00 голосов за00 голосов против

23 просмотров233 стр.Feb 02, 2020

© © All Rights Reserved

DOCX, PDF, TXT или читайте онлайн в Scribd

© All Rights Reserved

23 просмотров

00 голосов за00 голосов против

© All Rights Reserved

Вы находитесь на странице: 1из 233

Real Analysis

Submitted by:

BS MATH 3A

Submitted to:

Dr. Louida P. Patac

0

Table of Contents

Sets and Functions ………………………. 2

Special Types of Function ………………………. 18

Mathematical Induction ………………………. 33

Finite and Infinite Sets ………………………. 48

Countable and Uncountable Sets ………………………. 61

The Algebraic and Order Property of Real Number ………………. 72

Absolute Value and Real Line ………………………. 85

Completeness Property of Real Number ………………………. 96

Applications of Supremum Property ………………………. 108

Intervals ………………………. 118

Sequences and their Limits ………………………. 129

Limit Theorem ………………………. 143

Monotone Sequences ………………………. 153

Subsequences ………………………. 164

The Cauchy Criterion ………………………. 177

Properly Divergent Sequence ………………………. 190

Introduction to Infinite Series ………………………. 201

Limits of Functions ………………………. 213

1

SETS and FUNCTIONS

2

SET

Definition:

A set is a well-defined collection of distinct objects. The objects that

make up a set (also known as the set's elements or members) can be

anything: numbers, people, letters of the alphabet, other sets, and so

on. A set can have any non-negative quantity of elements, ranging from

none (the empty set or null set) to infinitely many. The number of

elements in a set is called the cardinality, and can range from zero to

denumerably infinite (for the sets of natural numbers, integers, or q

rational numbers) to non-denumerably infinite for the sets of irrational

numbers, real numbers, imaginary numbers, or complex numbers).

Set Operations:

The set { x | some property of x } is the set of all objects x that satisfy the given

property. Formally, we have that w ∈ { x | some property of x } if and only if the

specified property holds for w.

The set A ∪ B is the set { x | x ∈ A or x ∈ B }. Equivalently, x ∈ A ∪ B if and only

if x ∈ A or x ∈ B.

The set A ∩ B is the set { x | x ∈ A and x ∈ B }. Equivalently, x ∈ A ∩ B precisely

if x ∈ A and x ∈ B.

The set A – B is the set { x | x ∈ A and x ∉ B }. This set is also sometimes

denoted A \ B.

The set A Δ B is the set { x | exactly one of x ∈ A and x ∈ B is true }.

Examples:

N : the set of natural numbers

W : the set of whole numbers

𝑍 + or I : the set of integers Z + :

𝑍 − : the set of negative integers

Q : the set of rational numbers

R : the set of real numbers

C : the set of complex number

REPRESENTATION OF A SET

(i) Roster method (Tabular form)

In this method a set is represented by listing all its elements, separating

these by commas and enclosing these in curly bracket. If V be the set of

vowels of English alphabet, it can be written in Roster form as : V = { a,

e, i, o, u}.

If A be the set of natural numbers less than 7. then A={1, 2, 3, 4, 5, 6}, is in the

Roster form.

3

Note:

To write a set in Roster form elements are not to be repeated i.e. all

elements are taken as distinct.

For example if A be the set of letters used in the word mathematics, then

A = {m, a, t, h, e, i, c, s}.

Examples:

(a) C = { x : x ∈ 𝑁 𝑎𝑛𝑑 50 ≤ 𝑥 ≤ 60}

(b) D = {x: x ∈ ℝ 𝑎𝑛𝑑 𝑥 2 − 5𝑥 + 6 = 0}

In this form elements of the set are not listed but these are represented by

some common property.

Let V be the set of vowels of English alphabet then V can be written in the

set builder form as: V = {x : x is a vowel of English alphabet}

Let A be the set of natural numbers less than 7. then A = {x : x ∈ 𝑁 and 1

≤x <7}.

CLASSIFICATION OF SETS

Definition:

Let A and B be two sets where A = {x : x is a natural number} B = {x : x

is a student of your school} As it is clear that the number of elements in

set A is not finite (infinite) while number of elements in set B is finite. A

is said to be an infinite set and B is said to be is finite set. A set is said

to be finite if its elements can be counted and it is said to be infinite if it

is not possible to count upto its last element.

Definition:

Consider the following sets. A = {x: x ∈ ℝ 𝑎𝑛𝑑 𝑥 2 + 1 = 0 } B = {x : x is

number which is greater than 7 and less than 5}.Set A consists of real

numbers but there is no real number whose square is -1. Therefore, this

set consists of no element. Similarly, there is no such number which is

less than 5 and greater than 7. Such a set is said to be a null (empty)

set. It is denoted by the symbol void, 𝜙 or {}. A set which has no element

is said to be a null/empty/void set and is denoted by.

Singleton Set

Definition:

Consider the following set: A = {x : x is an even prime number} As there

is only one even prime number namely 2, so set A will have only one

4

element. Such a set is said to be singleton. Here A = {2}. A set which

has only one element is known as singleton.

Definition:

Consider the following examples.

(i) A = {1,2,3}, B = {2,1,3}, (ii) D = {1,2,3} , E = {a,b,c} .

(i) Sets A and B have the same elements. Such sets are said to be equal sets

and it is written as A = B. In example (ii) set D and E have the same number

of elements but elements are different. Such sets are said to be equivalent

sets and are written as A » B. Two sets A and B are said to be equivalent

sets if they have same number of elements, but they are said to be equal if

they have not only the same number of elements, but elements are also the

same.

Definition:

The cardinality of a set is its size. For a finite set, the cardinality of a set is

the number of members it contains. In symbolic notation the size of a set

S is written |S|. We will deal with the idea of the cardinality of an infinite set

later.

Example:

Set cardinality for the set S = {1, 2, 3} we show cardinality by writing |S| = 3.

We now move on to a number of operations on sets. You are already familiar

with several operations on numbers such as addition, multiplication, and

negation.

Definition:

The intersection of two sets S and T is the collection of all objects that are in

both sets. It is written S ∩ T. Using curly brace notation S ∩ T = {x : (x ∈ S) and

(x ∈ T )} The symbol and in the above definition is an example of a Boolean or

logical operation. It is only true when both the propositions it joins are also true.

It has a symbolic equivalent ∧. This lets us write the formal definition of

intersection more compactly: S ∩ T = {x : (x ∈ S) ∧ (x ∈ T )

Example:

Intersections of sets Suppose S = {1, 2, 3, 5}, T = {1, 3, 4, 5}, and U = {2, 3, 4,

5}. Then:

S ∩ T = {1, 3, 5}, S ∩ U = {2, 3, 5}, and T ∩ U = {3, 4, 5}

Disjoint Sets

Definition:

Two sets are said to be disjoint if they do not have any common element. For

example, sets A= { 1,3,5} and B = { 2,4,6 } are disjoint sets.

Sub- Set:

Definition:

5

Let set A be a set containing all students of your school and B be a set

containing all students of class XII of the school. In this example each element

of set B is also an element of set A. Such a set B is said to be subset of the

set A. It is written as B ⊆ A.

Consider:

D ={1, 2, 3, 4,........}

E = {.....-3-2,-1, 0, 1, 2, 3, .......}

Clearly each element of set D is an element of set E also ∴ 𝐷 ⊆ E If A and B

are any two sets such that each element of the set A is an element of the set B

also, then A is said to be a subset of B.

Remarks

(i) Each set is a subset of itself i.e. A ⊆ A .

(ii) Null set has no element so the condition of becoming a subset is

automatically satisfied. Therefore, null set is a subset of every set.

(iii) If A ⊆ B and B⊆ A then A = B.

(iv) If A ⊆ B and A≠ B, then A is said to be a proper subset of B and B is said to

be a super set of A. i.e. A ⊂ B or B ⊃ A.

Examples:

1. If A = {x : x is a prime number less than 5} and B = {y : y is an even prime

number} then is B a proper subset of A ?

Solution : It is given that A = {2, 3 }, B = {2}. Clearly B ⊆ A and B A ¹ We

write B ≠ A and say that B is a proper subset of A.

2. If A = {1, 2, 3, 4}, B = {2, 3, 4, 5}. is A⊆ B or B ⊆ A ?

Solution : Here 1 ∉ A but 1 ∉ B⟹ A⊈ B. Also 5∈ B but 5 ∉ A ⟹ B ⊈ A

. Hence neither A is a subset of B nor B is a subset of A.

3. If A = {a, e, i, o, u} B = {e, i, o, u, a } Is A ⊆ B orB ⊆ A or both ?

Solution : Here in the given sets each element of set A is an element of set

B also ∴ A ⊆ B (i) and each element of set B is an element of set A also. ∴

B⊆ A. (ii) From (i) and (ii) A = B

Power Set

Definition:

Let A = {a, b} Subset of A are f , {a}, {b} and {a, b}. If we consider these subsets

as elements of a new set B (say) then B = {𝜙,{a},{b},{a,b}} B is said to be the

power set of A.

Notation:

Power set of a set A is denoted by P(A). Power set of a set A is the set of all

subsets of the given set.

Examples:

1. A={𝑥: 𝑥 ∈ ℝ 𝑎𝑛𝑑 𝑥 2 + 7 = 0}

2. B={𝑦: 𝑦 ℕ 𝑎𝑛𝑑 1 ≤ 𝑦 ≤ 3}

Solution:

(i) Clearly A = 𝜙 (Null set)

∴ 𝜙 is the only subset of given set

6

∴ P (A)={ 𝜙 }

(ii) The set B can be written as {1, 2, 3} Subsets of B are 𝜙 , {1}, {2}, {3},

{1, 2}, {1, 3}, {2, 3}, {1, 2, 3}. \ P (B) = { 𝜙 , {1}, {2}, {3}, {1, 2}, {1, 3}, {2, 3},

{1, 2, 3} } .

Universal Set

Consider the following sets.

A = {x : x is a student of your school}

B = {y : y is a male student of your school}

C = {z : z is a female student of your school}

D = {a : a is a student of class XII in your school}

Clearly the set B, C, D are all subsets of A.

A can be considered as the universal set for this particular example. Universal

set is generally denoted by U. In a particular problem a set U is said to be a

universal set if all the sets in that problem are subsets of U.

Remarks:

(i) Universal set does not mean a set containing all objects of the universe. (ii)

A set which is a universal set for one problem may not be a universal set for

another problem.

Example:

Which of the following set can be considered as a universal set?

X = {x : x is a real number}

Y = {y : y is a negative integer}

Z = {z : z is a natural number}

Solution:

As it is clear that both sets Y and Z are subset of X. ∴ X is the universal set for

this problem.

VENN DIAGRAM\

Definition:

British mathematician John Venn (1834-1883 AD) introduced the concept of

diagrams to represent sets. According to him universal set is represented by

the interior of a rectangle and other sets are represented by interior of circles.

