Академический Документы
Профессиональный Документы
Культура Документы
The concept of Derivative is at the core of Calculus and modern mathematics. The
definition of the derivative can be approached in two different ways. One is
geometrical (as a slope of a curve) and the other one is physical (as a rate of
change). Historically there was (and maybe still is) a fight between mathematicians
which of the two illustrates the concept of the derivative best and which one is
more useful. We will not dwell on this and will introduce both concepts. Our
emphasis will be on the use of the derivative as a tool.
If we now assume that A and B are very close to each other, we get close to what is
called the instantaneous velocity. Of course, if A and B are close to each other,
then the time it takes to travel from A to B will also be small. Indeed, assume that
at time t=a, we are at A. If the time elapsed to get to B is
, then we will be at B
at time
is
. If
. In this case,
Example. Consider a parabolic motion given by the function f(t) = t2. The
instantaneous velocity at t=a is given by
Since
, where
Notation. Now we get to the hardest part. Since we can not keep on writing
"Instantaneous Velocity" while doing computations, we need to come up with a
suitable notation for it. If we write dxfor
small, then we can use the notation
draw the graph in a plane with a horizontal axis (usually called the x-axis) and a
vertical axis (usually called the y-axis).
Fix a point on the graph, say (x0, f(x0)). If the graph as a geometric figure is "nice"
(i.e. smooth) around this point, it is natural to ask whether one can find the
equation of the straight line "touching" the graph at that point. Such a straight line
is called the tangent line at the point in question. The concept of tangent may be
viewed in a more general framework.
(Note that the tangent line may not exist. We will discuss this case later on.) One
way to find the tangent line is to consider points (x,f(x)) on the graph, where x is
very close to x0. Then draw the straight-line joining both points (see the picture
below):
As you can see, when x get closer and closer to x0, the lines get closer and closer to
the tangent line. Since all these lines pass through the point (x0,f(x0)), their
equations will be determined by finding their slope: The slope of the line passing
through the points (x0,f(x0)) and (x,f(x)) (where
) is given by
The tangent itself will have a slope m, which is very close to m(x) when x itself is
very close to x0. This is the concept of limit once again!
In other words, we have
Notation. Writing "m" for the slope of the tangent line does not carry enough
information; we want to keep track of the function f(x) and the point x0 in our
notation. The common notation used is
m = f'(x0).
In this case, the equation of the tangent line becomes
y - f(x0) = f'(x0) (x-x0)
where
One last remark: Sometimes it is more convenient to compute limits when the
variable approaches 0. One way to do that is to make a translation along the x-axis.
Indeed, if we set h=x-x0, we get
Techniques of Differentiation
Maybe the easiest and most useful formulas are the ones that say that the derivative
is linear:
Combined with the formula (xn)' = n xn-1, we see that every polynomial function has
a derivative at any point.
Example. For P(x) = 1-2x + 3x4 -5 x6, we have
The next two formulas are the most powerful ones. They deal with the derivative
of a product and a quotient. They are commonly called the product rule and
the quotient rule. We have
In particular, we have
So, we have
which means that the formula (xr)' = r xr-1 is also valid for negative exponents.
Before we discuss the derivative of trigonometric functions, let us stop here and
reflect a little bit more on polynomial functions. Indeed, we saw that the derivative
of a polynomial function is also a polynomial function. So we can take another
derivative and generate a new function. This function is called the second
derivative. We can keep doing this as long as we want to. The functions obtained
are called higher derivatives. The common notations used for them are
which gives
The answer to this problem came through a very nice idea. Indeed, let us split the
region
into small subregions which we can approximate by rectangles or other
simple geometrical figures (whose areas we know how to compute). This is how it
goes: split the interval [a, b] into subintervals, preferably with the same width x,
x=
, for i = 0, 1, ... , n - 1.
Let
be the subregion bounded above by the graph of f (x), bounded below by
the x-axis, bounded to the left by x = xi - 1, and to the right by x = xi, for i = 1, ... , n.
Clearly we have
Area(
) = Area(
) + Area(
) + ... + Area(
) .
Let Li be the rectangle defined by the left-end point and Ri be the rectangle defined
by the right-end point. Then an approximation to Area( ) is given by
Area(L1) + Area(L2) + ... + Area(Ln) = xf (x0) + xf (x1) + ... +
which we will call the left-sum denoted LEFT(n), and
Area(R1) + Area(R2) + ... + Area(Rn) = xf (x1) +
which we will call the right-sum denoted RIGHT(n)
Example. Consider the function
f (x) = x2
for x
xf (x2) + ... +
xf (xn - 1)
xf (xn)
We have
x=
, x2 =
, x3 =
, x4 = 1 .
and
LEFT(4) =
02 +
and
RIGHT(4) =
12 =
Note that Area( ) equals , a result which we will prove in later pages.
