999 (iv) 0.9999. (b) What limiting value does the slope of the secant line appear to be approaching as Q approaches P? In Exercises 12 17, determine the derivative at the given point on the curve using equation (2). 12. y = x2 at the point (3,9). 13. f1x2 = 3 - 2x - x 2 at the point 1 - 1, 42. 14. f(x) as defined in Exercise 4. 15. f(x) as defined in Exercise 5. 16. f(x) as defined in Exercise 6. 17. f(x) as defined in Exercise 7. Find f 1x2 in Exercises 18 - 25. 18. f1x2 = 2x2 - 7x + 9. 19. f1x2 = 1x 20. f1x2 = - 3x2 + 7x - 11. 21. (a) f1x2 = 53 (b) Give a geometric explanation for your result. 22. f1x2 = mx + b

28. Given f1x2 = 3x2 - 12x + 5. At which point will the curve have slope (a) 0; (b) 6; (c) - 6? 29. Given f1x2 = x3 - 12x. At which points will its tangent line (a) be horizontal; (b) have slope 15; (c) have slope 36? 30. Given f1x2 = 1x (a) At which point will the tangent line be vertical? (b) What can you say about the derivative at this point? 31. f1x2 = e 4x - 2 if x 1 x + 1 if x 7 1 (a) Sketch the graph of this function. (b) Determine f 1x2 if x 6 1. (c) Determine f 1x2 if x 7 1 (d) What can you conclude about f 112? x2 if x 0 x if x 6 0

32. f1x2 = e

Section 2.2

(a) Sketch the graph of this function. (b) Determine f 1x2 if x 6 0 (c) Determine f 1x2 if x 7 0 (d) What can you say about f 102? 33. Given f1x2 = 2x - 5 . (a) At what point is the function not differentiable? (b) What is the derivative to the left of this point? (c) What is the derivative to the right of this point? 34. f1x2 = e 9x + 5 if x 1 (a) What is f 1x2 if x 7 1? (b) What is x2 + 7x + 6 if x 6 1 f 1x2 if x 6 1? (c) What is the slope of the curve just to the left of x = 1?

Derivative Rules 1

* **

179

37. Suppose that, in the development of the definition of the derivative, we wrote (x2, f1x22 for Q instead of 1x + h, f1x + h2). Show that the definition of the derivative will then have the following alternate form: f 1x2 = lim f1x22 - f1x2 x2 - x

x2 : x

(d) What is f 112? 35. Find the point on the curve y = x at which the tangent line at (2,8) crosses the curve. (You may want to use the result of Exercise 17.) 36. Consider the two functions: f1x2 = x1/3 and g1x2 = x4/3 near x = 0. Using h = - 0.1, - .0.01, - 0.001 and - 0.0001, (as h approaches 0 from the left), and h = 0.1, .0.01, 0.001 and 0.0001 (as h approaches 0 from the right). Find the slope of the secant lines passing through P(0,0) and Q(h, f(h)). Does mtan1x2 exist at (0, 0)? Why not? (b) Now repeat the process for g(x). What is the difference in the behavior at P(0,0) for the two functions?

3

38. Use the alternate form of the derivative given in Exercise 37, to compute f 1x2 for the function defined in: (a) Exercise 21; (b) Exercise 22; (c) Exercise 23; (d) Exercise 24. 39. Let y11x2 = x2 + 1, determine the equation of the secant line through each of the following x-values and x * 2: (a) x = 2.1, call the equation y2(x) and enter it the Y = screen (b) x = 2.05, call the equation y3(x) and enter it the Y = screen (c) x = 2.025, call the equation y4(x) and enter it the Y = screen (d) choose an appropriate window so the curve and all these secant lines can be seen. (e) Have the calculator add the tangent line at x = 2. What is happening to the secant lines as x approaches 2?

2.2

Derivative Rules 1

Derivative of a Linear Function The Simple Power Rule The Constant Multiplier Rule The Sum Rule Calculator Tips

At this point, the only way that we can compute the derivative is by applying the definition, equation (2) of the previous section. We have seen that in order to apply the definition, it may be necessary to go through a great deal of algebraic manipulation. However, when you realize that the calculus has been used for over three centuries, you must expect that some shortcuts would have been discovered. If nothing else, you would assume that tables of derivatives would have been put together by working mathematicians. In fact, something better has been done. A collection of simple rules for producing derivatives, without resorting to the definition, have been developed. Our objective in this section is to begin to determine those rules that will let us calculate derivatives without using the definition. Of course, the justification for these rules will rest upon the definition. In the previous section we indicated that the derivative of a linear function is its slope. This followed directly from the interpretation of the derivative as the slope of a tangent line. Of course, this may be also be proved from equation (2) of the previous section, as you were asked to do in the preceding exercise set. Let us state this as a rule.

d 1mx + b2 = m dx Note that in the special case in which m = 0, the line is horizontal (it has zero slope) and the derivative is zero. We call this special case Rule 2. Derivative of a Linear Function

180

* **

Section 2.2

Derivative Rules 1

d 1b2 = 0 dx In words, the derivative of a constant is zero. So far, so good but, we have not really expanded our knowledge. Rules 1 and 2 just summarized our observation that the slope of a curve reduces to the slope of the line if the function is linear. Let us now look at some simple non-linear functions, and develop a derivative formula for these functions. Consider the problem of finding the derivative of the function whose equation is f1x2 = x100. We apply the definition and obtain f 1x2 = lim 1x + h2100 - x 100 h :0 h

We must expand the first expression in the numerator and cancel the h from the denominator. Otherwise, how will we be able to let h approach zero without encountering the form 0/0 ? Of course, we want more than the solution to this specific problem. We would like to be able to determine the derivative of any function whose equation is of the form f1x2 = xN. Consider expressions of the form 1x + h2N, where N is a positive integer. (Actually, we need only consider N 7 1, since we have Rule 1.) If N = 2, If N = 3, If N = 4, 1x + h22 = x 2 + 2xh + h2 1x + h23 = x 3 + 3x 2h + 3xh2 + h3 = x3 + 3x 2h + h213x + h2 = x3 + 3x2h + h2T3 where T3 = 13x + h2 1x + h24 = x 4 + 4x 3h + 6x 2h2 + 4xh3 + h4 = x 4 + 4x 3h + h2T4, where T4 = 16x2 + 4xh + h22 Do you see the developing pattern? In general, we have that

1x + h2N = x N + Nx N - 1h + h2TN

(1)

where TN is a polynomial in x and h. You may recognize (1) as another way of writing the binomial expansion of 1x + h2N. In which case, you know the exact form of TN.

d N 1x 2 = Nx N - 1 where N is any real number. dx We prove this rule in the special case in which N is a non-negative integer. We already know that Rule 3 is correct in the case N = 0, or N = 1 (why?). Thus, we need only prove it for any integer N 2. Using the definition of the derivative, (2) of the previous section, with f1x2 = xN, we have, f 1x2 = lim 1x + h2N - x N h :0 h

We now use (1) from above to replace the first term in the numerator and obtain,

Section 2.2

Derivative Rules 1

181

f 1x2 = lim

x N + Nx N - 1h + h2TN - x N Nx N - 1h + h2TN = lim = h:0 h:0 h h h1NxN - 1 + hTN2 lim = lim 1Nx N - 1 + hTN2 = Nx N - 1 h:0 h:0 h

Notice that because TN, which is a polynomial in x and h, is multiplied by h, that entire term goes to zero as h goes to zero.

In later sections we shall show why this rule is valid for values of N which are not nonnegative integers. For now, assume its truth for any constant N. We illustrate its use in the following example. Example 1 Determine f 1x2 for each of the following. (a) f1x2 = x43 (b) f1x2 = x 3/5 1 (c) f1x2 = 1x (d) f1x2 = 2 . x Solution. d 43 1x 2 = 43x 43 - 1 = 43x 42 dx d 3/5 3 3 5 3 2 3 (b) 1x 2 = x 5 - 5 = x -5 = . 5 5 dx 5x 2/5 d d 1/2 1 1 2 1 1 1 1 (c) 1 1x2 = 1x 2 = x 2 - 2 = x - 2 = = 1/2 dx dx 2 2 2 1x 2x d 1 d -2 2 (d) a b = 1x 2 = - 2x -2 - 1 = - 2x -3 = - 3 dx x 2 dx x (a) Notice that in (c) and (d) we used various rules involving exponents and rewriting radicals in terms of fractional exponents. It might be a good idea for you to review these rules which are summarized in Table 1. Also note the form of our answers, if we start with a radical, as in (c),we usually leave the answer in radical form, similarly, we rarely leave answers with negative exponents as in (d).

bn = b # b b 1bm2n = bm n 1ab2n = a nbn b -n = 1 bn b m # b n = bm + n b0 = 1 bm/n = 2bm = A 2b B m

n n

bm = bm - n bn a n an a b = n b b

Given the differentiable function defined by y * f(x), suppose we multiply this function by a constant C. How will the derivative of Cf(x) be related to the derivative of f(x)? Rule 4 answers this question.

d 1Cf1x22 = Cf 1x2 dx

182

Section 2.2

Derivative Rules 1

In words, the derivative of a constant times a function is the constant times the derivative of the function. It is not hard to see why this rule is true. Multiplying a function by a constant means that each y-value is multiplied by the constant. Thus, if g1x2 = Cf1x2, then g1x + h2 - g1x2 = Cf1x + h2 - Cf1x2 = C[f1x + h2 - f1x2] Thus, the entire numerator of the difference quotient in equation (2) of the previous section is multiplied by C (without any change in the denominator) resulting in the derivative being multiplied by C. Example 2 Determine Solution. Applying Rule 4, we have d d 15x 72 = 5 1x 72 = 517x 62 = 35x 6. dx dx d 15x 72. dx

After doing a few examples using this rule, you should find yourself doing the intermediate steps mentally and writing d 15x 72 = 5 # 7x 6 = 35x 6. dx The Sum Rule Suppose we have two differentiable functions f(x) and g(x). We can form a new function, called the sum, s1x2 = f1x2 + g1x2. The next rule shows how the derivative of s(x) is related to the derivative of the two component functions.

d 1f1x2 + g1x22 = f 1x2 + g 1x2 dx d 1f1x2 - g1x22 = f 1x2 - g 1x2 dx In words, this rule states that the derivative of a sum is the sum of the derivatives and the derivative of the difference is the difference of the derivatives. A little thought should indicate why this rule is true. When we sum two functions, s1x2 = f1x2 + g1x2, for a given x-value, the y-value of the sum is the sum of the two individual y-values. Is that clear? Suppose f132 = 5 and g132 = 7. Certainly s132 = 5 + 7 = 12. Therefore, the numerator of equation (2) of Section 2.1 becomes s1x + h2 - s1x2 = f1x + h2 + g1x + h2 - [f1x2 + g1x2] = [f1x + h2 - f1x2] + [g1x + h2 - g1x2]. The denominator remains simply h. Thus the quotient becomes a sum of the individual components, which in the limit becomes the sum of the individual derivatives. The details of the formal proof are given at the end of Section 3.3. Of course, we may handle f1x2 - g1x2 by simply noting that f1x2 - g1x2 = f1x2 + 1 - 12g1x2 and applying the constant multiple rule. This rule generalizes to more than two functions and we may restate it in words by stating that the derivative of a sum of functions is the sum of their derivatives.

Section 2.2

Derivative Rules 1

183

Example 3 If f1x2 = 5x4 + 2x -3, find f 1x2. Solution. f 1 x2 = d d d 15x4 + 2x -32 = 15x 42 + 12x -32 = 4 # 5x3 - 3 # 2x -4 = 20x3 - 6x -4 dx dx dx

Note that in applying Rule 3 to the first term, we multiplied the coefficient, 5, by the power of x, 4, and decreased the power of x by 1, rather than writing it out in detail. The second term was handled similarly. Remember, reducing the number - 3 by 1, yields - 4 not - 2!

Let us look at some applications of these rules. Example 4 Find the equation of the line tangent to y = x3 - 3x 2 + 5x - 6, at the point where x = 3. Solution. In order to find the equation of the line, we need the slope and one point. The point is easy to find. If x = 3, then on the curve, y = 1323 - 31322 + 5132 - 6 = 9. Therefore, the point of tangency is (3, 9). To find the slope, we need the derivative. Using our new rules, y = 3x 2 - 6x + 5. At the point with x = 3, we have mtan132 = 31322 - 6132 + 5 = 14. Substituting this information into the point-slope equation for the line, we have y - 9 = 141x - 32 = 14x - 42, so, y = 14x - 33

Example 5 We remarked in Section 3.1 that the tangent line is the straight line that best approximates the curve near the point of tangency. Find the equation of the tangent line to y = 6x 2/3 at the point (8, 24). Find the exact value of y at x = 9 on the curve and on the tangent line. Find the error and the percent error that you have by using the straight line approximation. Solution. Since we are given the point (8, 24) on the curve, we need only find the slope of the tangent line in order to get its equation. For this we need the derivative. The derivative is, y = 612/32x2/3 - 1 = 4x -1/3. At x = 8, mtan182 = 4182-1/3 = 4/2 = 2. Hence, the equation of the tangent line is y - 24 = 21x - 82; y = 2x + 8. At y = 9, on the tangent line, we have y = 2192 + 8 = 26. On the curve, we have y = 61922/3 = 25.9605 (to four decimal places on a calculator). The error in the approximation is 0.0395, which represents a percent error of 10.0395/25.960521100%2 = 1.52%. As you can see, in the days before calculators, an error of less than 2 parts per hundred was a small price to pay for the greater simplicity of computing the value on the tangent line function rather than the value of f(x).

184

* **

Section 2.2

Derivative Rules 1

Calculator Tips

The TI 89 can find derivatives very easily. Our notation for derivative is d/dx (f(x)) which tells us to take the derivative of f with respect to x. The TI 89 does it a little differently. On the keypad, above the number 8, in orange is the letter d. To access this d, you press the 2nd key followed by the number 8. Doing so presents on the screen d( which means the derivative of . . . . We must add the function we want the derivative of, the variable we are d differentiating with respect to, and the closing parenthesis. For example, 15x 82 would dx be entered as d15x 8, x2 as in Figure 1.

Figure 1: Using the TI 89 to Find Derivatives Pressing Enter gives the result as seen in Figure 2. Using the with key we can have the calculator compute the value of the derivative at a given value for x. For example d 15x 82 x = 2 is illustrated in Figure 3. Note the with key is the symbol located to the dx left of 7 on the keypad.

Be careful, the d used for differentiation is above the number 8 key (and first requires you press the 2nd key), is different from the alphabetical d which requires you first press the alpha key and then the comma key. Only the d above the number 8 key is used for differentiation, the alphabetic d will not work!

Section 2.2

Derivative Rules 1

* **

185

In Exercises 1 11, use the appropriate rules to determine the derivative. 1. y = 3x2 - 2x + 1, find

3

dy dx

` x=2

23. Determine the equations of the tangent lines from the point (0,13) to the curve y = 4 - x2. 24. Determine the equations of the tangent lines from the point (a) (1,10) (b) (4,17) to the curve y = 2x - x2. 25. One of the most commonly used mathematical models for a demand function in micro-economics is x = Mp -k, where p is the price of a commodity and x is the quantity of the commodity that can be sold. Suppose that M = 4, k = 1/2, p is the price per ride on a New York City subway, and x is the number of riders per day (in millions). (a) Find the ridership when the price is two dollars. (b) Find the drop in ridership if the price is raised to $2.25. (c) Find the equation of the tangent line to the demand curve at p = 2. (d) Use the value of x on the straight line to find the approximate drop in ridership if the price is raised to $2.25. This is called the marginal demand. (e) Find the ratio of the value found in (d) to the change in price. Use your knowledge of the derivative to compute the limit given in Exercises 26 31. 26. lim 27. lim 1x + h257 - x 57 . h 2/3 2/3 1x + h2 - x h 1 28. lim 29. lim 1x + h22 h

h:0

2. f1x2 = 3x - 7x + 9, find f 1 - 12. 3. y = 3x2/3 - 3x4 + 2/x2 - 7, find y . d 7 a 3x5 - 2x3/2 - 5 + 293 b dx x dy 3x5 - 7x 2 + 2 5. y = , find . Hint: Divide first. dx 3x3 4. Compute 3 7 + 3/4 + 5, find f 1x2. x4 x 7. f1x2 = 2 3x - 1 4 x + 7x + 32x4 + 71, find f 1x2. 6. f1x2 = 2x7 8. w = 32v1/4 16 v2 + 7v2 + 2, find ds . dt dw . dv

h:0

11. f1x2 = 1x - 221x + 22, find f 1x2. In Exercises 12 15, determine the equation of the tangent line at the indicated x-value. 12. y = 3x4 - 2x 2 + 7; 13. f1x2 = 3x2 64 x = 2

h:0

1 x2 .

+ 2 3 x + 2; x = 8 x2 14. y = 2 3 x + 2x - 20; x = 8 15. y = 4x + 2/x - 1; x = 2 16. Find the points on the curve y = 2x3 - 54x + 1 at which the tangent line is horizontal. 17. Find the points on the curve y = x - 8x + 3 at which the tangent line is horizontal. 18. Find the point on the parabola y = ax2 + bx + c where the tangent line is horizontal. (Note that this exercise gives an alternative method for locating the vertex of a parabola.) 19. Find the equation of the line perpendicular to the tangent line to the curve y = x3 - x + 1 at x = 2. Recall that lines are perpendicular if the product of their slopes is - 1. 20. Consider the function defined by f1x2 = x4 + 8x3. (a) For what values of x is f 1x2 7 0? (b) For what values is f 1x2 6 0? (c) At which point(s) will the tangent line be horizontal? 21. Find the point(s) on the curve y = 6x1/3 at which (a) the slope is 1*2. (b) the tangent line is vertical. 22. Show that the area of the triangle in the first quadrant formed by the tangent line to any point on the curve y = k/x, k 7 0 and the coordinate axes is a constant.

4 2

h:0

31. lim

h:0

2 38 + h - 2 . h

32. Consider f1x2 = 4x2/3, suppose you do not have a calculator and want to approximate f(7.9). (a) Find the equation of the tangent line at x = 8, (b) Compute the y-value on the tangent line at x = 7.9. (c) Compare your result with the answer given by your calculator. 33. Approximate, using the method of the previous exercise, f(31.99), if f1x2 = x2/5. 34. Given the parabola, y = x2. (a) Find the slope of the line connecting the points (1, 1) and (3, 9) on the curve. This line is called a chord. (b) Find the slope of the tangent line to the curve at (2, 4). (c) Compare the slopes of the tangent and the chord. 35. Given the parabola, y = x2 + 3x - 5. (a) Find the slope of the line connecting the points 11, - 12 and (3, 13) on the curve. (b) Find the point on the curve at which the slope of the tangent line is the same as the slope of the chord. 36. Let y = ax2 + bx + c. Find the slope of the chord (see Exercises 34 and 35) connecting the points with x-coordinates r and s. Find the x-coordinate on the parabola where the tangent line to the curve has the same slope as the chord.

186

Section 2.3

2.3

The Limit Limits by Substitution One Sided Limits Jumps and Holes Continuity Removable Discontinuities Differentiability and Continuity Calculator Tips

The limit is one of the most commonly used tools of the calculus. At this level, it will only be necessary for us to understand it intuitively. We leave its formal definition to more advanced courses. In the preceding two sections, you have already seen one of its applications, and even know how to determine limits in many cases. In the following definition, we assume y is a function of x and given by the equation y = f1x2.

DEFINITION 1

x :a

it exists) is the unique number that the y-values are very near when the x-values are very close to a, either just to its left or just to its right. Now you are probably wondering what this means, as well as the meaning of the words very near, very close, just to the right or just to the left. A few examples should illustrate the meaning of the definition. These terms will all be illustrated in the examples. In more advanced courses, these terms are more carefully defined. However, once you understand what they mean intuitively, the transition to a more rigorous definition is not very difficult. Example 1 Find lim 12x + 42.

x :3

The Limit

Solution. We want to find the y value that f1x2 = 2x + 4 is very near when x is very near 3. So we pick some values of x that we would think of as very near 3. For example, suppose x is 2.999. Now f12.9992 = 212.9992 + 3 = 9.998. If x is 2.9999, f12.99992 = 9.9998. If x = 3.0001, f13.00012 = 10.0002. If x = 3.00001, f13.000012 = 10.00002, and so on. The closer the x-values are to 3 (either just to its left, or just to its right), the closer the y-values get to 10. Thus, 10 is that unique number that the y-values are very close to, when the x-values are very close to 3. Therefore,

x :3

You might be asking yourself why not just substitute x = 3 into the expression and obtain the value 10 for the limit? In many cases, that is all you have to do to get the correct answer. However, there are times when the function is not defined at the point at which x = a. Other times the function is defined in pieces, and the rule changes at the point x = a. In these or other situations, special care must be taken. Let us consider examples that illustrate these cases. Of particular concern are fractions in which the denominator is zero. Remember, division by zero is always undefined.

Section 2.3

187

Example 2 x2 - 9 Find lim . x :3 x - 3 Solution. At x = 3, the function is undefined. (Why? Because the denominator of the fraction is zero!) If you tried to substitute x = 3 directly into this expression, you would obtain the form 0/0 . The form 0/0 is called an indeterminate form and requires closer examination. Therefore, let us proceed as above, choosing values of x just to the left and right of 3. We show our results in Table 1. (The entries in the y1 column are the y-values corresponding to the x-values in the first column.) Table 1: Examining the Value of the Function Near x = 3

Observe that as x approaches 3 from either direction, the values of f(x) get very close to 6. Therefore, the limit is 6. It did not matter that the function was not defined at x = 3, all that matters in determining the limit is what happens to the y-values near x = 3. In fact, the function could have been defined to have any y-value at all at x = 3, it would not change the limit. For example, if we were to define f132 = 21, the limit would remain 6. (Note: For details in setting up a Table, refer to the Calculator Tips in Section 1.5)

You might be convinced that substituting x = 3 directly into the given expression in the previous example is not useful and you would be correct. But observe that 1x - 321x + 32 x2 - 9 = = x + 3 x - 3 1x - 32 If we let x = 3, x - 3 = 0 and we could not cancel the common factor 1x - 32. However, the cancellation is valid, since, when we consider the limit, x is near but never equal to 3. Therefore, x2 - 9 = lim 1x + 32 x :3 x - 3 x :3 lim Thus, the problem reduces to the kind of situation considered in Example 1. When x is near 3, x + 3 is near 6. Again, you see that substitution (in this case, after the quotient is simplified) gives us the correct answer. If the function is nice , it will usually turn out that all you have to do is substitute to determine the limit. We will discover that nice usually means that the function has no

188

* **

Section 2.3

jumps or holes . We again stress that most functions that you shall encounter in this text will be nice except at an occasional isolated point. Suppose that we try to sketch the graph of y = f1x2 from Example 2. Except at the point x = 3, the function is defined by the equation y = x + 3. The graph of this function is simply a straight line of slope 1 with y-intercept 3. However, we see that since the function is not defined at x = 3, there is a hole in its graph at the point (3,6) (see Figure 1). Nonetheless, when x is close to x = 3, the y-values are all close to 6, and that is the limit as x approaches 3.

x=3

Figure 1: f1x2 =

x2 - 9 = x + 3 if x Z 3 x - 3

Limits by Substitution

In general, the following is true: If the function defined by the equation y = f1x2 is a nice function, then lim f1x2 = f1a2. To see what we mean by nice , let us state x :a some theorems (without proof) that will make the notion more precise. All of these theorems should seem obvious, although the proofs of some of them require several steps. In all of these theorems, we assume that

x :a

x :a

each exist.

THEOREM 1

x :a

lim c = c

THEOREM 2

x :a

lim x = a

(To see why these two limits are true, draw the graph of y = c and y = x, and use the definition of the limit.)

THEOREM 3 If k is a constant,

x :a

x :a

In words, the limit of a constant times a function is the constant times the limit of the function.

Section 2.3

189

THEOREM 4

x :a

x :a x :a

In words, the limit of a sum (difference) is the sum (difference) of the limits.

THEOREM 5

x :a

x :a x :a

THEOREM 6

If M Z 0, lim lim f1x2 f1x2 L x :a = = x : a g1x2 lim g1x2 M

x :a

THEOREM 7

If N is any real number, lim 1f1x22N = A lim f1x2 B N = LN

x :a x :a

(if L 6 0, and if N = p/q where q is even, then L is not a real number. Therefore, we exclude this case from the theorem.) In words, the limit of a function (raised) to a power is the power of the limit (subject to the restrictions indicated above). Let us look at some examples of how these theorems could be used. Example 3 Apply Theorems 1 through 7 to evaluate the following limits. (a) lim 6x

x :2

(b) lim 1x 2 - 32

x :2

x :2

(d) lim 1x + 42

x :2

(e) lim

6x1x 2 - 32 x :2 x + 4

x :2 x :2

x :2 x :2 x :2 x :2 2

x :2 2 x :2

thus, lim 1x - 32 = 4 - 3 = 1 (Theorem 4) (c) lim 16x1x 2 - 322 = lim 6x # lim 1x 2 - 32 1Theorem 52 = 12 # 1 = 12

x :2 x :2 x :2

x :2

(e) In part (c) we found the limit of the numerator is 12. In part (d) we found the limit of the denominator is 6. Therefore, by Theorem 6, the limit of the quotient is 2.

The point of the last example is to evaluate the limit in part (e). It seems like a lot of steps to find the limit of one function. However, every algebraic function is defined by a similar

190

Section 2.3

series of component steps. All of these steps are made up of the operations of addition, subtraction, multiplication, division, powers, and roots. Therefore, by application of these seven theorems, any algebraic function is nice. The limit as x approaches a of f(x) will be just f(a) as long as nothing special (such as division by zero) occurs at x = a. Thus, we could have simply evaluated the limit in part (e) above by just substituting 2 directly for x wherever it appeared in the fraction. Example 2 above is a bit more complicated because of the appearance of a zero denominator. However, once we cancelled the 0/0 term, substitution was appropriate. This is a good trick to look for. Any time you are faced with an expression in which substitution yields something of the form 0/0 , try some algebraic manipulations first, and then substitute. Example 4 Compute lim

h :0

29 + h - 3 . h

Solution. If we try to substitute 0 for h in the expression, we obtain the form 0/0 . Instead, we first perform some algebraic manipulations rationalizing the numerator by multiplying by the conjugate. Observe that a 19 + h2 - 9 29 + h - 3 29 + h + 3 ba b = h 29 + h + 3 h A 29 + h + 3 B = Therefore, we have,

h :0

h h A 29 + h + 3 B

1 29 + h + 3

lim

29 + h - 3 1 1 = lim = . h : 0 29 + h + 3 h 6

In finding limits we often use this technique. (We could have also taken values for h just to the left and right of 0 and constructed a table to estimate the limit.)

When considering a limit, we need to consider what happens just to the left and right of the point in questions. It proves useful to introduce symbols which represent the behavior of the function to the left and right of the point. Again, assuming y = f1x2, we define lim f1x2 read the limit of f(x) as x approaches a from its left the left-handed limit, x : awhich, when it exists, is the unique number that the y-values are very near when the x-values are just to the left of a, and lim+ f1x2, read the limit of f(x) as x approaches a from its x :a right the right-handed limit, which, when it exists, is the unique number that the y-values are very near when the x-values are just to the right of a. It then follows from the definition of the limit, that the limit exits at a if and only if these two one sided limits are equal to each other, that is lim f1x2 exits if and only if lim- f1x2 = lim+ f1x2. x :a x :a x :a Our next example illustrates one-sided limit through the use of the notion of piecewise linear functions. Example 5 Given f1x2 = e - 2x + 6 3x + 1 Determine lim f1x2.

x :1

if x 6 1 if x 1

Section 2.3

191

Solution. If x is just to the left of 1, then x 6 1, and we are on the piece of the function defined by f1x2 = - 2x + 6. Just to the left of x = 1, we have lim- f1x2 = lim-1 - 2x + 62 = 4. Similarly, if x is just to the right of 1, then x 7 1,

x:1 x:1

and we are on the piece of the function defined by f1x2 = 3x + 1, so we have lim+ f1x2 = lim+ 13x + 12 = 4, Since the left and right handed limits are the same,

x:1 x:1

number, we have lim f1x2 = 4. Figure 2 illustrates graphically why the limit is 4.

x :1

y=4 x=2

Figure 2: f1x2 = e

- 2x + 6 3x + 1

if x 6 1 if x 1

The next example illustrates the notion of a jump in a function. You will see that when a function has a jump, it is impossible for the y-values just to the left of the jump to be near the y-values just to its right. Therefore, the limit does not exist. Example 6 Given, f1x2 = e - 2x + 6 if x 6 2 3x + 4 if x 2 Determine lim f1x2, if it exists.

x :2

Solution. Just to the left of x = 2, we are on the piece of the function defined by f1x2 = - 2x + 6. Therefore the left-hand limit, lim- f1x2 = lim- 1 - 2x + 62 = 2. Just

x :2 x :2

to the right of x = 2, we are on the piece of the function defined by f1x2 = 3x + 4, and the right hand limit, lim+ f1x2 = lim+ 13x + 42 = 10. Since we are not close to the

x :2 x :2

same y-value on either side of x = 2, that is, the left and right handed limits are not equal, the limit does not exist. From Figure 3 it is evident why the limit does not exist at x = 2. There is a jump in the y-value at x = 2.

In general, at any point in the domain of the function at which it has a jump, the limit does not exist. However, there is a difference between a jump and a hole. Example 2 has a hole, but the limit exists because the limiting y-values on either side of the hole are the same. In general, a function will have a limit at any point at which its graph has a hole. In fact, the limit is precisely the y-value needed to plug the hole.

192

Section 2.3

y = 10 x=2 y=2

Figure 3: f1x2 = e

- 2x + 6 3x + 4

if x 6 2 if x 2

The next example illustrates how to determine a limit when there are no obvious ways of cancelling the 0/0 term. It requires the use of the calculator. Example 7 3x - 1 . x x :0

Determine lim

Solution. If you try to substitute x = 0 into this expression, you will find that we obtain the indeterminate form 0/0. There are no obvious algebraic tricks to try, so we resort to the definition, and use the calculator. We shall choose x-values just to the right and just to the left of x = 0, to see what y-value, if any, we are near. We summarize our calculations in Table 2. We see from Table 2, that as x gets very close to 0, (on either side), the y-value gets very close to 1.099 (to three decimal places). In fact, we shall see, the exact answer is ln 3, the natural logarithm of 3, which is 1.098612289 to 9 decimal places.

Since a numerical approach is not a rigorous proof, whenever possible we shall show how to determine a limit algebraically and use the calculator as a check. However, when we are not able to handle the problem algebraically, a numerical approach may be very useful

Section 2.3

* **

193

and will usually suffice for our needs. In more advanced courses, techniques are developed so that one could actually prove the validity of such numerical results. Essentially, we have demonstrated that to compute a limit, when there are no jumps, substitution is an effective method. But, you must be careful when to substitute if you obtain the form 0/0. There are functions which do not have limits for reasons other than jumps but they will not be considered here. As we indicated earlier, if f(x) is a nice function, for example, a polynomial, or a rational function (the ratio of two polynomials), or a radical function, then lim f1x2 = f1a2. In the case of a rational function, we assume that x :a the denominator is non-zero at x = a. In the case of a function that involves a radical with an even index, we assume that any expression under the radical is non-negative. (In Exercise 27 of Section 3.5, we consider another method for determining the limit when you obtain the indeterminate form 0/0 it is called L Hpital s rule.) The time has come to attach the correct name to a nice function. The kind of function that we have in mind is what we call a continuous function. Roughly speaking, any function whose graph is without holes or jumps is a continuous function. Any point in the domain at which there is a hole or jump in the function is called a discontinuity. There are other types of discontinuities, a discussion of them is left to more advanced courses.

Continuity

Since polynomials are functions without holes or jumps, they are functions which are continuous everywhere. Similarly, at any point at which its denominator is non-zero, a rational function is continuous. To illustrate a discontinuous function, we need only construct a function which has either a hole or a jump at some point in its domain. In Figure 4, we have a function which is undefined at x = 1, has a hole at x = 2, and has a jump at x = 4. At each of these x-values, the function is discontinuous. Thus, if we have a function without undefined points, holes, or jumps, we have a continuous function. Now that we have a feel for the physical definition of continuity, we give a formal definition.