For example, Let X = {1, 2, 3} and Let Y = {3, 4, 5}. Draw and label a Venn

diagram to show the intersection of sets X and Y.

The Venn Diagram in example 4 makes it easy to see that the number 3 is

common to both sets. So, the intersection of X and Y is 3. This is

what X and Y have in common. The intersection of X and Y is written as X ∩

7

𝑌 and is read as "X intersect Y". So, Intersection means "X and Y". In example

5 below, we will find the union of two sets. The union of two sets is the set

obtained by combining the elements of each.

Definition:

The compliment of a set S is the collection of objects in the universal set that

are not in S. The compliment is written 𝑆 𝑐 . In curly brace notation 𝑆 𝑐 = {x : (x ∈

U) ∧ (x /∈ S)} or more compactly as 𝑆 𝑐 = {x : x /∈ S} however it should be

apparent that the compliment of a set always depends on which universal set

is chosen. There is also a Boolean symbol associated with the

complementation operation: the not operation. The notation for not is ¬. There

is not much savings in space as the definition of compliment becomes 𝑆 𝑐 = {x :

¬(x ∈ S)}.

Example: Set Compliments

(i) Let the universal set be the integers. Then the compliment of the even

integers is the odd integers.

(ii) Let the universal set be {1, 2, 3, 4, 5}, then the compliment of S = {1, 2, 3}

is S c = {4, 5} while the compliment of T = {1, 3, 5} is T c = {2, 4}.

Definition:

The difference of two sets S and T is the collection of objects in S that are not

in T. The difference is written S − T . In curly brace notation S − T = {x : x ∈ (S

∩ (T c ))}, or alternately S − T = {x : (x ∈ S) ∧ (x /∈ T )}

Notice how intersection and complementation can be used together to create

the difference operation and that the definition can be rephrased to use Boolean

operations. There is a set of rules that reduces the number of parenthesis

required.

These are called operator precedence rules.

(i) Other things being equal, operations are performed left-to-right.

(ii) Operations between parenthesis are done first, starting with the innermost

of nested parenthesis.

(iii) All complementation’s are computed next.

(iv) All intersections are done next.

(v) All unions are performed next.

(vi) Tests of set membership and computations, equality or inequality are

performed last. Special operations like the set difference or the symmetric

difference, defined below, are not included in the precedence rules and thus

always use parenthesis.

Example: Operator precedence

Since complementation is done before intersection the symbolic definition of

the difference of sets can be rewritten: S − T = {x : x ∈ S ∩ T c }

If we were to take the set operations A ∪ B ∩ C c and put in the parenthesis we

would get

(A ∪ (B ∩ (C c )))

Definition:

8

The symmetric difference of two sets S and T is the set of objects that are in

one and only one of the sets. The symmetric difference is written S∆T. In curly

brace notation: S∆T = {(S − T ) ∪ (T − S)}

Example:

Symmetric differences Let S be the set of non-negative multiples of two that are

no more than twenty four.

Let T be the nonnegative multiples of three that are no more than twenty four.

Then S∆T = {2, 3, 4, 8, 9, 10, 14, 15, 16, 20, 21, 22}.

Another way to think about this is that we need numbers that are positive

multiples of 2 or 3 (but not both) that are no more than 24.

Another important tool for working with sets is the ability to compare them. We

have already defined what it means for two sets to be equal, and so by

implication what it means for them to be unequal. We now define another

comparator for sets.

Definition: For two sets S and T we say that S is a subset of T if each element

of S is also an element of T . In formal notation S ⊆ T if for all x ∈ S we have x

∈ T . If S ⊆ T then we also say T contains S which can be written T ⊇ S. If S ⊆

T and S 6= T then we write S ⊂ T and we say S is a proper subset of T .

Example:

If A = {a, b, c} then A has eight different subsets: ∅ {a} {b} {c} {a, b} {a, c} {b, c}

{a, b, c}

Notice that A ⊆ A and in fact each set is a subset of itself. The empty set ∅ is a

subset of every set. We are now ready to prove our first proposition. Some new

notation is required, and we must introduce an important piece of mathematical

culture. If we say “A if and only if B” then we mean that either A and B are both

true or they are both false in any given circumstance.

For example: “an integer x is even if and only if it is a multiple of 2”. The phrase

“if and only if” is used to establish logical equivalence. Mathematically, “A if and

only if B” is a way of stating that A and B are simply different ways of saying the

same thing. The phrase “if and only if” is abbreviated iff and is represented

symbolically as the double arrow ⇔. Proving an iff statement is done by

independently demonstrating that each may be deduced from the other.

Proposition 1:

Two sets are equal if and only if each is a subset of the other. In symbolic

notation:

(A = B) ⇔ (A ⊆ B) ∧ (B ⊆ A)

Proof:

Let the two sets in question be A and B. Begin by assuming that A = B. We

know that every set is a subset of itself so A ⊆ A. Since A = B we may substitute

into this expression on the left and obtain B ⊆ A. Similarly, we may substitute

on the right and obtain A ⊆ B. We have thus demonstrated that if A = B then A

and B are both subsets of each other, giving us the first half of the iff. Assume

now that A ⊆ B and B ⊆ A. Then the definition of subset tells us that any element

of A is an element of B. Similarly, any element of B is an element of A. This

9

means that A and B have the same elements which satisfies the definition of

set equality. We deduce A = B and we have the second half of the iff. ∎

Proposition 2:

De Morgan’s Laws Suppose that S and T are sets. DeMorgan’s Laws state that

(i) (S ∪ T) c = S c ∩ T c, and (ii) (S ∩ T) c = S c ∪ T c.

Proof:

Let x ∈ (S ∪ T) c; then x is not a member of S or T. Since x is not a member of

S we see that x ∈ S c. Similarly, x ∈ T c. Since x is a member of both these

sets, we see that x ∈ S c ∩ T c and we see that (S ∪ T) c ⊆ S c ∩ T c. Let y ∈

S c ∩ T c. Then the definition of intersection tells us that y ∈ S c and y ∈ T c.

This in turn lets us deduce that y is not a member of S ∪ T, since it is not in

either set, and so we see that y ∈ (S ∪ T) c. This demonstrates that S c ∩ T c

⊆ (S ∪ T) c. Applying Proposition 2.1 we get that (S ∪ T) c = S c ∩ T c and we

have proven part (i). The proof of part (ii) is left as an exercise. ✷ In order to

prove a mathematical statement you must prove it is always true. In order to

disprove a mathematical statement, you need only find a single instance where

it is false. It is thus possible for a false mathematical statement to be “true most

of the time”. In the next chapter we will develop the theory of prime numbers.

For now, we will assume the reader has a modest familiarity with the primes.

The statement “Prime numbers are odd” is false once, because 2 is a prime

number. All the other prime numbers are odd. The statement is a false one.

This very strict definition of what makes a statement true is a convention in

mathematics. We call 2 a counter example. It is thus necessary to find only one

counterexample to demonstrate a statement is (mathematically) false.

Disproof by counter example:

Prove that the statement A ∪ B = A ∩ B is false.

Let A = {1, 2} and B = {3, 4}. Then A ∩ B = ∅ while A ∪ B = {1, 2, 3, 4}. The sets

A and B form a counterexample to the statement.

Functions

Definition:

Let X and Y be sets. A function f from X to Y is a rule that assigns every element

x of X to a unique y in Y . We write f : X → Y and f(x) = y. Formally, using

predicate logic:

(∀x ∈ X, ∃y ∈ Y, y = f(x)) ∧ (∀x1, x2 ∈ X, f(x1) 6= f(x2) → x1 6= x2).

Then X is called the domain of f, and Y is called the codomain of f. The element

y is the image of x under f, while x is the preimage of y under f. Finally, we call

range the subset of Y with preimages.

Example:

Consider the assignment rule f : X = {a, b, c} → Y = {1, 2, 3, 4} which is defined

by: f = {(a, 2),(b, 4),(c, 2)}. We first check that this is a function. For every

element in X, we do have an assignment: f(a) = 2, f(b) = 4, f(c) = 2. Then the

condition that whenever f(x1) 6= f(x2) it must be that x1 6= x2 is also satisfied.

The the domain of f is X, the codomain of f is Y . The preimage of 2 is {a, c}

10

because f(a) = f(c) = 2. For the range, we look at Y , and among 1, 2, 3, 4, only

2 and 4 have a preimage, therefore the range is {2, 4}.

Example:

The rule f that assigns the square of an integer to this integer is a function.

Indeed, every integer has an image: its square. Also whenever two squares

are different, it must be that their square roots were different. We write

f : ℤ → ℤ, f(x) = 𝑥 2 .

Its domain is Z, its codomain is Z as well, but its range is {0, 1, 4, 9, 16, . . .},

that is the set of squares in Z.

Definition:

Let f be a function from X to Y , X, Y two sets, and consider the subset S ⊂ X.

The image of the subset S is the subset of Y that consists of the images of the

elements of S: f(S) = {f(s), s ∈ S}

Definition:

A function f is one-to-one or injective if and only if f(x) = f(y) implies x = y for

all x, y in the domain X of f. Formally: ∀x, y ∈ X(f(x) = f(y) → x = y). In words,

this says that all elements in the domain of f have different images.

Example.

Consider the function f : R → R, f(x) = 4x − 1. We want to know whether each

element of R has a different image. Yes, this is the case, why? well, visually,

this function is a line, so one may “see” that two distinct elements have distinct

images but let us try a proof of this. We have to show that f(x) = f(y) implies x

= y. Ok, let us take f(x) = f(y), that is two images that are the same. Then f(x) =

4x − 1, f(y) = 4y − 1, and thus we must have 4x − 1 = 4y − 1. But then 4x = 4y

and it must be that x = y, as we wanted. Therefore, f is injective.

Example.

Consider the function g: R → R, g(x) = x 2. Do we also have that two distinct

reals have distinct images? Well no... because 1 and −1 are both sent to 1...so

this function is not injective! If g(x) = g(y) = 1, we cannot conclude that x = y, in

fact this is wrong, it could be that x = −y.

Definition:

A function f is onto or surjective if and only if for every element y ∈ Y, there

is an element x ∈ X with f(x) = y: ∀y ∈ Y, ∃x ∈ X, f(x) = y. In words, each element

in the co-domain of f has a pre-image.

Example.

Consider again the function f: R → R, f(x) = 4x − 1. We want to know whether

each element of R has a preimage. Yes, it has, let us see why: we want to

show that there exists x such that f(x) = 4x − 1 = y. Given y, we have the

relation x = (y + 1)/4 thus this x is indeed sent to y by f.

Example.

Consider again the function g : R → R, g(x) = x 2 . Does each element in R

have a preimage? well, again no... Because R contains all the negative real

11

numbers, and it is not possible to square a real number and get something

negative... Formally, if y = −1, there is no x ∈ R such that g(x) = x 2 = −1.