Indeed if the function f (x) is not too badly behaved, we will show that when n gets
larger, the numbers LEFT(n) and RIGHT(n) get closer to Area( ), i.e.
Area( ) =
LEFT(n) =
RIGHT(n) .
This is the main idea described above. The number Area( ) is called the definite
integral (or more simply the integral) of f (x) from a to b and is denoted by
f (x) dx .
0. Then we have
dx = (b - a) .
is simply a rectangle.
x dx = (b2 - a2) .
is simply the union of two regions: one rectangle and one
The rectangle (depicted in red) is bounded above by x = a and its area is a(b - a).
The triangle (in blue) is determined by the points: (a, a), (a, b), and (b, b). Its area
is
(b - a)2. So we have
Theorem. We have
Lf(P)
Area( )
Uf(P)
for any partition P of [a, b]. Moreover if f (x) is continuous on [a, b], except maybe
at a finite number of points, and I is a number such that
Lf(P)
I
Uf(P)
for any partition P of [a, b], then I = Area( ).
This theorem is fundamental. Let us illustrate this with the following example.
Example. Use the above theorem to show
x2dx =
(b3 - a3)
where b
a
0. Let P = {x0, x1, ... , xn} be a partition of [a, b]. Since f (x) = x2 is
increasing on [a, b], then mi = xi - 12 and Mi = xi2. So we have
Uf(P) = x21(x1 - x0) + x22(x2 - x1) + ... + x2n(xn - xn - 1)
and
Lf(P) = x20(x1 - x0) + x21(x2 - x1) + ... + x2n - 1(xn - xn - 1) .
For each i, we have
xi - 12
since xi - 1
xi - 12 + xi - 1xi + xi2
xi 2
xi2(xi - xi - 1) .
But
xi - 12 + xi - 1xi + xi2 (xi - xi - 1) = xi3 - xi - 13
which implies
xi - 12(xi - xi - 1)
(xi3 - xi - 13)
xi2(xi - xi - 1) .
Hence
Lf(P)
(b3 - a3)
Uf(P)
since
(x13 - x03) + (x23 - x13) + ... + (xn3 - xn - 13) = b3 - a3 .
f (x) dx +
g(x) dx;
(ii)
f (x) dx =
f (x) dx, for any arbitrary number .
The next results are very useful in many problems.
Theorem. If f (x) is defined and continuous on [a, b], except maybe at a finite
number of points, then we have
(i)
f (x) dx = 0;
(ii)
f (x) dx =
f (x) dx +
f (x) dx;
(iii)
f (x) dx = -
f (x) dx;
Remark. It is easy to see from the definition of lower and upper sums that if f (x)
is positive then
If f (x)
f (x) dx
g(x) for x
f (x) dx
g(x) dx .
Example. We have
(x2 - 2x)dx =
We have seen previously that
x2 dx =
x2 dx - 2
and
x dx .
x dx =
Hence
(x2 - 2x)dx =
-2
=-
f (x)dx = - 1 ,
f (x)dx.
We have
f (x)dx =
f (x)dx +
f (x)dx
which implies
f (x)dx =
f (x)dx -
f (x)dx = - 1 - 2 = - 3
Techniques of Integration
Integration by Parts
One of very common mistake students usually do is
To convince yourself that it is a wrong formula, take f(x) = x and g(x)=1. Therefore,
one may wonder what to do in this case. A partial answer is given by what is
called Integration by Parts. In order to understand this technique, recall the
formula
which implies
and
is easy to
evaluate, we can use it to get the other one. This is the main idea behind Integration
by Parts. Let us give the practical steps how to perform this technique:
1
Write the given integral
where you identify the two functions f(x) and g(x). Note that if you are given
only one function, then set the second one to be the constant
function g(x)=1.
2
Introduce the intermediary functions u(x) and v(x) as:
Then you need to make one derivative (of f(x)) and one integration (of g(x))
to get
Note that at this step, you have the choice whether to differentiate f(x)
or g(x). We will discuss this in little more details later.
3
Use the formula
4
Take care of the new integral
The first problem one faces when dealing with this technique is the choice that we
encountered in Step 2. There is no general rule to follow. It is truly a matter of
experience. But we do suggest not to waste time thinking about the best choice,
just go for any choice and do the calculations. In order to appreciate whether your
choice was the best one, go to Step 3: if the new integral (you will be handling) is
easier than the initial one, then your choice was a good one, otherwise go back to
Step 2 and make the switch. It is after many integrals that you will start to have a
feeling for the right choice.
In the above discussion, we only considered indefinite integrals. For the definite
integral
1
Evaluate the indefinite integral
which gives
Then you need to make one derivative (of f(x)) and one integration (of g(x))
to get
(iii)
Use the formula
(iv)
Take care of the new integral
References
http://www.sosmath.com/calculus/calculus.html
http://hyperphysics.phy-astr.gsu.edu/hbase/math/derint.html