DEFINITION 2 A function defined by the equation y = f1x2 is continuous at x = a, if the following three conditions are all satisfied:

194

Section 2.3

x:a

(3) f1a2 = L Condition (1) means that at x = a, we know how to compute the corresponding y-value, f(a). Condition (2) implies that there is no jump in the graph at x = a, and Condition (3) means that there is no hole at x = a. Many texts shorten the definition of continuity at x = a by just writing that the function is continuous at x = a, if lim f1x2 = f1a2. The x :a assumption being that by writing f(a) and its limit, we understand that both are defined. We could replace the statement of continuity at a

x :a

h :0

If h is near zero, (that is, a + h is near a) the y-values are near f(a); so the limit is f(a). We leave the verification of their equivalence as an exercise. Now that we have a formal definition of continuity at a point, we can say that a function is continuous on a given interval if it is continuous at every point of the interval. Let us apply the definition to a function to determine where it is or is not continuous. Example 8 Determine those points, if any, at which the function defined by y = f1x2 = e is not continuous. Solution. For x 2, the rule is y = x 2 + 1, which is a polynomial and therefore continuous. Similarly, for x 7 2, the graph is a line and therefore continuous. The only questionable point is at x = 2. Let us check the three conditions that must be satisfied in order that the function be continuous at x = 2. 1. f122 = 1222 + 1 = 5. 2. lim f1x2 = 5. (Did you check the limit from both sides?)

x:2

x2 + 1 3x - 1

if x 2 if x 7 2

x:2

Thus, the three conditions are satisfied and the function is continuous at x = 2, and therefore continuous for all x. Figure 5, is a calculator sketch of the graph, it is not always clear from such sketches if the function has a hole or not at a given point.

Section 2.3

195

Figure 5: The Graph of x2 + 1 if x 2 y = f1x2 = e 3x - 1 if x 7 2 At what values, if any, is it discontinuous. Solution. We have that Recalling that |x| = e - x if x 6 0 x if x 0

The only point at which continuity is questionable is at x = 0 (why?). Since f102 = 3, the function is defined at x = 3. However, just to the left of x = 0, the y-values are - 1, and just to its right the y-values are + 1. Thus, there is no common y-value that both are near and so the limit does not exist. Condition 2 is violated, so the function is not continuous at x = 0. Note that the geometrical evidence for the discontinuity is the jump in the graph at x = 0 (see Figure 6).

(0,3) +*

Figure 6: f1x2 = c

-1 3 1

if x 6 0 if x = 0 if x 7 0

196

Section 2.3

x 2 - 16 Determine what f(4) should be if for x Z 4 if the function defined by f1x2 = is x - 4 to be continuous. Solution. If x Z 4 f1x2 may be rewritten as f1x2 = x + 4, (Why?). It then follows that for continuity, we must satisfy condition(3), lim f1x2 = lim 1x + 42 = 8, therex :4 x :4 fore, for continuity, we need f142 = 8.

Example 10

Removable Discontinuities

Hole discontinuities are often called removable discontinuities, because defining or redefining the function at the hole so that its y-value equals the limit at the point eliminates the hole. All other types of discontinuities are called non-removable. For example, in the last example by defining y to be equal to 8 at x = 4, that is, f142 = 8. The function becomes continuous because the hole that would have appeared without this definition is removed. The following theorem is an immediate consequence of the alternate formulation and the definition of differentiability.

THEOREM 9 At any point in its domain at which a function is differentiable, it is continuous. PROOF Because f(x) is differentiable at x = a, we know that f 1a2 = lim f(a + h) -f(a) exists. This implies that f(a) exists. We need only show that h:0 h lim f1a + h2 = f1a2 to prove that the function is continuous at x = a. We proceed with

h:0

h:0

h:0

= lim c

h:0

(Note the application of the product and sum rule for limits.) The first limit on the right is f 1a2 the second is 0, and the limit of a constant is the constant. Thus, we have lim f1a + h2 = f1a2 # 0 + f1a2 or, lim f1a + h2 = f1a2, h :0 h :0 and by the alternate formulation of continuity, we have that at any point at which the function is differentiable, the function is continuous.

Theorem 9 gives a quick method of determining if a function is continuous. If you can determine its derivative, then it is continuous. The rules developed in the previous section of this chapter may prove useful in this connection, as the next example illustrates. Example 11 Prove the quartic function defined by f1x2 = 3x 4 + 2x 3 - 5x 2 + 8x + 12 is everywhere continuous.

Section 2.3

197

Solution. We have f 1x2 = 12x 3 + 6x 2 - 10x + 8. That is, the derivative exists for every value of x, therefore the function is everywhere continuous.

However, you must be warned that the converse of the theorem is not true. That is, a function may be continuous but not differentiable. We have already considered examples of this in Section 3.1, where we observed that a sharp point on a curve (a function is continuous at a sharp point), the derivative does not exist, consider the next example. Example 12 x - 3 (a) At what points is f discontinuous? (b) One point is a removable x2 - 3x discontinuity. How would you define y at that point so as to remove the discontinuity? (c) At what points does f not have a derivative? (d) If you redefine f so as to remove the removable discontinuity, is the function differentiable there? Let f1x2 = Solution. (a) The denominator is x 2 - 3x = x1x - 32. This is equal to zero at x = 0 and x = 3. So the function is discontinuous at x = 0 and x = 3 because it is undefined at these two places. (b) No matter what you do as x approaches 0, f(x) increases without bound. For example at x = 0.0001, f10.0012 = 1000, as x gets closer to 0, the y-values grow even larger. Therefore, there is no way to define f(0) that will make the function continuous. However, the discontinuity at x = 3 is removable. If we simply divide out the factor 1x - 32 from numerator and denominator, we have f1x2 = 1/x, for x Z 3. Hence, lim f1x2 = 1/3. If we define f132 = 1/3, the function as redefined is continuous at x = 3.

x :3

(c) The function is not differentiable at x = 0 or x = 3 because it is not continuous at these two points. (d) Once the value has been properly defined at x = 3, we have simply f1x2 = 1/x = x -1, which has derivative f 1x2 = - x -2 = - 1/x 2, except at x = 0 where f remains undefined and, therefore, neither continuous nor differentiable. Before you jump to any conclusions, let us say that it is not automatic that when a discontinuity is removed the function becomes differentiable at that point. In fact, examples to the contrary may be found in the exercises.

As one further indication of how the limit definitions may be used, we give a formal proof of Rule 5, Section 2.2, the Sum Rule. We are given that f(x) and g(x) have f 1x2 and g 1x2 as their respective derivatives. We must show that s1x2 = f1x2 + g1x2 has derivative s 1x2 = f 1x2 + g 1x2. We use the definition of the derivative. It is useful to note that s1 2 = f1 2 + g1 2, therefore, 1. s1x + h2 = f1x + h2 + g1x + h2 g1x2 s1x2 = f1x2 + 2. s1x + h2 - s1x2 = f1x + h2 - f1x2 + g1x + h2 - g1x2 Therefore, s1x + h2 - s1x2 f1x + h2 - f1x2 + g1x + h2 - g1x2 = h h f1x + h2 - f1x2 g1x + h2 - g1x2 = + h h s1x + h2 - s1x2 4. s 1x2 = lim h:0 h 3.

198

Section 2.3

f1x + h2 - f1x2 g1x + h2 - g1x2 + lim = f 1x2 + g 1x2. h:0 h:0 h h Note that the proof really follows directly from the fact that the limit of a sum is equal to the sum of the limits. = lim

We can compute limits with functions of several variables, consider the next example. Example 13 Given f1x, y2 = 2x 2y 3 determine f1x + h, y2 - f1x, y2 (a) lim , h :0 h f1x, y + k2 - f1x, y2 (b) lim . k :0 k Solution. f1x + h, y2 - f1x, y2 21x + h22y 3 - 2x 2y 3 (a) lim = lim = h:0 h:0 h h 2 3 3 2 2 3 2x y + 4xhy + 2h y - 2x y lim = h:0 h h14xy 3 + 2hy 32 4xhy 3 + 2h2y 3 = lim 14xy 3 + 2hy 32 = 4xy 3 lim = lim h:0 h:0 h:0 h h f1x, y + k2 - f1x, y2 2x 21y + k23 - 2x 2y 3 = lim (b) lim = k:0 k:0 k k 2x21y 3 + 3y 2k + 3yk2 + k32 - 22y 3 lim = k:0 k 2 3 2 2 2 2 2 3 2x y + 6x y k + 6x yk + 2x k - 2x2y 3 lim = k:0 k 6x2y 2k + 6x2yk2 + 2x 2k3 lim = k:0 k 2 2 2 2 2 k16x y + 6x yk + 2x k 2 = lim 16x 2y 2 + 6x 2yk + 2x 2k22 = 6x 2y 2 lim k :0 k :0 k

Calculator Tips

Notice in the above example that the limit requested in (a) looks almost like a derivative with respect to x. In fact, observe that if you look at the expression, y does not change, so if you treated y as a constant it looks like the derivative with respect to the variable x. Similarly, the limit in (b) looks almost like a derivative with respect to y. We will have more to say about this when we study differentiation of functions of two or more variables. The TI 89 calculator can compute limits directly, but some care needs to be taken. The syntax for having the calculator compute a limit is limit (expression, variable, point [, direction]). The [,direction] portion of the command is optional and is only used for one-sided limit. For example, to compute 3x - 1 x x :0 lim we enter limit 113x - 12/x, x, 02, Figure 7 indicates the input and result. The calculator gives the answer as ln(3), which we shall see is the natural logarithm of 3, which is approximately 1.09861 (press * Enter).

Section 2.3

* **

199

Figure 7: Computing Limits with the Calculator To determine a one-sided limit we need to add the direction option, - 1 for a leftx handed limit and + 1 for a right-handed limit. For example, to compute lim- , we prox:0 x ceed as in Figure 8 (note abs(x) is the way we enter x on the calculator).

In Exercises 1 26, find the indicated limit. 1. lim 5 2. lim 9

x:7 x:0

x:1 x:2

3.

lim a 5x +

1 b 2

x:3

x : -2

x : -2 2

(b) lim f1x2, (c) lim f1x2. x2 - 9 11. Let f1x2 = , determine each of the following limits: (a) lim f1x2, x : -3 x - 3 (b) lim f1x2.

x:3

200

* **

Section 2.3

x

6x - 5 x2 36. If f1x2 = d 5x + 14 70

x:5

12. lim

x:1 3

14. f1x2 = 7x2 + 2h - 3, determine (a) lim f1x2 (b) lim f1x2. 15. lim 16. lim x2 - 7x + 12 x:3 x - 3 2t2 + 1

t : 2 3t2 h:0 x:0

if if if if

x:5

x 6 5 5 6 x 6 7 x 9 7 x 7 9

x:7 x : 77

Determine (a) lim- f1x2 (b) lim+ f1x2 (c) lim- f1x2 (d) lim + f1x2 (e) lim- f1x2 (f) lim+ f1x2 (g) at which x-values is f discontinuous, and classify

x:9 x:9

the discontinuities. (h) graph the function. In Exercises 37 39, decide whether the function whose graph is shown is continuous. If it is not continuous, identify the x-values at which it is discontinuous and classify the discontinuity.

a:9

2a - 5 a - 5 x2 + 5x + 2

19. lim

2 21. lim A 1 2 t + 1t B t:0

x : 0 2x2

22. lim

t:3

23. lim

h:0

21 + x + h - 21 + x 24. lim h:0 h 25. f1x2 = x , determine (a) lim f1x2 (b) lim f1x2 (c) lim f1x2 26. f1x2 = |x| , determine (a) lim f1x2 (b) lim f1x2 (c) lim f1x2 x:4 x : -4 x : -a x (d) lim f1x2 where a 7 0.

x:a x:1 x:0 x : -1

Figure 9: Ex. 37

27. (a) Draw the graph of a function which is continuous at each point in its domain. (b) Draw the graph of a function which is continuous at every point in its domain but is not differentiable at x = 0. 28. Determine any function which is discontinuous at x = 1 and x = 5 but which has a derivative at every other point. 29. Where is the function f1x2 = 1/1x - 52 discontinuous. Why? 1 if x 0 30. (a) Graph the function defined by f1x2 = e 2 if x 7 0 (b) Where is the function discontinuous? Why? 31. Is the function defined by f1x2 = e x 1 (b) x = 0?

1

37. Figure 9 38. Figure 10 39. Figure 11 In Exercises 40 46 decide whether or not the function is continuous. If it is not continuous, identify the points at which it is discontinuous. 40. f1x2 = x2 + 2x + 5 41. f1x2 = x2 - 3 x + 2

1 x

if x Z 0 continuous at (a) x = 2? if x = 0 2 if x 6 1 4 if x = 1 3 - x if x 7 1

32. (a) Graph the function defined by f1x2 = c (b) Where is it discontinuous?

(c) What kind of discontinuities does it have? 33. Discuss the continuity of the function defined by f1x2 = 34. Determine (a) lim+

x:3

x if x 6 5 x2 if x 5 0 if x 6 0 x if x 0

x2 - 9 x - 3

x:0

x3 x

Section 2.3

(a) Sketch the graph of the function. (b) Where is it discontinuous? (c) Redefine the function so as to remove the discontinuity. (d) Is the function as redefined differentiable at this point? In Exercises 48 52, each function is discontinuous at x = a. If the discontinuity may be removed, redefine the function so as to make it continuous at x = a and decide if the function as redefined is differentiable at x = a. 48. f1x2 = x2 - 4 x - 2 a = 2 a = 2 a = 2

* **

201

64. Suppose f1x, y2 = 3x2 + 2y2 + 5x - 11y + 7, determine f1x + h, y2 - f1x, y2 f1x, y + k2 - f1x, y2 . (a) lim (b) lim h:0 k:0 h k 65. (a) Give an example of a function which is defined for all values of x except x = 1 and x = 2; and discontinuous at x = 1 and x = 2. (b) Give an example if both discontinuities are removable. 66. Can a jump discontinuity be removed? Explain! 67. Suppose f is continuous at x = a and that f1a2 7 0. Show, by means of a sketch, that f1x2 7 0 in some vicinity of x = a. If f is not continuous at x = a, show the conclusion need not be true. 68. Discuss the discontinuities of D1x2 = e 0 1 if x is irrational if x is rational

x2 if x 6 2 3 if x = 2 6 - x if x 7 2

if x 6 3 x3 51. f1x2 = c 15 if x = 3 5x + 12 if x 6 3

a = 3

x2 - 5x + 6 52. f1x2 = a = 2 x - 2 In Exercises 53 62, discuss (i) the continuity and; (ii) the differentiability of the given function. (iii) If a discontinuity is removable, redefine the function so as to make it continuous there. (iv) Is it now differentiable? x2 - 9 53. (a) f1x2 = x - 3 x + 3 (b) g1x2 = e 4 if x 2 if 2 6 x if x 7 4 if x Z 3 if x = 3 (c) h1x2 = x + 3.

69. (a) Suppose f is continuous on 1 x 3, with f112 = 7, and f132 = 12. Show that every y-value between 7 and 12 is assumed at least once. (Hint: draw a sketch) (b) More generally, if f is continuous on a x b, with f1a2 = M, and f1b2 = N, then each y-value between M and N is assumed at least once. (This is known as the Intermediate Value Theorem.) 70. (a) Using the previous exercise, prove that there is some x-value between 2 and 3 at which f1x2 = x2 - 1 equals 6. (b) Find this value of x. 71. Using Exercise 69, prove that x 5 + x - 5 = 0 has a root between x = 1 and x = 2. 72. On July 2, Sam weighed 120 lbs. On August 2, he weighed 130 lbs. (a) Prove that at some time between these two dates his weight was 125 lbs. (b) Could his weight ever have been 135 lbs between those dates? 73. At 7 AM the outdoor temperature was 68*F and at 11 AM the temperature was 76*F, is there a time between 7 and 11 AM when the temperature was (a) 73*F? (b) 71*F? (c) 65*F?

x - 1 2 x - 1 2

if x 6 3 if x 7 3 if x 3 if x 7 3

(b) M1x2 = c

x - 1 1 2

if x 6 3 if x = 3 if x 7 3 if x 6 3 if x 3

(c) N1x2 = e

(d) P1x2 = e

x - 1 2

74. Mary begins her climb at the bottom of the mountain at 8 AM Monday morning and reaches the top of the mountain at 1 PM that afternoon. She camps there overnight and begins her descent on Tuesday at 8 AM and reaches the bottom of the mountain at 1 PM that afternoon. Is there a common time on each day when she is at the same elevation on the mountain? Explain. x b. Show that the function attains both a 75. Suppose f is continuous for a maximum and minimum value. That is, there are at least two numbers, c1 and c2, between a and b, such that, f1c12 = M and f1c22 = N, where f1x2 M, and f1x2 N, for all x such that a x b. This is called the Extreme Value Theorem x 3. (a) What is the maximum value 76. Consider f1x2 = x2 - 1 on - 2 that this function attains? (b) Its minimum? (c) For which x-values does it attain the maximum and minimum? 77. A manufacturer estimates that his daily cost, in dollars, of producing x television sets is C1x2 = 2000 + 500x + 8000/x. He must manufacture at least 2 sets, and due to storage limitations, he cannot produce more than 250 sets. Prove, by using Exercise 75 that there is some number of sets which will minimize his cost. x b. 78. Suppose y = f1x2 is a non-constant continuous function on a Also suppose that f1a2 = f1b2. Why must the graph of this function have at least one turning point between a and b?

57. f1x2 =

x2 - 2 x - 5 x x2 if x 6 1 x3 (b) g1x2 = e 2 if x 1 x if x 6 2 if x 2

2x + 1 62. g1x2 = c 3x - 7 4

63. Suppose f1x, y2 = 3x2y 2 determine (a) lim (b) lim f1x, y + k2 - f1x, y2 k

k:0

h:0

202

* **

Section 2.4

82. (a) Show lim x2 = 4. How close to 2 should the x-values be chosen to guarantee that the y-values are (b) between 4 - 0.1 and 4 + 0.1? (c) between 4 - .01 and 4 + .01? (d) between 4 - H and 4 + P? where P is any positive number less than 4. 83. Show lim f1x2 = f1a2 is equivalent to lim f1a + h2 = f1a2.

x:a h:0 x:2

79. Show that an alternate definition for the derivative at a is given by f (x) = lim f1x2 - f1a2 x - a

x:a

the derivative. 80. Using the alternate definition of the derivative from the preceding exercise, compute determine the derivative of each of the following: (a) f1x2 = 2x2 (b) f1x2 = 4/x. 81. Suppose in Example 2 we want to select x-values so the y-values lie (a) between 6 - .01 and 6 + .01, (b) between 6 - .001 and 6 + .001, (c) between 6 - P and 6 + P, where P is any positive number less than 6, determine how close to 3 the x-values need to be?

2.4

Limits at Infinity Dominant Terms Horizontal Asymptotes Infinite Limits Vertical Asymptotes Calculator Tips

Limits at Infinity

In this section, we shall re-examine the notion of a rational function which was discussed in Section 1.7, using the limit as a means of defining a horizontal and vertical asymptote. In Section 1.7, we saw that much of the information about such graphs can be determined by examining two facets of its behavior; the behavior of the function as the x-values get very large (in absolute value), and considering its values near the zeros of both its numerator and denominator. We saw that this investigation coupled with the sign analysis of the function itself will reveal the essential shape of the graph. Recall that a rational function has the form r1x2 = p1x2/q1x2, where both p and q 5x4 - 2x 3 - 7 are polynomials. For example, r1x2 = is a rational function. 3x7 + 8x 2 - 11 We begin with the investigation of the behavior of a function for large values of x. For example, what happens to the values of the function whose equation is f1x2 = 3x 2 - 7 2x2 + 1

as x assumes arbitrarily large values? This question is reduced in mathematical shorthand to, What is lim f1x2? Here, to say that lim f1x2 = L, means that L is the unique x:q x:q number (if one exists) that the y-values get very near, when the x-values get arbitrarily large. If the values of f(x) become unbounded, we say that the limit is infinite. Let us examine 3x 2 - 7 x : q 2x 2 + 1 lim for large values of x. Consider Table 1 which examines what happens to the y-values as the x-values get large. As x becomes arbitrarily large, the y-values approach 1.5, therefore,

Section 2.4

203

L = 1.5. We remark that for larger values of x, say 100,000 or higher, most calculators will give 1.5 as the answer! In fact, as far as your calculator is concerned, when it does the calculations for large values of x, the 7 in the numerator is Table 1: Examination of the Behavior as x Becomes Large

x f(x) 100 1.4996 1000 1.499996 10000 1.49999996

insignificant compared to the term 3x 2. Similarly, in the denominator the 1 is negligible compared to the 2x 2 term. Hence the calculator is eventually just evaluating 3x2/2x 2 = 3/2 = 1.5. The following observation gives us a very simple method for computing limits, often called limits at infinity of rational functions (when they exist). When x is very large only the terms having the largest exponents in the numerator and denominator have any effect upon the limit. In Section 1.7, we referred to these terms as the dominant or leading terms. (Sometimes we shall say for x being infinite when we mean for x very large or arbitrarily large.) We illustrate this observation with several examples. Example 1 8x 3 - 2x + 3 . x : q 4x3 - 5x 2 + 1

Dominant Terms

The dominant term in the numerator is 8x3, and the dominant term in the 8x 3 = 2. denominator is 4x3. Thus, the problem reduces to computing lim x : q 4x 3 It is important to realize that only the dominant term matters when x becomes arbitrarily large. All other terms, no matter how large their coefficients may be, are insignificant in comparison to the dominant term when x is large. Example 2 Evaluate lim Solution. 5x2 - 300 - 3x3 + 2

x : q 2x5

Ignoring all but the dominant terms, we have lim 5x2 - 300

5 3

x : q 2x

- 3x + 2

= lim

5x 2

5

x : q 2x

= lim

5 . x : q 2x 3

As x becomes arbitrarily large, the denominator gets very large, the entire fraction gets arbitrarily close to zero, and the given limit is zero.

The observation made in the preceding example generalizes, and we have the following:

THEOREM 1

lim k

p

x:q x

= 0

The above theorem, which is nothing more than a limit formulation of an observation we made in Section 1.7, provides an alternative method for computing limits as x becomes

204

Section 2.4

large. To compute such limits, factor the dominant term from both the numerator and denominator and then apply Theorem 1. In Example 1, we may write 2 3 2 + 3b 8 - 2 + 2 8x - 2x + 3 x x x lim = lim = lim x : q 4x3 - 5x 2 + 1 x:q x:q 5 1 5 4 + + 3b x3 a 4 x x x

3

x3 a 8 -

3 x3 1 x3

Using Theorem 1, we see that the second and third terms in both the numerator and denominator approach zero and the limit is 8/4 = 2. Example 3

2x 3 - 5x + 3 . x:q 3x2 + 5 Looking only at the dominant terms, we have 2x 3 - 5x + 3 2x . = lim x:q x:q 3 3x2 + 5 lim

As x becomes arbitrarily large, 2x/3 is itself unbounded. That is, it too becomes arbitrarily large number. Furthermore, it is always positive. We indicate this behavior by saying that the limit is plus infinity 1 + q 2. Thus, the required limit is + q . When we write + q for a limit, it means that the function is increasing without bound. In other words, the limit does not exist. When a limit does exist, it means that when x becomes arbitrarily large, the y-values tend to stabilize, or approach a finite equilibrium value. This equilibrium value is the limit. We shall call this the limit at infinity. Conversely, when the function has no limit as x takes on arbitrarily large values, it means that the function does not stabilize to an equilibrium value. Sometimes we find different behavior for a function as x decreases to - q . That is, when x takes on negative values which are, in absolute value, arbitrarily large. If f is such a function, and we want to examine such a limit, we write, lim f1x2. However, the procedure for finding such limits is x: -q exactly the same as when x is a large positive number. Example 4 Evaluate lim Solution. 3x 2 - x . x : - q 2x2 + 1 Keeping the dominant terms, we have, 3x2 - x = x : - q 2x 2 + 1 lim 3x 2 = 3/2. x : - q 2x 2 lim

We summarize the results found in the above examples with the following theorem which generalizes Theorem 1

Section 2.4

205

THEOREM 2

a nx n + a n - 1 x n - 1 bkx k + bk - 1x k - 1 an bn + + a0 = d0 + + b0 ;q if k = n if k 7 n if k 6 n

x: ; q

lim

where n and k are positive integers. Note that the theorem tells us that the limit at infinity of the ratio of two polynomials is determined by the degrees of the polynomials. Remember, the degree of the polynomial is just the exponent of the dominant term. Thus, that if the numerator and denominator are of the same degree, then for (absolutely) large x the limit is the ratio of the coefficients of the dominant terms. If the denominator is of higher degree than the numerator, the limit is 0. Finally, if the degree of the numerator exceeds the degree of the denominator, then the limit is infinite. The sign depends upon the sign of the coefficients of the dominant terms, whether k - n is an odd or even integer, and whether x is positively large or is an absolutely large negative number. The eight possibilities are examined in Exercise 37. It can be shown that, under appropriate conditions, the theorem generalizes, and is true even when n and k are not integers. Example 5 Let P be the population of a certain species of fish. P is a function of time, t. When t is measured in months and P in hundreds of thousands, P is given by the equation P = f1t2 8t2 + 3 = 2 . Determine the equilibrium population of the fish, that is, the limiting popula4t + 9 tion of the fish. By the equilibrium or limiting population, we mean the population after 8t2 + 3 many months. That is, when t is arbitrarily large. We have, lim 2 = 2. Thus, the t : q 4t + 9 limiting or equilibrium population is two hundred thousand fish. Solution.

For simplicity, in what follows we shall use the expression x gets large to mean x gets large in absolute value . Where there is a difference in behavior for positive and negative values of x, we shall point it out carefully. With that understanding, let us examine the behavior of a function as its x-values get large and see how its graph looks. In Example 1 we found that lim 8x 3 - 2x + 3 8x 3 - 2x + 3 Similarly, lim = 2. = 2. This means x : q 4x3 - 5x 2 + 1 x : - q 4x3 - 5x 2 + 1 that in a sketch of the function, when x is any very large value, the associated y-value must be essentially equal to the equilibrium value of 2. That is, the graph should begin to look like the line y = 2. We recall, from Section 1.7 that the horizontal line y = 2 is called a horizontal asymptote for the curve. In Figure 1 the horizontal asymptote y = 2 is drawn as a dashed line. Note that the graph is sketched only for large values of x. (The connection of the various component pieces is accomplished using the method of Section 1.7.) We can now redefine a horizontal asymptote using the limit.

Horizontal Asymptotes

206

* **

Section 2.4

y=2

x:q

a horizontal asymptote of f as x approaches q . If lim f1x2 = N, then y = N is a horix:-q zontal asymptote of f as x approaches - q . = 0. As x approaches - q we ob- 3x + 2 tain the same limit. (See Theorem 2.) This means that as x gets arbitrarily large the x-values approach 0. That is, y = 0 is the horizontal asymptote. Figure 2 gives a sketch of the graph of the function for absolutely large values of x, Figure 2a gives a sketch as x approaches - q , and Figure 2b gives a sketch as x approaches q ; note that in each case, the curve approaches the x-axis - y = 0, as the horizontal asymptote. To join these two components into one graph, you will need to determine the vertical asymptote and zeros, and proceed as in Section 1.6. This will be left to the exercises, where you will need a calculator to help you determine the approximate location of the vertical asymptote. A reasonable question at this time is, how do we know whether the curve is approaching the horizontal asymptote from above or below? Choosing a large positive (negative) number and then calculating the y-value at this number will indicate the direction. For example, in Figure 2, choose x = 100 and use your calculator to show that the corresponding y-value is about 2.0253. In Example 2 we found that lim

x : q 2x 5

5x 2 - 3

Section 2.4

* **

207

This means that the curve is a little above the asymptote. Similarly, if we let x = - 100 in Figure 3 we find that x is about - 0.0000025, just below the asymptote y = 0. The next question we examine in this section is, What happens if we seek a limit for a function which, after substitution, has the form N/0 and N is not zero? To understand what happens, we need to consider several examples. We shall see that the limit is undefined (sometimes said to be infinite). By infinite or undefined, we shall mean that the value of the function is (in absolute value) arbitrarily large. 1 Consider, for example, lim . Clearly the function is undefined at x = 2. Let x :2 x - 2 us examine the behavior of the function just to the right of x = 2. Suppose we choose x to be 2 + .01. Substituting, we obtain y = 1/12 + .01 - 22 = 100. Similarly, if we choose a number even closer to 2, say 2 + .00001, we get y = 10,000 as the result. Thus, if we choose any number just to the right of 2, the result is an arbitrarily large positive number, which we write as + q . Thus, we write, lim 1 = + q. - 2

Infinite Limits

x : 2+ x

That is, if we approach 2 from the right, (indicated by 2+), then the y-values obtained are arbitrarily large positive numbers, indicated by + q . Actually, our reasoning goes something like this. The numerator is always 1. If x is slightly larger than 2, then x - 2 is a very small positive number. Now 1 divided by a very small number gives a very large result. Since both the numerator and denominator are positive, we say that the limit is + q . Similarly, if we choose a number just to the left of 2, say 1.99999, we obtain - 100,000 as our result. Hence, as we approach 2 just from the left, (indicated by 2-) we 1 = - q . Since the problem asks us what happens as x approaches 2, have, limx :2 x - 2 without regard to direction, we say the limit does not exist. If the direction was indicated, say as x approached 2+, then we would indicate the limit by + q . Visualizing what is happening is easy. Since x = 2 is not in the domain of the function, its graph may not cross the vertical line x = 2. (If it did, at the crossing, there would be a y-value defined for x = 2.) To the left of this line, the y-values are arbitrarily large and negative, while to the right of x = 2 the y-values are arbitrarily large and positive. The line x = 2, which is drawn as a dotted line in Figure 3 is recognized to be a vertical asymptote. Note that to the left of x = 2 the y-values are negative and (absolutely) large, while to its right the y-values are positively large. More formally, we may redefine a vertical asymptote as follows:

Vertical Asymptotes

x :a x :a

vertical asymptote at x = a. Note that those x-values at which the denominator is zero (and the numerator is not zero) usually produce the vertical asymptotes. Therefore, particular attention must be paid to the zeros of the denominator. Example 6 Determine lim+

x :3

x . x - 3

Solution. Choose a value just to the right of 3, say 3.00001. We have, upon substitution, 3.00001/13.00001 - 32 = 300,001. This is a large positive number. Thus, we con-

208

Section 2.4

x :3

mptote.

We may generalize the above example to obtain the following theorem, in which we assume n is a positive integer.

THEOREM 3

1 q n = + x : a 1x - a2 1 + q if n is an even integer limn = e q x : a 1x - a2 if n is an odd integer lim+ With appropriate modifications, the various methods used for evaluation of limits are valid for infinite limits and limits at infinity. However, it should be clear that for continuous functions, such as polynomials, one can calculate the limits intuitively as follows: Suppose f1x2 = P1x2/Q1x2, and P1a2 = N Z 0 while Q1a2 = 0. If we try to evaluate lim f1x2 by substitution we will have something of the form N/0. This really means that x :a in the limit, we are dividing a non-zero number very close to zero. This will give a result of very large magnitude. Hence we need only determine the sign of the quotient to know if the limit is + q or - q . Example 7 Determine:

0.5x

(a) lim+

x :2

1 ; 1x - 223

(b) lim+

x :2

-x ; 1x - 223

2x - 7 1 - 3x + 425

Solution. (a) This follows immediately from Theorem 3 and we conclude that the limit is + q . (b) As x approaches 2+, the numerator approaches - 2. The denominator is 1x - 223, which must be positive since x 7 2. Therefore, we have - 2 divided by a very small positive number. The quotient is negative and, hence, the limit is - q . (c) As x approaches 4/3, the numerator approaches - 13/3. The denominator is 1 - 3x + 425, which is positive since x 6 4/3. Therefore, we have a negative number divided by a very small positive number, and the quotient is negative. The limit is thus - q.