Definition:

A function f is a one-to-one correpondence or bijection if and only if it is

both one-to-one and onto (or both injective and surjective). An important

example of bijection is the identity function.

Definition:

The identity function iA on the set A is defined by: iA : A → A, iA(x) = x.

Example 102. Consider the function f : R → R, f(x) = 4x − 1, which we have

just studied in two examples. We know it is both injective and surjective ,

therefore it is a bijection. Bijections have a special feature: they are invertible,

formally:

Definition:

Let f : A → B be a bijection. Then the inverse function of f, f −1 : B → B is

defined elementwise by: f −1 (b) is the unique element a ∈ A such that f(a) =

b. We say that f is invertible. Note the importance of the hypothesis: f must be

a bijection, otherwise the inverse function is not well defined. For example, if f

is not one-to-one, then f −1 (b) will have more than one value, and thus is not

properly defined. Note that given a bijection f : A → B and its inverse f −1 : B

→ A, we can write formally the above definition as: ∀b ∈ B, ∀a ∈ A(f −1 (b) =

a ⇐⇒ b = f(a)).

Example.

Let us look again at our two previous examples, namely, f(x) = 4x − 1 and g(x)

= x 2 . Then g(x), for g : R → R is not a bijection, so it cannot have an inverse.

Now f(x) is a bijection, so we can compute its inverse. Suppose that y = f(x),

𝑦+1 𝑦+1

then y = 4x − 1 ⇐⇒ y + 1 = 4x ⇐⇒ x = 4 , and f −1 (y) = 4 . We saw that

for the notion of inverse 𝑓 −1 to be defined, we need f to be a bijection. The

next result shows that 𝑓 −1 is a bijection as well.

Proposition 1.

If f : X → Y is a one-to-one correspondence, then f −1 : Y → X is a one-to-one

correspondence. Proof. To prove this, we just apply the definition of bijection,

namely, we need to show that f −1 is an injection, and a surjection. Let us start

with injection.

•𝑓 −1 is an injection: we have to prove that if f −1 (𝑦1 ) = f −1 (𝑦2 ), then y1 = 𝑦2 .

All right, then f −1 (𝑦1 ) = f −1 (𝑦2 ) = x for some x in X. But f −1 (𝑦1 ) = x means

that 𝑦1 = f(x), and 𝑓 −1 (𝑦2 ) = x means that 𝑦2 = f(x), by definition of the inverse

of function. But this shows that 𝑦1 = 𝑦2 , as needed.

• 𝑓 −1 is an surjection: by definition, we need to prove that any x ∈ X has a

preimage, that is, there exists y such that f −1 (y) = x. Because f is a bijection,

there is some y such that y = f(x), therefore x = f −1 (y) as needed. Suppose

that you have two functions f and g. It may be possible to compose them to

obtain a third function, here is how:

12

Definition:

Let f : A → B be a function, and g : B → C be a function. Then the composition

of f and g is a new function denoted by g ◦ f, and defined by: g ◦ f : A → C, (g

◦ f)(a) = g(f(a)). Note that the codomain of f is B, which is the domain of g.

Under this condition, the composition g ◦ f consists of applying first f, and then

apply g on the result. Therefore, g ◦ f 6= f ◦ g in general!

Example. Consider f : Z → Z and g : Z → Z defined by f(n) = 2n+3, g(n) = 3n

+ 2. We have (f ◦ g)(n) = f(g(n)) = f(3n + 2) = 2(3n + 2) + 3 = 6n + 7, while (g ◦

f)(n) = g(f(n)) = g(2n + 3) = 3(2n + 3) + 2 = 6n + 11. Suppose now that you

compose two functions f, g, and both of them turn out to be injective. The next

result tells us that the combination will be as well.

Proposition 2.

Let f : X → Y and g : Y → Z be two injective functions. Then g ◦ f is also

injective. Proof. What we need to do is check the injectivity of a function, so

we do this as usual: we check that g ◦ f(𝑥1 ) = g ◦ f(𝑥2 ) implies 𝑥1 = 𝑥2 . Typically,

to be able to prove this, you will have to keep in mind assumptions, namely

that both f and g are injective. So let us start. We have g ◦f(𝑥1 ) = g ◦f(𝑥2 ) or

equivalently g(f(𝑥1 )) = g(f(𝑥2 )). But we know that g is injective, so this implies

f(𝑥1 ) = f(𝑥2 ). Next we use that f is injective, thus 𝑥1 = 𝑥2 , as needed! Let us

ask the same question with surjectivity, namely whether the composition of

two surjective functions gives a function which is surjective too.

Proposition 3.

Let f : X → Y and g : Y → Z be two surjective functions. Then g ◦ f is also

surjective. Proof. The codomain of g ◦ f is Z, therefore we need to show that

every z ∈ Z has a preimage x, namely that there always exists an x such that

g ◦ f(x) = z. Again, we keep in mind that f and g are both surjective. Since g is

surjective, we know there exists y ∈ Y such that g(y) = z. Now again, since f

is surjective, we know there exists x ∈ X such that f(x) = y. Therefore there

exist x, y such that z = g(y) = g(f(x)) as needed.

This section defines and gives examples of domains and ranges of functions.

These are important properties of a function and we will meet them in

subsequent sections. We write a function using the notation f(x). The notation

means that given a number x then the function gives another unique number

f(x). If we write y = f(x) then we say that x is the independent variable and y the

dependent variable. Note that a function can be written as f(x) = x 3 or as h(t)

= t 3 or as c(y) = y 3. These are all the same function - they all do the same

thing, cube a number.

Domain

Definition:

13

The domain of a function is the complete set of possible values of the

independent variable. The domain is the set of all possible x-values which will

make the function "work", and will output real y-values. When finding the

domain, remember:

The denominator (bottom) of a fraction cannot be zero,

The number under a square root sign must be positive in this section

Examples:

1. f(x) = x 2 . Defined for all values of x i.e. the domain is R.

2. f(t) = 2t . Defined for all values of t i.e. the domain is R.

3. f(z) = √𝑧. Defined for z ≥ 0 i.e. the domain is [0,∞) or R+.

4. f(x) =√4 + 3x − x 2 . We need 4 + 3x − x 2 ≥ 0 and solving this inequality

we see that this is only true for −1 ≤ x ≤ 4. So this function is only defined in this

range of values for x i.e. the domain is [−1, 4].

5. f(x) = sin(x). Defined for all values so the domain ℝ is .

6. f(u) = 1/u. The domain is defined for all u ≠= 0 and is denoted by ℝ − {0}.

Range

Definition:

The range of a function is the complete set of all possible resulting values of

the dependent variable (y, usually), after we have substituted the domain.

For example, consider the function defined by the rule that we take an input

and raise it to the third power. This can be

represented in equation form as y = x^3, and when this function is given input

values of {-2, -1, 0, 1, 2}, we can find the corresponding outputs by plugging

those inputs in for 'x' in the equation. For instance, if we input -2, we have y =

(-2)^3 = -8, so when the input is -2, the output is -8, and -8 is in our range. When

we find each of the corresponding outputs to our inputs, we have our range. In

our example, the range is {-8, -1, 0, 1, 8}, because these are the outputs

corresponding to -2, -1, 0, 1, and 2 respectively.

Graph of Domain and Range example:

14

15

Exercises on Sets and Functions

Exercise 2. If S = {{a}, {b}}, and T = {{a}, b}, is it true that S = T?

Exercise 3. If S = {A, B}, what is S ∪ ∅?

Exercise 3. If S = {A, B}, what is S ∪ ∅?

Exercise 4. For S as in the previous exercise, what is S ∪ S?

Exercise 5. If S = {A, B}, what is S ∩ ∅?

Exercise 6. If S = {A, B}, what is S ∩ S?

Exercise 7. Is it true that if S ⊂ T and T ⊂ S, then S = T?

Exercise 8. Find all the subsets of {{A, B}, {}}.

Exercise 9. For S = {1, 2}, what is S × S?

Exercise 10. What is the range of the identity function Id : S → S?

Exercise 11. For S = {A, B}, T = {A, B, C}, how many functions f : S → T are there with

the range of f equal to the range of the inclusion j : S → T?

Exercise 12. The example above is a bit special because the right hand side of the

equation is 0. Generalizing the above, explain how the solution to the equation f(x) =

g(x) can be interpreted as a subset of Z for any functions f : Z → Z, g : Z → Z.

Exercise 13. Let S = {1, 2} and T = {A, B, C}.

(i) How many functions are there from S to T?

(ii) How many 1-1 functions are there from S to T?

(iii) How many onto functions are there from S to T?

(iv) How many bijections are there from S to T?

Exercise 14. For which integers a does the function f : Z → Z given by f(x) = x + a

have an inverse? For each such a, what is the inverse?

Exercise 15. For which integers a does the function f : Z → Z given by f(x) = ax have

an inverse? For each such a, what is the inverse?

Exercise 16. Is {1, 2, 3} a partition of {1, 2, 3}?

Exercise 17. Find all partitions of {1, 2, 3}.

16

Answer Keys: Sets and Functions

1. Answer: No, because {{0}} is the set whose only element is {0}. And 0 6= {0}.

2. Answer: No, because the sets have different elements.

3. Answer: S ∪ ∅ = S. Many people wrote S = {A, B, {}}, but this is wrong; {} is not

an element of either S or {}.

4. Answer: S.

5. Answer: ∅.

6. Answer: S.

7. Answer: Yes

8. Answer: {{A, B}, {}}, {{A, B}}, {{}}, {}.

9. Answer: {(1, 1),(1, 2),(2, 1),(2, 2)}

10. Answer: S. (Some students assumed that S = {A, B}. Because of where the

question was placed, this was reasonable, though not what I intended, so I also

accepted the answer {A, B}.)

11. Answer: Most students got this wrong, at least in part. There are two functions.

One is given by f = j. The other is given by f(A) = B, f(B) = A.

12. Answer: the solutions are the set {z ∈ Z | f(z) = g(z)}.

13. (i) Answer: 9, because there are 3 options for where to map 1 to, and 3options

for where to map 2 to. 3 × 3 = 9.

(ii) Answer: 6. Of the 9 functions from (i), 3 are not 1-1.

(iii) Answer: 0

(iv) Answer: 0

14. Answer: All integers a. The inverse g : Z → Z is given by g(z) = z − a.

15. Answer: a ∈ {1, −1}. The inverse g : Z → Z is given by g(z) = z/a.

16. Answer: No. It is not even a set of subsets of {1, 2, 3}.

17. Answer: {{1, 2, 3}}, {{1}, {2}, {3}}, {{1, 2}, {3}}, {{1}, {2, 3}}, {{2}, {1, 3}}.

17

SPECIAL TYPES OF FUNCTIONS

18

Inverse and Composite Functions

To find the inverse function, switch the x and y values, and then solve for y.

Learning Objectives

Calculate the formula of a function’s inverse by switching x and y and then

solving for y.

Key Points

An inverse function reverses the inputs and outputs.

To find the inverse formula of a function, write it in the form of y and x, switch

y and x, and then solve for y. Some functions have no inverse function, as a

function cannot have multiple outputs.

Key Terms

Inverse function: A function that does exactly the opposite of another.

An inverse function, which is notated f−1(x), is defined as the inverse function

of f(x) if it consistently reverses the f(x) process. That is, if f(x) turns a into b,

then f−1(x) must turn b into a. More concisely and formally, f−1(x) is the

inverse function of f(x)

Notice that the ordered pairs are reversed from the original function to its

inverse. Because f(x) maps a to 3, the inverse f−1(x) maps 3 back to a.

Thus, the graph of f−1(x) can be obtained from the graph of f(x) by switching

the positions of the x and y-axes. This is equivalent to reflecting the graph

across the line y=x, an increasing diagonal line through the origin.

Inverse functions:

mapping

representation: An

inverse function

reverses the inputs and

outputs.

In general, given a function, how do you find its inverse function? Remember

that an inverse function reverses the inputs and outputs. So, to find the

inverse function, switch the x and y values of a given function, and then solve

for y.

19

Inverse functions:

Graphic representation: The function graph (red) and its inverse function

graph (blue) are reflections of each other about the line y=x (dotted black

line). Notice that any ordered pair on the red curve has its reversed ordered

pair on the blue line. For example, (0,1) on the red (function) curve is reflected

over the line y=x and becomes (1,0) on the blue (inverse function) curve.

Where one curve is on the line y=x, the curves intersect, as a reflection over

the line leaves the point unchanged.

Example 1

b.: Switch the x and y variables: x = y2

c.: Solve for y, x = y2 ± √x = y

Since the function f(x) = x2 has multiple outputs, its inverse is not a function.

Notice the graphs in the picture below. Even though the blue curve is a

function (passes the vertical line test), its inverse would not be. The red curve

for the function f(x) = √x is not the full inverse of the function f(x) = x2

f(x)=x2, includes the points (−1,1) and (1,1). Therefore, the inverse would

include the points: (1,−1) and (1,1) which the input value repeats, and

therefore is not a function. For f(x)=√x to be a function, it must be defined as

positive.

20

Example 2

Find the inverse function of: f(x) = 2x . As soon as the problem includes an

exponential function, we know that the logarithm reverses

exponentiation. The complex logarithm is the inverse function of the

exponential function applied to complex numbers.

The graphs of y=2x (blue) and x=2y (red) are inverses of one another. The

black line represents the line of reflection, in which is y=x.Test to make sure

this solution fills the definition of an inverse function.

Functional composition allows for the application of one function to another;

this step can be undone by using functional decomposition.

Learning Objectives

Practice functional composition by applying the rules of one function to the

results of another function.

21

Key Takeaways

Key Points:

Functional composition applies one function to the results of another.

Functional decomposition resolves a functional relationship into its constituent

parts so that the original function can be reconstructed from those parts by

functional composition.

Decomposition of a function into non-interacting components generally

permits more economical representations of the function.

The process of combining functions so that the output of one function

becomes the input of another is known as a composition of functions. The

resulting function is known as a composite function. We represent this

combination by the following notation: (f∘ g (x)=f (g(x) )

The domain of the composite function (f∘ g) is all x such that x is in the domain

of g and g(x) is in the domain of f.

Key Terms:

Codomain: The target space into which a function maps element of its

domain. It always contains the range of the function but can be

larger than the range if the function is not subjective.

Domain: The set of all points over which a function is defined.

Function Composition

The process of combining functions so that the output of one function

becomes the input of another is known as a composition of functions. The

resulting function is known as a composite function. We represent this

combination by this notation:

( f ∘ g ) (x) = f ( g (x) )

We read the left-hand side as “f “composed with g at x, and the right-hand

side as “f of g of x.” The two sides of the equation have the same

mathematical meaning and are equal. The open circle symbol, is called

the composition operator. Composition is a binary operation that takes two

functions and forms a new function, much as addition or multiplication takes

two numbers and gives a new number.

It is important to understand the order of operations in evaluating a composite

function. We follow the usual convention with parentheses by starting with the

innermost parentheses first, and then working to the outside.

In general, (f∘g) and (g∘f) are different functions. In other words, in many

cases f(g(x))≠g(f(x)) for all x.

Note that the range of the inside function (the first function to be evaluated)

needs to be within the domain of the outside function. Less formally, the

composition has to make sense in terms of inputs and outputs.

22

When evaluating a composite function where we have either created or been

given formulas, the rule of working from the inside out remains the same. The

input value to the outer function will be the output of the inner function, which

may be a numerical value, a variable name, or a more complicated

expression.

Example

To evaluate f (g(3) ), first substitute, or input the value of 3 into g(x) and find

the output. Then substitute that value into the f(x) function, and simplify:

To evaluate g (f (3)), find f (3) and then use that output value as the input

value into the g(x) function:

f (3) =−2(3) = −6

g(−6) = (−6)2 −1 = 36 −1 = 35 . Therefore, g(f(3))=35

While we can compose the functions for each individual input value, it is

sometimes helpful to find a single formula that will calculate the result of a

composition f(g(x)) or g(f(x)). To do this, we will extend our idea of function

evaluation.

In the next example we are given a formula for two composite functions and

asked to evaluate the function. Evaluate the inside function using the input

value or variable provided. Use the resulting output as the input to the

outside function.

Example

If f(x)=−2x and g(x)=x2−1, evaluate f(g(x)) and g(f(x)).

First substitute, or input the function g(x), x2−1 into the f(x) function, and then

simplify:

f(g(x))=−2(x2−1)

f(g(x))=−2x2+2

For g(f(x)), input the f(x) function, −2x into the g(x) function, and then simplify:

g(f(x))=(−2x)2−1

g(f(x))=4x2−1

Functional Decomposition

Functional decomposition broadly refers to the process of resolving a

functional relationship into its constituent parts in such a way that the original

function can be reconstructed (i.e., recomposed) from those parts by function

23

composition. In general, this process of decomposition is undertaken either for

the purpose of gaining insight into the identity of the constituent components

(which may reflect individual physical processes of interest), or for the

purpose of obtaining a compressed representation of the global function; a

task which is feasible only when the constituent processes possess a certain

level of modularity (i.e., independence or non-interaction).

“sparseness” in the dependency structure; i.e. when constituent functions are

found to depend on approximately disjointed sets of variables. Also,

decomposition of a function into non-interacting components generally permits

more economical representations of the function.

Domain restriction is important for inverse functions of exponents and

logarithms because sometimes we need to find an unique inverse.

Learning Objectives

Determine inverses of functions by restricting their domains

Key Takeaways

Key Points:

f−1(x) is defined as the inverse function of f(x) if it consistently reverses the

f(x) process.

f (x)=x2, without any domain restriction, does not have an inverse function, as

it fails the horizontal line test.

Key Terms:

domain: The set of points over which a function is defined.

Inverse Functions

f(x) process. That is, if f(x) turns a into b, then f−1x must turn b into a. More

concisely and formally, f−1x is the inverse function of f(x) if f(f−1(x))=x.

If f maps X to Y, then f−1 maps Y back to X.

Remember that:

If f maps X to Y, then f−1 maps Y back to X. This is not true of the function

f(x)=x2.Without any domain restriction, f(x)=x2 does not have an inverse

24

function as it fails the horizontal line test. But if we restrict the domain to be

x>0

then we find that it passes the horizontal line test and therefore has an inverse

function. Below is the graph of the parabola and its “inverse.” Notice that the

parabola does not have a “true” inverse because the original function fails the

horizontal line test and must have a restricted

domain to have an inverse.

Graph of a parabola with the equation y=x2, the

U-Shaped curve opening up. This function fails

the horizontal line test, and therefore does not

have an inverse. The inverse equation,

y=√x (other graph) only includes the positive

input values of the parabola’s domain.

However, if we restrict the domain to be x>0,

then we find that it passes the horizontal line

test and will match the inverse function.

Functions

logarithms because sometimes we need to find a unique inverse. The inverse

of an exponential function is a logarithmic function, and the inverse of a

logarithmic function is an exponential function.

Example 1

Is x = 0 in the domain of the function f(x)=log(x)? If so, what is the value of

the function when x=0? Verify the result.

No, the function has no defined value for x = 0. To verify, suppose x = 0 is in

the domain of the function f(x) = log (x). Then there is some number n such

that n = log (0). Rewriting as an exponential equation gives: 10n = 0, which is

impossible since no such real number n exists. Therefore x = 0 is not in the

domain of the function f(x) = log(x).

A composite function represents, in one function, the results of an entire chain

of dependent functions.

Learning Objectives

Solve for the inverse of a composite function

Key Takeaways

25

Key Points:

The composition of functions is always associative. That is, if f, g, and h are

three functions with suitably chosen domains and co-domains, then f∘ (g∘ h) =

(f∘ g)∘ h.

Functions can be inverted and then composed, giving the notation of: ( f′∘ g′)

(x).

Functions can be composed and then inverted, yielding the following notation:

(f ∘ g)′ (x)

Key Terms

Composite function: A function of one or more independent variables,

at least one of which is itself a function of one or

more other independent variables; a function of a

function.

In mathematics, function composition is the application of one function to the

results of another.

The functions g and f are said to commute with each other if g ∘ f = f ∘ g. In

general, the composition of functions will not be commutative.

of dependent functions. For example, if a school becomes larger, the supply

of food in the cafeteria must become larger. This would entail ordering more

sandwiches, which means ordering more ingredients, drinks, plates, etc. The

entire chain of dependent functions are the ingredients, drinks, plates, etc.,

and the one composite function would be putting the entire chain together in

order to calculate a larger population at the school.

example. let’s take two functions, compose and invert them.

function g:

√x−1 = y√x −1 = (f ∘ g)′ (x).

You can think of a function as a way of matching the members of a set "A" to

a set "B":

26

Injective means that every member of "A" has its own unique matching

member in "B". You won't get two "A"s pointing to one "B", but you could have

a "B" without a matching "A"

Surjective means that every "B" has at least one matching "A" (maybe more

than one).

Bijective means Injective and Surjective together. So there is a perfect "one-

to-one correspondence" between the members of the sets.

Formal Definitions

Injective

Theorem:

If f: A→B and g:B→C are injective functions, then g∘f:A→C is injective

also.

Proof:

Suppose g (f (a)) =g (f (a′))

Since g is injective, f (a) = f(a′). Since f is injective, a=a′. Thus, (g ∘ f)

(a) = (g ∘ f) (a′) implies a=a′, so (g ∘ f) is injective.

Theorem:

Suppose f1, f2:A→B, g:B→C, h1,h2:C→D are functions.

a) If g is injective and g∘f1=g∘f2 then f1=f2

. b) If g is surjective and h1∘g=h2∘g then h1=h2.

Proof.