Calculator Tips

Using horizontal an vertical asymptotes, along with the zeros and sign of the function we were able to produce a rough sketch of the graph of the function. However, missing from this sketch are the determination of any turning points of the functions, its peaks and valleys. We shall examine how to include such points, when they exist, in the next chapter. The Calculator can be used to determine limits, even as x becomes infinite (note that the q symbol is inserted by pressing *CATALOG). However, you need to be careful, as sometimes the calculator may give the wrong answer. Consider the following limit, lim 11 + .07/x20.5x. Using the calculator, we enter the limit and press Enter, see Figure 4. Note the calculator gives zero as the incorrect answer, the correct answer is approximately, 1.4191. How could you check to see if the calculator gives the correct result?

x: q

Section 2.4

* **

209

In Exercises 1 19 compute the indicated limit. 1. lim 2. lim 3x - 7x 2x - 3 5x2 + x + 3 6x2 + 7 3x4 + 7x - 11

2 2 x: q

26. f1x2 = x/1x2 - 12 27. f1x2 = x/11 + x22 28. f1x2 = 4x2/1x2 + 42 29. f1x2 = 4x2/1x2 - 42 30. f1x2 = 3x2/1x + 122 In Exercises 31-32 compute the indicated limit. 31. (a) lim 2x 2x 2 + 8 2x 2 - 1 x + 3

x: q

x: q

3. lim

x: q

2x 2x 2 + 8

x: -q

5. (a) lim x2

x: q

(b) lim x 2

x: -q 3

(c) lim x3

x: q

(d) lim x 3

x: -q 3

- 5x + x + 3

x: q

14 4 3

(b) lim

- 5x + x + 3 6x2 + 7

x: q

x: -q

2x 2 - 1 . x + 3

x: -q

x: q

In Exercises 33-36: (a) Find the x-intercept(s); (b) Find the vertical asymptotes; (c) Find the horizontal asymptotes. (d) Sketch the graph 33. f1x2 = 13x - 2212x + 322

x: -q

lim

9. (a) lim

x: q

+ 7x + 2

(b) lim

x: -q

x : q 12x

x215x - 32 2x 34. f1x2 = . 2x 2 + 1 2x 35. f1x2 = . 2x 2 - 1 1x - 121x + 2221x - 322 . 36. f1x2 = 1x + 12212x - 323 37. Normally if we say a limit L exists, we mean that L is a finite number. In what follows, we allow L to stand for either + q or - q . Let anxn + a n - 1xn - 1 + + a 0 lim = L If the integer n - k 7 0, show q x : bkx k + bk - 1x k - 1 + + b0 that L is as given in Table 2 for each of the eight cases indicated:

x: q 1

x : q 13x

x:0

1 x3

. 1 1x - 322 x 1x - 22

2

x:3

1 1x - 322 x 1x - 22 - 32

2

(b) limx:3

Table 2

.

x:2

(b) limx:2

n - k

even even even even odd odd odd odd

Sign of a nbk

positive positive negative negative positive positive negative negative

x approaches

+q -q +q -q +q -q +q -q

L

+q +q -q -q +q -q -q +q

16.

x : 3 - 1x

lim

1x + 22

x2 - 11x + 30 x2 - 11x + 30 17. (a) lim+ (b) lim + x:5 x : -5 x + 5 x + 5 1 1 . 18. (a) lim+ (b) lim + x : 5 1x + 523 x : - 5 1x + 523 19.

x:2

lim-

x - 2 x2 - 4

In Exercises 20 30: (a) Find the x-intercept(s); (b) Find the vertical asymptotes; (c) Find the horizontal asymptotes. (d) Sketch the graph. 20. f1x2 = 2x/1x - 12 21. f1x2 = 12 - x2/1x + 12 22. f1x2 = 12x - 32/1x - 12 23. f1x2 = 11 + 2x2/1x - 32 24. f1x2 = 12x + 42/14 - x2 25. f1x2 = x/12x - 32 38. (a) Determine the x-intercepts, (b) the vertical asymptotes, (c) the horizontal asymptotes and (d) Sketch the graph of the function whose equation is 8x3 - 2x + 3 f1x2 = . 4x3 - 5x2 + 1

210

* **

Section 2.5

Derivative Rules 2

yields f1x2 = 2x - 4 + 14x2 - 44x + 5 2x3 - 7x + 9

39. (a) Determine the x-intercepts, (b) the vertical asymptotes, (c) the horizontal asymptotes and (d) Sketch the graph of the function whose equation is 5x2 - 3 f1x2 = . You ll need to use your calculator to determine the 2x5 - 3x + 2 vertical asymptotes. Exercises 40 44 deal with oblique asymptotes. The line y = mx + b is called a slant or oblique linear asymptote of the function whose equation is y = f1x2 if either

x: q

This implies that the oblique asymptote is y = 2x - 4. Using long division, find the oblique asymptotes in Exercises 41 42. 41. f1x2 = 12x2 + x - 62/1x - 12 42. f1x2 = 13x4 - 2x3 + 2x2 - 3x + 22/1x 3 - 12 43. Sketch the graph of the function in Exercise 41. 44. Sketch the graph of the function in Exercise 42. 45. The calculator gave the wrong limit for lim 11 + .07/x25x, how could you obtain an approximately accurate result?

x: q

or

x: -q

40. Show that the given linear equation is a slant asymptote of y = f1x2 for (a) f1x2 = x2/1x + 12, y = x - 1; (b) f1x2 = 1x3 + 2x + 32/ 1x2 + 5x - 32, y = x - 5 Long division of polynomials may be used to determine the oblique asymptotes of a rational function in which the degree of the numerator exceeds the degree of the denominator by one. For example, performing long division on f1x2 = 4x4 - 8x 3 + 2x + 5 2x3 - 7x + 9

46. Find the x-values at which the graph in Exercise 33 crosses its horizontal asymptote. 47. Find the x-values at which the graph in Exercise 36 crosses its horizontal asymptote.

2.5

Derivative Rules 2

In Section 2.2, we began to develop the rules for finding the derivatives of simple functions. We can find the derivatives of powers of x and constant multiples of these powers. We can also handle sums and differences of functions whose derivatives we know. We shall now consider finding the derivatives of products and quotients of functions. Unlike the derivative of a sum, the derivative of a product is not the product of the derivatives nor is the derivative of a quotient the quotient of the derivatives. For example, consider y = x2. We know that y = 2x. However, we could write y = x2 = xx. If the derivative of the product were the product of the derivatives, then y = 1 # 1 = 1, which is clearly wrong. Therefore, let us start with the product rule. Assume that we have two differentiable functions, F(x) and S(x). (F for first , S for second .)

d 1F1x2S1x22 = F1x2S 1x2 + S1x2F 1x2 dx In words, the derivative of a product is the first (function) times the derivative of the second (function) plus the second times the derivative of the first. (The proof of the product rule can be found at the end of this section.)

Section 2.5

Derivative Rules 2

211

Example 1 Find f 1x2 if f1x2 = x512x 3 + 12 Solution. We do not have to use the product rule to find f 1x2, all we need do is distribute the x5 term and use the rules developed in Section 2.2. As an exercise, you should do it that way and compare your solution to the one obtained using the product rule. f 1 x2 = x 5 d d 12x 3 + 12 + 12x 3 + 12 1x 52 = x5[6x 2] + 12x3 + 12[5x 4] dx dx

(We enclose the factor that was differentiated in square brackets [...], for easy detection.) Now simplifying, we have f 1x2 = 6x 7 + 10x 7 + 5x 4 = 16x7 + 5x 4.

As we already mentioned, Example 1 could have been done without using the product rule. However, there are products which can not be done any other way, as we shall soon see. Beginners often confuse the product rule with the constant multiplier rule. While it is true that you may apply the product rule in finding the derivative of an expression like 3x5, it should not be done. By the constant multiplier rule and power rule, its derivative is simply 5 # 3x 4 = 15x 4. The product rule is meant to be applied when both factors involve the variable. If one of the factors is a constant, use the constant multiplier rule! Let us now consider the quotient rule. Again, assume that we have two differentiable functions, N(x) and D(x)

D1x2N 1x2 - N1x2D 1x2 d N1x2 a b = dx D1x2 [D1x2]2 In words, the quotient rule says the derivative of the quotient is the bottom (denominator) times the derivative of the top (numerator) minus the top times the derivative of the bottom divided by the bottom squared. (The proof of the quotient rule is left for you as an exercise.) Example 2 Find f 1x2 if f1x2 = Solution. 3x - 2 4x - 5

Applying the quotient rule, we have, 14x - 52[3] - 13x - 22[4] 14x - 52

2

f 1x2 =

(Again note that the square brackets [ ] indicate the expression that was differentiated.) Note that the sign in front of the 8 is + . To avoid any distribution errors, observe that - 13x - 22142 = - 413x - 22. It is absolutely essential that each term be enclosed within its own parenthesis. Not doing so is incorrect, and will often result in algebraic errors.

x2 . 4x 2 + 1

212

Section 2.5

Derivative Rules 2

Solution.

Applying the quotient rule, we have, 14x2 + 12[2x] - 1x 22[8x] 14x2 + 122 = 8x 3 + 2x - 8x 3 2x = 2 2 2 14x + 12 14x + 122

f 1x2 =

(Again note that the square brackets [ ] indicate the expression that was differentiated.)

Example 4 1x2 - 621x - 32 Find the equation of the line tangent to y = at the point when x = 4. 2 1x + 1 Solution. First, substitution of x = 4, yields y = 2. Now take the derivative to find the slope of the tangent line. Notice that the numerator of the fraction is itself a product. First use the quotient rule, 52 1x + 16 y = d d 51x 2 - 621x - 326 - 1x 2 - 621x - 32 52 1x + 16 dx dx 12 1x + 122

Now, the first term in the numerator has a factor that is the derivative of a product. For this we need the product rule. In addition, in order to take the derivative of +x, in the second -1/2 . numerator term, remember that 1x = x1/2. Thus, 1 1x2 = 1 2x Thus, y 1x2 = 12 1x + 1251x2 - 62[1] + 1x - 32[2x]6 - 51x 2 - 621x - 326[x -1/2] 12 1x + 122

We would ordinarily simplify this expression algebraically, but here we only need the value of the derivative at x = 4. Since it is easier to do arithmetic than algebra, let us sub-1 stitute x = 4 directly. Note that 24 = 2 and 4 *2 = 1/ 24 = 1*2. mtan142 = 5152[1102 + 112182] - A 1*2 B 11021126/1522 = 85/25 = 17/5 Hence, the equation of the tangent line is y - 2 = 117/521x - 42 or 5y = 17x - 58.

d -2 -2 1x 2 = - 2x -3 = 3 however, dx x we have not yet proven the power rule when the power is negative. We can use the quotient rule along with the power rule for positive powers to establish its validity. In this particular case, We know, by direct application of the power rule, that d -2 x2[0] - 1[2x] d 1 - 2x -2 1x 2 = a 2b = = = - 2x -3 = 3 4 4 dx dx x x x x More generally, let N be any positive integer, then we shall show, that d -N 1x 2 = - Nx -N - 1 dx That is, we bring down the power and decrease the power by one.

Section 2.5

Derivative Rules 2

213

When the power rule was applied, it was on the denominator term, xN, which has a posibM tive exponent. In the next to the last step we used the law of exponents N = bM - N. b

Thus, we have proven that the power rule is valid for any integer. We have yet to show that it is true for all rational and irrational powers. Example 5 x4 - 2x 5 + 7 , find f 1x2. If f1x2 = 3x2 Solution. We could apply the quotient rule, but it is not necessary. x4 2x 5 7 2 3 7 -2 2 + = 1 3x - 3x + 3x 2 2 3x 3x 3x 2

14 -3 3 x

2 f 1 x2 = 2 3 x - 2x -

The quotient rule is usually applied when the denominator is a sum (or difference) of two or more terms. If the denominator consists of one term, as in the above example, it is often more efficient to rewrite the expression and apply the other rules. We conclude this section with a derivation of the product rule. The proof uses the definition along with a very clever trick of adding and subtracting a function that gives us recognizable difference quotients. The same kind of trick may be used to prove the quotient rule, which we leave to the exercises. Consider P1x2 = F1x2S1x2, that is, P12 = F12S12. Then: Step 1. P1x + h2 = F1x + h2S1x + h2 P1x2 = F1x2S1x2 Step 2. P1x + h2 - P1x2 = F1x + h2S1x + h2 - F1x2S1x2 Step 3. P1x + h2 - P1x2 F1x + h2S1x + h2 - F1x2S1x2 = h h

We do not yet recognize the difference quotient. However, if we add zero written as the terms - F1x + h2S1x2 + F1x + h2S1x2 then we may rewrite the numerator of the difference quotient as

214

* **

Section 2.5

Derivative Rules 2

F1x + h2S1x + h2 - F1x + h2S1x2 + F1x + h2S1x2 - F1x2S1x2. Factoring we have the numerator as, F1x + h25S1x + h2 - S1x26 + S1x25F1x + h2 - F1x26 Dividing the numerator by h, we have, F1x + h2 a S1x + h2 - S1x2 h b + S1x2a F1x + h2 - F1x2 b h

If we now allow h to approach zero, then the difference quotients enclosed in the large parentheses become S 1x2 and F 1x2 respectively. Since F is continuous, F1x + h2 approaches F(x) as h approaches 0, thus, Step 4. P 1x2 = F1x2S 1x2 + S1x2F 1x2. * It certainly was a clever trick to add on and subtract off the term F1x + h2S1x2. It is surprising how often mathematical proofs hinge upon just such tricks. However, we shall see another way of deriving this result when we study logarithmic differentiation. It may happen that your answer for a derivative appears different from the answer given in the text. Sometimes it may take some clever algebraic manipulations to show the answers are equivalent. There is a way around this using your calculator, using the solve command. Suppose text answer is the answer given by the text, and your answer is the one you obtained. We assume the independent variable is x. Enter solve1text answer = your answer, x2 If, after pressing Enter, the calculator responds with TRUE then you know the answers are equivalent, if it responds FALSE then they are not. For example, suppose your an1x + 12-2 1 s answer is swer is and the text , then we need only enter x - 1 1x + 123/21x - 121/2 Ax + 1 solve11x + 121-22/*11x - 12/1x + 122 = 1/1x + 1213/22*1x - 1211/22, x2 If this is entered properly the TI 89 responds with TRUE.

Calculator Tips

1. f1x2 = 17x3213x4 - 92. Find f 1x2 in two different ways. 2. f1x2 = 15x + 2212x + 92. Find f 1x2 in two different ways. 3. f1x2 = 1x2 + 221x4 - 72. Find f 1x2 in two different ways. 4. f1x2 = x612x4 - 3x + 42. Find f 1x2 in two different ways. 5. Find f 1x2 if f1x2 = 6. Find y if y = dy x2

2 6 4

4x3 + 5 2x2 + 7

d 3 a b by (a) using the power rule and; (b) using the quotient rule. dx x2

2x + 3 3x - 2

d x5 - 2x 2 + 3 a b by: (a) writing it as a sum of powers and; 4x dx (b) using the quotient rule. d 4x6 + 3x 3 - 8x a b by: (a) writing it as a sum of powers and; dx 6x5 (b) using the quotient rule.

x + 1 3 - 5x 7. Find if y = . dx x + 3

11. Find

Section 2.6

In Exercises 12 15, determine the equation of the tangent line to the given curve at the indicated x-value. 12. f1x2 = 13. f1x2 = x = 1. ; x2 - 3 1x + 121x2 - 3x + 22 x4 + 1 4 x + 3x2 - 8 x2 + 3

* **

215

line crosses the x-axis. (c) Write the expression x3 - 4x in factored form and find the roots of x3 - 4x = 0. Do you see any relationship between these roots and the intercept of the tangent line? [See Exercise 24 below.] 24. A cubic equation of the form y = ax3 + bx 2 + cx + d having three distinct roots (x-intercepts) x1, x2, x3, may be written as y = a1x - x12 1x - x221x - x32 (why?). Show that the tangent line to the curve at 1x0, y02, where 1x0, y02 is midway between two successive roots, has the third root as its x-intercept. 25. (a) Find the cubic y = f1x2 with roots - 3, 1, and 5 and f102 = 30. (b) Verify the result obtained in Exercise 24 for the tangent lines drawn to the curve at x = 3 and x = - 1.

x = 2 x = 2 x = 1

14. f1x2 = x2 a

16. Find those points at which the tangent line to y = f1x2 is horizontal for: (a) f1x2 = x/1x2 + 12. 17. Repeat Exercise 16, for: (a) f1x2 = x

1

*2/1x

*2.

26. Explain where the fact that a differentiable function must be continuous was assumed in the proof of the product rule for derivatives. 27. This exercise illustrates how derivatives may be used to compute limits in one f1x2 very special and important case. Consider lim , where f and g are differx : a g1x2 entiable functions with f1a2 = g1a2 = 0. (a) Show that lim f1x2 = f 1a2 g 1a2

18. Find all points at which the derivative of y = 1x2 - 3x - 422 is equal to zero. [Hint: To find the derivative, remember that A2 = A # A.] 19. For the function y = x/1x2 + 12: (a) Find the derivative. (b) Find the equations of the tangent lines at x = 0, x = 1, x = 2, and x = 3. (c) Sketch short segments of the tangent lines found in (b), use them to get an idea of the shape of the graph of y = f1x2 between x = 0 and x = 4. 20. Prove the quotient rule using the same kind of trick that was used for the product rule. 21. Here is another way to derive the quotient rule. Observe that N1x2 = N1x2 d N1x2 D1x2a b . Now differentiate both sides and solve for a b. D1x2 dx D1x2 22. (a) What happens to the product rule if F1x2 = S1x2? (b) Derive a product rule for y = F1x2S1x2T1x2, where F, S, and T are each differentiable functions. Show that in words it says: to multiply two factors by the derivative of the third factor, and do this in all three possible ways and add the results. [Hint: Consider 5F1x2S1x26T1x2 and apply the product rule twice.] (c) Guess the product rule for y = F1x2S1x2T1x2W1x2, where F, S, T and W are each differentiable functions. (d) Rewrite (b) if F = S = T. (e) Using (c), what do you think y is, if y = [F1x2]N, where N is a positive integer? 23. Let y = x3 - 4x. (a) Find the equation of the line tangent to this curve at x = 1. (b)Find the x-intercept of the tangent line, that is, the point where the

x : a g1x2

provided g (a) Z 0. Hint: Use the alternate definition of the derivative as given in Exercise 37 in Section 3.1. (b) Show that if the derivatives of both f and g are continuous functions then the above result may be written as

x : a g1x2

lim

f1x2

= lim

f 1x2

x : a g 1x2

This result is a special case of a theorem known as L Hpital s rule. 28. Use the results of the previous exercise to compute each of the following limx3 - 8 x4 - 1 . its: (a) lim (b) lim x:2 x - 2 x:1 x - 1 29. Explain the relationship between the answers in parts (a) and (b) of Exercises 16 and 17. Hint: let g1x2 = 1/f1x2 and find g 1x2 using the quotient rule.

The Chain Rule The General Power Rule Calculator Tips

Even with the product and quotient rules, we are severely limited in the types of functions that we can differentiate. We can only handle functions like x N or sums, products and quotients of such functions. However, how do we find the derivative of F1x2 = 12x3 + 5234? It is important that you understand why we cannot, as yet, find the derivative of this expression. It is not simply of the form, x to a power. F(x) can be thought of as [g1x2]34, where g1x2 = 2x3 + 5. That is, it is an expression raised to a power. Thus, to compute F(x), you first find 2x 3 + 5, and then raise the result to the 34th power. This two stage process is typical of a composite function.

216

* **

Section 2.6

Before we give the rule which will enable us to differentiate such functions, we review the notion of a composite function as discussed in Section 1.2 where we saw that it is sometimes convenient to represent a function by a diagram as given in Figure 1.

x+

+ f(x) = y

Figure 1: Function Diagram y = f1x2 Thus, if x is in the domain, the corresponding range value is y = f1x2. The boxlike figure indicates the rule or correspondence used in determining y. For example, we might have f1x2 = x34. Suppose now that we have a second function, y = g1x2 and we represent the process by box-like figures which are attached in series as in Figure 2. (For example, g1x2 = 2x3 + 5.)

x+

g (x) +

f(g(x)) +

We note that in order to obtain the final output, x must be such that it is in the domain of g, and g(x) is in the domain of f. All such x will give us the domain of the combined function y = f1g1x22, which is called a composite function. In our example, x goes into the g-box and 2x 3 + 5 comes out. This then goes into the f-box and comes out 12x3 + 5234. We note that if the boxes were switched, the f box first followed by the g box, we would obtain a different composite function, namely y = g1f1x22. It is easy to determine the composite function once we are given the component functions, as the next example illustrates. Example 1 Find the composite functions f(g(x)) and g(f(x)) if: (a) f1x2 = x23 and g1x2 = x 3 + 11; (b) f1x2 = 1x and g1x2 = 3x2 + 5. Solution. (a) f12 = 1223 and g12 = 123 + 11. To find f(g(x)), we fill the parentheses in f() with g(x). Thus f1g1x22 = 1g1x2223. But, we know that g1x2 = x 3 + 11, and so, f1g1x22 = 1x3 + 11223. Similarly, to find g(f(x)), we fill the parentheses in g() with f(x). Thus, g1f1x22 = 1f1x223 + 11 = 1x2323 + 11 = x69 + 11. Notice the answers are not the same.

Section 2.6

217

(b) In this case, f12 = 212 and g12 = 3122 + 5. Now, f1g1x22 = 21g1x2) = 213x2 + 112. Similarly, g1f1x22 = 31f1x222 + 5 = 31*x22 + 5 = 3x + 5 Now we can pose the reverse question, given a composite function, can we decompose it into its component functions? The answer is yes, but the decomposition may not be unique. We will indicate why this is so shortly, but this will not be a problem for us. There is almost always a natural or suggested decomposition that will be obvious. For example, given the function H1x2 = 1x2 + 9234, we can write H(x) as the composition of two very simple functions. Think of H1x2 = f1g1x22, then H1x2 = 1g1x2234, where g1x2 = x2 + 9, so let f12 = 1234, that is, f1x2 = x 34. It was very natural to choose this as the decomposition, since the function H(x) was some expression to a power, the expression being g(x) and the power function being f(x). Example 2 Rewrite each function as the composition of two functions. (a) D1x2 = 134x 3 + 9x + 11254 (b) T1x2 = 1x3 - 222 + 51x 3 - 226 Solution. (a) D(x) looks like something to the 54th power so we call the expression within the parentheses, g(x). f(x) is the power function, that is g1x2 = 34x3 + 9x + 11 and f1x2 = x54. Now, D1x2 = f1g1x22. (b) Inspection of T(x) shows that it is the sum of two terms. Each term involves within the parentheses, the expression x 3 - 2. If we let g1x2 = x3 - 2, then T1x2 = [g1x2]2 + 5[g1x2]6. Hence, we let f1x2 = x 2 + 5x6, and T1x2 = f1g1x22. Since we can handle the derivatives of sums of functions independently, you will soon see that for purposes of finding derivatives, this kind of decomposition is not a natural one. That is, it is more useful to think of this one as T1x2 = f1g1x22 + 5h1g1x22; where g1x2 = x3 - 2, f1x2 = x 2 and h1x2 = x 6 The choice of decomposition soon becomes very natural. As a guideline, try the following. If we have a function H(x) that we want to decompose as f(g(x)), choose f(x) to be a simple function and g(x) to be the messy or inner part. We shall sometimes refer to g(x) as the core or inner function. In fact, for now, the only good choice for f(x), the outer function, will be the power function x N. There are other decompositions that one could choose, as in (b) above but they are not very helpful. Even when using the power function, you should look for the natural decomposition. For example, consider the function H1x2 = 1x2 + 228 we could write this as 21x 2 + 2216. This is not the natural thing to do, but it could be done, then we could choose f1x2 = *x and g1x2 = 1x 2 + 2216. Of course, the natural decomposition is f1x2 = x 8 and g1x2 = x2 + 2. Now we can return to the original question posed at the opening of this section. Functions defined by expressions such as 12x 3 + 5234, that we want to be able to differentiate are composite functions. We shall give a rule that will enable us to determine the derivative of the composite function by thinking of it in terms of its components. The rule is called the chain rule. We assume that all component functions are differentiable over common domains.

RULE 8 THE CHAIN RULE Given H1x2 = f1g1x22, then H 1x2 = f 1g1x22g (x)

Sometimes the chain rule is written a little differently. Let u = g1x2, and y = H1x2 = f1u2. Then we may write dy dy du dH = = dx dx du dx

218

Section 2.6

The chain rule states the following: To find the derivative of a composite function, you first apply the usual rule on f. However, this is not all there is to it, since f is a composite function. We must multiply by a correction factor, which is g 1x2, the derivative of the core function. At this point, the most important application of the chain rule is to functions of the form [g1x2]N. This is known as the General Power Rule. It is derived by simply letting f1x2 = xN. Since f 1x2 = Nx N - 1, we have: The General Power Rule

RULE 8A THE GENERAL POWER RULE If H1x2 = 1g1x22N then H 1x2 = N1g1x22N - 1g 1x2

or more simply as du d N u = NuN - 1 dx dx where u is a differentiable function of x. We illustrate the chain rule (general power rule) with several examples. Example 3 Find H 1x2 if H1x2 = 13x2 + 7223 Solution. We may write H1x2 = f1g1x22 where f1x2 = x23 and g1x2 = 3x2 + 7, f 1x2 = 23x22. Therefore, f 1g1x22 = 231g1x2222 = 2313x 2 + 7222. In addition, g 1x2 = 6x. Thus, we have H 1x2 = f 1g1x22 g 1x2 = 2313x 2 + 7222 6x = 138x13x 2 + 7222. Note that we never really need to do the decomposition in this detail. Basically, the general power rule states that we use the simple rule on the power, that is bring down the power and decrease the power by one and then multiply this expression by the derivative of the inner part, thus we immediately have, H 1x2 = 2313x 2 + 7222 6x = 138x13x 2 + 7222.

Example 4 Find the equation of the line tangent to y = 4 at the point where x = 1. 3x - 1

Solution. Of course, when x = 1, we find y = 2 (verify!). The required equation is then y - 2 = m1x - 12. Thus, we need to find m, the slope of the tangent line. To find the slope we must find the derivative. Now, we could think of the function as a quotient and use the quotient rule. Instead, let us rewrite it as y = 413x - 12-1 and use the chain rule. By the general power rule y = 1 - 4213x - 12-2132, where the factor 3 is the derivative of the inner or core function 3x - 1. Thus, y = - 1213x - 12-2 = - 12 13x - 122

Section 2.6

219

Now that we have the chain rule, we may use it along with some of the other rules to differentiate more complicated functions. The application of the various differentiation rules is usually the easier part. The more complicated part could well be performing the algebraic simplifications. Of course, if the problem is numerical, that is, asking us to compute the derivative, or to determine the equation of the tangent line at a particular point, then the algebra is not necessary. All that has to be done is to use the appropriate rules and then substitute a number for the independent variable immediately after the differentiation. We illustrate both approaches in the next few examples. Example 5 If f1x2 = x41x 2 + 3x + 525, find f 1x2. Solution. we have Applying the product rule with F1x2 = x4 and S1x2 = 1x2 + 3x + 525, d 2 d 1x + 3x + 525 + 1x 2 + 3x + 525 1x 42 dx dx

f 1x2 = x4

Applying the chain rule (general power rule) to the derivative of S(x), we get f 1x2 = x4 51x 2 + 3x + 524 12x + 32 + 1x 2 + 3x + 525 4x 3. We have concluded the calculus portion of the problem. However, in many problems, you might need to work with the derivative function. For example, as you will see in the following chapters, you might want to know where the derivative is equal to zero. You might even need to know the derivative of the derivative function. In such cases, you would have to simplify f 1x2. Let us do that now. Factoring the common factor x31x2 + 3x + 524, yields f 1x2 = x31x 2 + 3x + 52455x12x + 32 + 41x 2 + 3x + 526. Distributing the 5x, combining like terms within the braces, and factoring out a 2, reduces the expression to f 1x2 = 2x31x2 + 3x + 524114x2 + 27x + 202.

The algebra probably takes as least as long as the differentiation. With practice, you will be doing the differentiation part of the problem very quickly. To show you how easy it is to find the derivative of a complicated expression if it is to be evaluated at a particular point, consider the next example. Example 6 If f1x2 = Solution. x2 + 3 , find f 112. B x2 + 8 We write f1x2 = a x 2 + 3 1/2 b , therefore, we have, by the chain rule, x2 + 8

220

Section 2.6

1 x 2 + 3 -1/2 # d x 2 + 3 a a b b dx x2 + 8 2 x2 + 8

We could have simplified before we did the evaluation, but that would have involved steps that are not needed in determining the derivative at the given x-value.

There is another way of describing when the chain rule must be applied. Consider d N 1x 2; there is no question that the chain rule is not applied because we are differentiatdx ing xN with respect to x. However, if the differentiation were with respect to any other variable, then we would need to use the chain rule. More precisely, if we want to differentiate f(g(x)) with respect to x, then the chain rule must be applied whenever g1x2 Z x; that is, whenever the inner expression is different from the independent variable, the chain rule must be applied. d Consider 1x22. We want to find the derivative, with respect to the variable t, of dt x2. It is understood that x is a function of t, therefore, the chain rule must be applied and dx d 5 the derivative is 2x . Similarly, if we want to determine 1y 2, the chain rule must be dt dx dy applied since the variables are not in agreement. We have, as the derivative, 5y 4 . More dx generally, we have dy d N 1y 2 = Ny N - 1 dx dx we recognize this derivative as a restatement of the extended power rule. Example 7 Determine: (a) d 16 + 2x2 + y 52 dx (b) d 2 3 1x y 2 dx

Solution. Observe that we are differentiating with respect to the variable x. Thus, any time we take the derivative of a variable that is different from x, the chain rule must be applied. (a) The derivative of 6 will be zero and the derivative of 2x 2 is 4x, as usual, but the derivative of y5 is not simply 5y 4. To differentiate y 5 properly, remember that y is some function of x and use the chain rule. dy d 16 + 2x 2 + y 52 = 4x + 5y 4 dx dx

Section 2.6

221

(b) In this case, we must start with the product rule, thinking of x 2 as F(x) and y3 as S(x). dy dy d d d 2 3 1x y 2 = x 2 1y 32 + y 3 1x22 = x 23y 2 + y 32x = 3x2y 2 + 2xy 3 dx dx dx dx dx Calculator Tips

We shall have a lot more to say about this in Section 2.8, where we will want to find the slope of a curve whose equation is not in the explicit form y = f1x2. The key idea is the following: if you are differentiating a function of one variable with respect to a different variable, then the chain rule must be applied. As we mentioned before, the calculator can differentiate functions very easily. The syntax is d(expression, variable) where d is the derivative symbol above the number 8, expression is the function to be differentiated, and variable is the differentiation variable. Given any differentiable function, we can find its derivative with respect to x without any thought to the process using the calculator, it does all the work for us. The answer it gives may not always be in the form we want it, but that can be easily resolved. Consider x2 + 3 d a b , see Figure 3. dx B x2 + 8 Figure 3: Differentiation Using the TI 89

In Exercises 1 7, find: (a) f(g(x)) and; (b) g(f(x)). 1. f1x2 = x + 5, 2. f1x2 = 3x - 2, x - 2 , x + 4 g1x2 = x 2. g1x2 = 1/x. 1 . x - 1 20 . 25u2 + 9 18. Find dy/dx if y = x61x2 - 928. 1x2 - 123 19. Find y if y = 2 . 1x + 123 2 2 2 20. Find y if y = 1x - 32 1x + 2x + 828. 17. Find dr/du if r = 21. Find dy/dx if y = x + 1 . Ax - 1

22. Find dw/dr if w = 15r2 - 222/31r2 - 121/2. 23. Find ds/dt if s = 13t2 + 1221t3 - 6t + 127. 24. Find dv/dt if v = t # 2 3 t2 - 1 25. Find dv/dt if v = t2 # 2 3 12t + 822 In Exercises 26 29, find the equation of the tangent line to the curve at the given x-value. 26. y = 1x2 - 123; 1x2 + 123 27. y = ; 1x + 124 28. y = 12x2 - 7x + 1241 - 6x + 223; 29. y = 1x + 822/315x3 - 7x + 1621/2 30. Compute: (a) d 7 1y 2, dy (b) d 7 1y 2, dx (c) x = 2 x = 1. x = 0 x = 0 d A 1x5 + 12y7 B dx

1 x - 1 7. f1x2 = . , g1x2 = x 1 - x In Exercises 8 12, find two functions f and g, whose composition f(g(x)) will result in the given function h(x). 8. h1x2 = 13x2 - 729. 9. h1x2 = 15x - 7x + 32 . 10. h1x2 = a 2x + 3 3 b . 2x - 4

3 3/5

11. h1x2 = 12x - 123 + 512x - 122 + 5. 4 . 2x2 + 4 13. f1x2 = 1x2 + 3x + 1217, find f 1x2. 12. h1x2 = 14. y = 15x2 - 7x + 223/4, find dy/dx. 15. y = 13t2 + 2t2-1/2, find dy/dt. 16. Find f 1x2 if f1x2 = 22x3 + 3x + 2.

222

* **

Section 2.7

(a)

(b) d 3 1y + 3y 2 + 42; dx 40. (a) Given f1x2 = x2 - 4, and g1x2 = 1x2 - 4210. Find the points of horizontal tangency for f(x) and g(x). (b) Repeat for f1x2 = x 2 - 3x + 2 and g1x2 = 1x2 - 3x + 2215. (c) What can you deduce about the relationship of the zeroes of f(x) and g(x) and the zeroes of their derivatives? (d) In general, if f(x) is any function, what can you say about the points of horizontal tangency of f(x) and [f1x2]N? 41. Given H1x2 = 551x2 - 423 + 1x2 - 42106 *2; (a) Rewrite H(x) in the form f(g(h(x))). (b) Find the derivative, (c) Find a general rule for the derivative of this, so-called, triple composition. 42. Suppose you are given a function defined by the equation y = T1x2 whose d derivative T 1x2 = S21x2, find A T12x32 B . dx 43. Differentiate 12x + x422 with respect to x3. Hint: let y = 12x + x422, u = x3 dy dy du and use = . dx du dx 44. Find d A 32x + 23x + 14x B . dx

3 2 5 3 d 13x + 122 12x - 92 + * when x = 1. 5 dx 1x + 12

1

d 3 1y + 3y2 + 42; dy

d 3 1y + 3y2 + 4210 dx u2 - 1

2

32. If f 1u2 = u2 - 1, and y = f1x22, find dy/dx. u + 1 While we have stated the chain rule, for the most part we examined the special case of the generalized power rule, as we have been dealing only with algebraic type functions. Later on in the text, we shall consider other kinds of functions which will utilize the chain rule. In Exercises 34 36 we anticipate other applications of the chain rule. 34. Suppose y = E1x2 where 35. Suppose y = L1x2 where d d 1E1x22 = E1x2, determine A E1x22 B . dx dx d d 1 A L1x2 B = , determine A L1x2 + 12 B . dx x dx 33. If f 1u2 = , and y = f1x32, find dy/dx.

d d 36. Suppose y = T1x2 where A T1x2 B = S21x2, determine A T1x22 B . dx dx 37. Given f 1x2 = 18x13x2 + 922, try to find f(x) by guessing. 38. Given f 1x2 = 48x31x 4 + 5212, try to find f(x) by guessing. 39. Given f 1x2 = 6x 51x 6 + 5210, try to find f(x) by guessing.