We prove part (b), leaving part (a) as an exercise. Suppose c ∈C. We

wish to show h1(c) =h2(c). By hypothesis g is surjective, so there is a b

∈B such that g (b)=c. So

h1(c)=h1(g(b))=(h1∘g)(b)=(h2∘g)(b)=h2(g(b))=h2(c),

as desired.

Example 1:

f(x) = x+5 from the set of real numbers to is an injective function.

27

This function can be easily reversed. for example:

f(3) = 8

Given 8 we can go back to 3

Example 2:

f(x) = x2 from the set of real numbers to is not an injective function

because of this kind of thing:

f(2) = 4 and

f(-2) = 4

This is against the definition f(x) = f(y), x = y, because f(2) = f(-2) but 2

≠ -2

In other words there are two values of "A" that point to one "B", and

this function could not be reversed (given the value "4" ... what

produced it?)

BUT if we made it from the set of natural numbers then it is injective,

because:

f(2) = 4 there is no f(-2), because -2 is not a natural number.

because it makes it sound like it is actually bijective which has "one-to-one

correspondence".

least one x in A such that f(x) = y, in other words f is surjective if and only if f(A)

= B.

So, every element of the range corresponds to at least one member of the domain.

Theorem:

Suppose f:A→B and g:B→C are functions.

b) If g∘f is surjective then g is surjective.

Proof.

We prove part (a), leaving part (b) as an exercise. Suppose a,a ′∈ A

and f(a) = f(a′). We wish to prove a = a′. We have (g ∘ f) (a) = g (f (a) ) = g (f (a′) ) = (g

∘ f) (a′) and since g ∘ f is injective, we conclude that,

A = a′, as desired.

Bijective

one x in A such that f(x) = y

28

Alternatively, f is bijective if it is a one-to-one correspondence between

those sets, in other words both injective and surjective.

Theorem:

A function f:A→B has an inverse if and only if it is bijective.

Proof:

Suppose g is an inverse for f (we are proving the implication ⇒). Since g∘f=iA

is injective, so is f (by 4.4.1(a)). Since f∘g=iB is surjective, so is f (by 4.4.1(b)).

Therefore, f is injective and surjective, that is, bijective.

the proof of theorem, we know that since f is surjective, f ∘ g = iB, and since f

is injective, g ∘ f = iA.

We have talked about "an'' inverse of f, but really there is only one.

Example:

The function f(x) = x2 from the set of positive real numbers to positive real

numbers is injective and surjective. Thus it is also bijective.

But not from the set of real numbers because you could have, for example,

both

f(2)=4 and f(-2)=4.

29

EXERCISES

1.) Let 𝐴 and 𝐵 be two nonempty sets where 𝐴 = {1 ,2,3,4} and 𝐵 = {𝑎,𝑏,𝑐}. Consider

each of the following relations: 𝑇 = {(1,𝑎),(2,𝑏),(2,𝑐),(3,𝑐),(4,𝑏)} 𝑈 =

{(1,𝑎),(2,𝑏),(4,𝑏)} 𝑉 = {(1,𝑎),(2,𝑏),(3,𝑐),(4,𝑏)} Which of these relations (𝑇,𝑈 and

𝑉) qualify as functions?

injective) function. Why?

(ii) 𝑉 is an onto (surjective) function. Why?

(iii) Does the function 𝑓, defined by the relation 𝑉, have an inverse?

3.) For each of the relations {𝑄,,𝑆,𝑇,𝑈,𝑉} below, determine whether the relation is a

function. If the relation is a function, determine whether the function is injective

and/or surjective.

(i) 𝐴 = {1,2,3}, 𝐵 = {𝑎,𝑏,𝑐,𝑑} 𝑄 = {(1,𝑎),(2,𝑑),(3,𝑏)}

(ii) 𝐴 = {1,2,3}, 𝐵 = {𝑎,𝑏,𝑐} 𝑅 = {(1,𝑎),(2,𝑏),(3,𝑐)}

(iii) 𝐴 = {1,2,3}, 𝐵 = {𝑎,𝑏,𝑐} 𝑆 = {(1,𝑎),(2,𝑏),(3,𝑏)} (iv) 𝐴 = {1,2,3}, 𝐵=

{𝑎,𝑏,𝑐,𝑑}

𝑇 = {(1,𝑎),(2,𝑏),(2,𝑐),(3,𝑑)} (v) 𝐴 = {1,2,3}, 𝐵 = {𝑎,𝑏} 𝑈 =

{(1,𝑎),(2,𝑏),(3,𝑏)}

(vi) 𝐴 = {1,2,3}, 𝐵 = {𝑎,𝑏} 𝑉 = {(1,𝑎),(2,𝑏)}

(ii) For the relation that is a bijection, write down the elements of the inverse

function.

(i) Give an example to show that 𝑓 is not injective.

(ii) Give an example to show that 𝑓 is not surjective.

(i) Give an example to show that 𝑓 is not injective.

(ii) Give an example to show that 𝑓 is not surjective

7.) For each of the functions below determine which of the properties hold, injective,

surjective, bijective. Briefly explain your reasoning.

(i) The function 𝑓:ℝ → ℝ defined by (𝑥) = 𝑒𝑥. (ii) The function 𝑓:ℝ → ℝ+

defined by (𝑥) = 𝑒𝑥. (iii) The function 𝑓:ℝ → ℝ defined by (𝑥) = (𝑥 + 1)𝑥(𝑥 − 1).

(iv) The function 𝑓:ℝ → ℝ defined by (𝑥) = (𝑥2 −9)(𝑥2 − 4)

8.) Which of the following cubic functions have an inverse? [Hint: Finding the

derivative of the function may help!] (i) (𝑥) = 𝑥3 −6𝑥2 +3𝑥 +7, (ii) 𝑓(𝑥) = −𝑥3 −6𝑥2

−13𝑥 +4, (iii) 𝑓(𝑥) = 𝑥3 +3𝑥2 +4𝑥 +3, (iv) 𝑓(𝑥) = −𝑥3 +3𝑥2 −𝑥 −1

30

ANSWERS KEY

1.) Relation 𝑇, maps 2 (∈ 𝐴) to both 𝑏 and 𝑐 (∈ 𝐵). This violates condition 2 of the

definition.

Relation 𝑇, is not a function.

Relation 𝑈 is not defined for all elements of 𝐴.This violates condition 1 of the

definition. Relation 𝑈 is not a function.

Relation satisfies both conditions of the definition of a function.

Relation 𝑉 is a function.

If we call the function 𝑓 we have (1) = 𝑎,𝑓(2) = 𝑏,𝑓(3) = 𝑐 and 𝑓(4) = 𝑏.

(ii) The range of 𝑓 is equal to the set 𝐵 (the codomain)

(iii) No, a function must be both injective and surjective to have an inverse.

3.) (i) The relation is a function. The function is injective. The function is not

surjective since 𝑐 is not an element of the range.

(ii) The relation is a function.The function is both injective and surjective.

(iii) The relation is a function.The function is not injective since (2) = 𝑓(3) but 2

≠ 3.The function is not surjective since 𝑐 is not an element of the range.

(iv) The relation is a not a function since the relation is not uniquely defined for

2.

(v) The relation is a function. The function is not injective since (2) = 𝑓(3) but 2

≠ 3. The function is surjective.

(vi) The relation is a not a function since the relation is not defined for 2.

(ii) 𝑓−1 = {(𝑎,1),(𝑏,2),(𝑐,3)}

5.) (i) 𝑓(−1) = 𝑓(1) = 3 but −1 ≠ 1, therefore the function is not injective.

(ii) There is no real number, 𝑥 such that (𝑥) = 1 therefore the function is not

surjective. Or the range of the function is 𝑦 ≥ 2. The range of the

function is not ℝ (the codomain) therefore the function is not surjective.

6.) (i) 𝑓(6) = 𝑓(0) = 0 but 6 ≠ 0, therefore the function is not injective.

(ii) (𝑥) = (𝑥 −3)2 −9 [by completing the square] There is no real number, 𝑥

such that 𝑓(𝑥) = −10 the function is not surjective. Or the range of the function

is 𝑦 ≥ 2. The range of the function is not ℝ (the codomain) therefore the function

is not surjective.

7.) (i) This function is injective, since 𝑒𝑥 takes on each nonnegative real value for

exactly one 𝑥. However, the function is not surjective, because 𝑒𝑥

never takes on negative values. Therefore, the function is not

bijective either.

(ii) The function 𝑒𝑥 takes on every nonnegative value for exactly one 𝑥, so it is

injective, surjective, and bijective.

(iii) This function is surjective, since it is continuous, it tends to +∞ for large

positive 𝑥, and tends to −∞ for large negative 𝑥. The function takes on

each real value for at least one 𝑥. However, this function is not

31

injective, since it takes on the value 0 at 𝑥 = −1, 𝑥 = 0 and 𝑥= 1.

Therefore, the function is not bijective either.

(iv) This function is not surjective, it tends to +∞ for large positive 𝑥, and also

tends to +∞ for large negative 𝑥. Also this function is not injective,

since it takes on the value 0 at 𝑥 = 3,

𝑥 = −3, 𝑥 = 4 and 𝑥 = −4. Therefore, the function is not bijective either.

8.) All cubic functions are surjective by their nature. So we check injective by seeing

if the function is always increasing or decreasing

(i) 𝑓′(𝑥) = 3(𝑥 −2)2 −9 this function is not always increasing or decreasing,

hence not injective

(ii) 𝑓′(𝑥) = −[3(𝑥 +2)2 +1] this function is always decreasing

(iii) 𝑓′(𝑥) = 3(𝑥 +2)2 +1 this function is always increasing

(iv) 𝑓′(𝑥) = −3(𝑥 −1)2 +2 this function is always increasing

32

MATHEMATICAL INDUCTION

SHAIRA S. BUCIO

33

Mathematical Induction

Definition:

This is a method of "pulling oneself up by one's bootstraps" and is

regarded with suspicion by non-mathematicians.

Example:

1) Suppose we want to sum an Arithmetic Progression: 1 + 2 +

1

3 +. . . + 𝑛 = 2 𝑛(𝑛 + 1).

Proof:

We define a sequence of "propositions" 𝑃(1), 𝑃(2), ... where 𝑃(𝑛) is "1 +

1

2 + 3 +. . . + 𝑛 = 2 𝑛(𝑛 + 1)"

First, we'll prove 𝑃(1); this is called "anchoring the induction".

Then we will prove that if 𝑷(𝒌) is true for some value of k, then so is

𝑃(𝑘 + 1) ; this is called "the inductive step".

If 𝑃(1) is OK, then we can use this to deduce that 𝑃(2) is true and then

use this to show that 𝑃(3) is true and so on.

So, if 𝑛 is the first value for which the result is false, then 𝑃(𝑛 – 1) is true

and we would get a contradiction.

So, let's look hard at the above example.

1

𝑃(1) is certainly OK: 1 = × 1 × 2.

2

Now suppose that 𝑃(𝑘) is true for some value of k.

Then try and prove 𝑃(𝑘 + 1):

1

Now 1 + 2 + 3 + . . . + 𝑘 + (𝑘 + 1) = 𝑘( 𝑘 + 1) + (𝑘 + 1) (using

2

𝑃(𝑘), which we are allowed to assume).

1 1

But this simplifies to (𝑘 + 1)( 2 𝑘 + 1) = (𝑘 + 1)(𝑘 + 2) and this is

2

exactly what 𝑃(𝑘 + 1) says.

Hence the result is true for all values.∎

Let r be a fixed real number.

1−𝑟 𝑛+1

Then 1 + 𝑟 + 𝑟 2 + 𝑟 3 + . . . + 𝑟 𝑛 = . This is 𝑃(𝑛).

1−𝑟

Proof:

Clearly 𝑃(0) is true. (Note that we can anchor the induction where we

like.

So, we suppose that 𝑃(𝑘) is true and we'll try and prove 𝑃(𝑘 + 1).

So, look at 1 + 𝑟 + 𝑟 2 + 𝑟 3 + . . . + 𝑟 𝑘 + 𝑟 𝑘+1 .

1−𝑟 𝑘+1

By 𝑃(𝑘) the term in brackets is and so we can simplify this to

1−𝑟

1−𝑟 𝑘+1 1−𝑟 𝑘+1 −𝑟 𝑘+1 1−𝑟 𝑘+2

1−𝑟

+ 𝑟 𝑘+1 = 1−𝑟

= 1−𝑟

which is what 𝑃(𝑘 + 1) predicts.∎

34

Theorem:

Mathematical induction ⇒ the well-ordering property.

Proof:

Suppose that the well-ordering property were false.

Let 𝑆 be a counterexample: a nonempty set of nonnegative integers that

contains no smallest element.

Let 𝑃(𝑛) be the statement “𝑖 ∉ 𝑆 for all 𝑖 ≤ 𝑛.”

We will show that 𝑃(𝑛) is true for all n (which will contradict the assertion

that 𝑆 is nonempty).

Now 𝑃(0) must be true, because if 0 ∈ 𝑆 then clearly S would have a

smallest element, namely 0.

Suppose now that 𝑃(𝑛) is true, so that 𝑖 ∉ 𝑆 for all 𝑖 = 0, 1, 2, . . . , 𝑛.

We must show that 𝑃(𝑛 + 1) is true, which amounts to showing that

𝑛 + 1 ∉ 𝑆. If 𝑛 + 1 ∈ 𝑆, then 𝑛 + 1 would be the smallest element of

S, and this would contradict our assumption.

Therefore 𝑛 + 1 ∉ 𝑆.

Thus, we have shown by the principle of mathematical induction that

𝑃(𝑛) is true for all 𝑛, which means that there can be no elements of S.

This contradicts our assumption that 𝑆 = ∅ , and our proof by

contradiction is complete.∎

Theorem:

The well-ordering property ⇒ strong induction.

Proof:

Let us suppose that we have a proposition ∀𝑛, 𝑃(𝑛) that has been

proved using it.

We must show that in fact ∀𝑛, 𝑃(𝑛) is true (to say that a principle of proof

is valid means that it proves only true propositions).

Let 𝑆 be the set of counterexamples, 𝑖. 𝑒, 𝑆 = {𝑛 | ¬𝑃(𝑛)}.

We want to show that 𝑆 = ∅ .

We argue by contradiction.

Assume that 𝑆 ≠ ∅.

Then by the well-ordering property, 𝑆 has a smallest element.

Since part of the method of strong induction is to show that 𝑃(1) is true,

this smallest counterexample must be greater than 1.

Let us call it 𝑘 + 1.

35

Since 𝑘 + 1 is the smallest element of 𝑆, it must be the case that

𝑃(1) ∧ 𝑃(2) ∧ · · · ∧ 𝑃(𝑘) is true.

But the rest of the proof using strong induction involved showing

that 𝑃(1) ∧ 𝑃(2) ∧ · · · ∧ 𝑃(𝑘) implied 𝑃(𝑘 + 1) ; therefore since the

hypothesis is true, the conclusion must be true as well, i.e., 𝑃(𝑘 + 1) is

true.

This contradicts our assumption that 𝑘 + 1 ∈ 𝑆.

Therefore, we conclude that 𝑆 = ∅, so ∀𝑛, 𝑃(𝑛) is true. ■

Theorem:

Strong induction ⇔ mathematical induction

Proof:

If one has shown that 𝑃(𝑘) → 𝑃(𝑘 + 1), then it automatically follows

that [𝑃(1) ∧ · · · ∧ 𝑃(𝑘)] → 𝑃(𝑘 + 1).

In other words, ordinary induction ⇒ strong induction.

Conversely, suppose that 𝑃(𝑛) is a statement that one can prove using

strong induction.

Let 𝑄(𝑛) be 𝑃(1) ∧ · · · ∧ 𝑃(𝑛).

Clearly ∀𝑛, 𝑃(𝑛) is logically equivalent to ∀𝑛, 𝑄(𝑛).

We show how ∀𝑛, 𝑄(𝑛) can be proved using ordinary induction.

First, 𝑄(1) is true because 𝑄(1) = 𝑃(1) and 𝑃(1) is true by the basis

step for the proof of ∀𝑛, 𝑃(𝑛) by strong induction.

Now suppose that 𝑄(𝑘) is true, i.e., 𝑃(1) ∧ · · · ∧ 𝑃(𝑘) is true.

By the proof of ∀𝑛, 𝑃(𝑛) by strong induction it follows that 𝑃(𝑘 + 1) is

true. But 𝑄(𝑘) ∧ 𝑃(𝑘 + 1) is just 𝑄(𝑘 + 1).

Thus, we have proved ∀𝑛, 𝑄(𝑛) by ordinary induction.∎

mathematical induction, and strong induction are all equivalent.

Definition:

The well-ordering principle is a concept which is equivalent to

mathematical induction. In your textbook, there is a proof for how the

well-ordering principle implies the validity of mathematical induction.

However, because of the very way in which we constructed the set of

natural numbers and its arithmetic, we deduced, in class, the validity of

mathematical induction directly from the axioms of set theory. In this

note, we show how mathematical induction, in turn, implies the well-

ordering principle.

36

Example:

1) The set ℕ is well-ordered.

The following sets are well-ordered:

(1) ℕ ∪ {0}

(2) ℕ ∪ {−1, 0}

(3) ℕ ∪ {−3, −2, −1}

(4) {𝑛 ∈ ℕ ∶ 𝑛 > 5}

smallest element)

(2) ℤ (the set of negative integers is a non-empty subset of Z but with

no smallest element)

(3) the interval [0, 1] (because (0, 1) is a non-empty subset of [0, 1]

without smallest element)

Theorem:

Every non-empty subset of the natural numbers has a least element.

Proof:

Let A be a non-empty subset of ℕ.

We wish to show that A has a least element, that is, that there is an

element 𝑎 ∈ 𝐴 such that 𝑎 ≤ 𝑛 for all 𝑛 ∈ 𝐴. We will do this by strong

induction on the following predicate:

𝑃(𝑛): “If 𝑛 ∈ 𝐴, then 𝐴 has a least element.”

Basic Step: 𝑃(0) is clearly true, since 0 ≤ 𝑛 for all 𝑛 ∈ 𝑁.

Strong Inductive Step: We want to show that [𝑃(0) ∧ 𝑃(1) ∧···∧ 𝑃(𝑛)] →

𝑃(𝑛 + 1).

To this end, suppose that 𝑃(0), 𝑃(1),···, 𝑃(𝑛) are all true and that 𝑛 +

1 ∈ 𝐴.

We consider two cases.

CASE 1: ¬∃𝑚(𝑚 ∈ 𝐴 ∧ 𝑚 < 𝑛 + 1).

In this case, n + 1 is the least element of A.

CASE 2: ∃𝑚(𝑚 ∈ 𝐴 ∧ 𝑚 < 𝑛 + 1).

In this case, since P(m) is true, A has a least element.

Either way, we conclude that 𝑃(𝑛 + 1) is true.

So, by strong mathematical induction, we obtain that 𝑃(𝑛) is true for all

𝑛 ∈ 𝑁.

Since A is not empty, we can pick an 𝑛 ∈ 𝐴.

Moreover, since 𝑃(𝑛) is true, this implies that A has a least element.∎

37

Definition:

The validity of both the principle of mathematical induction and strong

induction follows from a fundamental axiom of the set of integers, the

well-ordering property. The well-ordering property states that every

nonempty set of nonnegative integers has a least element. We will show

how we can directly use the well-ordering property in proofs.

true. Following four axioms about Z +:

1. The number 1 is a positive integer.

2. If n ∈ Z +, then n + 1, the successor of n, is also a positive integer.

3. Every positive integer other than 1 is the successor of a positive

integer.

4. (The Well-Ordering Property) Every nonempty subset of the set of

positive integers has a least element.

Theorem:

(The Division Algorithm) Let 𝑎 be an integer and 𝑑 a positive integer.

Then there are unique integers 𝑞 and 𝑟, with 0 ≤ 𝑟 < 𝑑, such that 𝑎 =

𝑑𝑞 + 𝑟.

Proof:

Let 𝑆 be the set of nonnegative integers of the form 𝑎 − 𝑑𝑞, where 𝑞 is

an integer.

This set is nonempty because −𝑑𝑞 can be made as large as desired

(taking 𝑞 to be a negative integer with large absolute value).

By the well-ordering property, S has a least element 𝑟 = 𝑎 − 𝑑𝑞0 . The

integer r is nonnegative. It is also the case that 𝑟 < 𝑑.

If it were not, then there would be a smaller nonnegative element in 𝑆,

namely, 𝑎 − 𝑑(𝑞0 + 1). To see this, suppose that 𝑟 ≥ 𝑑. Because 𝑎 =

𝑑𝑞0 + 𝑟, it follows that 𝑎 − 𝑑(𝑞0 + 1) = (𝑎 − 𝑑𝑞0 ) − 𝑑 = 𝑟 − 𝑑 ≥ 0.

Consequently, there are integers q and r with 0 ≤ 𝑟 < 𝑑.

The proof that 𝑞 and 𝑟 are unique:

Suppose that we have two such pairs, say (𝑞, 𝑟) and (𝑞′, 𝑟′ ), so that 𝑎 =

𝑑𝑞 + 𝑟 = 𝑑𝑞′ + 𝑟′ , with 0 ≤ 𝑟, 𝑟′ < 𝑑. We will show that the pairs are

really the same, that is, 𝑞 = 𝑞′ and 𝑟 = 𝑟′ .

𝑑𝑞 + 𝑟 = 𝑑𝑞′ + 𝑟′ ⇒ 𝑑(𝑞 − 𝑞′ ) = 𝑟′ − 𝑟.