45. Find

B 1x2 + 227

2.7

Marginal Functions Average Cost Velocity Average and Instantaneous Rates of Change Calculator Tips

Marginal Functions

The slope of a curve is only one interpretation of the derivative. In many disciplines, such as in the natural or social sciences, you will find the derivative being used as a quantitative tool. We will first examine how the derivative is used in microeconomics. You will see that when an economist discusses marginal functions, he or she is doing nothing more than examining the derivative of a function. Let us begin by looking at a very simple, but possible, economic situation. A man sells pretzels at a small refreshment stand in a shopping mall. His rent for the stall, utility bills, insurance, and other fixed costs total $800 per month. He buys the pretzels from a wholesaler for 50 or 1*2 dollar apiece. Thus, his total cost, in dollars, of doing business each month is given by the function C1x2 = 1*2 x + 800, where x = the number of pretels that he buys. This expression is called a linear total cost function. You can see that his overhead, $800, is the y-intercept and the cost per item, 1*2, is the slope. The slope is known as the marginal cost. It is the cost of acquiring (or in the case of a manufacturer, producing) one additional item. For the linear function, of course, the slope and hence the marginal cost, is a constant. However, for more complex situations, the total cost will not be a linear function. Suppose, for example, that for a manufacturer, the total cost, in dollars, of producing x items is given by the equation C1x2 = x 2 + 70x + 2000. Notice that the overhead cost is C102 = $2000. That is, it costs the producer money even if he

Section 2.7

* **

223

does not produce anything. The cost of producing 100 items is C11002 = 1002 + 7011002 + 2000 = $19,000, while the cost of producing 99 items is C1992 = 992 + 701992 + 2000 = $18,731. Consider the expression C11002 - C1992. It represents the cost of producing the 100th item, which is $19,000 - $18,731 = $269. Economists call this difference the marginal cost when the production level is 99. If we denote the total cost function by C(x), the total revenue function by R(x), and the total profit function by P(x), then we have the following definition.

DEFINITION 1

C1x + 12 - C1x2, the cost of producing the x + 1st item, is called the marginal cost at production level x. R1x + 12 - R1x2, the revenue derived by the producer when x + 1 items are sold, is called the marginal revenue at x. P1x + 12 - P1x2, the producers profit due to the x + 1st item, is called the marginal profit at x. If we let E(x) represent any of these economic functions, then we shall show that the marginal economic function, E1x + 12 - E1x2 L E 1x2, that is, the marginal cost is approximately equal to the derivative of the cost function, the marginal revenue is approximately equal to the derivative of the revenue function, and the marginal profit is approximately equal to the derivative of the profit function. Assuming that each of these economic functions is differentiable, then we have, from the definition of the derivative that E 1x2 = lim E1x + h2 - E1x2 h:0 h E1x + h2 - E1x2 h

In most real cases it happens that h = 1 makes this approximation valid. Thus, we have E 1x2 L E1x + 12 - E1x2 That is, the derivative of the economic function is approximately equal to the marginal function. We should also observe that if we allow h = - 1 (remember, we can approach the point on the curve from either the left or right), then we would obtain the approximation, E 1x2 L E1x - 12 - E1x2 = E1x2 - E1x - 12 -1

Thus, the derivative, E 1x2, is approximately equal to the marginal function at level x or level x - 1. Example 1 Given that the cost, in dollars, of producing x short wave radios is C1x2 = x2 + 80x + 3500, (a) Find the cost of producing: (i) the 100th radio; (ii) the 101st radio. (b) Find C 11002 and interpret this derivative. Solution. (a)(i) We need to compute C11002 - C1992. C11002 = 1002 + 8011002 + 3500 = 21,500.

224

Section 2.7

C1992 = 992 + 801992 + 3500 = 21,221. Thus, the marginal cost at level 99 is C11002 - C1992 = $279. (ii) We need to compute C11012 - C11002. C11012 = 21,781. Thus, C11012 - C11002 = $281. That is, the marginal cost at production level 100 is $281. (b) Now, C 1x2 = 2x + 80. Therefore, C 11002 = 280. C 11002 may be used to approximate the marginal cost at level 99, or level 100. (Notice, that the error in either case is $1.00. The percent error in the approximation is less than 0.4%.)

In the above example, we found that the cost of producing 100 radios was $21,500. What is the average cost per radio? All we have to do is to divide the total cost by 100 to obtain the average cost per radio as $215. More generally, we have the following definition.

Average Cost C1x2 = C1x2 x

There are natural questions that can be posed at this time. If you were a producer, what should the level of production be to minimize your total cost? How do you maximize your revenue? How do you maximize your profit? How do you minimize your average cost? Notice that there are certain functions that you want to minimize (cost and average cost), while there are others that you seek to maximize (revenue and profit). We shall answer these questions in the next chapter. Example 2 below examines one such question when the demand and cost functions are linear. We shall assume that the derivative E 1x2 is such a good approximation for the marginal function E1x + 12 - E1x2, (or for E1x2 - E1x - 12) that we shall not distinguish between them. Thus, we shall call C 1x2 the marginal cost at x, R 1x2 the marginal revenue at x, P 1x2 the marginal profit at x, and C 1x2 the marginal average cost at x. Recall, from Chapter 1 that the total revenue, R1x2 = 1price per item2 # 1number of items2 = p # x Example 2 Suppose that the relationship between price of and demand for a certain type of large color television set is given by the demand equation 10p + x = 10,000, where p is the per unit price in dollars, and x is the number of sets demanded. If the producer s cost is C1x2 = 600x + 3000: (a) Determine the revenue function; (b) Determine the marginal revenue function. (c) Determine the profit function. (d) Determine the marginal profit function. (e) What is the price for each color television when the marginal profit is zero? (f) Sketch the profit function, drawing the tangent line when the marginal profit is zero. What does this profit represent? (g) What can you conclude about the profit at the price obtained in (e)? Solution. (a) If we solve the demand equation for price, we find p = - 0.1x + 1000.

Section 2.7

225

Therefore, R1x2 = p # x = 1 - 0.1x + 10002x = - 0.1x 2 + 1000x. (b) The marginal revenue function, R 1x2 = - 0.2x + 1000. (c) The profit function, P1x2 = R1x2 - C1x2 = - 0.1x2 + 1000x - 1600x + 30002 = - 0.1x 2 + 400x - 3000. (d) P 1x2 = - 0.2x + 400. (e) When P 1x2 = 0, we have - 0.2x + 400 = 0, or x = 400/0.2 = 2000. Thus the production level when the marginal profit is zero is 2000, and the corresponding price per color television is p = - 0.1120002 + 1000 = 800. That is, the price per color television is $800. (f) Since the profit function is P1x2 = - 0.1x 2 + 400x - 3000, we recognize it to be a parabola opening downwards. We are interested in those values of x in its domain for which x 0 (why?). We sketch the parabola in Figure 1. (g) Observe that the vertex (turning point) of the parabola occurs at the point at which the tangent line is horizontal. (This is true, in general for all parabolas, see Section 1.4.) At the vertex, when x = 2000 and p = 800, the point at which the marginal profit 1P 1x22 is zero, the profit is maximized. Its maximum value is P120002 = $397,000.

(2000, 397,000)

Figure 1: P1x2 = - 0.1x2 + 400x - 3000 Do you think that is an accident that the profit is maximized where the marginal profit is zero? Since the profit is P1x2 = R1x2 - C1x2. the marginal profit is P 1x2 = R 1x2 - C 1x2. When R 1x2 = 0, R 1x2 - C 1x2 = 0. That is, R 1x2 = C 1x2. In words, the marginal revenue equals the marginal cost. That means that the cost of producing one more item exactly equals the revenue obtained by producing it. In other words, there is no gain to be made by making the next item. As long as marginal cost is less than marginal revenue, it pays to keep producing more items. When the marginal cost becomes equal to the marginal revenue, it is time to stop. Let us now consider an application of the derivative that occurs in the natural sciences. We shall see that the derivative may be interpreted as an instantaneous rate of

Velocity

226

* **

Section 2.7

change. In this application, we assume that a particle is moving along a straight line. The distance of the particle measured from some fixed point, is given by the equation, s = f1t2, where s is position and t is time. More specifically, assume a ball is thrown vertically upwards with an initial velocity of 64 feet per second, from the ledge of a building 512 feet high. We shall later show that its height s at time t is governed by the equation, s = f1t2 = - 16t2 + 64t + 512. At the end of one second, the height of the ball is f112 = - 161122 + 64112 + 512 = 560 feet. At the end of two seconds, the height is f122 = - 161222 + 64122 + 512 = 576 feet. The distance traveled during the time interval [1,2], that is, from t = 1 to t = 2, is f122 - f112 = 576 - 560 = 16 feet. Its average velocity during this one second time interval is the total distance traveled divided by the total time elapsed. That is, f122 - f112 2 - 1 = 576 - 560 = 16 ft/sec. 2 - 1

Suppose we reduce the time interval from [1, 2] to [1, 1.1]. That is, from a time interval of length one second to one of length one-tenth of a second. Let us determine the average velocity during this one-tenth of a second interval. To begin, f11.12 = - 1611.122 + 6411.12 + 512 = 563.04. Therefore, the average velocity over the one-tenth of a second time interval [1, 1.1] is, f11.12 - f112 564.04 - 560 = = 30.4 ft/sec. 0.1 0.1 Does this difference quotient look familiar? By writing f11 + 0.12 in place of f(1.1) we have the average velocity over this one-tenth of a second time interval as f11 + 0.12 - f112 0.1 You are probably anticipating our next step. Suppose we make the time interval even smaller, say 0.01 of a second. That is, [1, 1.01], then the average velocity over this time interval is f11 + 0.012 - f112 560.3184 - 560 = = 31.84 ft/sec 0.01 0.01 Suppose now, that the time interval is [1, 1 + h]. That is, the time interval is h seconds. The average velocity over this h second time interval is f11 + h2 - f112 h If we ask what the velocity of the ball is at the instant when t = 1, it should be clear how we proceed. We let h approach zero in the above difference quotient. That is, the instantaneous velocity when t = 1, v112 = lim f11 + h2 - f112 h:0 h

Section 2.7

* **

227

v112 =

ds = f 112 ` dt t = 1 f1t + h2 - f1t2 h f1t + h2 - f1t2 h:0 h Average and Instantaneous Rates of Change

More generally, if s = f1t2 then the average velocity over the time interval [t, t + h] is vavg =

Observe that in the above example, the average velocity over the h second time interval corresponds to the slope of the secant line, and the instantaneous velocity corresponds to the slope of the tangent line. Thus, we may interpret the slope of the secant line as the average rate of change of y with respect to x on the interval [x, x + h], and the slope of the tangent line as the instantaneous rate of change of y with respect to x. Similarly, we may interpret the marginal cost as the instantaneous rate of change of cost with respect to production level. In general, we may interpret a derivative as an instantaneous rate of change. This is such a useful interpretation that we often delete the adjective instantaneous. Thus, when we ask for velocity, we mean the instantaneous velocity. Similarly, when we ask for a rate of change, we mean the instantaneous rate of change (the derivative). Summarizing, we have the following.

DEFINITION 3 Given the differentiable function defined by the equation y = f1x2, the average rate of change of the function on the interval [x, x + h] is

yavg = f1x + h2 - f1x2 h

the (instantaneous) rate of change at x is the derivative, f 1x2. Example 3 A ball is thrown vertically upwards with an initial velocity of 64 ft/sec from the top of a building 512 feet high. Assume the altitude of the ball (measured in feet) at time t (in seconds) is governed by the equation s1t2 = - 16t2 + 64t + 512. (a) Determine the height at the end of 3 seconds. (b) Determine its velocity at the end of 3 seconds. (c) At what time will its velocity be zero? (d) What is the maximum altitude reached by the ball? (e) At what time will its height be zero? (f) What is the impact velocity of the ball when it hits the ground. (That is, what is its velocity the instant before it strikes?) (g) Sketch the graph of s. Solution. (a) The height at the end of 3 seconds is s132 = - 161322 + 64132 + 512 = 560 ft. (b)The velocity is v1t2 = s 1t2 = - 32t + 64. At the end of 3 seconds, its velocity is v132 = - 32132 + 64 = - 32 ft/sec. The negative sign indicates that the ball is moving downward. (c) If the velocity is zero we have 0 = - 32t + 64. Solving for t, we find that t = 2 sec. (d) A positive velocity means it is moving upward. A negative velocity means that it is moving downward. Therefore, the ball will reach its maximum height at the instant when its velocity is zero. That is, it is at its maximum height when t = 2. If the velocity were positive, the ball would continue to rise. The height is governed by the equation s1t2 = - 16t2 + 64t + 512. Therefore, the maximum height, which occurs when v = 0, or t = 2, is s122 = - 161222 + 64122 + 512 = 576 ft. (e) To determine when the height is zero, we set s = 0 and solve the quadratic equation

228

Section 2.7

- 16t2 + 64t + 512 = 0. Dividing each side of the equation by - 16, we obtain, t2 - 4t - 32 = 0. Factoring, we have 1t - 821t + 42 = 0. Therefore, the solution is t = 8. (We reject the root t = - 4. Why?) Thus, it takes the ball 8 seconds to reach the ground. (f) The velocity at any time is v1t2 = - 32t + 64. The impact velocity, the velocity when s = 0, that is, when t = 8, is v182 = - 32182 + 64 = - 192 ft/sec. Again, the negative sign indicates that the ball is moving downward. (g) The graph of s versus t is a parabola and is given in Figure 2. Figure 2: s1t2 = - 16t2 + 64t + 512

When we study integral calculus, we shall show that the height, s(t), of any object that moves vertically, subject only to the force of gravity, with an initial velocity v0, and an initial height s0, is given by the equation

(1)

The constant g is called the acceleration due to gravity and is approximately equal to 32 ft/sec 2, or using metric units, 9.8 meters/sec 2. If the object is projected upwards, v0 is positive, while v0 is negative if the object is projected downward. You may assume equation (1) in the exercises. Of course, particles may move in many other ways. That is, it is possible for s(t) to take different forms. Consider the following example. Example 4 The location s of a certain particle from its starting position is given by the equation s1t2 = t3 - 3t2 for t 7 0. For what values of t is it moving in the positive direction? Solution. The particle is moving in the positive direction when the velocity, s 1t2, is positive. Since s 1t2 = 3t2 - 6t, we need only determine where 3t2 - 6t 7 0 3t1t - 22 7 0 The inequality is solved easily using sign analysis (see Section 0.5). For positive values of t, if t is between 0 and 2, the product of 3t and t - 2 will be negative, and for t 7 2 the product will be positive, thus s is moving in the positive direction when t 7 2. Figure 3 shows the path of the particle along the s-axis.

One additional remark. The average rate of change on the interval [x, x + h] was defined as Figure 3: s1t2 = t3 - 3t2 for t 7 0 yavg = f1x + h2 - f1x2 h

Section 2.7

* **

229

suppose we rename the interval [a, b], that is, we set x = a and x + h = b then h = b - a and we have yavg f1b2 - f1a2 = b - a

Calculator Tips

This is called the average rate of change of the function over the interval [a, b]. We have seen that the difference quotient has various interpretations. Geometrically, it represents the slope of the secant line, and in this section, it was interpreted as an average rate of change. We can use the calculator to compute the difference quotient by pressing the Catalog key and scrolling to avgRC (which stands for average rate of change). The syntax is avgRC(expression, var, h). Suppose y11x2 = 3x 2 - 4x + 7, has been entered in the Y = screen, then avgRC(y1(x), x, h) returns 6x + 3h - 4, as it should, see Figure 3. (Note that h is optional, if it is omitted, the calculator assumes you are using h = 0.001.)

1. The cost, in dollars, of producing x bicycles is given by C1x2 = 60 + 10x + 1000/x 1x 102. (a) What does it cost to produce: (i) 99; (ii) 100 bicycles? (b) What is the cost of producing the 100th bicycle? (i) Find it exactly; (ii) Use the derivative to find it approximately. (c) What is the error in using the derivative to approximate the marginal cost of the 100th bicycle? (d) What is the average cost per bicycle when producing the 100 bicycles? 2. Let C1x2 = 2x2 + 300x + 50 be the cost of producing x items. What is the cost of producing the 89th item? Give both the exact and approximate answers. 3. Let 10p + x = 100 be the demand equation, where p is the price per item when x items are demanded. (a) Find the total revenue when the level of production is: (i) 40; (ii) 41; (b) Find the exact revenue derived from the 41st item. (c) Find the approximate revenue derived from the 41st item. (d) What is the error if the derivative is used to approximate the marginal revenue? 4. Suppose that the number of riders per day on the New York City subway system is 4 (million) when the fare is $2.00. Suppose that the ridership will drop to 3.8 (million) when the fare is raised to $2.25. Assuming that the relationship between demand and price is linear: (a) Find the demand equation. (b) Find the revenue and marginal revenue functions. (c) Graph the revenue function and show that the maximum revenue is at the point where the marginal revenue is zero. (d) Find the price and number of riders that will maximize the total revenue. In Exercises 5 and 6, the cost and demand equations are given. (a) Determine the profit function. (b) What is the profit when the level of production is 100? (c) Find the marginal profit function. (d) What is the price per item when the marginal profit is zero? (e) Sketch the profit function. (f) What is the level of production when the marginal profit is zero? What does it represent? 5. C1x2 = 20x + 500, 20p + x = 1000. 6. C1x2 = 30x + 1000, 10p + x = 700. In Exercises 7 and 8, find the average velocity over the given time interval, if s = f1t2 is the equation of height as a function of time. 7. s = - 16t2 + 128t: (a) [0, 3]; (b) [3, 4]; (c) [0, 4]. 8. s = - 16t2 + 160t: (a) [0, 3]; (b) [3, 6]; (c) [3, 10]. 9. A ball is thrown vertically upward from the ground with an initial velocity of 176 ft/sec. (a) Determine its height s, as a function of time (use equation (1)). (b) Find the average velocity of the ball during the 2 second time interval [1, 3]. (c) Find the distance traveled over the time interval [t, t + h]. (d) Find the average velocity during the time interval [t, t + h]. (e) What limiting average velocity do you obtain if h approaches zero in (e)? (f) What is the velocity when: (i) t = 1? (ii) t = 3? (g) When will the velocity be zero? What does this mean? (h) How high will the ball go? 10. Suppose in Exercise 9, the ball was thrown vertically upward from a ledge 192 feet high. (a) How high will the ball go? (b) How long does it take for the ball to hit the ground? (c) What is its impact velocity with the ground? 11. A rocket is launched vertically upward with an initial velocity of 6400 ft/sec. (a) When will the velocity of the rocket be zero? (b) What is the maximum altitude the rocket will attain? 12. A ball is thrown vertically upward from the ledge of a building 256 feet high with an initial velocity of 16 ft/sec. What is the ball s impact velocity with the ground? 13. A helicopter is stationary at an altitude of 512 feet. A package falls vertically from it. Determine the time it takes the package to hit the ground and its impact velocity with the ground. In Exercises 14 17, find the average rate of change of y with respect to x on the given interval. 14. y = 3x + 2 (a) [0, 1]; (b) [3, 5]; (c) [x1, x2] interpret this answer 15. y = mx + b;

2

16. y = 2x - 7x + 2; [2, 3]. 17. y = 3x2 - 7x - 5; [3, 6]. In Exercises 18 22, find the instantaneous rate of change of y with respect to x at the given x-value 18. y = 3x + 2; 19. y = mx + b; x = 5 x = a x = 2

20. y = 3x2 - 7x + 2;

230

* **

Section 2.8

Implicit Differentiation

31. If y = 13x2 + 923/2, find the rate of change of y with respect to (a) x; 3x (b) 2 . x - 1 32. If y = 1x3 - 122, find the rate of change of y with respect to (a) 2x; 2x (b) mx + b; (c) 2 . x - 1 33. Find the rate of change of the volume of a cube with side x with respect to (a) a side; (b) its (surface) area. 34. Find the rate of change of the area of an equilateral triangle of side x with rex2 spect to its: (a) side; (b) perimeter. (Hint: A = 23 ) 4 The relative rate of change of y = f1x2 with respect to x at x = a is defined by f 1a2 . Find the relative rate of change of f(x) with respect to x, (or f(t) with respect f1a2 to t) in Exercises 35 38. 35. f1t2 = 3t3 - 4t2 - 9; 36. f1t2 = - 16t2 - 128t; x + 3 ; Ax - 2 a = 2. a = 8. a = 0.

21. y =

x + 3 ; Ax - 2

x = 6 x = 0

In Exercises 23 25, (a) Find the instantaneous rate of change of s as a function of t. (b) Find the value(s) of t for which the instantaneous rate of change is positive. 23. s1t2 = 3t2 - 12t 24. s1t2 = 4t2 - 8t 25. s1t2 = t3 - 3t2 - 9t 26. Let C1x2 = ax2 + bx + d represent the cost of producing x items. Find the error if C 1x2 is used for the marginal cost of the xth item. 27. Given the demand equation p + mx = b 1m 7 0, b 7 02. (a) On the same set of axes, graph the demand and revenue equations. (b) Show that the x-intercept of the marginal revenue function occurs at that x-value for which the revenue is maximum and is 1*2 the x-intercept of the demand equation. 28. An object is thrown vertically upward with an initial velocity v0 from a height s0. (a) Show that it takes the object the same time to go from s0 to its maximum height as to fall from its maximum height to s0. (b) Show that the velocity of the object when it returns to height s0 is - v0. (c) Why must (a) and (b) be true physically? (d) Give an expression for the maximum height assuming that v0 7 0. 29. A ball is to thrown vertically upward from the ground so that it may be caught by an individual on a ledge 128 feet high. What is the minimum initial velocity with which it may be thrown? 30. With what initial velocity must an object be projected vertically upward from the ground to reach a height of 600 feet?

39. Suppose the horizontal position x a particle has traveled in t seconds, 0 t 5 is given by the equation x = 5 - 25 - t. (a) How far will the particle have traveled when t = 1? (b) What is the velocity of the particle after t seconds? (c) What happens to the velocity as t gets close to 5?

2.8

Implicit Differentiation

With the exception of some of the examples on related rates, we have only determined the derivative of functions of the form y = f1x2. That is, we have y explicitly as a function of x. What if x and y are related by some equation for which it is not possible or desirable to write y as a function of x. For example, suppose we have x = y3 + y + 1 In this case, x is a function of y. You may draw its graph by substituting values for y and finding the corresponding x-values, see Figure 1. From the graph you can see that y is also a function of x. But could you solve this equation for y in terms of x? Probably not.* However, for each x, it is possible to determine y graphically or by various numerical schemes. The important question is, can we easily determine the derivative dy/dx? The answer is yes. How? By the chain rule! Recall, from Section 3.6, when the dependent and independent variables are the same, there is no need for the chain rule, but when they are different,

*There is a formula by which one can solve a cubic, but it is very cumbersome to use in practice.

Section 2.8

Implicit Differentiation

* **

231

d 5 1x 2 = 5x 4 (note the independent and dependent variwe must use it. For example, dx ables agree),

Figure 1: x = y3 + y + 1 dy d 5 1y 2 = 5y 4 . So, how do we find dy/dx dx dx when we do not have y explicitly in terms of x? The idea behind the method is simple. Even though we may not be able to solve for y in terms of x, we assume that y is a function of x. Then we differentiate the equation that gives the relationship between x and y, and solve for the derivative. This method is called implicit differentiation. One main difference is that when we find the derivative, it will usually be a function of both x and y, unlike the explicit case, where the derivative is obtained as a function of x alone. Remember, the purpose of the derivative is to determine the slope of a curve, the velocity, or a particular marginal value. To compute the numerical value of the derivative we only need to substitute. It should not bother us that we are substituting for both x and y. The next examples illustrates the idea. but if y is a function of x then by the chain rule, Example 1 Find dy /dx if x = y3 + y + 1 Solution. Since the two sides of the equation are equal to one another, their derivatives must be equal. We take the derivative of both sides with respect to x. On the left side we have just x, and the derivative of x with respect to x is 1. However, on the right side, we have an expression in y, not x. Remember that y is a function of x even though we do not know its exact form. Thus, we must use the chain rule on the right hand side, and get 1 = 3y 2 dy dy + dx dx

(Do you see what happened on the right? The derivative of y3 is 3y 2 dy/dx by the chain rule. The derivative of y is just dy/dx and the derivative of 1 is zero.) Now we solve for dy/dx. Factoring the dy/dx, gives 1 = 13y 2 + 12 and dividing by 13y 2 + 12 yields dy dx

232

Section 2.8

Implicit Differentiation

Example 2 Given the equation x3 + 3y 5 + 2x = 3y 7 + 2x 2 + 4. Determine the equation of the tangent line to the curve defined by this equation at the point (2, 1). Solution. We know that the answer to the problem is y - 1 = m1x - 22, where m, the slope of the tangent line is the derivative at the point (2, 1). Since we have the equation relating x and y, all we have to do is differentiate each side of the equation to find an equation involving x, y, and dy/dx from which we can determine the derivative. Differentiating each term of the equation we have, 3x2 + 15y 4 dy dy + 2 = 21y 6 + 4x dx dx

(Do you understand why the derivative factor is present after differentiating each term involving y? y is a function of x, so the chain rule had to be applied.) To complete the solution, we need only substitute x = 2 and y = 1. We obtain, 31222 + 151124 dy dy + 2 = 211126 + 4122 dx dx

Simplifying, combining like terms, and transposing, we have, - 6dy/dx = - 6, or dy/dx = 1 Thus, the equation of the tangent line is y - 1 = 11x - 22, or y = x - 1

Example 2 was a numerical problem. In any numerical problem, where you are given a point at which to compute the derivative (slope, velocity, or marginal value) substitute immediately after the differentiation step, that will make the computations the simplest. Suppose you choose instead, to find the derivative in terms of x and y, and then substitute afterwards. Then you must first group together all terms involving the derivative on one side of the equation, and everything else on the other side of the equation. Example 3 Find dy/dx for the function in Example 2. Solution. Let us save some writing by using y for dy/dx. In the above example, we would obtain, (writing y instead of dy/dx). 15y4y - 21y 6y = 4x - 3x 2 - 2 Factoring the left hand side of the equation, we obtain 115y 4 - 21y 62y = 4x - 3x 2 - 2 dividing by the term in parentheses, yields, y = dy 4x - 3x2 - 2 = dx 15y4 - 21y 6

Section 2.8

Implicit Differentiation

233

As you can see, it is far easier to do the substitution immediately after the differentiation, rather than after the algebra. However, if you need to find the derivative at several points, or to find a general expression for the derivative, then perform the algebraic manipulations. The next example illustrates the use of the product rule in finding the derivative implicitly. We will use y to indicate dy/dx whenever we find it convenient to do so. Example 4 Find y if 3x2y 4 - 2x 2 + 3y 5 = 17. Solution. We differentiate each side of the equation. Note that the first term involves a product, so the product (and chain rule) must be applied. We have, 3x2[4y 3y ] + y 4[6x] - 4x + 15y 4y = 0. or, y 112x 2y 3 + 15y 42 = 4x - 6xy 4. Dividing, we find that y = 4x - 6xy4 12x2y3 + 15y4

Even when we can solve for y in terms of x, we may prefer to find the derivative by implicit differentiation. The next example illustrates this. Example 5 The equation x2 + p2 = 25 represents a demand equation when x and p are each in the first quadrant, that is, x 0, and p 0. Find: (a) dp/dx; (b) dx/dp; (c) dR/dx (d) dR/dp. Solution. We could solve for p as a function of x and find p = 225 - x 2, and then differentiate with respect to x. It is simpler to use implicit differentiation. (a) Differentiating with respect to x, we have, 2x + 2p dp/dx = 0, or dp/dx = - x/p (b) Differentiating with respect to p, we have, 2x dx/dp + 2p = 0, or dx/dp = - p/x (c) The revenue function is R = xp; therefore, by the product rule, dR/dx = x dp/dx + p. Substituting from (a) for dp/dx, we have, dR/dx = x1 - x/p2 + p = 1p2 - x22/p. (d) To find dR/dp. we again use the product rule. R = xp, dR/dp = x + p dx/dp. Substituting from (b) for dx/dp, we have dR/dp = x + p1 - p/x2 = 1x 2 - p22/x

234

* **

Section 2.8

Implicit Differentiation

Note that dx/dp = 1/1dp/dx2. This will always be true as long as neither of these derivatives is zero. The reason is very simple. If we think of p as a function of x, say p = P1x2, then dp/dx represents the slope of the curve at the point (x, p). If we take the same curve and think of x as a function of p, say x = D1p2, then the slope is the same. But now suppose we think of x as the vertical axis, then the slope at the point (p, x) is dx/dp. Since slope is change in y divided by change in x, interchanging the axes makes the slope into the change in x divided by change in y. Thus, interchanging axes has the effect of taking the reciprocal of the slope. Hence, the derivatives are reciprocals. Now go back and look at the solution to Example 1. Compare the result there to what you would get by just taking dx/dy. We close this section with a proof of the power rule for rational exponents. To this point, we have proved the rule only for positive and negative integers.

d m Axn B = dx where m and n are integers.

m m n -1 nx

Differentiate each side of this equation with respect to x, giving, nyn - 1y = mxm - 1. Solving for y , we have, y = Substitute for y, using, y = xm/n, we have, y = * bM # = bM - N and 1bM2N = bM N.) N b Thus, once again, we have shown that the power rule states that to find the derivative of x to a power, multiply by the power and decrease the power by one, that is for any rational (Note the two rules of exponents that were used, d r 1x 2 = rx r - 1. We shall examine the case when the exponent is irrational dx when we consider logarithmic and exponential functions. The TI 89 can find the derivative, implicitly, but we have to do it two stages. First, when the equation is entered, we need to let the calculator know that y is a function of x so we must always write y(x) in place of y, wherever it appears. Thus, if we wanted to find dy/dx when the equation is x2 + y 2 = 9, we would enter d1x 2 + y1x22 = 9, x2 and then press enter, what the screen displays is the equation differentiated, see Figure 2. We can now have the calculator solve this equation for the derivative, but we need to give the derivative a new name otherwise an error message is produced. Suppose we call the derivative D, then we enter number r, solve12y*D + 2x = 0, D2 This is indicated in Figure 3, along with the solution.

m m xm - 1 m xm - 1 m m m = = x m - 1 - Am - n B = x n - 1 m n A x n B 1n - 12 n xm - m n n n

m xm - 1 . n yn - 1

Calculator Tips

Section 2.8

Implicit Differentiation

* **

235

Figure 3: Finding the Derivative Thus, we have dy/dx = - x/y. (Be careful, the D entered as uppercase and exhibited on the screen as lowercase is the letter D located above the comma key and requires the pressing the alpha key first. This is different from the d used for differentiation which is above the number 8 key and requires pressing the 2nd key first.)

In Exercises 1 4, find dy/dx at the indicated point in two different ways: (a) Solve for y as a function of x and differentiate; (b) Differentiate implicitly. 1. y2 = 3x + 1; 3. x2y + 1 = 2x; 15, - 42 12, 12 11, 12 2. 3x2 - 4xy - 4 = 0; 10. x2/3 + y 2/3 = 1 11. 3x2y - 4y2x + 7 = 2x3y3 12. 2/x - 3/y = x 2y2 = 2x + y3 x2 + y2 14. 3x 3/4 - 2y 2/3 + 7y2 - 3x = 9x2 - 5y 4 13. 15. 2 1x + y - 1x - y = 2. Find dy dx x2 - y2

4. 2x2 - 3xy + 4y2 - 5y = 2; 11, 22 (Hint: Use the quadratic formula to solve for y.) In Exercises 5 14 find dy/dx using any method. 5. x3 + y 3 = 10 6.

x2 9 y2 4

` 14,02

16. 5s21v3 - 12 = 7. Find (a) ds/dv; (b) dv/ds. 17. 1/t + 1/s = 1. Find (a) ds/dt; (b) dt/ds, (c) Show that 1ds/dt21dt/ds2 = 1. 18. 1x + y + 1x - y = to the curve at (5/2, 3/2). x2 y3 + 31 . Determine the equation of the tangent line 27

= 1

7. xy = 7 8. x - xy + 7 = 0 9. 5x 2 + 6x2y 2 = y 2 + 15

2

236

* **

Section 2.9

Elements of Geometry

23. (a) Determine the equation of the tangent line at any point (h, k) which lies on the curve x2/3 + y2/3 = a 2/3 1a 7 02. (b) Let the tangent line to this curve intersect the x and y axes at the points 1x1, 02 and 10, y12 respectively. Show that the length of the line segment joining these two points is a constant. 24. Show that the tangent line to any point on the circle x2 + y 2 = a 2 is always perpendicular to the radius drawn to the point of tangency.

19. x2 + y 2 = 25. (a) Determine the equation of the tangent line at each of the following points: (i) (3, 4); (ii) 13, - 42; (b) Find the equation of the tangent line at (3, 4) by finding the equation of the line passing through (3, 4) which is perpendicular to the (radius) line segment whose end points are (0, 0) and (3, 4). 20. 2x2 + 3y 2 = 14. Determine the equation of the tangent line at each of the following points: (a) (1, 2); (b) 11, - 22. 21. A demand equation is given by p = 29 - x, where p is the price per item when x items are demanded. Find dR/dx when x = 1. 22. (a) Find the equation of the tangent line to the curve x3 + 3y2 - 12x - 1 = 0 at (1, 2). (b) At what point(s) on the curve will the tangent line be horizontal? (c) Show that the tangent line is vertical for some x0 where 3 6 x0 6 4.

2.9

Elements of Geometry

This section examines some of the elementary geometric concepts that may arise while solving applied problems in the calculus. Consider the angle illustrated in Figure 1, As OA rotates counterclockwise, the angle increases. When AOB becomes a line, the angle is said to be a straight angle whose measure is 180*, the symbol for degree is *.

A A

A __________________ B O

A C B E D

Section 2.9

Elements of Geometry

* **

237

Since CD is a straight line, in degrees, the measure of 1 AED + 1 AEC = 180 ( 1 is the symbol for the measure of an angle, also note that when three letters are used to describe the angle, the middle letter will always represent the vertex of the angle.) similarly, since AB is a straight line 1 AEC + 1 CEB = 180 so it follows that 1 AED = 1 CEB and similarly, 1 AEC = 1 BED These equal pairs of angles are sometimes called vertical angles. Thus, when two lines intersect, their vertical angles are equal. We also note that when two angles sum to 180*, the angles are said to be supplementary angles. Thus, the equality of vertical angles is a consequence of the fact that supplements of equal angles are equal to each other. In Section 1.1, we discussed equations of parallel lines from an intuitive point of view. More precisely, two non-vertical lines are said to be parallel if the angle they make with the positive x-axis is the same. This angle is sometimes called the inclination of the line. See Figure 3, where lines L1 and L2 intersect the x-axis at the same angle a, and are therefore parallel. (Note that a 6 180) 0

L1 L2

Vertical Angles

Parallel Lines

x-axis

Figure 3: Parallel Lines Let us now examine Figure 4, where the parallel lines L1 and L2 are intersected by the line T

L2 L1 T

, . 0 /

+ - * 2

238

* **

Section 2.9

Elements of Geometry

Since the lines are parallel the angles measured by a and b are the same. To convince yourself, rotate the figure so the line T, sometimes called a transversal, is horizontal (i.e., is the x-axis), then these angles are the inclination angles and since the lines L1 and L2 are parallel, they must be the same. Since T is a straight line, it then follows that angles g and d are also equal. Thus, when two parallel lines are cut by a transversal, the corresponding angles are equal, that is, in Figure 2, 0a 0g 0m 0P = = = = 0b 0d 0s 0h

The angles m and d are called alternate interior angles as are a and h. It is easy to see that as a consequence of vertical angles being equal as well as the supplements of two equal angles are equal, that when two parallel lines are cut by a transversal, the alternate interior angles are equal, that is 0m = 0d and 0a = 0h Example 1 Given the parallel lines in Figure 5, and the given angle, determine remaining angles. Solution.

L2 L1

140o

, / .

+ - * 1

Figure 5: Determining the Angles We have, 140 + 0 b = 180, therefore, 0 b = 40 a is a corresponding angle to b , h, P are vertical angles, therefore 0 a = 0 P = 0 h = 40 g is an angle corresponding to the 140 angle, m is its vertical angle, and s its corresponding angle, thus,

Section 2.9

Elements of Geometry

239

* g = * m = * s = 140

The next notion we discuss is similarity of triangles. Two triangles are said to be similar if two angles of one triangle are equal to two of the other. For example, in Figure 6, given the triangles joining the vertices A, B and C, and A , B and C written ABC and A B C , suppose the angle at vertex A equals the angle at vertex A and the angle at vertex B equals the angle at vertex B then the triangles are similar, sometimes written as ABC ' A B C . It follows from the fact that the sum of the three angles in any triangles is 180* that the remaining two angles are also equal to each other.

B B'

Similarity

A' A C

C'

Figure 6: Similar Triangles Essentially, two triangles are similar to each other if one is a reduced version of the other, or one can be obtained from the other by a uniform compression without distorting any of its angles. Consider Figure 7,

A C' B'

A'

Figure 7: Similar Triangles Assume the angles at vertices A and A equal each other and the angles at vertices at C and C are also equal. Then these two triangles are similar to each other. Note that one is a uniform compression of the other, the fact that one is also rotated in no way changes the fact that they are similar. Thus, physically, we can describe similarity of triangles (or of any geometric figure) as a uniform compression, possibly along with a rotation, or reflection of one figure. One would expect similar figures to have other related geometric properties as well. The fact that one is a uniform compression of the other, implies that there is a simple relationship between their sides. Consider the two triangles in Figure 8 Assuming the angles at vertices A, B (and C) equal the angles at vertices A , B , (and C ), then the lengths of their corresponding sides, a, b, c, and a , b , and c are in proportion, that is, a b c = = a b c This follows intuitively from the visualization of one triangle being a uniform compression of the other. Since the triangle is being compressed uniformly, it follows that the

240

* **

Section 2.9

Elements of Geometry

B c' c A b a A' C

sides are as well, resulting in the ratio of corresponding sides being in proportion. (This can be proven formally using the notions of parallelism. We omit the details.) The fact that corresponding sides of similar triangles are in proportion is a very useful property that may be used in a variety of applications. Jules Verne, in his novel The Mysterious Island, has the engineer, Cyrus Harding, use this property to compute the height of a cliff.

Example 2 Cyrus Harding had provided himself with a straight stick, twelve feet long, which he had measured as exactly as possible by comparing it with his own height, which he knew to a hair. Herbert carried a plumb-line which Harding had given him, that is to say, a simple stone fastened to the end of a flexible fiber. Having reached a spot about twenty feet from the edge of the beach, and nearly five hundred feet from the cliff, which rose perpendicularly, Harding thrust the pole two feet into the sand, and wedging it up carefully, he managed, by means of the plumb-line, to erect it perpendicularly with the plane of the horizon. That done, he retired the necessary distance, when, lying on the sand, his eye glanced at the same time at the top of the pole and the crest of the cliff. He carefully marked the place with a little stick. Then addressing Herbert Do you know the first principles of geometry? he asked. Slightly, captain, replied Herbert, who did not wish to put himself forward. You remember what are the properties of two similar triangles? Yes, replied Herbert; their homologous sides are proportional. Well, my boy, I have just constructed two similar right-angled triangles; the first, the smallest, has for its sides the perpendicular pole, the distance which separates the little stick from the foot of the pole and my visual ray for hypothenuse; the second has for its sides the perpendicular cliff, the height of which we wish to measure, the distance which separates the little stick from the bottom of the cliff, and my visual ray also forms its hypothenuse, which proves to be prolongation of that of the first triangle. Ah, captain, I understand! cried Herbert. As the distance from the stick to the pole is to the distance from the stick to the base of the cliff, so is the height of the pole to the height of the cliff. Just so, Herbert, replied the engineer; and when we have measured the two first distances, knowing the height of the pole, we shall only have a sum in proportion to do, which will give us the height of the cliff, and will save us the trouble of measuring it directly. The two horizontal distances were found out by means of the pole, whose length above the sand was exactly ten feet. The first distance was fifteen feet between the stick and the place where the pole was thrust into the sand. The second distance between the stick and the bottom of the cliff was five hundred feet. These measurements finished, Cyrus Harding and the lad returned to the Chimneys. The engineer then took a flat stone which he had brought back from one of his previous excursions, a sort of slate, on which it was easy to trace figures with a sharp shell. He then proved the following proportions: 15:500 * 10:x From which it was proved that the granite cliff measured 333 feet in height.

Section 2.9

Elements of Geometry

241

Solution. The engineer made some obvious approximations, but the result certainly produced an approximate determination of the cliff s height. Pictorially, the situation is illustrated in Figure 9.

B x (Cliff s Height) D 10 A 15 E |---------- 500 -----------| C

Figure 9: Harding s Determination of the Cliff s Height Note that angle at vertex A in ABC is the same angle in ADE, and each triangle has a right angle, so the two triangles are similar. Therefore, their corresponding sides are similar, so we have 15 10 = x 500 Giving, 15x = 5000 or to the nearest foot, x = 333 feet.

Example 3 A water tank in the shape of a right circular cone is 20 feet high, with base radius 4 feet (see Figure 10). Water is entering the tank at some given rate. When the water depth is 12 feet, determine the radius at the top of the water. Solution.

8 x 12 4 20

242

Section 2.9

Elements of Geometry

Figure 10, which is not drawn to scale, illustrates the water tank and all given information. By similarity (Verify!), we have 8 20 = x 4 20x = 32 x = 1.6 feet

Congruence

If two triangles are similar, and if in addition the connecting side in one triangle is equal to the corresponding connecting side in the other, then the triangles are said to be congruent. Therefore, it follows that if two triangles are congruent, their corresponding sides (and angles) are equal to each other (the similarity constant of proportionality is one). It turns out that we can prove two triangles to be congruent with less information. For example, if three sides of one triangle are equal to three sides of another triangle they are congruent, or if two angles and their connecting side in one triangle are equal two angles and their connecting side in another triangle, then the triangles are congruent. We can use the geometric concepts developed above to determine the midpoint of any line segment. Suppose we wish to determine the midpoint of the line AB joining the points 1x1, y12 to 1x2, y22. We proceed as in Figure 11.

B(x2, y2) M

A(x1, y1)

- x1

C(x2, y1)

Midpoint Formula

Let the midpoint of the line be denoted by 1x, y2. We draw the dotted lines as indicated and label the points as shown (verify!). Since M is the midpoint of AB, it follows that AM = MB. Moreover, AME is similar to MBD (verify this hint: use parallel lines.) Thus, we have two similar triangles with corresponding connecting sides equal to each other 1AM = MB2, therefore these triangles are congruent. It then follows that corresponding sides are equal, giving x2 - x = x - x1 solving for x, we have x = similarly y2 - y = y - y1 x1 + x2 2

Section 2.9

Elements of Geometry

* **

243

and we have y = y1 + y2 2

Thus, the coordinates of the midpoint of a line segment is the average of the coordinates of the endpoints. *

1. Consider Figure 12, which shows the intersection of two lines. Determine the indicated angles. 4. The sides of a triangle measure 5, 6 and 8 inches. Find the sides of a similar triangle if the side corresponding to the 6inch side of the given triangle is 9 inches. 5. The sides of a triangle measure 9, 12 and 15 inches. Find the sides of a similar triangle if the side corresponding to the 15inch side of the given triangle is 5 inches.

70o

+ * ,

6. Show that a line parallel to the base of a triangle intersecting the other two sides, forms a traingle similar to the given triangle. 7. Suppose two triangles are similar to each other and the first triangle has a side of length 8 inches and its area is 24 square inches. If the corresponding side of the other triangle is 12, what is its area? 8. Determine the coordinates of the midpoint of the line segment joining the points (a) 1 - 2, 52 and (6, 8), (b) 13, - 72 and 1 - 12, 152. 9. Determine the coordinates of the midpoint of the line segment joining the points (a) (9, 11) and 116, - 282, (b) 1 - 5, 132 and 12, - 192 10. Determine the equation of the perpendicular bisector of the line joining the points 11, - 32 and (7, 13).

Figure 12

2. Consider Figure 13, which shows the intersection of two lines. Determine the indicated angles.

120o , *

11. A 6 foot man walks away from a light sitting atop a pole 16 feet above ground. (a) How long is his shadow when he is 8 feet from the pole? (b) What is the distance from the base of the pole to the tip of his shadow? 12. The sides of a triangular parking lot are 120 yards, 150 yards and 200 yards. An architect is drawing a scale drawing of the lot with 1 inch representing 50 yards, determine the scaled measurements of the drawing.

Figure 13

3. The two parallel lines L and M are intersected by T, see Figure 14. Find all the indicated angles.

13. The shadow of a tree is 40 feet long at the same time that the shadow of a 4foot flag pole is 2 feet long. What is the height of the tree. 14. Given the triangle with vertices A(1, 1), B(3, 7), and C(9,3). (a) Determine the equation of the three lines drawn from each vertex to the midpoint on the opposite side. These lines are called the medians of the triangle. (b) Determine the length of each median. (c) show these medians intersect in a point which is 2/3 of the way from each vertex to the opposite side. 15. Show that for any triangle, the medians intersect in a point which is 2/3 of the way from each vertex to the opposite side. Hint: Choose the vertices of the triangle to be (0, 0), (b, c) and (a, 0), 16. For the triangle given in the previous example, show that the line segment joining the midpoints of any two sides is equal in length to one-half the third side. 17. Determine the coordinates of the point P which divides the line segment AB where A is the point (2,4) and B is the point (5,12) so that AP/PB = 2/3.

65o /

+ . -

L M

18. Determine the coordinates of the point P which divides the line segment AB where A is the point 1x1, y12 and B is the point 1x2, y22 so that AP/PB = m/n. Verify that in the case m = n, you obtain midpoint formula.

Figure 14

244

* **

Section 2.10

Related Rates

Assume neither A nor B is zero. (If either was zero, the problem is trivial, why?) Draw the vertical line from P to T. (a) Show that * PSR = * TSW, (b) Show that * RPS = * SWT. (c) Show that PRS is similar to triangle WOQ. (d) Find the coordinates of Q, W and S. (e) By similarity, conclude that d PR = OW QW (e) Substitute for these distances and simplify, to show that d = Ax1 + By + C 2A2 + B2 (Note, the absolute value was included as distance must always be positive.) .

19. Show that the altitude h, drawn to the hypothenuse of a right triangle (the line drawn from the right angle and perpendicular to the hypothenuse) is the mean proportional between the two segments along the hypothenuse, that is c1/h = h/c2, see Figure 15

c1 h c2

21. The medians of a triangle are the lines drawn from a vertex to the opposite sides which it bisects. For the triangle with vertices at A(1, 3), B(5, 9) and C(9, 5) (a) find the equation of the medians and (b) show they intersect in a single point. Determine this point. 22. Show the statement of the preceding exercise is true for any triangle. 23. Show that the slope of a line is independent of the points on the line used to compute it.

Figure 15

20. To determine the distance d from the point P1x1, y12 to the line Ax + By + C = 0, proceed as follows: Consider Figure 15.

W Ax + By + C = 0

Figure 16

A Geometric Example An Ecological Example An Economic Example Using Similarity

In preceding sections we saw that a derivative may be interpreted as a rate of change. In particular, if the dependent variable represents position (distance), and the independent variable represents time, then the rate of change may be viewed as velocity. In many real situations, we shall encounter cases in which several variables are related to each other and are also changing with time. Since the variables are related their rates of change will

Section 2.10

Related Rates

* **

245

be related and hence problems of this type are usually referred to as related rates problems. In order to fix the central ideas, let us consider a typical case. Suppose that a space shuttle S is launched from Cape Canaveral. In Figure 1, we see a diagram showing the location of the shuttle at some time, t, after the launch. Its altitude, y, its range (ground distance from the launch site), x, and straight line distance from the launch site, z, are all changing with time. In other words, x, y, and z are all functions of t.

Figure 1: Shuttle Location at time t It should be obvious from the geometry of the problem, that x, y, and z are not independent of one another. Consequently, their rates of change, dx/dt, dy/dt, and dz/dt are not all independent either. The following specific example will illustrate the way in which the variables and their velocities are related. It is important for you to observe that we attack the problem in three distinct steps: 1. Determine an equation that relates the variables in general. 2. Differentiate each side of the equation to obtain an equation relating their rates. 3. Substitute the specific numerical data to solve for any unknown quantity. Example 1 A space shuttle is launched and is moving away from the launch site with a constant velocity of 8 miles per second. Suppose that the range is increasing at the rate of 3 miles per second. At what rate is the altitude changing when the shuttle is 1500 miles from the launch site and at an altitude of 900 miles? Solution. Using Figure 1, we let x be the range, y the altitude, and z the distance from launch at time t. It is a good idea to reformulate the problem as follows: Given: dz/dt = 8 mi/sec, and dx/dt = 3 mi/sec.

A Geometric Example

Find: dy/dt when z = 1500 miles and y = 900 miles. We need a relationship among the three variables. Looking at Figure 1, we see that at any time t, we have a right triangle. Thus, 1. We have the relationship x 2 + y 2 = z2. 2. We may differentiate the equation x2 + y 2 = z2 to find the relationship between the velocities. Remember, each variable is now a function of t, requiring the application of the chain rule. Differentiating this equation with respect to t, we find that 2x dy dx dz + 2y = 2z dt dt dt (1)

246

Section 2.10

Related Rates

Notice that each term is multiplied by a derivative term which resulted from the application of the chain rule. We may now complete the problem. 3. We note that when z = 1500 and y = 900, x = 215002 - 9002 = 1200. Substitution into (1) yields 2112002132 + 219002 dy/dt = 2115002182 or dy/dt = 9 1 3 mi/sec. That is, the altitude is increasing at a rate of 9 1 3 mi/sec. In the above example, we needed to use the geometry of the problem to find an equation which related the variables. When the equation was differentiated, it gave their related rates. Often, this is the case. That is, we must use the physical nature of a problem to determine a necessary equation relating the variables. Sometimes the equation is given, as the next ecological example illustrates. Example 2 The fish population, N, in a small pond depends upon the number of pounds of algae, a, it contains. The equation is given by N = 13a2 - 20a + 2624. If the quantity of algae is increasing at a rate of 2 lb/week, at what rate is the fish population changing when the pond contains 5 lb of algae? Solution. Given. N and a, are each functions of time.

An Ecological Example

da/dt = 2 lb/week.

Find. dN/dt when a = 5 lb. Step 1 has already been completed, since we have been given the relationship between N and a. 2. Differentiating the equation N = 13a2 - 20a + 2624, we have, using the chain rule, dN/dt = 413a2 - 20a + 262316a # da/dt - 20 # da/dt2 3. Substituting a = 5 and da/dt = 2, we have that

Thus, the fish population is increasing by 80 per week at the instant when the stream contains 5 lb of algae. What does this suggest to you about the amount of algae in the stream and about its rate of increase?

An Economic Example

Economic problems often fall into this category. For example, cost, revenue, and profits are related to one another and are always changing with time. The next example illustrates such a case. Example 3 The retail price per gallon of gasoline is increasing at $0.02 per week. The demand equation is given by 10p - 2356 - x2 = 0 where p is the price per gallon, in dollars, when

Section 2.10

Related Rates

247

x million gallons are demanded. At what rate is the revenue changing when 10 million gallons are demanded? Solution. Given: dp/dt = 0.02 dollars per week

Find: dR/dt when x = 10 million gallons. 1. The total revenue in millions of dollars is given by the equation

R = xp

(2)

We could solve the demand equation for p in terms of x, substitute into (2) and have R in terms of x. However, let us take a different tack. Let us differentiate using the product rule, thus, 2. dR/dt = x dp/dt + p dx/dt Now, we solve the demand equation for p, finding p = when x = 10 p = 1/10125621/2 = 1.6. We find dx/dt by differentiating the demand equation. Using the chain rule, we have, dp 1 1 dx = 1356 - x 22-1/2 a - 2x b dt 10 2 dt Thus, when dp/dt = 0.02 and x = 10, we find that dx/dt = - 0.32 million gallons per week. (The negative sign indicates that the demand is decreasing). We may now complete the solution. 3. Substituting, we have that dR/dt = 1010.022 + 1.61 - 0.322 = - 0.312 That is, the revenue is decreasing by about 0.31 million dollars per week.

1 10 1356

Using Similarity

There is a large class of geometrically-based related rates problems that use the properties of similar triangles as examined in Section 2.8. Our last example in this section is typical of this type.

B (Top of Lamppost)

Example 4 There is a light atop a 12 foot high lamppost. A five-foot-tall girl is walking in a straight line away from the light at a rate of 6 feet/sec. How is the length of her shadow changing? Solution. Figure 2 shows the essential features of the problem. Here y is the girl s distance from the base of the lamppost and x is the length of her shadow. The triangles ABC and AED are similar. Therefore, their sides are in proportion. That is, the ratio of x to 5 is the same as the ratio of 1x + y2 to 12. In symbols, x + y x = 5 12 Cross-multiplying, 12x = 5x + 5y,

12 5 C y D x A E

248

Section 2.10

Related Rates

7x = 5y Hence, we have the relationship between x and y. 1. 7x = 5y 2. Finding the derivative, 7 dx/dt = 5 dy/dt, and substituting dy/dt = 6 3. 7 dx/dt = 152162, dx/dt = 30/7 Since x is the length of the shadow, we see that the shadow is getting longer at a rate of 30/7 feet/sec.

It is usually very easy to determine the given rates in a specific example. Notice that a rate has units that are per unit time. For example, miles per hour, gallons per week, feet per second, cubic feet per minute, dollars per month etc.

1. If y = x3 - 3x, determine dy/dt when x = 2 and dx/dt = 3. 2. If y = 2x , determine dx/dt when x = 2 and dy/dt = 4. 3. If y = 23x2 - 3, find dy/dt when x = 2 and dx/dt = - 2. 4. If y = x3 - 22x, find dx/dt when x = 8 and dy/dt = 64. 5. If x2 + y2 = 169, find dy/dt when x = 5 and dx/dt = - 6 6. Suppose that Q = x y . If dQ/dt = 6 and dx/dt = 2, find dy/dt when x = 2 and y = 1. 7. If Q = x/y and R = xy, find dx/dt and dy/dt if dQ/dt = - 3 and dR/dt = 9, when x = 6 and y = 3. 8. Two cars leave an intersection at the same time. One travels north at 40 miles per hour, and the other travels east at 30 miles per hour. If z represents the distance between the cars, (a) describe how the rate at which the cars are separating is related to the rates at which they are traveling. (b) At what rate is the distance between the two cars changing at the end of two hours? 9. The radius of a circle is increasing at the rate of 1 inch per hour. (a) At what rate is the area of the circle changing when the radius is 4 inches? (b) At what rate is the circumference changing when the radius is: (i) 4 inches; (ii) 5 inches? 10. A large cube of ice is melting uniformly at the rate of 6 cubic inches per second. At what rate is a side changing when it is 1 inch long? 11. The area of a square is increasing at the rate of 1 square inch per minute. At what rate is the: (a) side of the square changing; (b) diagonal changing; when the side is 6 inches long? 12. A spherical balloon is being deflated at a rate of 10 cubic feet per second. At what rate is the radius of the balloon changing when the radius is 2 feet. (The 3 volume V of a sphere of radius r is given by the formula V = 4 3 pr .) 13. A kite maintains a constant altitude of 50 feet and is moving horizontally at the rate of 2 feet per second. Assume that there is no slack in the kite-string, at what rate is the string being paid out when its length is 130 feet? 14. A 26 foot ladder is leaning against a vertical wall and its base is on level ground. The top of the ladder slides down the wall at a rate of 2 feet per minute. At what rate is the bottom of the ladder moving away from the wall when the top of the ladder is 10 feet above the ground? 15. A pebble thrown into a pond creates a circular ripple whose radius is increasing at 2 feet per second. At what rate is the area of the circle changing when the radius is 12 feet?

2 -3 3 2

16. The demand for apples is given by the equation xp = 150, where x is the number of pounds demanded and p is the price per pound. If the price is increasing at a rate of $0.25 per week, at what rate is the: (a) demand changing; (b) revenue changing; when the demand is 10 pounds? (Can you solve (b) by inspection?) 17. The number, N (in million) of VCR s sold in the United States for the years 1990 to 2001 may be approximated by the formula N = 112t + 12/1t + 22, where t is the time in years, with 1990 being t = 0. (a) Find the number sold in 1990, 1994, and 2000. (b) Find the rate of change of sales in 1990, 1994, and 2000. 18. (Continuation of previous exercise.) Suppose that the average price of a VCR during the same time period was governed by the rule p = 820 - 50t. (a) Find the price and total revenue from sales for 1990, 1994, and 2000. (b) Find the rate of change of total revenue for 1990, 1994, and 2000. 19. Boyle s law for gasses states that PV = constant, where P is the pressure in pounds per square inch (psi) and V is the volume of the gas. For a particular gas it is known that when the pressure is 10,000 psi, its volume is 4 cubic inches. If the volume is increasing at 8 cubic inches per second, find the rate at which the pressure is changing when the pressure is 20,000 psi. 20. Suppose that the oil spill from the damaged hull of a ship forms a circular slick whose thickness is uniformly 1 millimeter and whose radius is increasing at a rate of 20 kilometers per day. At the instant when the radius of the oil slick is 15 kilometers: (a) Determine the rate at which the spill is flowing out of the ship. (b) What effect does a 12 hour time delay in plugging the leak have on the spill? (Note: the volume of a circular disc of radius r and thickness d is p r2d, and 1 millimeter = 10-6 kilometer.) 21. A point moves along the circle x2 + y2 = 5 in such a way that the distance from the point (4, 6) is increasing at a rate of 7 units per second. Find dx/dt at (1, 2). 22. Water is flowing into a conical reservoir 50 feet high with a top radius of 10 feet, at the rate of 16 cubic feet per second. At what rate is the depth of the water in the reservoir changing when the water is 10 feet deep? (The volume 2 of a cone is 1 3 pr h, where r is the radius and h its height.) 23. A 6 foot tall man is walking towards a 24-foot lamppost at the rate of 8 feet per second. At what rate is the tip of his shadow moving towards the lamppost when he is 5 feet from the lamppost?

Section 2.11

24. For the case described in the previous exercise, find the rate at which the length of the man s shadow is changing. Is it getting longer or shorter? 25. The perimeter of a rectangle is growing at the rate of 6 cm/sec. Its area is growing at the rate of 10 cm2/sec. At a certain time one side is 50 cm long and the other side is 20 cm long. Find the rate of change of each side. Are they growing or shrinking? 26. The perimeter of a rectangle is growing at the rate of 12 m/sec. Its diagonal is growing at the rate of 4 m/sec. At a certain time one side is 4 m long and the other side is 3 m long. Find the rate of change of each side. Are they growing or shrinking?

Newtons Method

* **

249

27. An alternative way to do implicit differentiation is to introduce a mythical extra variable t. Think of x and y as both functions of t. Now, use the fact that dy dt = dy dx dx dt

(a) Solve for dy/dx. (b) For the equation x2 + y 3 = 6, find the relationship between the rates of change of x and y as in the previous section. (c) Use the result of parts (a) and (b) to find y . 28. Use the method developed in the previous exercise to find the equation of the line tangent to x2y 2 - 4xy + 7y3 = 5 at 12, - 12.

2.11

Newton s Method

This section deals with Newton s Method for determining the roots of the equation f1x2 = 0. Consider the smooth function defined by the equation y = f1x2. To solve the equation f1x2 = 0 we need only find those x-values at which y = 0, that is the x-intercepts of the curve. We begin by selecting any value that is near a root (x-intercept) and call this value x0. (This first choice may be obtained from a sketch of the curve y = f1x2 and choosing x0 as the approximation to the actual intercept.) The equation of the tangent line to the curve at the point 1x0, f1x022 is given by the equation, y = f 1x021x - x02 + f1x02 The x-intercept of this line, called the first iterate, x1 is found by setting y = 0 (see Figure 1), we obtain x1 = x0 - f1x02/f 1x02 (Note that if f 1x02 = 0, this tangent line has no x-intercept, why?)

Newton s Method

y = f( x )

x3

x2

x1

x0

250

* **

Section 2.11

Newtons Method

Now we determine the equation of the tangent line to the curve passing through the point 1x1, f1x122, the calculation is identical to the one above with x0 being replaced by x1. We obtain y = f 1x121x - x12 + f1x12 and its x-intercept, the second iterate, x2 is x2 = x1 - f1x12/f 1x12 We repeat (iterate) this process, generating a sequence of x-intercepts (iterates), x1, x2, x3, , xN where

(1)

For many well behaved functions xN is an excellent approximation to the root of f1x2 = 0. In fact, it can be shown in a more advanced course that this method will produce a root f1x2f 1x2 if ` ` 6 1 near the root, where f 1x2 is the derivative of the derivative, (called 1f1x222 the second derivative). When the sequence of iterates approaches the root (converges), we need to decide when to stop, that is to how many decimal places do we want the approximate root? Suppose we want the root to five decimal places, then we stop when the sixth decimal place no longer changes with the next iteration. This is a repetitive process, so the easiest way of performing the calculations is with a calculator or even more easily with a spreadsheet, like Excel, where we can copy one row to the next and automate the calculations. We illustrate the method in the following example.

Example 1 Determine the positive root of the equation x 2 - 2 = 0. Solution. Let f1x2 = x2 - 2, then we have f 1x2 = 2x and (1) becomes xN = xN - 1 x2 N-1 - 2 2xN - 1

Suppose we choose as our first guess, x0 = 1.5, then we generate the following spreadsheet indicated in Table 1. Notice, the cell C4 is x0, the next iterate (approximation), x1 is obtained by entering in the spread sheet in the cell C5, the formula (1) in the form * C4 - 1C42 - 22/12*C42. We copy this formula to the succeeding cells in its column to automate the process. Note that when N = 3 the fifth decimal place no longer changes, thus to four places, the zero is 1.4142. It is clear that an excellent approximation to the zero of the equation is 1.41412356. (Note that we just computed the square root of 2.)

Section 2.11

Newtons Method

251

Example 2 Determine the positive root of f1x2 = 5x4 - 2x 3 - 7x 2 + 11x - 21. Solution. We need a first guess. Observe, f112 = - 14 and f122 = 37, so there must be a root between 1 and 2 (why?). We choose x0 = 1. f 1x2 = 20x3 - 6x 2 - 14x + 11 so we have

3 2 5x4 N - 1 - 2x N - 1 - 7x N - 1 + 11xN - 1 - 21 2 20x3 N - 1 - 6x N - 1 - 14xN - 1 + 11

xN = xN - 1 -

(1)

We put this into a spreadsheet, see Table 2. Note that the entry which does all the calculations is cell C5 its formula is = C4 - 15*C44 - 2*C43 - 7*C42 + 11*C4 - 212/ 120*C43 - 6*C42 - 14*C4 + 112. This entry is then copied into the cells beneath it. Thus, we have, to seven decimal places, the root is 1.5367405

The function defined in Example 2 also has a negative root, we leave it as an exercise for you to locate it using Newton s method, we shall locate it below using the TI 89. Newton s method doesn t always work. Sometimes, the sequence generated by the method moves away from the root, rather than towards it. It can also happen that the sequence cycles, again not approaching a root. We examine these cases in the exercises. Table 2: Finding a Root of f1x2 = 5x 4 - 2x 3 - 7x 2 + 11x - 21

Newton s method may be used directly on the calculator by realizing that each iterate is a composition of the previous. In Example 2, define y1(x) as

252

* **

Section 2.11

Newtons Method

Calculator Tips

y11x2 = x -

this is essentially (1) with the subscripts deleted. Then if x0 = 1., then x1 = y111.2, x2 = y11y111.22, x3 = y11y11y111222 and so on, see Figure 2. Continuing this way will improve the accuracy of the iterates to the root. Another way of using the calculator to locate the root is to first have it sketch the graph, and then press F5 to locate the required zero. For the lower bound, you move the cursor to the left of the desired root and press enter, and for upper bound you move the cursor to the right of the desired root and press enter. The required approximation to the root is then displayed. The TI 89 has a variation of Newton s method built into its solve command. However, you need to be careful that it locates the root you are looking for. It is usually a good idea to have it sketch the graph first to get a visual location of the root you want to find and then use the solve command along with the with feature to obtain the desired root. For example, suppose we want to find the negative root of the function given in Example 2, that is, f1x2 = 5x4 - 2x 3 - 7x 2 + 11x - 21. We enter Figure 2: Newton s Method and the TI 89 solve15x 4 - 2x 3 - 7x 2 + 11x - 21 = 0, x2 x 6 0 The calculator gives, after a few seconds be patient it can take a few seconds (or longer) for more complicated functions, - 1.7032256. Actually, if you forget to enter the with condition, it produces both roots, but this will not always be the case.