Therefore 𝑑 | (𝑟′ − 𝑟).

(0 ≤ 𝑟 < 𝑑) ∧ (0 ≤ 𝑟′ < 𝑑) ⇒ |𝑟′ − 𝑟| < 𝑑.

The only multiple of d in that range is 0, so we are forced to

conclude that 𝑟 ′ = 𝑟.

To show that 𝑞 = 𝑞′ :

𝑎 − 𝑟 𝑎 − 𝑟′

𝑞 = = = 𝑞′ . ∎

𝑑 𝑑

38

Principle of Mathematical Induction

Definition:

A mathematical technique which is used to prove a statement, a formula

or a theorem is true for every natural number.

Step 1(Base step) − It proves that a statement is true for the initial value.

Step 2(Inductive step) − It proves that if the statement is true for the

nth iteration (or number n), then it is also true for (𝑛 + 1)𝑡ℎ iteration (or

number n+1).

Example:

1. 3𝑛 − 1 is a multiple of 2 for n = 1, 2, ...

Solution:

Step 1 − For 𝑛 = 1, 31 − 1 = 3 − 1 = 2 which is a multiple of 2

Step 2 − Let us assume3𝑛 − 1 is true for 𝑛 = 𝑘 , Hence, 3𝑘 − 1 is true

(It is an assumption)

We have to prove that 3𝑘+1 − 1 is also a multiple of 2

3𝑘+1 − 1 = 3 × 3𝑘 − 1 = (2 × 3𝑘 ) + (3𝑘 − 1)

The first part (2 × 3𝑘 ) is certain to be a multiple of @ and the second

part (3𝑘 − 1) is also true as our previous assumption.

Hence, 3𝑘+1 − 1 is a multiple of 2.

So, it is proved that 3𝑛 − 1 is a multiple of 2.∎

Solution:

Step 1 − For 𝑛 = 1,1 = 12 , Hence, step 1 is satisfied.

Step 2 − Let us assume the statement is true for 𝑛 = 𝑘.

Hence, 1 + 3 + 5 + ⋯ + (2(𝑘 + 1) − 1) = (𝑘 + 1)2 also holds

1 + 3 + 5 + ⋯ + (2(𝑘 + 1) − 1 )

= 1 + 3 + 5 + ⋯ + (2𝑘 + 2 − 1)

= 1 + 3 + 5 + ⋯ + (2𝑘 + 1)

= 1 + 3 + 5 + ⋯ + (2𝑘 − 1) + (2𝑘 + 1)

= 𝑘 2 + (2𝑘 + 1)

= (𝑘 + 1)2

So, 1 + 3 + 5 + ⋯ + (2(𝑘 + 1) − 1) = (𝑘 + 1)2 hold which

satisfies the step 2.

Hence, 1 + 3 + 5+. . . +(2𝑛 − 1) = 𝑛2 is proved.∎

Definition:

39

Let 𝑃(𝑛) be a given statement involving the natural number n such that

(𝑖) The statement is true for 𝑛 = 1, i.e., 𝑃(1) is true (or true for any fixed

natural number) and

(𝑖𝑖) If the statement is true for 𝑛 = 𝑘 (where k is a particular but

arbitrary natural number), then the statement is also true for 𝑛 = 𝑘 + 1,

i.e., truth of 𝑃(𝑘) implies the truth of 𝑃(𝑘 + 1). Then 𝑃(𝑛) is true for all-

natural number n.

Definition:

The difference between weak induction and strong induction only

appears in induction hypothesis. In weak induction, we only assume that

statement holds at k-th step, while in strong induction, we assume that

the statement holds at all the steps from the base case to k-th step.

Example:

(Weak Induction)

Prove the following statement is true for all integers 𝑛. The statement

𝑃(𝑛) can be expressed as below:

𝑛

𝑛(𝑛 + 1)

∑𝑖 =

2

𝑖=1

1) Base Case: Prove that the statement holds when n=1

1

1(1 + 1)

∑𝑖 = =1

2

𝑖=1

2) Induction Hypothesis: Assume that the statement holds when 𝑛 = 𝑘

𝑘

𝑘(𝑘 + 1)

∑𝑖 =

2

𝑖=1

3) Inductive Step: Prove that the statement holds when 𝑛 = 𝑘 + 1 using

the assumption above.

We already assumed the following fact:

𝑘

𝑘(𝑘 + 1)

∑𝑖 =

2

𝑖=1

Now we would like to consider the case of finding:

40

𝑘

∑𝑖

𝑖=1

We can interpret the formula above in this way:

𝑘+1 𝑘

∑ 𝑖 = ∑ 𝑖 + (𝑘 + 1)

𝑖=1 𝑖=1

From the induction hypothesis above, the right hand side of the

equation above can be written as:

𝑘+1 𝑘

𝑘(𝑘 + 1)

∑ 𝑖 = ∑ 𝑖 + (𝑘 + 1) = + (𝑘 + 1)

2

𝑖=1 𝑖=1

We can simplify this re-formatted right-hand side:

𝑘+1

𝑘(𝑘 + 1) (𝑘 + 1)(𝑘 + 2)

∑𝑖 = + (𝑘 + 1) =

2 2

𝑖=1

Now, we can conclude that when 𝑖 = 𝑘 + 1, the proposition we are

trying to prove holds, because under this case, 𝑛 = 𝑘 + 1 and

𝑛 + 1 = 𝑘 + 2.

Therefore, the proposition holds for all integers n.

(Strong Induction)

factored into primes. The statement P(n) is that an integer n greater

than or equal to 2 can be factored into primes.

proving 𝑃(2). 2 itself is a prime number, so the prime factorization of

2 is 2. Trivially, the statement 𝑃(2) holds.

2) Induction Hypothesis: Assume that for all integers less than or equal

to k, the statement holds.

Note: In the previous example, the assumption was only about the

case when n = k.

3) Inductive Step: Consider the number 𝑘 + 1.

Case 1: 𝑘 + 1 is a prime number.

When 𝑘 + 1 is a prime number, the number is a prime factorization

of itself.

Therefore, the statement 𝑃(𝑘 + 1) holds.

41

Case 2: 𝑘 + 1 is not a prime number.

We know that 𝑘 + 1 is a composite, so 𝑘 + 1 = 𝑝 × 𝑞(𝑝, 𝑞 ∈ 𝑍 + ).

Intuitively, we can conclude that p and q are less than or equal to

𝑘 + 1.

From the induction hypothesis stated above, for all integers less than

or equal to k, the statement holds, which means both p and q can be

expressed as prime factorizations.

In this sense, because 𝑘 + 1 is a product of p and q, by multiplying

the prime factorizations of p and q, we can get the prime factorization

for 𝑘 + 1 as well.

Therefore, the statement that every integer greater than or equal to

2 can be factored into primes holds for all such integers.

The common mistake in this question was to prove the Case 2 in the

inductive step without using induction hypothesis by dividing the

cases further into even number and odd number, etc.

It works but does not ﬁt into the notion of inductive proof that we

wanted you to learn.

For inductive step in inductive proof, you must reason your argument

based on induction hypothesis you yourself state.

Definition:

The strong induction principle says that you can prove a statement of

the form:

P(n) for each positive integer n.

As follows:

Base case: 𝑃(1) is true.

Suppose K is a positive integer such that 𝑃(1), 𝑃(2), . . . , 𝑃(𝑘) are all true.

𝑃(1), 𝑃(2), . . . , 𝑎𝑛𝑑 𝑃(𝑘) =⇒ 𝑃 (𝑘 + 1).

(1) The first domino has fallen.

(2) If k is such that the first k dominos have fallen, then the (𝑘 + 1)𝑡ℎ

domino has fallen. We now give an example of a proof related to prime

numbers using strong induction.

Example:

Every integer greater than 1 can be written as the product of prime

numbers.

42

Let 𝑃(𝑛) be the statement that 𝑛 can be written as the product of prime

numbers.

Then the proposition says: 𝑃(𝑛) is true for each integer greater or equal

to 2.

(1) Base case: 2 is a prime, so it is the product of a single prime.

(2) Strong inductive step: Suppose for some 𝑘 ≥ 2 that each integer n

with 2 ≤ n ≤ k may be written as a product of primes.

We need to prove that 𝑘 + 1 is a product of primes.

Case (a): Suppose 𝑘 + 1 is a prime. Then we are done.

Case (b): Suppose 𝑘 + 1 is a not prime. Then by the fact stated above,

there exist integers a and b with 2 ≤ 𝑎, 𝑏 ≤ 𝑘 such that

𝑘 + 1 = 𝑎 · 𝑏.

By the strong inductive hypothesis, since 2 ≤ 𝑎, 𝑏 ≤ 𝑘, both a and b

are the product of primes.

Thus 𝑘 + 1 = 𝑎 · 𝑏 is the product of primes.

(3) We are done by strong induction.

43

Exercises on Mathematical Induction

𝑛(𝑛−1)(𝑛+1)

1. 𝑃(𝑛): ∑𝑛−1

𝑡=1 𝑡(𝑡 + 1) = , for all natural numbers 𝑛 ≥ 2.

3

1

2. Prove: 12 + 22 + ⋯ + 𝑛2 = 6 𝑛(𝑛 + 1)(2𝑛 + 1)

is any real number greater than −1.

Note this means that (1 + 𝑥) > 0.

44

Answers Key: Mathematical Induction

𝑛(𝑛−1)(𝑛+1)

𝑃(𝑛): ∑𝑛−1

𝑡=1 𝑡(𝑡 + 1) = , for all natural numbers 𝑛 ≥ 2.

3

We observe that

2−1 1

1∙2∙3

𝑃(2): ∑ 𝑡(𝑡 + 1) = ∑ 𝑡(𝑡 + 1) = 1 ∙ 2 =

3

𝑡=1 𝑡=1

2 ∙ (2 − 1) ∙ (2 + 1)

=

3

Thus, 𝑃(𝑛) is true for 𝑛 = 2.

Assume that 𝑃(𝑛) is true for 𝑛 = 𝑘 ∈ ℕ.

𝑘(𝑘−1)(𝑘+1)

i.e., 𝑃(𝑘): ∑𝑘−1

𝑡=1 𝑡(𝑡 + 1) = 3

To prove that 𝑃(𝑘 + 1) is true, we have

(𝑘+1−1) 𝑘

∑ 𝑡(𝑡 + 1) = ∑ 𝑡(𝑡 + 1)

𝑡=1 𝑡=1

𝑘−1

𝑘(𝑘 − 1)(𝑘 + 1)

∑ 𝑡(𝑡 + 1) + 𝑘(𝑘 + 1) = + 𝑘(𝑘 + 1)

3

𝑡=1

𝑘(𝑘 + 1) [ ]=[ ]

3 3

(𝑘 + 1)][(𝑘 + 1) + 1]

3

Thus, 𝑃(𝑘 + 1) is true, whenever 𝑃(𝑘) is true.