1. Let f1x2 = x3 - 2x - 7, (a) show that a root of this function lies in the interval (1, 3), (b) Determine the equation of the tangent line to this curve at x = 2, and determine its x-intercept. (c) Determine the equation of the tangent line to the curve at the x-intercept found in (b) and determine the x-intercept of this tangent line. (d) Determine the equation of the tangent line to the curve at the x-intercept found in (c) and its x-intercept. (e) Continue the above and produce a root of the given function to six decimal places. 2. Find the fourth root of 7 to six decimal places. Hint: find a zero of x 4 - 7. Why does this produce the root? 3. Find the fifth root of 25, to six decimal places. 4. Find the roots of f1x2 = x4 + x - 3 to six decimal places. Observe that this function has more than one root, they each need to be determined. 5. Find all the roots of f1x2 = x3 + 3x2 - 3 to six decimal places. 6. (a) Find all the roots of f1x2 = 5x4 - 3x 3 - 20x2 + 4x + 1 to four decimal places. (b) Using the roots, give an approximate factoring of the quartic polynomial. 7. Consider f1x2 = 1x - 121/3, obviously, x = 1 is the root. Try Newton s method with x0 = 1.1. What seems to be happening? Try to explain it. 8. Given f1x2 = x2 - 4x + 3, suppose you take x0 = 2, what happens why? Suppose you take x0 7 2 what happens, what if x0 6 2? 9. Suppose you make a lucky starting choice, that is, x0 is actually the root of f(x). What happens to the successive iterates? 10. Explain why x0 = 3 will not produce a root of f1x2 = x 3 - 9x2 + 27x + 5. Determine the root of this function, to six decimal places. 11. f1x2 = x3 - 6x2 + 12x - 10, what happens if you try to use x0 = 2 to locate the root of this function, why? Find its root, to six decimal places. 12. Newton s method can be used to approximate reciprocals. Suppose you want to approximate 1/b, then let f1x2 = 1/x - b. Use this observation to compute (a) 1/6, (b) 1/9, (c) 1/13. 13. An $8,000 car loan is paid off with $250 monthly payments over three years, what monthly interest rate r is the bank charging? It can be shown that this problem translates into the equation 8000 = 250 a 1 - 11 + i2-36 i b,

where r = 12i. Find r. Hint: rewrite the above expression as a polynomial, with x = 1 + i. A variation of Newton s Method is called the Secant Method. It avoids the use of the derivative. It replaces the derivative by the approximation, f1xN2 - f1xN12 , that is, the derivative is replaced by the slope of the f (xN) L xN - xN - 1 (secant) line through the points 1xN, f1xN22 and 1xN - 1, f1xN - 122. Thus, we have xN + 1 = xN f1xN21xN - xN - 12 f1xN2 - f1xN - 12 for N 1

Note that this means we need two starting values near the root, x0 and x1. 14. Use the Secant Method to find the roots of the function of f1x2 = x3 + 3x2 - 3, compare with the results in Exercise 5.

Chapter Review

* **

253

CHAPTER REVIEW

Key Ideas Slope of a curve Tangent Line Derivative Constant Multiplier Rule Sum Rule Product Rule Quotient Rule Limit Continuity Discontinuity Removable Discontinuity Differentiable Composite Function Chain Rule Extended Power Rule Marginal Function Geometric Concepts Rate of Change Related Rates Implicit Differentiation Newton s Method

1. (a) By using the definition of the derivative, find y if y = 6 - 2x - x2. (b) Find the equation of the tangent line to the curve at the point (1, 3). (c) Find the point(s) at which the tangent line is horizontal. 2. (a) By using the definition of the derivative, find y if y = equation of the tangent line to the curve at the point (3,1). 3. (a) Find y for y = 2x4 - 3x2 + 1x gent line at 11, - 52. 4. (a) Find y for y = line at (1,9). 5. Calculate the indicated limits: (a) lim x + 2 x ; (b) lim 2 x:5 x + 5 x + 1 x 2 + 3x 4 - x2 6. Calculate the indicated limits: (a) lim 2 ; (b) lim 2 x : -3 x - 9 x:2 x - x - 2 7. (a) Find f 1x2 if f1x2 = x2 2x 3 + 1 (b) Find the equation of the tangent line to the curve at the point where x = 2.

x:2 2 3

13. If a ball is thrown upward with initial velocity 28 feet/sec. and it strikes the ground 10 seconds later: (a) How far above the ground was the ball when it was first thrown? (b) What was the highest point the ball ever reached? (c) How fast was the ball moving when it hit the ground? 14. A 20 foot high grain silo in the shape of a cone, has a 10 foot base radius. When the radius of the enclosed grain is 8 feet, how high is the grain piled? 15. The line segment connecting the points A(5, 9) and B( 12, 20) is to be divided at a point P so that the ratio of the distances AP/BP is 5/8, find the coordinates of the point P. 16. A 15 foot ladder is leaning against a vertical wall and its base is on level ground. The bottom of the ladder is being pulled away from the wall at a rate of 4 feet per minute. At what rate is the top of the ladder descending when the top of the ladder is 12 feet above the ground? 17. A 61*2 foot tall man is walking away a 13-foot lamp post at the rate of 5 feet per second. At what rate is his shadow changing when he is 15 feet from the lamp post? 18. Find the equation of the tangent line to x 2y - 3xy 2 = y 3 - 24 at the point 12, - 22. In Exercises 19 21: (a) Sketch the graph of the given function. (b) For what values of x is the function discontinuous? (c) For what values of x is the function not differentiable? (d) If any point of discontinuity is removable, redefine the function at that point to make it continuous and determine whether the function as so redefined is differentiable at that point. 19. g1x2 = x + 3 if if if if if x 6 1 1 6 x x 0 0 6 x 6 1 1 x x2 - 9 x 20. f1x2 = c 2x - 1 x2 21. f1x2 = c 0 1 - x

5 x2

5 x4

(b) Find the equation of the tangent 2x + 10 line to the curve at the point where x = - 1.

3

x2

9. Find the derivative of the indicated function. ( a) g1x2 = (b) h1x2 = 1x - 12 13x + 42

2 5 6

x3 + 1

B 2x3 + 5

10. If the total cost of producing x toaster-ovens is given by the formula C1x2 = 40 + 12x + 0.2x2, find the approximate cost of producing the 30th toaster-oven. 11. If the demand for VCR s is described by the equation 5p + 3x = 60, where x is the number of items sold at price p, where p is in hundreds of dollars: (a) Find the total revenue function. (b) Find the approximate revenue obtained by producing and selling the 7th VCR. 12. If the position of a particle at time t is given by the equation s1t2 = 6 + 3t2 - 4t3: (a) Find the velocity as a function of time. (b) Find the times and locations at which the velocity is zero. (c) Find those values of t for which the velocity is positive.

22. Find the equation of the tangent line to 1x2 + y221/2 - 2x + 4y = 15 at the point (3, 4). 23. Using Newton s Method, find the root of f1x2 = 3x5 - 2x4 + 4x 3 7x 2 + 11x - 15 to five decimal places.

In Chapter 2, we developed rules for finding the derivatives and began to see some of the uses to which the derivative may be put. In this chapter, we begin a more detailed investigation of some of these applications. In particular, we shall be concerned with finding extreme values. That is, we shall be looking for ways to find the largest and smallest possible values attained by a given function. These methods will be applied to geometric problems and examples from Economics and Finance. We shall also see how limits and derivatives can be used to help us in curve sketching. In Section 3.6, we introduce the notion of the differential and see how it is used for approximations.

256

* **

Section 3.1

Extrema of a Function

Continuity Maximum and Minimum Values Extreme Value Theorem Relative Maxima and Minima Critical Numbers and Critical Points Calculator Tips

Continuity

The dictionary defines optimum to be the best or most favorable degree or amount. To optimize is to find an optimum or optimal value. To optimize a function means to find its largest or smallest possible value. In this section, we shall consider the problem of optimizing a non-linear function of a single variable. Later, we shall generalize the problem of optimization to cases in which a function may depend upon several variables. There are numerous applications of optimization. For example, in designing containers of fixed volume, one would want to use the least amount of material. In the business world it is usually desirable to maximize profit and/or minimize cost. We shall discover that once we model an application by a mathematical equation, it is often a straight-forward matter to determine the optimal solution. We must first develop the necessary tools with which to perform the analysis. Before we begin it is useful for us to recall what is meant by a continuous function. In Section 2.3 a precise definition was given. It is sufficient for this chapter to think of a continuous function as one whose graph has no holes or jumps. Each point in the domain at which there is a hole or jump in the function s graph is called a discontinuity. Since polynomials are functions without holes or jumps, they are everywhere continuous functions. Similarly, a rational function (the quotient of two polynomials) is continuous at every point at which its denominator is not zero. To illustrate a discontinuous function, we need only construct a function which has either a hole or a jump at some point in its domain. In Figure 1, we have a function which has a hole at x = 2, and a jump at x = 4. At

Section 3.1

Extrema of a Function

* **

257

each of these x-values, the function is discontinuous. Its precise mathematical definition may seem confusing at first. However, intuitively, continuity is a simple notion. Consider the graph of the function given in Figure 2. Notice that we have labeled the points A(a, f(a)), B(b, f(b)), C(c, f(c)), D(d, f(d)), E(e, f(e)) and G(g, f(g)). You can see that A is the highest point on the graph and G is the lowest. Thus, f(a) is the highest y-value attained by the function, and f(g) the lowest y-value. We say that this function has its maximum value occurring at x = a and its minimum value at x = g. Its maximum value is f(a) and its minimum is f(g). More generally, we have the following definition.

DEFINITION 1 A function f has f (a) as its maximum value if f1a2 f1x2 for all x in its domain. A function f has f (g) as its minimum value if f1g2 f1x2 for all x in its domain.

This means that the maximum value is the largest possible y-value attained by the function over its domain. Similarly, the minimum value is the smallest possible y-value attained by the function over its domain. Sometimes, when we want to refer to either a maximum or minimum we use the term extremum. (The plural of extremum is extrema.) We note that some texts use the terminology absolute extrema, to distinguish an extremum from a relative extremum (which we consider below), we shall not do so here. There are cases in which we shall be concerned with the extrema of a function on some interval not necessarily the entire domain of the function. The interval may be open or closed, or half-open. For such cases, we shall make a small change in Definition 1. For simplicity we shall state the revised definition for an open interval.

DEFINITION 2 A function f has f(c) as its maximum value on the interval a 6 x 6 b, if f1c2 f1x2 for all x in this interval. A function f has f(g) as its minimum value on the interval a 6 x 6 b, if f1g2 f1x2 for all x in this interval.

Not every function has a maximum and/or minimum value. This may at first surprise you, but it is possible that a function may never achieve its extrema. Consider f1x2 = 2x2 on the open interval 1 6 x 6 2. We sketch this function in Figure 3. Note that this interval does not contain x = 1 or 2. Therefore, this function never attains a y-value of two. It gets very close, for example it attains the y-values 2.00000001 and 2.000000000001, and in fact, it attains any y-value with as many zeros as you like after the decimal point followed by a 1, but it never attains the y-value 2. In order for y to reach 2, we would have to have x be 1 but then x would not be in the interval. Similarly, this function never attains the y-value 8. It gets arbitrarily close, but never there. Of course, you see why this happens. The function is defined and continuous on the open interval, that is, between 1 and 2, but it is not defined at either end point. When can we be certain that a function will actually attain both its extrema? The above example

258

* **

Section 3.1

Extrema of a Function

o (2, 8)

(1, 2) o

Figure 3: f1x2 = 2x2 on 1 6 x 6 2 indicates the answer. We need to keep the function continuous and include the end points of the interval, that is, close the interval. This suggests the following theorem, known as the Extreme Value Theorem.

THEOREM 1: THE EXTREME VALUE THEOREM If the function defined by y = f1x2 is continuous on the closed, bounded interval a x b, then it attains both a maximum and minimum value on this interval.

This theorem says there are x-values between a and b, or possibly a or b themselves, at which the function has a maximum and a minimum. It does not tell us how to find the extreme values, only that they exist. We shall soon see that they are easily found. Example 1 Show that the function f1x2 = 1 - x2 has a maximum value. Solution We sketch the function, which is the parabola in Figure 4.

V(0,1)

1 -2 -1 0 -1 -2 -3 0 x 1 2

Figure 4: f1x2 = 1 - x2 It is clear from our sketch, that at the vertex of the parabola V(0,1), f has its maximum value. That is, its maximum value is 1. Note that this function does not have a minimum.

Section 3.1

Extrema of a Function

* **

259

Why not? Note that the Extreme Value Theorem is not applicable here. While the function is everywhere continuous, its domain is not a closed interval.

A function may even be discontinuous on an open interval and yet attain a maximum and minimum. The function whose graph is given in Figure 5 attains its maximum value at its turning point A and its minimum at its endpoint B. Theorem 1 does not exclude such situations, it just tells us what conditions are needed to guarantee attainment of the extreme values. In other words, sufficient conditions to guarantee extrema. If the conditions of the theorem are not met, then the function may or may not attain extreme values.

We would like to find those points at which the function attains its extrema, as asserted by Theorem 1. Before we can, we must first examine Figure 2 a little more closely and introduce a few more definitions. Points C and E resemble mountain peaks. Near C, f(c) is locally a maximum. For example, if you were standing on top of a mountain, nearby (locally), no one is higher than you. However, further away, there may be someone at a higher elevation. The function in Figure 2 is said to have a relative maximum at C, since locally or relative to its neighboring points, it is the highest point. Similarly, the function also has a relative maximum at E. Following the same reasoning, each of the points B and D is a relative minimum, because each is the lowest point in some vicinity of itself. Thus, at any point at which the function has a peak, it has a relative maximum, and at any point at which it has a valley it has a relative minimum. As above, if we wish to refer to either a relative maximum or relative minimum we shall use the term relative extremum. The precise definition of relative extrema for a function defined on a 6 x 6 b follows.

DEFINITION 3 A function defined by the equation y = f1x2 has f(c) as a relative maximum value if there is an open interval a 6 x 6 b which contains c, and f1c2 f1x2 for all x in this interval. It has a relative minimum value at f(c) if there is an open interval a 6 x 6 b which contains c, and f1c2 f1x2 for all x in this interval.

We would really like to have a simpler way to pin down relative extrema. Perhaps the easiest way to do this is by noticing where a function is increasing or decreasing. Recall that a function is said to increase on an interval, if as we move from left to right on that interval, its y-values increase (that is, get larger). Similarly, a function is said to decrease on an interval, if as we move from left to right on that interval, its y-values

260

* **

Section 3.1

Extrema of a Function

decrease (that is, get smaller). In Figure 2, the function is decreasing between A and B, between C and D, and between E and G. It is increasing between B and C, and between D and E. Notice that just to the left of a peak the function is increasing, and just to its right it is decreasing. At a valley , the reverse occurs, on its left it is decreasing and increasing on its right.

RULE OF THUMB A function defined by the equation y = f1x2 has a relative maximum at (M, f(M)) if the function is increasing just to the left of M and decreasing just to the right of M. The function has a relative minimum at (m, f(m)) if the function is decreasing just to the left of m and increasing just to the right of m.

Except for extraordinary cases where the graph actually has a plateau, such as in Figure 6 , this Rule of Thumb could serve as an equivalent definition of a relative maximum and a relative minimum. Indeed, it is the way most of us think of them. (Why must we exclude functions with plateaus?)

Figure 6: A Function with a Plateau Now that we have defined a relative extremum we may re-examine Theorem 1. Where may the extrema of a continuous function occur? From Figure 2 we see that they may occur at the end points of its interval of definition. However, a glance at Figure 7a, b, or c reveals that they may also occur at interior points of the interval. If an extremum occurs at an interior point, then it will also be a relative extremum. Thus, to find the maximum and minimum values of a continuous function over a closed interval, as asserted by the Extreme Value Theorem, we need only examine the y-values at the endpoints and at the relative extrema.

4 5 4 3 2 2 1 1 0 y 2 1 0 5 4 3

2 x

0.5

1 x

1.5

2 x

Figure 7a

Figure 7b

Figure 7c

Section 3.1

Extrema of a Function

* **

261

If we know the interval, it is a simple matter to check the end points. But how do we find the relative extrema? Observe that at a relative extremum one of two things occurs: The tangent line is horizontal, as at points B and C (Figure 2), or, the curve has a sharp corner where a tangent line does not exist, as at points D and E (Figure 2). The tangent line is horizontal at any point where the derivative is zero. The tangent line is vertical or does not exist at any point where the derivative fails to exist. However, we must not conclude that all such points are relative extrema. In fact, Figure 8 illustrates the contrary. The function in Figure 8a has a horizontal tangent line at the point P, and P is not a relative extremum. At the point P in Figure 8b the function has a vertical tangent line, and P is not a relative extremum. At the point P in Figure 8c the derivative fails to exist, and P is a relative extremum. However, all is not lost. First we shall locate all points where the derivative is zero or fails to exist. Then, we need only examine those points to determine if they are relative extrema, as we shall see.

12 10 8 6 4 2 0 4 3.5 3 2.5

2 1.5 1

0.5

1.5

2.5

8 4 0 -4 -8

0.5

1.5

2.5

Figure 8a

-6 -4 -2

Figure 8b

Figure 8c

DEFINITION 4 Any number in the domain of the function at which the derivative is either zero or fails to exist is called a critical number. If c is a critical number for the function defined by y = f1x2 then the point (c, f(c)) is called a critical point.

It follows from this definition that every relative extremum is a critical point. However, from Figure 8 we see that not every critical point is a relative extremum. Example 2 Find the critical points of f1x2 = x3 - 3x2 - 9x + 15 Solution f 1x2 = 3x - 6x - 9. Since the derivative always exists we need only find those points at which it is zero. Factoring, we have 3x2 - 6x - 9 = 31x2 - 2x - 32 = 31x - 321x + 12. This will be zero if x = - 1 or 3. These are the critical numbers. They are the x-coordinates at the critical points. To find the y-coordinates, we must substitute each x-value into

2

262

* **

Section 3.1

Extrema of a Function

the original equation. We find f1 - 12 = 20, and f132 = - 12. Thus the critical points are 1 - 1, 202 and 13, - 122.

Example 3 Find the critical points of f1x2 = 2x2 - 4. Solution Writing f1x2 = 1x2 - 421/2, we take its derivative using the chain rule. We have that f1x2 = 1 2 x 1x - 42-1/2 # 2x = 2 2 2x - 4

Here the derivative is a fraction. A fraction can only be zero when the numerator is zero and its denominator is not zero. Therefore, whenever f 1x2 is a quotient, we determine those values at which it is zero by setting the numerator equal to zero. In this case, that occurs when x = 0. However, x = 0 is not in the domain of the given function, (why?). Therefore, it is rejected from consideration. The derivative will fail to exist at those x-values at which the denominator is zero. In this example, when x2 - 4 = 0, or x = - 2 or x = 2. Thus, we have two critical numbers: x = - 2, and 2. The corresponding y-values are f1 - 22 = 0 and f122 = 0. Thus, the critical points of the given function are 1 - 2, 02 and (2, 0).

Solution The domain of this function is all x except x = 1. Using the quotient rule, we -3 find f 1x2 = . 1x - 122 Since the numerator is a constant, this function can never be zero. The denominator is positive for all x in its domain, therefore, the function has no critical points.

Now that we can locate critical points we may finally give a procedure with which to obtain the extrema of a continuous function on the closed interval a x b. Observe that the maximum and minimum values of a continuous function on a closed interval can occur either at an endpoint of the interval or at critical point. Therefore, we need only compute the y- values at each of these points and compare to determine which is the maximum and which is the minimum.

Procedure for Determining the Extreme Values for a Continuous Function on [a, b]

1. Compute f(a) and f(b). 2. Differentiate and determine the critical numbers, c1, c2, c3, , cn 3. Compute the y-values at each of these critical numbers obtaining, f1c12, f1c22, f1c32, , f1cn2 4. The largest y-value in Steps 1 and 3 is the maximum value, the smallest is the minimum value.

Section 3.1

Extrema of a Function

* **

263

4.

Solution 1. We find f102 = 15 and f142 = - 5. 2. The critical numbers were found in Example 2. They are x = - 1 or x = 3. x = - 1 is rejected since it lies outside the given domain. Thus, x = 3 is the only critical number. 3. f132 = - 12. 4. The largest of the y-values in steps 1 and 2 is 15 and the smallest is - 12. Thus the maximum is 15 (the extreme point is (0, 15)), and the minimum is - 12 (the extreme point is 13, - 122).

We see that if we have a continuous function defined on a closed interval we can easily determine the extrema. What if the interval is not closed or even unbounded? What should we do then? In such situations there is no guarantee that the function has extrema. However, there is one case that often arises in applications for which the question is easily resolved. We shall refer to the key notion as The Only Critical Point Test.

THEOREM 2: THE ONLY CRITICAL POINT TEST Suppose the differentiable function defined on an open or unbounded interval by the equation y = f1x2 and it has a relative extremum at x = c, and x = c is the only critical number for this function, then f(c) is the extremum of the function.

Theorem 2 says that if the only critical point is a relative maximum, then it is the maximum for the function. Similarly, if the only critical point is a relative minimum, then it is the minimum for the function. This theorem applies to any case in which end points are not present. It says that any time a function has exactly one critical point which is a relative extremum it is the (only) extremum. To understand why the theorem is true, we need only examine Figure 9b which shows a relative minimum at P. Since it is the only critical point, the curve cannot turn around and go lower than it is at P. If it did, it would have to have another critical point. In addition, since the function is assumed differentiable it is also continuous. Thus, it cannot jump up or down to a new extreme value. Thus, the function has a minimum at P. Similarly, in Figure 9a, Q is a relative maximum. There can be no higher point on the curve since it is the only critical point. (This theorem can be

5 4 3

Q

6

4 2 1 0 2

-0.5

0.5

1.5

2.5

-0.5

0.5

1.5

2.5

Figure 9a

Figure 9b

264

* **

Section 3.1

Extrema of a Function

generalized to the case of a continuous function where the derivative fails to exist at the single critical point, Exercise 38.) Example 6 x on [0, q 2 has a single x + 1 critical point which is a maximum. Does it have a minimum? Show that the function defined by the equation f1x2 =

2

Solution Note that since the domain of the function is x 0, both the numerator of this function (and the denominator) are never negative, and the graph of this function lies entirely in the first quadrant. Moreover, it will be zero only when x = 0, therefore, we can easily answer the second question, the minimum of this function is 0, (which occurs when x = 0). x We also note that lim 2 = 0, as x gets large, thus, y = 0 is a horizontal asymptote. x: q x + 1 Therefore, the graph begins at (0, 0) rises and must turn at least once at some critical point (which is necessarily a relative maximum) and eventually approach the x-axis as its horizontal asymptote. 1 - x2 , this derivative is zero only when the numerator is The derivative is f 1x2 = 2 1x + 122 zero, namely 1 - x2 = 0, or x = - 1 or 1 Since - 1 is not in our domain, we have a single critical number at x = 1. From the discussion above, it is clear that at the critical point A 1, 1*2 B the graph has a relative maximum, and by the Only One Critical Point Test, it is a maximum. Thus, the maximum of this function is 1*2. We close this section with a reminder of some notation used to represent intervals (see Section 0.5). Instead of writing a x b, it is common to write [a, b]. In place of a 6 x 6 b we often use (a, b). Observe that the bracket indicates inclusion of the end point, while a parenthesis indicates exclusion of the end point. Thus, (a, b] would stand for a 6 x b. This notation will be used in the exercises. The TI 89 calculator can locate relative extrema. Suppose the function is stored as y1(x). We illustrate with the function of Example 6. This function is entered as y1(x) and the with command is used to include its domain. (Recall the is obtained by pressing *) See Figure 10.

Calculator Tips

x x2 + 1

Section 3.1

Extrema of a Function

* **

265

We next have the TI 89 plot the graph for us, (we choose a window containing our domain), see Figure 11.

x x2 + 1

We now press the MATH (F5) key which gives us various options; number 3 is MINIMUM and number 4 is MAXIMUM. We choose maximum. The TI asks you for a Lower Bound, move the cursor to any point to the left of the maximum and press enter, it then asks for an Upper Bound, move the cursor to any point to the right of the maximum and press enter. It then gives xc: 1. Yc: .5, that is the maximum is 0.5 at the point (1, 0.5).

1. Sketch the graph of a function which is increasing to the left of x = 1 and decreasing to its right. 2. Sketch the graph of a function which is decreasing to the left of x = 1, increasing to its right and passes through 10, - 32. 3. Sketch the graph of a continuous function increasing on - 1 6 x 6 2 and decreasing on 2 6 x 6 4. Indicate the point M on your graph which is a relative maximum. 4. Sketch the graph of a continuous function decreasing on - 5 6 x 6 3 and increasing on 3 6 x 6 7. Indicate the point m on your graph which is a relative minimum. 5. Sketch the graph of a continuous function defined on - 5 x 3, that has relative maxima at 1x = - 12 and 1x = 32 and a relative minimum at (x = 0). 6. Sketch the graph of a continuous function that has a relative maximum at A 1, 1*2 B a minimum at 1 - 2, - 52 and a maximum at 1 - 3, 12. 7. Which of the points in Figure 12 are: (a) maxima? (b) minima? (c) relative maxima? (d) relative minima? 8. (a) Draw the graph of a function which has two maxima and two relative maxima. (b) What must be true about the y-values at the maxima? 9. Draw the graph of a function such that the minimum is also a relative minimum. 10. Draw the graph of a function such that the maximum is also a relative maximum. In each of Exercises 11 27 locate all critical points. 11. f1x2 = x2 - 2x + 3 12. f1x2 = 4x5 13. g1x2 = ax2 + bx + c 14. s1t2 = 2t3 - 9t2 - 60t + 5 15. h1x2 = 4x3 - 13x2 + 12x + 9 16. f1x2 = 2x4 + 2x3 - x2 - 7 17. h1x2 = 112 - x22 *2 18. r1x2 = 4x3/4 + 2 19. v1t2 = t6 - 3t2 + 5 D B E A C G F 20. s1t2 = 1t - 1241t + 323 21. f1x2 = 22. w1x2 = 23. h1x2 = x + 3 . x - 3 3x . 4x2 + 9 1 x2 - x - 2

1

Figure 12

266

* **

Section 3.2

36. Let f1x2 = x- *2 on the interval (0, 1]. (a) Does f satisfy the conditions of Theorem 1? (b) Does it have a maximum value? (c) Does it have a minimum value? 37. Let f1x2 = x-2/3 on the interval [ - 1, 8]. Does f have extrema on this interval? 38. Show, by means of a sketch, that if a continuous function with a single critical point that is a relative extremum, then this critical point is also an extremum. 39. Determine the extrema of the function defined by the equation x on 1 - q , 0]. Justify your conclusions. f(x) = 2 x + 1 40. (a) Determine the extrema of the function defined by the equation x2 f(x) = 3 on [0, q 2. Justify your conclusions. x + 1 (b) Does this function have extrema on 1 - q , 0]?

1

24. f1x2 = 1x2 - 922/3. 25. w1x2 = x 2x - 4 x2 26. g1x2 = 2 . x - 9 27. f1x2 = 4 - x2/3. In Exercises 28 34 determine the extrema on the given interval. 28. f1x2 = 3x + 9 on: (a) [ - 1, 3]; (b) 1 - 1, 32 29. f1x2 = x2 - 2x + 3 on: (a) [0, 2]; (b) [2, 3]; (c) (2, 3] 30. f1x2 = 2x3 + 3x2 - 12x - 6 on: (a) [ - 3, 2]; (b) [ - 5, 3] 31. f1x2 = 4x3/4 + 2 on: (a) [0, 16]; (b) (0, 16); (c) [0, 16) 32. f1x2 = x/1x2 + 12 on: (a) [0, 2]; (b) [ - 2, 2]; (c) (0, 2); (d) 1 - 2, 22 33. f1x2 = x4 - 6x3 + 12x2 + 2 on: (a) [1, 2]; (b) [ - 1, 2] 34. f1x2 = x 22 - x on [0, 3/2]. 35. Let f(x) be defined on the closed interval [0, 1] by the rule: f1x2 = b 2x2 1 if 0 6 x 6 1 if x = 0 or x = 1 .

(a) Does f have extrema on [0, 1]? (b) Is this a violation of Theorem 1?

Increasing and Decreasing Functions The First Derivative Test Sign Diagrams Calculator Tips

In the preceding section we defined a critical point as any point in the domain of the function at which the derivative of the function is zero or does not exist. There is one problem that we have not resolved. How do we classify a critical point? For example, the Only Critical Point Test requires that we classify the critical point as either a relative maximum or relative minimum before it can be applied. Let us now consider the question of how to classify the critical points. Of course, there are three possibilities: The point is a relative maximum, a relative minimum, or neither. Fortunately, there is a simple procedure by which we may make the classification. We shall see that we need only examine the sign of the derivative. Recall that it is generally the case that just to the left of a relative maximum the function is increasing and just to the right it is decreasing. At a relative minimum the reverse occurs. The function is decreasing as we approach the minimum from the left and increasing to its right. In Figure 1 we show the graph of an increasing function. We have indicated the tangent lines at a few points on the graph. Notice that each tangent line has positive slope. That is, at every point where a function is increasing and the derivative exists it is non-negative*. More importantly for our purposes, wherever the derivative is pos*It is possible for a function to be increasing (or decreasing) on an interval and yet have its derivative equal to zero at one or more points on the interval. Consider f1x2 = x3, which has f 102 = 0. The function is increasing on any interval containing x = 0. Such critical points pose no problem, as we shall see.

Section 3.2

itive the function is increasing. Similarly, Figure 2 is the graph of a function whose derivative is always negative. Observe that its tangent line has negative slope at each point at which it exists, and so the function is decreasing. If a function is neither increasing nor decreasing, then it is a constant. We know that the derivative of a constant is zero everywhere and it can be shown that the converse is true**.

These observations give rise to the following theorem. (We assume f is a differentiable function on a given domain.)

THEOREM 1 In any interval on which f 1x2 7 0 the function is increasing; in any interval on which f 1x2 6 0 the function is decreasing, and in any interval on which f 1x2 = 0 the function is constant.

But, what does this tell us about relative maxima and minima? It means that if M is a relative maximum, the derivative of the function is positive just to its left and negative just to its right (Figure 3a). Similarly, if m is a relative minimum, the derivative is negative just to its left and positive just to its right (Figure 3b). We summarize this in the following theorem known as The First Derivative Test, where we assume f is continuous at the critical point.

M m

**With the exception of certain singular functions which are analyzed in more advanced courses.

268

* **

Section 3.2

(1) Suppose M is a critical number for f, and f 1x2 is positive just to the left of M and negative just to the right of M, then f(M) is a relative maximum. (2) Suppose m is a critical number for f, and f 1x2 is negative just to the left of m and positive just to the right of m, then f(m) is a relative minimum. (3) If c is a critical number of f and f 1x2 does not change its sign around c then f(c) is neither a relative maximum nor a relative minimum. This theorem is graphically illustrated by the sign diagrams of Figure 4.

f is increasing + M f is decreasing

sign of f '(x)

f is increasing +

Figure 4b: m is a relative minimum The First Derivative Test when used in conjunction with The Only Critical Point Test can be an extremely powerful tool for analyzing the behavior of a function. Consider Example 1 below. Example 1 Find and classify the critical point(s) of f1x2 = x2 - 6x. Solution We find the derivative f 1x2 = 2x - 6 Since this exists for all x, the only critical points are where f 1x2 = 0. Putting 2x - 6 = 0, yields x = 3 as the only critical number. Now factoring f 1x2, we have, f 1x2 = 21x - 32. It is not hard to see that 21x - 32 is negative if x 6 3 and 21x - 32 is positive if x 7 3. Hence, we immediately conclude that f(3) is a relative minimum. Since f132 = 32 - 6132 = - 9, the critical point is 13, - 92 and this is a relative minimum. By The Only Critical Point Test, we now know that the minimum value for f1x2 = x2 - 6x is - 9. (Of course, the graph of y = f1x2 is a parabola opening upward.)

Thus, we have reduced the classification of relative extrema to examination of the sign of the derivative. We shall examine the sign of the derivative by using a sign diagram. In Section 0.5 we give a detailed discussion of this procedure and recommend you review it before proceeding. An abridged discussion follows.

Section 3.2

* **

269

Consider the quotient 13x - 521x - 22/1x + 32. The zeros of the numerator (that is, those values of x at which the quotient becomes zero) are x = 5/3 and x = 2. The zero of the denominator (that is, the value at which the quotient is undefined) is x = - 3. Observe that the sign of the quotient is determined by the signs of the three expressions: 13x - 52, 1x - 22, and 1x + 32. If an odd number of these are negative then the sign of the quotient is negative, otherwise the sign of the quotient is positive. The quotient can change sign only where one of the three is zero. We shall refer to these as key numbers. Suppose we indicate these key numbers on the number line (Figure 4).

+ 5/3 2 +

Figure 4c: sign of 13x - 521x - 22/1x + 32 Observe that for any x 6 - 3 (choose any number you wish as long as it is less than - 3, say - 5,) each factor is negative. Thus, the quotient is negative. If x is any number between - 3 and 5/3 (say 0) both factors in the numerator are negative and the denominator is positive, resulting in a positive quotient. If x is any number between 5/3 and 2 (say 1.9), 1x - 22 is negative the other factors are positive, and the sign of the quotient is negative. If x is any number larger than 2 (say 5), all factors are positive and the sign of the quotient is positive. The sign of the quotient in each interval is indicated in Figure 4. Again, note that the possible sign changes of the function can occur only at these key numbers. Thus, once we test any point in an interval between the key numbers, we have determined the sign of the function throughout the interval. This method is quite general and we use it to draw the sign diagram for any function whose key numbers have been determined. Sign Diagrams

1. Locate the key numbers of the function 2. Draw them on the number line and test any point between each pair of successive key numbers to determine the sign of the function at the test point. Record its sign on the number line. Now that we are able to determine the sign of a function it is a simple matter to classify critical points. As you can see, usually, the critical numbers correspond to the key numbers for f 1x2. The classification follows immediately from the first derivative test and analysis of the sign diagram of the derivative. Let us look at some examples that involve curve sketching. Example 2 Determine where f1x2 = x3 - 3x2 - 9x + 15 is increasing and decreasing. Use this information to classify its critical points and sketch its graph. Solution We must first determine the critical numbers f 1x2 = 3x2 - 6x - 9. Factoring, we have, f 1x2 = 31x2 - 2x - 32 = 31x - 321x + 12 There are two critical numbers, they occur at x = - 1 and x = 3. (The critical numbers are the zeros of the derivative.)

270

* **

Section 3.2

We proceed as follows: 1. Draw the number line indicating the critical numbers. Place a 0 above each critical number to remind us that the derivative is zero at that value of x (that is, the tangent line is horizontal at these values of x). See Figure 5a.

0 *1 0 3

Figure 5a: sign of f 1x2 = 31x2 - 2x - 32 = 31x - 321x + 12 Choose any number less than - 1 (say - 2) and evaluate the sign of the derivative at this value of x. (You may substitute into either the factored or non-factored form of the derivative, it does not matter which is used.) The sign is positive. Choose any number between - 1 and 3 (say 0), the sign of the derivative is negative. Choosing any number larger than 3 (say 4) we find the sign is now positive. We record the signs giving Figure 5b.

+ 0 *1 0 3 +

Figure 5b: sign of f 1x2 = 31x 2 - 2x - 32 = 31x - 321x + 12 The function is increasing over any interval on which its derivative is positive. Thus, from Figure 5b we see that the function is increasing if x 6 - 1 or x 7 3. Similarly, it is decreasing when f 1x2 6 0, which occurs when - 1 6 x 6 3. Figure 5b indicates that the function is increasing to the left of x = - 1 and decreasing to its right. Therefore, at x = - 1 the function has a relative maximum. To the left of x = 3 the function is decreasing and to its right it is increasing. Thus, the function has a relative minimum at x = 3. Since f1 - 12 = 20 and f132 = - 12 we conclude that M1 - 1, 202 is the relative maximum and m13, - 122 is the relative minimum. To sketch the curve we begin by drawing Figure 6a which illustrates that the tangent line is horizontal both at the peak (M) and the valley (m). In Figure 6b, the sketch is completed by joining the relative extrema with a continuous curve. Note that we have labeled an additional point, (0, 15) which is the y-intercept of the function. It is found by setting x = 0, that is f102 = 15. (We remind you that the y-coordinate of any point on a graph is obtained by substituting the appropriate x-coordinate into the equation of the function.)

Section 3.2

* **

271

Example 3 Sketch the graph of f1x2 = 2x4 - 4x3 + 7, indicating its relative extrema. Label the intervals on which it is increasing and decreasing. Solution We start with the derivative, f 1x2 = 8x3 - 12x2 = 4x212x - 32 The critical numbers occur where the derivative is zero, that is, when x = 0 or x = 3/2. We label these critical numbers on the number line as indicated in Figure 7a.

0 0 0 3/2

Figure 7a: sign of f 1x2 = 8x3 - 12x2 = 4x212x - 32 Choosing any number less than 0, we find that f 1x2 6 0. Choosing any number between 0 and 3/2 we find that f 1x2 6 0 on this interval as well. Choosing any number greater than 3/2, we find that f 1x2 7 0. We record this information and obtain Figure 7b.

0 0 0 3/2 +

Figure 7b: sign of f 1x2 = 8x 3 - 12x 2 = 4x 212x - 32 The function is decreasing if x 6 0 or if 0 6 x 6 3/2. The function is increasing if x 7 3/2. Since f 1x2 does not change sign at x = 0, it is neither a relative maximum nor a relative minimum. Nevertheless, the tangent line is horizontal at this point. At x = 3/2 the function has a relative minimum since at this point it changes from a decreasing to increasing function. We label both these points in Figure 8a, also indicating that we have a valley at x = 3/2. Note that f102 = 7 and f13/22 = 29/8 = 3.625. We complete the sketch in Figure 8b.

(0, 7) (0, 7)

272

* **

Section 3.2

Note in the previous example, 29/8 is the minimum of this function. Example 4 Sketch the graph of f1x2 = 6x2/3 - 2x. Solution f 1x2 = 4x-1/3 - 2 = 212 - x1/32 4 4 - 2x1/3 2 = = x1/3 x1/3 x1/3

This function has two critical numbers. First, we see that the derivative fails to exist at the zero of the denominator, x = 0. Second, the derivative is zero at the zero of the numerator, when x1/3 2 - x1/3 = 0 = 2; x = 23 = 8

Since f102 = 0 and f182 = 8, the critical points are (0, 0) and (8, 8). Since the derivative does not exist at x = 0, we place DNE (Does Not Exist) above it in the sign diagram, Figure 9. If any number less than 0 is tested, the sign of the derivative is negative (remember the cube root of a negative number is negative). If any number between 0 and 8 is tested, the sign of the derivative is positive. If any number greater than 8 is tested the sign is negative. This information is summarized in Figure 9.

DNE 0 + 0 8

Figure 9: sign of f 1x2 = 212 - x1/32/x From Figure 9 we see that the function has a relative maximum at x = 8 and a relative minimum at x = 0. In this case, the derivative does not exist at x = 0, because the tangent line is vertical there. We first sketch Figure 10a and then complete the sketch in Figure 10b. Note that f102 = 0 and f182 = 8.

M (8, 8)

M(8, 8)

m(0, 0) m(0, 0)

Section 3.2

* **

273

Note that the above graph has two roots, the first at the origin and the second not shown on the sketch. How could you determine that this second root is x = 27? We next consider some examples from Economics. Example 5 Suppose that the relationship between price and demand for a certain brand of color television set is given by the equation p = - 3x + 1200, where p is the price of a set in dollars and x is the number of sets demanded. What price should be charged per set if the total revenue is to be maximized? Solution The total revenue is given by R = xp = x1 - 3x + 12002 or R = - 3x2 + 1200x. Now, we must have 0 x 400, in order that neither x nor p be negative (verify this!). At one end point, x = 0, that is the demand is zero and there is no revenue. At the other end point x = 400 gives p = 0. Here the price is zero so there is no revenue. Differentiating, we find R 1x2 = - 6x + 1200. The critical number occurs when - 6x + 1200 = 0, or x = 200. Using sign analysis on R 1x2 we obtain the following sign diagram. (Figure 11)

+ 0 200

Figure 11: Sign of R 1x2 = - 6x + 1200 Thus, the revenue has a relative maximum when x = 200 (200 television sets are demanded). This is the only critical number. Therefore, it is the value that maximizes the revenue. For x = 200, p = - 312002 + 1200 = $ 600. The revenue is maximized when the cost for each set is $600.00. The maximum revenue is $120,000.

Now you might be thinking that the calculus was not needed to solve this example. After all, the total revenue equation is R1x2 = - 3x2 + 1200x. Its graph is a parabola opening downward. Therefore, the maximum revenue is at the highest point on the parabola, at its vertex. However, it is not always an easy algebraic matter to determine the highest or lowest point on a curve. Consider the following example, which is similar to the previous one except that the demand equation is no longer linear. Example 6 The price of a bicycle is given by the equation p = 91300 - 10x21/2, where p is the price per bicycle in dollars when x bicycles are demanded. If the total revenue is to be maximized, what price should be charged for each bicycle? Solution Observe that x 30, otherwise the radicand would be negative. Since x stands for the number of bicycles, we must have x 0. Thus, R1x2 = xp = 9x1300 - 10x21/2 where 0 x 30. This is a continuous function defined on a closed interval. By the

274

* **

Section 3.2

Extreme Value Theorem, the maximum must occur at either a critical point or an end point. Since R102 = R1302 = 0, the maximum must occur at a critical point. Using the chain rule, we have that 1 R 1x2 = 9x # 1300 - 10x2-1/21 - 102 + 1300 - 10x21/2 # 9 2 - 45x R 1x2 = + 9 2300 - 10x 2300 - 10x The derivative fails to exist at x = 30. (The tangent line to the curve is vertical at (30, 0).) The other critical point occurs when R 1x2 = 0; - 45x 2300 - 10x That is, when 9 2300 - 10x = 45x 2300 - 10x + 9 2300 - 10x = 0

Clearing fractions and dividing by 9, we find 300 - 10x = 5x, or 15x = 300. Thus, x = 20. Now, R1202 = 200. It follows from the Extreme Value Theorem, that a maximum occurs when x = 20. The price per bicycle is then p = 9 2300 - 101202 = $ 90.

Calculator Tips

We shall examine additional applications of the derivative to optimization theory in Section 3.4. As mentioned before, the TI 89 can differentiate functions using the d key (located above the number 8). This means the TI 89 can determine the critical points of a function. Suppose the equation of the function is already stored as y1(x) in the calculator. To determine the x-coordinates where the derivative is zero, we need only enter solve1d1y11x2, x2 = 0, x2. To find the points at which the derivative does not exist, we can solve for the zeros of the denominator of the derivative, that is, we first determine d(y1(x), x), suppose we then save its denominator as y2(x), then we enter solve1y21x2 = 0, x2. The y-coordinates of the critical points are obtained by substituting the critical numbers just found into y1(x). Alternately, the relative maxima and minima may be found by having the calculator graph y1(x), press F5, choose maximum (or minimum) and for each relative extremum, enter an appropriate lower and upper bound. Both the x and y-coordinates are found this way, but it means you have to have the graph displayed. The problem with this approach is that the window used to display the graph may not clearly show all the relative extrema.

In Exercises 1 14 use the first derivative to determine where the given function is increasing and decreasing. 1. g1x2 = - 5x + 3 2. f1x2 = mx + b, if: (a) m 7 0; (b) m 6 0 3. f1x2 = 3x - 2x + 1

2

4. g1t2 = t2 - 1 5. h1x2 = 2x2 - 18 6. s1t2 = t2 + 4t - 21 7. g1x2 = 4x3 - 24x2 + 36x + 96 8. r1t2 = 4t3 - 15t2 + 18t + 2 9. v1x2 = x4 + 5 10. q1x2 = x4 - 32x (Hint: x3 - a3 = 1x - a21x2 + ax + a22.)

Section 3.2

11. f1x2 = 3x - x1/3 12. p1x2 = 1x - 12/1x + 12 13. j1x2 = 28 - 2x 14. f1x2 = x 2x + 1

* **

275

y = mx + b. (Hint: Substitute for y in the objective function and apply the first derivative test.) 41. Show that ax2 + bx + c always has the same sign as the constant term c if b2 - 4ac 6 0. 42. Suppose a function is continuous on [a, b] and differentiable on (a, b). Additionally, suppose that f1a2 = f1b2. Sketch various possibilities for the graph of f. Show from your graphs that there must be some point c, where a 6 c 6 b, such that the tangent line at x = c is parallel to the x-axis 1f 1c2 = 02. This result is known as Rolle s Theorem. 43. Suppose that a function satisfies the same conditions as in the previous exercise, except that f1a2 Z f1b2. One such function is illustrated in Figure 12. Show that if the figure is rotated so that the dotted line joining A(a, f(a)) to B(b, f(b)) is horizontal, then the rotated figure satisfies all the conditions of Rolle s Theorem. Therefore, there must be at least one point c, a 6 c 6 b, such that the tangent line at x = c is parallel to the line joining A to B. Thus, f1b2 - f1a2 show that f 1c2 = . This result is known as the Mean Value b - a Theorem, or The Law of the Mean. 44. Use the Mean Value Theorem to prove the First Derivative Test. Hint: let x1 6 x2 be any two points in the interval in question. Apply the Mean Value Theorem using these points to deduce that f1x22 - f1x12 = 1x2 - x12f 1c2.

In Exercises 15 24 classify all critical points. 15. f1x2 = x2 - 2x + 3 16. f1x2 = 4x5 17. g1x2 = ax2 + bx + c 18. s1t2 = 2t3 - 9t2 - 60t + 5 19. h1x2 = 4x3 - 13x2 + 12x + 9 20. f1x2 = 2x4 + 2x3 - x2 - 7 21. h1x2 = 112 - x221/2 22. r1x2 = 4x3/4 + 2 23. v1t2 = t6 - 3t2 + 5 24. s1t2 = 1t - 1241t + 323 In Exercises 25 35 sketch the graph of the function defined in the given exercise. Use all the information obtained from the first derivative. 25. Exercise 7. 26. Exercise 8. 27. Exercise 11. 28. Exercise 15. 29. Exercise 20. 30. Exercise 21. 31. Exercise 22. 32. Exercise 23. 33. Exercise 24. 34. f1x2 = 1x2 - 922/3 35. f1x2 = x/1x2 + 12 36. Find two numbers whose sum is 100 such that their product is as large as possible. Justify your conclusion! 37. A rectangular picture frame is to enclose an area of 72 in2. If the cost of the top and bottom is twice the cost of the sides, determine the frame s dimensions if the total cost is to be minimized. Justify your conclusion! 38. The price of a magazine, in cents, is given by the equation p = 101147 - 0.02x21/2, where x is the number of magazines demanded. What should the price be to maximize the publisher s total revenue? Justify your conclusion! 39. For the demand equation p = 1a - bx21/2, with a 7 0 and b 7 0, show that the price at which the total revenue is maximized is independent of b. 40. Show that the objective function P = Ax + By + C is either increasing, decreasing or remains the same when it is evaluated along the line

B(b, f (b))

A (a, f (a))

Figure 12

45. Show that if two functions f and g have the same derivative on the same interval, then their difference, f1x2 - g1x2 = constant on the interval. Hint: Define D1x2 = f1x2 - g1x2, show D 1x2 = 0 on the interval and then deduce the result as a consequence of the first derivative test.

276

* **

Section 3.3

The Second Derivative Higher Order Derivatives Velocity and Acceleration Concavity The Second Derivative Test for Concavity The Second Derivative Test for Relative Extrema Implicit Differentiation and Curve Sketching Calculator Tips

We have learned how examination of the first derivative can help us to sketch a curve. It tells us where the curve is increasing and decreasing and allows us to classify the critical points. What it does not do is tell us how the curve increases or decreases. For example, think about the way a curve looks as the function nears its minimum (for example, at point P in Figure 1). At first, it may be decreasing very quickly.

Figure 1 But, as it nears the low point, its rate of decrease must slow until it stops decreasing altogether at the minimum. After passing the minimum, the function must begin to increase; first slowly and then more rapidly. Thus, its rate of change is itself changing. To analyze this behavior, we need to discuss the notion of higher order derivatives. dy Suppose y = f1x2 = 4x3 - 2x2 - 7x + 9, then f 1x2 = = 12x2 - 4x - 7. dx We see that f 1x2 is itself a function of x. Furthermore, it is a differentiable function. We now ask, what is the derivative of the derivative? The answer is very simple. It is d 112x2 - 4x - 72 = 24x - 4. dx The derivative of the derivative is called the second derivative. The following symd2y bols are commonly used to denote the second derivative: f 1x2, y , 2 . (You might be dx

Section 3.3

* **

277

2

d2y dx2

. The reason

d1dy2 dy d2y d dy a b . It has the form or , written as . dx dx dx dx 1dx22 dx2 We may extend the above notion and ask for third, fourth or even higher order derivatives. In the above example suppose we want the derivative of the second derivative, d3y which is called the third derivative. It is denoted by f 1x2 = y = . In our example, dx3 f 1x2 = 24. We could go on in this manner defining fourth, fifth, and even higher order derivatives. (For our example the fourth and higher order derivatives are all zero.) However, in general, if we wanted to indicate, for instance, the 27th derivative it would not be wise to write f with 27 primes. Instead we would indicate it by f12721x2 or y1272. More generally, the following notation is used. The nth order derivative of y = f1x2 is denoted dny by any of the following: f1n21x2, y1n2, or n . dx Example 1 Find the first three derivatives for y = 4x5 - 5x-4 Solution The derivatives are taken successively by the usual rules. y = 20x4 + 20x-5 1First derivative.2 y = 80x3 - 100x-6 1Second derivative.2 y = 240x2 + 600x-7 1Third derivative.2

Example 2 1 Find the second derivative of f1x2 = 1x2 + 12 *2 Solution For the first derivative, we need the chain rule. f 1x2 = 1*21x2 + 12- *212x2 = x1x2 + 12- *2 For the second derivative, we need the product rule and the chain rule. f 1x2 = x C A - 1*2 B 1x2 + 12-3/212x2 D + 1x2 + 12- *2[1] f 1x2 = x21x2 + 12-3/2 + 1x2 + 12- *2 Factoring out 1x2 + 12-3/2, we have f 1x2 = 1x2 + 12-3/2[x2 + 1x2 + 12] = 1 . 1x + 123/2

2

1 1 1 1

Example 3 Find f 1x2 if f1x2 = x/1x2 + 12. Solution Applying the quotient rule once again, we have, f 1x2 = 1x2 + 12 - x12x2 1x2 + 122 = 1 - x2 . 1x2 + 122

278

* **

Section 3.3

f 1x2 =

1x2 + 122[ - 2x] - 11 - x22[21x2 + 12[2x]] 1x2 + 122 = - 2x1x2 + 12[1x2 + 12 + 211 - x22] 1x2 + 124 = - 2x13 - x22 1x2 + 123

The more complex the original function, the more involved it is to find the various derivatives. To find the second derivative in Example 3 required a double application of the quotient rule, along with the chain rule. We have several interpretations for the (first) derivative. What about higher order derivatives? We shall only be concerned with two interpretations of the second derivative. The first deals with particle motion and the second with curve sketching. First, we consider s = f1t2, where s is position and t is time. We know v, the velocity, is the instantaneous rate of change of position with respect to time and dv ds v = s 1t2 = . Now the derivative of velocity, , is the instantaneous rate of change of dt dt velocity with respect to time. This derivative is called the acceleration, and we write dv d2s a = = 2 . The acceleration measures how velocity changes with time. dt dt Example 4 The position (in centimeters) of a particle as a function of t (in seconds), is given by the equation s = t3 - 30t2 + 15t + 23. (a) Find the velocity and acceleration of the particle at time t = 2. (b) Find the time at which the acceleration is zero. (c) What is the velocity at that time? Solution We find the first two derivatives. v = v1t2 = s = 3t2 - 60t + 15 a = a1t2 = v = s = 6t - 60 (a) At t = 2, the velocity is v122 = 31222 - 60122 + 15 = - 93 cm/sec. The acceleration is a122 = 6122 - 60 = - 48 cm/sec.2 Thus, the velocity is negative. The particle is moving in the negative direction. The acceleration is also negative, which means the velocity is decreasing. That is, it is becoming more negative. Notice this interpretation. The velocity of the particle is becoming more negative. So we say that the velocity is decreasing but, actually the particle is going faster. That is, its speed, which is the absolute value of velocity, is getting larger. (b) The acceleration is zero when 6t - 60 = 0; at t = 10 sec. (c) At t = 10, the velocity is v1102 = 311022 - 601102 + 15 = - 285 cm/sec. You should notice that when a1t2 = 0, we have a critical point for v(t). You should verify that this point 110, - 2852 gives the minimum velocity.

Section 3.3

* **

279

What about the geometric interpretation of the second derivative? What shall see that it determines the shape or concavity of a function.

Concavity

C G B D F

E A

Figure 2: Illustrating Concavity Consider the function whose graph is sketched in Figure 2. Observe that the graph is increasing from A to C as well as from E to G. However, the way it increases is different. From A to B the graph is U-shaped, while from B to C the graph is upside down U-shaped. Similarly between E and F it is U-shaped and between F and G it is upside down U-shaped. Between C and E the function is decreasing, from C to D it is upside down U-shaped, and from D to E it is U-shaped. Mathematicians use the term concavity to indicate how the curve is shaped. Thus, the function in Figure 2 is concave upward from A to B, and from D to F. It is concave downward from B to D and from F to G. It should be pointed out that at any point in an interval on which the graph is concave upward, the tangent line lies beneath the graph. Similarly, in any interval on which the function is concave downward the tangent line lies above the graph. See Figure 3, where the graph is concave upward at P and concave downward at Q. You might be wondering what those points are called at which the curve changes its concavity. They are called inflection points. More precisely, we have the following definition.

P Q

Figure 3:

DEFINITION 1 Any point on the graph of a function at which the concavity changes is called an inflection point.

We now have a similar situation to the one we had in Section 3.2. We know what an inflection point is, but how do we locate it? That is, how do we determine the concavity of the function? We shall see that the answer to this question is analogous to the determination of where a function increases or decreases.

280

* **

Section 3.3

You might be thinking that concavity, while being a nice geometrical property, is not very useful. After all, the major focus of the previous two sections was the location of extrema, how can concavity aid us? Another look at Figure 2 should convince you otherwise. Point C is a relative maximum. Observe that at C the function is concave downward. Similarly, at E, a relative minimum, the function is concave upward. Thus we see that if we have a point at which f 1x2 = 0, and if we can determine the concavity at that point, then we can classify the critical point by its concavity. We have the following preliminary theorem, which we will revise shortly.

THEOREM 1(PRELIMINARY VERSION) Suppose f is a differentiable function with f 1c2 = 0. If the function is concave upward at x = c, then the function has a relative minimum at x = c. If the function is concave downward at x = c, then the function has a relative maximum at x = c.

Let us now look at the relationship between the second derivative and concavity. Consider the concave upward function whose graph is given in Figure 4a. Notice that as we follow the graph from left to right the slope of the tangent lines drawn at P, Q and R gets larger. Thus, along this curve, the slope is an increasing function (the slope increases). But the slope is the derivative. Thus f 1x2 is an increasing function. Therefore, d 1f 1x22 7 0, or f 1x2 7 0. by the first derivative test its derivative is positive, that is, dx Similarly, if we follow the slope of the tangent lines, as we move from left to right on the concave downward function in Figure 4b, we see that the slope is decreasing (the tangent lines become less steep).

Q Q P

d 1f 1x22 6 0 or f 1x2 6 0 along this curve. It can be shown that the condx verse, which is called the Second Derivative Test for Concavity, is also true. Thus,

THEOREM 2: THE SECOND DERIVATIVE TEST FOR CONCAVITY In any interval on which f 1x2 7 0, the function is concave upward. In any interval on which f 1x2 6 0, the function is concave downward.

Note how similar the uses of the first and second derivatives are. In each, we need only examine the sign of the appropriate derivative. However, the sign of the first derivative indicates where the function is increasing and decreasing, while the sign of the second derivative reveals where the function is concave upward or downward.

Section 3.3

* **

281

To determine the sign of the second derivative we need only find those points in the domain of the function at which the second derivative is zero or fails to exist. Usually, that means finding these key numbers for the second derivative function. We may then apply sign analysis on the second derivative. Example 5 Determine where f1x2 = x/1x2 + 12 is concave upward and where it is concave downward. Identify the points of inflection for the graph of y = f1x2. . The second derivative 1x2 + 123 always exists (Why?), so we need only consider values at which it is zero. The second derivative is zero when x = 0 or 3 - x2 = 0, that is, x = - 23, 0, 23. To analyze the sign of f 1x2, we begin with Figure 5.

sign of f" ( x) = - 2 x (3 - x 2)/( x 2+ 1) 3 0 0 0 0

- 2x13 - x22

Figure 5 Choosing any number to the left of x = - 23, say - 2, we find that the sign of f 1x2 is negative, thus in this interval, the function is concave downward (CD). Choose any number between - 23 and 0, say - 1, we find f 1x2 7 0. Therefore, it is concave upward (CU) in this interval. Continuing in this manner, we have the completed sign diagram for f 1x2, given in Figure 6.

CD sign of f " ( x ) = - 2 x (3 - x 2)/( x 2 + 1) 3 0 CU + CD 0 0 0 CU +

Figure 6: The Concavity of f 1x 2 = x /1x 2 + 12 The given function is concave upward if - 23 6 x 6 0 or x 7 23. It is concave downward if x 6 - 23 or 0 6 x 6 23. Note that the concavity changes when x = - 23, 0 and 23. These are the x-values at the inflection points. Substituting into the equation of the function to find the y-values, we have the inflection points, I1 A - 23, - 23/4 B , I210, 02 and I3 A 23, 23/4 B . Notice that determining the inflection points is analogous to classifying relative extrema. To find the inflection points we locate those points at which the second derivative is zero or fails to exists, since it is at precisely those points where the sign of f 1x2 may change. The Second Derivative Test for Relative Extrema

We are now ready to revise Theorem 1. Since we can determine the concavity of a function in terms of the second derivative, we restate Theorem 1 as a corollary to the Second Derivative Test, called the Second Derivative Test for Relative Extrema.

282

* **

Section 3.3

THEOREM 1: (REVISED) SECOND DERIVATIVE TEST FOR RELATIVE EXTREMA Let f be a differentiable function with f 1c2 = 0. If f 1c2 7 0, then (c, f(c)) is a relative minimum; If f 1c2 6 0, then (c, f(c)) is a relative maximum

Note that the case f 1c2 = 0 is omitted. If this occurs, the test is inconclusive, and the first derivative test (or other methods) could be used. Consider f1x2 = x3, g1x2 = x4 and h1x2 = - x4. Each of these functions has a critical point at x = 0 and the second derivative at x = 0 is also zero (verify!). However, at x = 0, f has an inflection point, g has a relative minimum, and h has a relative maximum (verify these statements). A slight modification of this test is sometimes used to classify the critical point when f (c) = 0. (See Exercises 71 and 72.) Example 6 Use the Second Derivative Test to classify the critical points of f1x2 = x3 - 3x2 - 9x + 15. Solution f 1x2 = 3x2 - 6x - 9 = 31x - 321x + 12. The critical points are at x = - 1 and 3. Since f 1x2 = 6x - 6, f 1 - 12 = - 12 6 0. Therefore, 1 - 1, 202 is a relative maximum. Similarly, f 132 = 12 7 0. Therefore 13, - 122 is a relative minimum.

Example 7 Sketch the graph of f1x2 = x3 - 3x2 - 9x + 15, indicating all relative extrema and inflection points. Solution In Example 6 we found that the function has a relative maximum at M1 - 1, 202 and a relative minimum at m13, - 122. In addition, we can use the second derivative to determine the concavity and the points of inflection for the curve. We see that f 1x2 = 6x - 6 = 61x - 12. Thus, the second derivative is zero at x = 1. Sign analysis on f 1x2 yields Figure 7.

CD 0 sign of f " ( x ) 6( x - 1) 1

CU +

Figure 7 Since the concavity changes at x = 1, we have that I(1, 4) is an inflection point. (Note that at f 1 - 12 6 0, and f 132 7 0 illustrating the second derivative test for classifying the two critical points.) To sketch the graph of the function, we first plot the relative extrema and inflection points as in Figure 8a. We then join the points, remembering that the concavity is downward until we reach the inflection point, where it changes. The completed graph is given in Figure 8b.

Section 3.3

M(-1,20) (0, 15) I (1, 4 M(-1, 20) (0, 15) I (1, 4)

* **

283

m(3, -12)

m(3, -12)

Example 8 Sketch the graph of f1x2 = x/1x2 + 12 indicating all relative extrema and inflection points. 1 - x2 Solution In Example 1 we found that f 1x2 = 2 and that 1x + 122 - 2x13 - x22 f (x) = . The critical numbers are x = - 1 and x = 1. The sign analysis of 1x2 + 123 the second derivative was done in Example 5 and is given again in Figure 9.

CU +

CD 0 0 0

CU +

Figure 9

From the sign diagram we see that f 1 - 12 7 0, thus, the function has a relative minimum at m A - 1, 1*2 B . f 112 6 0, therefore M A 1, 1*2 B is a relative maximum. The inflection points are at I1 A - 23, - 23/4 B , I210, 02 and I3 A 23, 23/4 B . We begin by plotting the points in Figure 10.

284

* **

Section 3.3

I3( 3 , 3 /4)

I2 (0,0)

I1(- 3, - 3 /4) Figure 10: Plotting the Relative Extrema and Inflection Points In Figure 11 we join the points, remembering that: Between I1 and I2 the function is concave upward; Between I2 and I3 it is concave downward; To the left of I1 it is concave downward, and; to the right of I3 it is concave upward.

I3

I2

I1

Figure 11 To finish the graph, we should observe that the x-axis is the horizontal asymptote for x approaching both + q and - q (verify this!). The completed graph is given in Figure 12.

M(1, ,+*) I3 (

3, 3/4)

I2 (0, 0)

I (- 3,- 3/4)

1

m(-1, -,+*)

Section 3.3

* **

285

Note that the function in the previous example is odd, so the graph is symmetric with respect to the origin. Therefore, we only really needed to determine how the graph looked for x 7 0, and the rest of the graph could be drawn by symmetry. Example 9 Sketch the graph of f1x2 = x4 - 4x3, indicating all relative extrema and inflection points. Solution Note that f1x2 = x4 - 4x3 = x31x - 42, so the two x-intercepts (zeros) are x = 0 and x = 4. We next calculate the first two derivatives. f 1x2 = 4x3 - 12x2 = 4x21x - 32 f 1x2 = 12x2 - 24x = 12x1x - 22. The critical numbers are found where f 1x2 = 0. Since f 1x2 has been factored above, it is easy to see that the critical numbers are x = 0 and x = 3. Let us try to apply the second derivative test in order to classify these points. We see that f 102 = 0 and f 132 = 36 7 0. Therefore, at x = 3, we have a relative minimum. The y-value at this point is f132 = 1324 - 41323 = - 27. However, at x = 0, where f102 = 0, the test fails. However, we can analyze the sign of f 1x2 in order to classify the point (0, 0). The sign analysis of f 1x2 is given in Figure 13.

0 sign of f '( x ) 4 x 2( x - 3) 0 0 3 +

I1 (0, 0) (4, 0)

The sign diagram confirms our conclusion that m13, - 272 is a relative minimum. On the other hand, it also reveals that (0, 0) is neither a relative maximum nor a relative minimum. Nevertheless, the tangent line there is horizontal and this will be indicated in our sketch. Let us analyze the concavity of the curve. The second derivative is zero when x = 0 or 2. The sign diagram is given in Figure 14.

I2 (2,-16)

m(3, -27

CU + sign of f " ( x ) 12 x ( x - 3) CD 0 0 0 2 CU +

I1 (0, 0)

We see that the function has an inflection point at I110, 02, where the tangent line is horizontal. It has a second inflection point at I212, - 162. We begin the sketch in Figure 15. Remember, the curve is decreasing until it reaches m. At the Inflection point I1 its tangent line is horizontal. From I1 to I2 the graph is concave downward, otherwise it is concave upward. The completed sketch is given in Figure 16.

286

* **

Section 3.3

For smooth functions the concavity around a relative maximum is downward, while around a relative minimum it is upward. Remember, by smooth we mean that a graph has no sharp points. That is, its derivative exists everywhere. When this is not the case, it is possible to have any kind of concavity near a relative extremum. See Figure 17 where we illustrate a relative minimum, in the graph on the left, the concavity is downward both to its left and right, while it is concave downward to the left and upward to the right of the relative minimum on the curve on the right.

Figure 17: Concavity Around a Non Smooth Critical Point It also follows that for smooth functions, a point cannot simultaneously be a relative extremum and an inflection point. However, if the point is not smooth, this could occur, see the graph on the right in Figure 17. Example 10 Use the first and second derivatives to sketch the graph of f1x2 = x - 3x2/3. Solution We start with the first derivative. f 1x2 = 1 - 2x-1/3 = 1 2 x1/3 - 2 = . 1/3 x x1/3

The critical points occur when the derivative is zero or when the derivative does not exist. That is, when the numerator or the denominator is zero. The numerator is zero when x1/3 - 2 = 0 whose solution is x = 8. The denominator is zero when x = 0. Sign analysis of the first derivative yields the sign diagram given in Figure 18.

+ sign of f '( x ) ( x 1/3 - 2)/ x 1/3 DNE 0 0 8 +

Figure 18: sign of f 1x2 = 1x1/3 - 22/x1/3 We find the corresponding y-values: f182 = - 4 and f102 = 0. Referring to the sign diagram, we see that we have a relative maximum at M(0, 0) and a relative minimum at m18, - 42. At M(0, 0), where the derivative does not exist the tangent line is vertical. The second derivative is given by f 1x2 = 2 -4/3 2 x = . 3 3x4/3

Note that f 1x2 is positive except at x = 0, where it fails to exist. Therefore, the graph is always concave upward. We first begin the sketch in Figure 19 and give the completed graph in Figure 20. Note that even though M(0,0) is a relative maximum, the graph is concave upward both on its left and right. This happens because of the graph not being smooth at this point.

Section 3.3

* **

287

Note that in the previous example, we can write f1x2 = x - 3x2/3 = x2/31x1/3 - 32, so the x-intercepts (zeros) occur when x = 0 or x1/3 - 3 = 0, this gives x1/3 = 3, cubing both sides, we find x = 27 as the second x-intercept. It should be noted that in applications, the second derivative test is often useful when finding the extrema of a function. It may often be complemented with the Only Critical Point Test, as the next example illustrates. Example 11 The average cost of producing a doll by a toy manufacturing company is given by the equation C1x2 = 2000 + 4x + 25/x. The average cost C1x2 is given in dollars per thousand dolls and the number of dolls, x, is given in thousands. How many dolls should be manufactured if the average cost is to be minimized? Solution The reality of the problem restricts the domain of the cost function. Obviously, the average cost of making no dolls makes no sense, and one cannot manufacture a negative number of dolls. Therefore, we must have x 7 0. Of course, in theory, there is no upper bound to the number of dolls we can build. Hence, the domain is the unbounded open interval, x 7 0. Finding the derivative, we obtain C 1x2 = 4 - 25/x2. We see that we need not worry about the denominator vanishing, since x Z 0. Setting C 1x2 = 0, we have that 4x2 - 25 = 0, 4x2 = 25 x = 25/4; x = 5/2.

2

M(0, 0)

m(8, -4)

M(0, 0)

m(8, -4)

That is, x = 2.5 thousand dolls. (Note, the negative square root is not considered since the domain is x 7 0.) The only critical point of the function is at x = 2.5. To prove it is the minimum, we need only show it is a relative minimum. We find C 1x2 = 50/x3, giving C 12.52 7 0. Thus, x = 2.5 gives a relative minimum, and by the Only Critical Point Test, the minimum.

In Sections 1.7 and 2.4, we considered rational functions. There we sketched the graph by examining the zeros, horizontal and vertical asymptotes of the function. We can now include examination of the first and second derivative to improve the sketch. We illustrate with an Example. Example 12 Sketch the graph of the function defined by f1x2 =

2x3 . x2 - 1

Solution You should verify that this function has no horizontal asymptote and has vertical asymptotes at x = ; 1. We take the derivative to obtain, f 1x2 = 2x2 1x2 - 32 1x2 - 122

We know that at x = ; 1 the function has vertical asymptotes, so the only critical numbers are x = 0, and x = ; 23. The second derivative is f 1x2 = 4x1x2 + 32 1x2 - 123

It will be zero when x = 0. Sign analysis of the second derivative is given in Figure 21. (Note that we need to consider the zeros of the denominator, as they determine the sign of the second derivative, even though the function has vertical asymptotes at these x-values.)

288

* **

Section 3.3

Figure 21: Sign of f 1x2 = 4x1x 2 + 32/1x 2 - 123 From the second derivative test, it follows that at x = 0, the function has an inflection point (the tangent line is horizontal at this point as well, why?). At x = - 23 the second derivative is negative so the graph is concave downward so we have a relative maximum at, M A - 23, - 3 23 B and at x = 23 the graph is concave upward so there is a relative minimum at m A 23, 3 23 B . The graph is sketched in Figure 22.

x = -1

x=1

I (0, 0) M

2x3 . x - 1

2

The final example of this section uses implicit differentiation as developed in Section 2.8. Example 13 Using the first and second derivatives, sketch the curve whose equation is 4x2 + 9y2 = 36. Solution The permissible x-values may be found by solving for y which gives 1 y = ; 236 - 4x2. Thus, - 3 x 3. 3 We note that the y-intercepts of the curve occur when x = 0. They are y = ; 2. Similarly, the y-intercepts occur when y = 0, they are x = ; 3. This curve is not the graph of a function. (Why?) We find the derivatives implicitly. 8x + 18yy = 0, or y = - 8x/18y = - 4x/9y The derivative is zero when x = 0; That is, at the y-intercepts 10, ; 22. It fails to exist when y = 0, at the x-intercepts 1 ; 3, 02. The x-intercepts occur at the endpoints of the curve. Therefore, they are not relative extrema. To find the second derivative, we again differentiate implicitly, using the quotient rule to find y .

Section 3.3

* **

289

y = -

Simplifying the complex fraction, we obtain, 4 4x2 + 9y2 d y = - c 9 9y3 However, from the equation of the curve, we have that 4x2 + 9y2 = 36. Substituting, we see that y = - 41362 919y 2

3

- 16 . 9y3

Note that the second derivative is positive (concave upward) when y 6 0, and it is negative (concave downward) when y 7 0. Thus, M(0, 2) is a relative maximum, and m10, - 22 is a relative minimum. We sketch the curve, first by plotting the critical points in Figure 23, and completing the curve in Figure 24. This curve is, as we saw in Section 1.5, an ellipse.

M(0, 2) M(0, 2)

(3, 0)

m(0, -2)

The second derivative may easily be found using the TI 89, assuming the function was named y1, the syntax is as follows: d(y1(x), x, 2), the 2 indicating the second derivative. To find the zeros of the second derivative (and possible inflection points), we need only enter solve 1d1y11x2, x, 22 = 0, x2. Inflection points may also be obtained from the GRAPH window by using F5, selecting inflection and then choosing lower and upper bounds. However, this only works if you know where to look for the inflection points. It is clear that for higher order polynomials it can be difficult to determine the zeros of the function, derivative and second derivative, or for other kinds of functions, the differentiations may be messy as well. It is precisely in these cases that the calculator is a useful tool in providing a qualitative sketch of a graph. A qualitative sketch is needed when given a function whose relative extrema, inflection, or zeros are outside or not visible in the usual window. Consider the following example.

Calculator Tips

290

* **

Section 3.3

Example 14 Sketching the graph of the function defined by 1 y11x2 = 12x6 - 24x5 - 825x4 + 6000x32. Show, on your sketch, all zeros, rela100000 tive extrema and inflection points. Solution The entire graph of this function is not viewable in any reasonable window, so we cannot expect to get anything reasonable on our calculator. However, we can find the zeros of the function, first and second derivative using solve 1y11x2 = 0, x2, solve 1d1y11x2, x2 = 0, x2 and solve 1d1y11x2, x,22 = 0, x2 respectively. For the xvalues found, we use the calculator to compute the corresponding y-values. It is an easy matter, to verify which zeros of the first derivative are indeed relative maxima and minima, and which zeros of the second derivative are inflection points. We will leave these to you as an exercise. After doing so, we can provide a qualitative sketch of the function as given in Figure 25. The breaks in the axes indicate the graph is not drawn to scale.

Figure 25: A Qualitative Sketch of 1 y11x2 = 12x6 - 24x5 - 825x4 + 6000x32 100000

1. f1x2 = x4, find f 1x2, f 1x2, f1321x2 and f1421x2. What do you observe for f1n21x2 when n 5? 2. Find f 1x2 if: (a) f1x2 = x + 3x + 5; (b) f1x2 = 2x + 3x + 1. 3. (a) Find f1321x2 if f1x2 = 1/x2; (b) Find s 1t2 if s1t2 = 21 - t. d2y (c) Find 2 if y = 3x2 - 7x + 9 dx 1x2 + 122 4. Find f 1x2 for: (a) f1x2 = (b) f1x2 = 1x2 + 1210 1x2 - 123

5 2 4 2

5. Find f 1x2 for f1x2 = 1x - 12/13x + 12. 6. f1x2 = 1x3 - 123, find (a) f 102 (b) f 102 (c) f132102 (d) f142102 (e) f1n2102 for n 7 9. 7. If f1x2 = a4x4 + a3x3 + a2x2 + a1x1 + a0. Find the first five derivatives of f. What happens for the sixth and higher order derivatives? 8. If f1x2 = a5x5 + a4x4 + a3x 3 + a2x 2 + a 1x + a0, find f1n21x2, where n is any positive integer. d2y 9. Find 2 , if x1/2 + y1/2 = 6. dx

Section 3.3

10. Find 11. Find 12. Find 13. Find d2y dx d2y

2

* **

291

2

I3 I4 I2 I1

dx2

14. The position of an object at any time is given by the equation s1t2 = 3t3 - 4t + 11. Find the equations of its velocity and acceleration. 15. The position of an object at any time is given by the equation s1t2 = 5t3 - 8t + 21. Find the equations of its velocity and acceleration. 16. Suppose that the position of a particle as a function of time is given by the formula s1t2 = 4t2 - t4 for t 7 0. (a) Find the velocity and acceleration as functions of time. (b) Find the time at which the velocity is zero and the time at which the acceleration is zero. (c) Find the time at which the velocity is maximum. (d) Find the time at which the acceleration is maximum. 17. Repeat Exercise 16 for s1t2 = 5t3 - 3t5, t 7 0. 18. (a) Draw the graph of a function which is increasing on (0, 2) and such that it is concave upward on (0, 1) and concave downward on (1, 2). (b) Does this function have an inflection point? 19. (a) Draw the graph of a continuous function increasing and concave downward on - 1 x 6 2 and decreasing and concave upward on 2 6 x 5. (b) Can such a function have any relative extrema or inflection points? (c) Suppose you have located the relative maximum, M, the relative minimum m, and the inflection points I1, I2, I3, I4, and I5. Draw the smooth graph through these points. (See Figure 26)

21. Sketch the graph of a function having a critical point that is simultaneously a relative extremum and an inflection point. In Exercises 22 39 determine where the function is concave upward and downward, and list all inflection points. 22. f1x2 = 4x2 23. f1x2 = - 3x2 + 2 24. f1x2 = ax2 + bx + c: (a) a 7 0; (b) a 6 0 25. f1x2 = 4x3 - 6x2 + 2x + 1 26. f1x2 = - 2x3 + 5x2 + 7 27. f1x2 = x4 - 6x2 + 9x - 2 28. f1x2 = 3x5 + 5x4 - 20x3 + 10x + 30 29. f1x2 = 1x - 12/1x + 12 30. f1x2 = 2x + 1. 31. f1x2 = x 2x - 4 3x 32. f1x2 = . 4x2 + 9 33. f1x2 = 4x2 x - 1

2

M I4 I3 I5 I2 I1 m

20. As in the preceding exercise, sketch the graph of the function passing through the relative maximum, M, the relative minimum, m and the inflection points I1, I2, I3, and I4, as given in Figure 27. Assume that the derivative fails to exist at M, but f 1x2 and f 1x2 exist everywhere else.

36. f1x2 = 29 + x2. 37. f1x2 = x4/5 - 2x3/5 38. f1x2 = 1 - x3/4. 39. f1x2 = x - x2/3 In Exercises 40 44 use the second derivative test to classify the critical points. 40. f1x2 = x2 - 2x + 3 41. f1x2 = 4x3 - 13x2 + 12x + 9 42. f1x2 = 3x4 + 16x3 + 6x2 - 72x + 20 43. f1x2 = 1x - 1221x + 322 44. f1x2 = 3x/14x2 + 92

292

* **

Section 3.3

68. Sketch a function satisfying the sign diagrams given in Figures 29a and 29b.

In Exercises 45 65 use the first and second derivatives to sketch the graph of the given equation. Also include the intercepts, whenever they are easily determined. 45. f1x2 = x2 - 5x + 6 46. f1x2 = 3x - 6x + 6 47. f1x2 = x3 - 12x 48. f1x2 = x3 + x + 2 49. f1x2 = x - 6x + 8x 50. f1x2 = x3 + 3x2 - 9x - 15 51. f1x2 = x4 - 24x2 52. f1x2 = x4 - 24x2 + 25 53. f1x2 = - x4 + 6x3 - 12x2 + 10 54. f1x2 = 5x - 3x 56. f1x2 = 1x - 42

4 5 3 3 2 2

+ sign of f '( x )

DNE 1

0 3

Figure 29a: Ex 68

+ sign of f " x) DNE 1 0 2 +

Figure 29b: Ex 68

69. Suppose the graph of f 1x2 versus x is given in Figure 30, sketch a possible graph of f(x) versus x. Include concavity in your sketch.

55. f1x2 = 1*41x4 - 4x3 + 12x2 - 8x + 112 57. f1x2 = 1*4x4 - 2x2 58. f1x2 = 3x/14x2 + 92 59. f1x2 = x 2x + 4 60. f1x2 = 6x - 5x 61. f1x2 = 5 + x2/3 62. f1x2 = x4/5 - 2x3/5 63. f1x2 = x - x2/3 64. x2 + 4y2 = 16 65. 4y2 - x2 = 16 66. Draw the graph of a continuous function which is concave downward on 1 - 1, 72, decreasing on 1 - 1, 52 and increasing on (5,7). 67. Sketch the graph of a function satisfying the following sign diagrams (Figure 28a and 28b.).

5 6

f ' (x)

70. Suppose the graph of f 1x2 versus x is given in Figure 31, sketch a possible graph of f(x) versus x. Include concavity in your sketch.

f '(x)

0 sign of f '( x ) -2

0 4

Figure 28a: Ex 67

additionally, f1 - 42 = 3, f1 - 22 = - 2, f1 - 12 = 1, f102 = 5, f122 = 7, f142 = 9, and f152 = 2.

0 sign of f " (x ) -4

0 -1

0 0

0 2

0 5

Figure 28b: Ex 67

71. Show that the second derivative test fails for: (a) f1x2 = 1x - 223; (b) f1x2 = 1x - 224, (c) f1x2 = - 1x - 224; (d) f1x2 = 1x - 1231x + 324 In each case, classify the critical point(s). 72. Suppose that f 1c2 = f 1c2 = 0, but f 1x2 is negative (positive) just to the left and right of x = c. Show that the function has a relative maximum (minimum) at x = c. Use this result to classify the critical points in Exercise 71.

Section 3.4

73. Assuming each limit exists, compute them. f 1x + h2 - f 1x2 (a) lim h:0 h f 1x + h2 - f 1x - h2 (b) lim h:0 h 74. Show that - x4 + 6x3 - 12x2 + 10 0 (Hint: Exercise 52.) 75. Determine the number of real roots of the equation x3 + 3x2 - 9x - 15 = 0. 76. (a) What condition must be imposed upon a, b, c and d such that the graph of f1x2 = ax3 + bx2 + cx + d will have both a relative maximum and relative minimum? (b) Show that, when this condition is satisfied, the coordinates of the inflection point are the averages of the corresponding coordinates of the relative extrema.

* **

293

77. Show that a smooth odd function passing through the origin has an inflection point there. 78. Show that any smooth even function has a relative extremum at its y-intercept. f1k2102 79. Referring to Exercise 7, show that ak = where the symbol k! # k! = k1k - 121k - 22 2 1. Thus, 4! = 4132122112 = 24. The symbol 0! is defined to be equal to 1. The symbol ! is read factorial. 80. If f1x2 = 11 + x25 = a0 + a1x + a2x2 + a3x3 + a4x4 + a5x5, use the preceding exercise to find the values for the ak s. 81. Show that y = - ac(m - 1)x m - 2 b 2 y 2m - 1 if ax m + bym = c.

Problems

Area and Perimeter Optimization Procedure Volume Distance and Velocity Calculator Tips

Given any function defined by the equation of the form y = f1x2, we wish to determine its maximum or minimum value, if they exist. Sometimes there will be an additional constraint on the possible values of the variables. Sometimes there will not be. We shall see that the solution to such optimization problems requires an examination of the derivatives of the function. We have found the solutions to several such problems in the previous two sections. The object of this section is to look at different types of problems which may be solved using those procedures developed in the last few sections. We shall call the function to be optimized the objective function. Sometimes, the function to be optimized involves more than one variable. In such cases, the methods of this chapter can only be used if you can determine equations, called constraint equations, that relate the variables so that the objective function is of the form y = f1x2, a function of a single variable. We begin with two simple examples which illustrate the basic procedure. Example 1 is a case in which there are no constraints. Example 2 illustrates a situation that involves two variables and one constraint. We shall see that many optimization problems may be reduced to problems of the form given in Example 1 or 2. Example 1 Find the minimum value of f1x2 = 3x2 - 12x + 17. Solution As we have learned previously, one good place to look for optimum values is at critical points. Therefore, we find the derivative, f 1x2 = 6x - 12. The only critical point is at x = 2. Calculating the second derivative, we get f 1x2 = 6. Since this is a positive number, by the second derivative test, the function has a relative minimum which

294

* **

Section 3.4

is its minimum (by the Only Critical Point Test) at x = 2. The minimum value is f122 = 5.

Note that we did not need calculus to solve the last example. Clearly a parabola opening upward has its minimum at its vertex. Example 2 Minimize D = 1x - 1622 + A y - 1*2 B 2 subject to the condition that y = x2. Solution Our objective is to minimize D. The equation y = x2 is a constraint, it tells us what the relationship between x and y is. Since the objective function contains two variables, x and y, we use the constraint equation to eliminate one of them. Substituting for y from the constraint equations, we obtain, D = 1x - 1622 + A x2 - 1*2 B 2 We now differentiate dD = 21x - 162 + 2 A x2 - 1*2 B 2x = 4x3 - 32 = 41x3 - 82. dx The only critical point occurs when x3 - 8 = 0, or x = 2. Applying the second derivative test to classify the critical point, we have, d2D = 12x2 7 0 if x = 2. dx2 Thus, D has a relative minimum at x = 2 and by the Only Critical Point Test, the minimum occurs when x = 2. From the constraint equation, y = x2, we have y = 4, and the minimum is D = 12 - 1622 + A 4 - 1*2 B 2 = 208.25.

Example 2 illustrates a very important principle. When the objective function involves two variables, we try to find a constraint equation, which we can use to eliminate one of the variables. One other remark. Suppose, in the preceding example you were asked to minimize 2 2 s = 31x - 1622 + A y - 1 2 B . Notice that S = D, and that the minimum of s will occur at the same place that D has its minimum, namely, at (2, 4). However the minimum value of S = 2208.25 L 14.43. Let us now look at some geometric problems. Area and Perimeter Example 3 A rancher has available 1600 feet of fencing to construct a rectangular corral. The corral is to be subdivided by two fences parallel to the sides. Determine the dimensions of the corral if its total area is to be maximized. Solution When looking at any geometric problem, you should usually do two things: First, sketch a figure illustrating the problem, and; Second, identify all the variables. Let x be the width of the corral and y its length (see Figure 1). Our objective is to determine x and y so that the total area is maximized.

Section 3.4 y

* **

295

Figure 1:

To achieve this objective we must maximize the total area, A = xy. The objective function has two variables. To eliminate one of them, we must find a relationship between them. The relationship is precisely the constraint equation. The rancher has 1600 feet of fencing available for the corral. Therefore, we have that 4x + 2y = 1600. Solving this constraint equation for y, we obtain y = 800 - 2x Thus, A = xy = x1800 - 2x2 = 800x - 2x2 A = 800 - 4x and we have the only critical point at x = 200. Using the second derivative test, we have A = - 4 6 0, implying that A achieves its maximum value at x = 200. Substituting for x in the constraint equation, we have, y = 800 - 212002 = 400. Therefore, to maximize the total area, the rancher should construct a corral with length 400 feet and width 200 feet. The previous examples suggests a general procedure that may be applied to many optimization problems.

OPTIMIZATION PROCEDURE

1. Define the variable or variables to be determined, when appropriate, sketch a figure illustrating the problem and the variables. 2. Determine the objective function to be maximized or minimized 3. When appropriate, determine the constraint equations. 4. Use the constraint equation and substitute into the objective function reducing it to a function of one variable. 5. Find the critical points of the objective function. 6. Justify your conclusion Example 4 A manufacturer wishes to produce a (rectangular) box with an open top with maximum volume from rectangular sheets 8 inches by 15 inches. The box will be constructed by cutting out an equal square from each corner, and then turning up the sides to form the box. Determine the dimensions of the box. Optimization Procedure

Volumes

296

* **

Section 3.4

Solution 1. Let x be the width of each square corner. Figure 2a shows the rectangular sheet with the square corner to be removed drawn dashed. Figure 2b shows the sheet with the corners cut out. The dotted lines in Figure 2b indicate where the sheet is to be folded to construct the box. Note that since x inches are cut from each corner, the width is now 8 - 2x, and the length is 15 - 2x. Also observe that a constraint is imposed upon x. That is, since the smaller of the two dimensions is 8, x must be less than 4, otherwise there is no material left to construct the sides of the box. This constraint dictates the domain of the function to be optimized. Folding along the dotted lines gives the completed box as illustrated in Figure 3. Note that the height of the box is x. The volume of a (rectangular) box is the product of the height, length and width. 1. 8 8 - 2x

x x x x x x x x

15 15 - 2 x

x x x

x x

x x

Figure 2a

8 - 2x

Figure 2b

15 - 2 x

Figure 3: The Box 2. 3. 4. 5. The problem reduces to maximizing V = LWH = x115 - 2x218 - 2x2. The constraint is 0 6 x 6 4. This step is not necessary as the objective function contains only one variable. Rather than apply the product rule we multiply the factors obtaining V = 4x3 - 46x2 + 120x Differentiating, V1x2 = 12x2 - 92x + 120. Setting V 1x2 = 0, we look for critical points. V 1x2 = 413x2 - 23x + 302 = 0 413x - 521x - 62 = 0 x = 5/3 or x = 6

Section 3.4

* **

297

Since x must be less than 4, the only critical point in the domain is at x = 5/3. 6. V 1x2 = 24x - 92, and V 15/32 = - 52. Thus, the volume is maximized when x = 5/3 inches. That is, we must cut out a square of side 5/3 inches from each corner. The resulting box will have height x = 5/3 inches, width 8 - 2x = 14/3 inches, and 14 35 2450 3 length 15 - 2x = 35/3 inches, and the volume = A 5 3 B A 3 B A 3 B = 27 in . The next example is easy to set up, but its solution requires some algebraic manipulation, as we shall see. Example 5 Towns A and B are located 3 and 6 miles respectively, from a straight river. Let C and D be the points on the river bank closest to A and B, respectively, as shown in Figure 4. The distance between points C and D is 9 miles. The two towns are to be supplied water from a pumping station situated on the river between them. Where on the river should the pumping station be located in order to minimize the length of piping used in connecting the towns? Solution 1. Let x be the distance along the river from town B to where the pumping station, P, is x 9. located (see Figure 5). Note that 0

C 3

River 9

C 3

9-x

6 A A B

Figure 4

Figure 5

2. By the Pythagorean Theorem, the distance from P to B is 2x2 + 62 = 2x2 + 36, and the distance from A to P is 232 + 19 - x22 = 2x2 - 18x + 90. Therefore, we must minimize L = 2x2 + 36 + 2x2 - 18x + 90 3. We have that 0 x (Step 4 is not necessary.) 5. Differentiating, 9.

L 1x2 = 1*21x2 + 362-1/22x + 1*21x2 - 18x + 902-1/212x - 182. Simplifying, and setting L 1x2 to zero, we have, L 1x2 = To solve for x, we write x 2x + 36

2

x 2x + 36

2

x - 9 2x - 18x + 90 9 - x

2

= 0

2x - 18x + 90

298

* **

Section 3.4

2

Clearing fractions (cross multiplying) we find, x4 - 18x3 + 90x2 = 81x2 - 18x3 + x4 + 2916 - 648x + 36x2. Thus, 27x2 - 648x + 2916 = 0 Dividing by 27 x2 - 24x + 108 = 0, or 1x - 621x - 182 = 0. Since x 9, x = 6 is the only critical point in the given domain. 6. To classify the critical point it is easiest to apply the first derivative test. A computation gives, L 15.92 = 0.701139952 - 0.71860214 6 0 and L 16.12 = 0.712926292 + 0.695022096 7 0 Thus, the sign of L 1x2 changes from negative to positive at x = 6, implying that at the only critical point L is a minimum. Therefore, the pumping station should be 6 miles from D and 3 miles from C.

(Note that we could have also determined the minimum using the Extreme Value Theorem, how?) It is interesting to observe that this problem may be solved without the use of calculus. Exercise 29 indicates a geometric method. Example 6 A ball is thrown vertically upward from the ledge of a building 192 feet high with an initial velocity of 96 feet per second. How high will the ball go? Assume that the height of the ball as a function of time is s1t2 = - 16t2 + 96t + 192. Solution Since we have s as a function of t already, we begin at step 5. The derivative is s 1t2 = - 32t + 96. Setting this to zero, we find that the critical point is at t = 3. 6. s 132 6 0, thus the critical point is a relative maximum, and by the Only Critical Point Test, a maximum. The maximum height of the ball is s132 = - 161322 + 96132 + 192 = 336 feet above the ground.

In Section 2.7 we solved similar problems by setting the velocity v = 0. From a physical point of view, the ball will reach its maximum height at the instant when its velocity is zero. Thus, the physical and mathematical explanations are in complete agreement. Example 7 A power station is on one side of a straight river which is five miles wide, and a factory is on the other side, 20 miles up-river. A power line is to be run from the power station, under the river to some point A, which is x miles up-river from P, and then over land to the

Section 3.4

* **

299

factory (Figure 6). If it costs $80 per mile to run the line under water and $50 per mile to run it overland, find x, the point on the other side of the river where the power line comes out of the river, if the total cost of the power line is to be as small as possible. Solution

P

Calculator Tips

Figure 6 If we let y1 represent the cost, then the problem is to minimize y11x2 = 80 2x2 + 25 + 50120 - x2 x 0 and x 20 This can be done fairly easily with pencil and paper, however, the calculations are considerably reduced using the calculator. By the Extreme Value Theorem, this function is continuous on a closed interval, therefore we need only compute the y-values at the two endpoint and critical points to determine the minimum cost. y1102 = $ 1400, y11202 = $ 1649.24. The critical point turns out to occur at x = (Verify!), and the cost at this value is approximately $1312.25.

25 239 39

L 4.0032 miles

Example 8 Suppose the power station in the previous example is moved one mile inland, now determine the point on the other side of the river where the power line is to come out of the river. Solution Figure 7 illustrates the problem to be solved.

P 1 x A 20 - x 5 F

Figure 7: The Power Station One Mile Inland The problem is then to minimize the cost function y21x2 = 50 2x2 + 1 + 80 225 + 120 - x22 x 0 and x 20

This problem is very difficult to do without a calculator. To determine when the derivative is zero requires solving a fourth degree equation. However, with a calculator, the solution is straight-forward and left as an exercise for you.

300

* **

Section 3.4

1. Find the dimensions of the largest rectangle whose perimeter is 144 feet. 2. Show that the rectangle of maximum area with fixed perimeter is a square. 3. The sum of two positive numbers is 100. what should the numbers be if their product is to be as large as possible? 4. Show that the answer to Exercise 3 follows from Exercise 2. 5. A farmer wishes to build a rectangular pig pen using as little fencing as possible. If the area of the pen is to be 225 square feet, what are its dimensions? 6. Show that the rectangle with fixed area and minimum perimeter is a square. 7. (a) A farmer has a field bordering a straight river. He wants to design a rectangular plot for an organic crop with the river acting as a natural border. If 400 feet of fencing will be used on the three other sides, what should the dimensions of the plot be if its area is to be maximized? (b) What are the dimensions if the river side is also to be fenced? 8. Consider the farmer in Exercise 7 whose field borders the river. Assume he wants the area of the plot to be 800 square feet. What should its dimensions be if he wants to use the least amount of fencing? 9. 180,000 square feet of land is to be enclosed in a rectangular plot which will then be subdivided into three plots by a pair of fences both parallel to the sides. What is the least amount of fencing needed? (see Figure 8.) 11. Minimize the amount of fencing needed to construct the region shown in Figure 10, if the total area is 480 square meters.

y /2 x /2 x

12. A Norman window is one in which the window is constructed by capping a rectangular region with a semicircular region (see Figure 11). If the total perimeter of the window is to be 30 feet, find the dimensions that maximizes its area.

Figure 8: Ex. 9

10. A rancher has divided a plot of ground into an L shaped region. The larger section to serve as a corral and the smaller as a training area (see Figure 9). If 480 feet of fencing is available, determine the dimensions x and y which maximizes the total area.

13. An aquarium is to be made in the shape of a rectangular solid with a square base and an open top. The volume of the aquarium is to be 108 cubic inches. What dimensions will minimize the amount of material needed to build it? 14. A holding pen for fish is to be made in the form of a rectangular solid with a square base and open top. The base will be slate that costs $4 per square foot and the sides will be glass that costs $5 per square foot. If the volume of the tank must be 50 cubic feet, what dimensions will minimize the cost of construction? 15. (a) A wholesale manufacturer of canned corn wants a tin can (a right circular cylinder) that will have a volume of 54 cubic inches. As little tin as possible as possible is to be used in the construction of the can. Determine its height and diameter. (b) Measure a can of corn from your local supermarket. Does its dimensions yield a minimum? (c) Why do you think manufacturers choose not to use this type of design?

2y x

16. Consider the can in the previous exercise. Suppose the tin for the top and bottom of the can cost 8 cents per square inch, while the tin for the remainder of the can costs 1 cent per square inch. What should the diameter of the can be if the cost of the can is to be minimized.

Figure 9: Ex. 10

17. A 12 centimeter square sheet of cardboard is to be made into a box by cutting equal squares from each corner and then folding up its sides. Determine what the side of the square cut from the corners should be if the box is to have as large a volume as possible.

Section 3.4

18. A right triangle is formed in the first quadrant by a line passing through the point (3, 5) and the coordinate axes. Find the coordinates of its vertices if its area is to be minimized. 19. A rectangular page is to contain 96 square inches of print. The top and bottom margins are each 1 inch wide, and the margins on each side are 1.5 inches. What should the dimensions of the page be if the least amount of material is to be used? 20. Suppose postal requirements are that the maximum of the length plus girth (cross sectional perimeter) of a rectangular package that may be sent is 144 inches. Find the dimensions of such a package with square ends whose volume is to be a maximum. 21. (a) John is at B, on a straight beach, 10 miles from A. Mary is in a boat in the sea at C, 4 miles from the beach (see Figure 12). Assuming Mary can row at 3 miles per hour and jog at 5 miles per hour, where along the beach should she land so that she may get to John in the least amount of time? Solve the problem if John is (b) 100 miles; (c) 1 mile from A.

* **

301

29. A low stone wall 27 inches high is 64 inches away from a building. What is the length of the shortest pole that passes over the fence and reaches the building? (Hint: The length of the pole is composed of two portions, the portion from the ground to the fence, and from the fence to the building. See Figure 13. Use the theorem of Pythagoras along with similar triangles to find the total length.)

Pole Building 27 64

A 10 B

30. Show that the solution to Example 5 may be found as follows: reflect A to the other side of the river to A as shown in Figure 14. Draw a line connecting A to B. Where this line intersects the horizontal line is the required point P, the location of the pumping station. (Hint: Assume that some other point gives the minimum length. Deduce a contradiction by using the fact that the sum of the lengths of two sides of a triangle exceeds the length of the third side.)

22. Find the point on the line 3x + 4y + 7 = 0 closest to the point 11, - 22. 23. Find the points on the curve whose equation is x + y = 16 nearest and farthest from (a) (8, 0) (b) (6, 2) (c) (2, 3). Hint: The computation may be easier if you try to minimize d2 rather than d. 24. Find the point on the curve y = 1x closest to (3/2, 0). 25. Find the points on the circle x + y = 9 (a) closest to and (b) farthest from the point (8, 12). 26. Find the dimensions of the largest rectangle with lower base on the x-axis and upper vertices on the curve whose equation is (a) x2 + y2 = 4 (b) 9x2 + 4y2 = 36. 27. A 12 inch piece of wire is to be cut into two pieces. One piece is to be used to form a square and the other to form a circle. How should the wire be cut if the sum of the areas is to be maximized? 28. (a) Show that f1x2 = 1x - a122 + 1x - a222 has its minimum value when x = 1a1 + a22/2. (b) Show that f1x2 = 1x - a122 + 1x - a222 + 1x - a322 has its minimum value when x = 1a1 + a2 + a32/3. (c) Generalize the above.

2 2 2 2

A' P

A B

31. Complete the solution to Exercise 8. Make sure to justify your conclusions.

302

* **

Section 3.5

Optimization Problems

Price, Demand and Revenue Cost and Average Cost Elasticity of Demand

In this section we shall apply the same methods used in the previous section. The only difference will be that the examples to be considered come from economics and business. We shall make some observations which follow directly from our knowledge of the derivative that are usually considered theorems in micro-economics. Our first example shall deal with the revenue available to a monopolist. That is, we consider a market situation in which there is no competition. Hence, the quantity and sales price of a commodity is not the market equilibrium point determined by supply and demand curves. Instead, we have only a demand curve, and the price is set by the monopolist producer. Example 1 Suppose that the price and demand for a particular luxury automobile are related by the demand equation p + 10x = 200,000; Where p is the price per car in dollars and x is the number of cars that will be purchased at that price. What price should be charged per car if the total revenue is to be maximized? Solution We follow the steps in the procedure indicated in the previous section. 1. We recall that the total revenue is defined as the product of the price per item times the number of items sold. That is, R = xp. Thus, we wish to maximize R 2 and 3. A sketch is not necessary here. However, if we sketch the demand equation we