Hence, by the Principle of Mathematical Induction, 𝑃(𝑛) is true for all natural

numbers 𝑛 ≥ 2.■

1

2. Prove: 12 + 22 + ⋯ + 𝑛2 = 6 𝑛(𝑛 + 1)(2𝑛 + 1)

Proof:

𝑃(1) is true, so we assume that 𝑃(𝑘) holds and work on 𝑃(𝑘 + 1).

1

12 + 22 + ⋯ + 𝑘 2 + (𝑘 + 1)2 = 6 𝑘(𝑘 + 1)(2𝑘 + 1) + (𝑘 + 1)2 using P(k).

1

(𝑘 + 1)(𝑘 + 2) + (2𝑘 + 3) which is what 𝑃(𝑘 + 1) says it should be.

6

45

3. Prove by mathematical induction: The 𝑛𝑡ℎ odd number is 2𝑛 − 1.

Initial step: The ﬁrst odd number is 1 which is 2(1) − 1 so the assertion is true

for 𝑛 = 1.

Inductive step:

Stage 1: The inductive hypothesis asserts that the 𝑛𝑡ℎ odd number is 2𝑛 − 1.

Stage 2: We want to prove that the (𝑛 + 1)𝑡ℎ odd number is 2( 𝑛 + 1) − 1, ie

2𝑛 + 1.

Stage 3: How can we get to Stage 2 from Stage 1?

By the inductive hypothesis, the 𝑛𝑡ℎ odd number is 2𝑛 − 1. \

The next odd number is 2 more than this, i.e. 2𝑛 − 1 + 2 = 2𝑛 + 1 as required.

Thus, the result is true for all 𝑛 ≥ 1.■

4. The sum of the ﬁrst 𝑛 odd numbers is 𝑛2 .

Initial step: For 𝑛 = 1, the sum is 1 which is12 .

Inductive step:

Stage 1: The inductive hypothesis asserts that

1 + 3 + 5 +··· + (2𝑛 − 1) = 𝑛2 .

(Note, we are using the result of question 3 here.)

Stage 2: We want to prove that the sum of the ﬁrst (𝑛 + 1) odd numbers is

(𝑛 + 1)2 .

Stage 3: Getting to stage 2 from stage 1.

The sum of the ﬁrst (𝑛 + 1) odd numbers is equal to the sum of the ﬁrst 𝑛 odd

numbers plus the next odd number, namely, 1 + 3 + 5 +··· + (2𝑛 − 1) +

(2𝑛 + 1), as each odd number is two more than the previous one.

So,

1 + 3 + 5 +··· + (2𝑛 − 1) + (2𝑛 + 1) = 𝑛2 + (2𝑛 + 1) the inductive

2

hypothesis = 𝑛 + (2𝑛 + 1)

= (𝑛 + 1)2 as required.

The result is therefore true for all n ≥ 1. ■

5. For all 𝑛 ≥ 1, (1 + 𝑥)𝑛 ≥ 1 + 𝑛𝑥, where 𝑥 is any real number greater than

−1.

Note this means that (1 + 𝑥) > 0.

Initial step: For 𝑛 = 1, (1 + 𝑥)1 = 1 + 𝑥 𝑎𝑛𝑑 1 + 𝑛𝑥 = 1 + 𝑥, so (1 + 𝑥)𝑏 =

1 + 𝑛𝑥 ≥ 1 + 𝑛𝑥 is true for 𝑛 = 1.

Inductive step: We assume that the proposition is true for 𝑛 = 𝑘. That is,

(1 + 𝑥) 𝑘 ≥ 1 + 𝑘𝑥, when 𝑥 ≥ −1.

We want to prove that(1 + 𝑥)𝑘+1 ≥ 1 + (𝑘 + 1)𝑥. (1)

Let us look at the inequality and try to relate it to the inductive hypothesis.

Now the left hand side of (1) is

(1 + 𝑥)𝑘+1 = (1 + 𝑥)(1 + 𝑥)𝑘

≥ (1 + 𝑥)(1 + 𝑘𝑥) using the inductive hypothesis and (1 + 𝑥) > 0

= 1 + 𝑘𝑥 + 𝑥 + 𝑘𝑥 2

≥ 1 + 𝑘𝑥 + 𝑥 since 𝑘𝑥 2 ≥ 0

46

= 1 + ( 𝑘 + 1)𝑥.

Hence (1 + 𝑥) 𝑘+1 ≥ 1 + (𝑘 + 1)𝑥,as required.

So, the result is true for all 𝑛 ≥ 1.■

47

FINITE and INFINITE SETS

FAVIOLA S. EDRADAN

48

Finite and Infinite Sets

Definition:

A set A is a finite set provided that A = ∅ or there exists a natural number

k such that A ≈ ℕk. A set is an infinite set provided that it is not a finite set.

If A ≈ ℕk, we say that the set A has cardinality k (or cardinal number k),

and we write card (A) = k.

In addition, we say that the empty set has cardinality 0 (or cardinal

number 0), and we write card (∅) = 0.

Notice that by this definition, the empty set is a finite set. In addition, for

each k ∊ ℕ, the identity function on ℕk is a bijection and hence, by definition,

the set ℕk is a finite set with cardinality k.

Theorem 1.1:

Any set equivalent to a finite nonempty set A is a finite set and has the

same cardinality as A.

Proof.

Suppose that A is a finite nonempty set, B is a set, and A≈B. Since A is a

finite set, there exists a k ∊ ℕ such that A ≈ ℕk. We also have assumed that

A≈B and so we can conclude that B≈A. Since A ≈ ℕk, we can conclude

that B ≈ ℕk. Thus, B is finite and has the same cardinality as A. ∎

The following two lemmas will be used to prove the theorem that states that

every subset of a finite set is finite.

Theorem 1.2:

If A is a finite set and x∈A, then A ⋃ {x} is a finite set and card A (A ⋃ {x})

= card (A)+1.

Proof.

Let A be a finite set and assume card (A) = k, where k = 0 or k ∊ ℕ. Assume

x ∊ A.

If A = ∅, then card(A) = 0 and A ⋃ {x} = {x}, which is equivalent to ℕ 1.

Thus, A ⋃ {x} is finite with cardinality 1, which equals card (A) + 1.

If A ≠ ∅ then A ≈ ℕk, for some k ∊ ℕ. This means that card(A) = k, and

there exists a bijection f: A ⟶ ℕk. We will now use this bijection to define

a function g: A ⋃ {x} ⟶ ℕk+1 and then prove that the function g is a

bijection. We define g: A ⋃ {x} ⟶ ℕk+1 as follows: For each t ∊ A ⋃ {x},

𝑓(𝑡) 𝑖𝑓 𝑡 ∊ 𝐴

g(t)= {

𝑘 + 1 𝑖𝑓 𝑡 = 𝑥.

that g(x1) ≠ g( x2 ).

49

If x1= x, then since x2 ≠ x1, we conclude that x2 ≠ x and hence x2 ∊

A. So g(x1) = 𝑘 + 1, and since f( x2 )= ℕk and g( x2 )= f( x2 ), we can

conclude that g(x1) ≠ g( x2 ).

Example 2:

For each natural number m, if A ⊆ ℕm, then A is a finite set and

card (A)≤ m.

Proof.

P(m) be, “If A ⊆ ℕm, then A is finite and card (A) ≤ m.”

which are finite and have cardinality less than or equal to the

cardinality of ℕ1. This proves that P(1) is true.

For the inductive step, let k ∊ ℕ and assume that P(k) is true. That is,

assume that if B ⊆ ℕk, then B is a finite set and card (B) ≤k . We need

to prove that P(k+1) is true.

Since P(k) is true, A⎯ {k+1} is a finite set and

card (A) ≤ k ≤ k + 1.

is a finite set and

card (A)= card (A ⎯ {k+1}) + 1

Since card (A ⎯ {k+1}) ≤ k, we can conclude that card (A) ≤ k + 1.

This means that we have proved the inductive step. Hence, by

mathematical induction, for each m ∊ ℕ, if A ⊆ ℕm, then A is finite

and card (A) ≤ m. ∎

Theorem 1.3:

If S is a finite set and A is a subset of S, then A is a finite set and card(A)

≤ card (S).

50

Proof.

Let S be a finite set and assume that A is a subset of S. If A = ∅,

then A is a finite set and card (A) ≤ card (S). So we assume that

A ≠ ∅.

Since S is finite, there exists a bijection f: S ⟶ ℕk for some k ∊ ℕ. In this

case, card (S)= k. We need to show that A is equivalent to a finite set.

To do this, we define g: A ⟶ f(A) by

g (x) = f (x) for each x ∊ A.

Since f is an injection, we conclude that g is an injection. Now let y ∊ f

(A). Then there exists an a ∊ A such that f(a)= y. But by the definition of

g, this means that g(a)= y, and hence g is a surjection. This proves that

g is a bijection.

Hence, we have proved that A ≈ f(A). Since f(A) is a subset of ℕk, we

use Theorem 3 to conclude that f(A) is finite and card (f(A)) ≤ k. In

addition, by Theorem 1.1, A is a finite set and card (A) =card( f(A) ).This

proves that A is a finite set and card (A) ≤ card (S).∎

Example 2:

A finite set is not equivalent to any of its proper subsets.

Proof.

Let B be a finite set and assume that A is a proper subset of B.

Since A is a proper subset of B, there exists an element x in B− A. This

means that A is a subset of B− {x}. Hence, by Theorem 1.4, card (A) ≤

card( B− {x})

Also, by Corollary 1.41, (If A is a finite set and x ∊ A, then A− {x} is a

finite set and card (A− {x})= card (A)−1.),

card (B− {x})= card (B)−1.

Hence, we may conclude that card (A) ≤ card (B)−1 and that

card (A) < card (B).

Theorem 1.1 implies that B≉A. This proves that a finite set is not

equivalent to any of its proper subsets. ∎

Definition 2:

One of the property of finite sets that we will consider is often called the

Pigeonhole Principle. The “pigeonhole” version of this property says, “If

m pigeons go into r pigeonholes and m > r, then at least one pigeonhole

has more than one pigeon.”

In this situation, we can think of the set of pigeons as being equivalent to a

set P with cardinality m and the set of pigeonholes as being equivalent to a

set H with cardinality r. We can then define a function f : P⟶ H that maps

each pigeon to its pigeonhole. The Pigeonhole Principle states that this

function is not an injection. (It is not one-to-one since there are at least two

pigeons “mapped” to the same pigeonhole.)

Theorem 2.1:

51

(The Pigeonhole Principle).

Let A and B be finite sets. If card (A) > card (B), then any function f : A⟶

B is not an injection.

Proof.

Let A and B be finite sets. We will prove the contrapositive of the theorem,

which is, if there exists a function f : A⟶ B that is an injection, then

card (A) ≤ card (B).

So assume that f : A⟶ B is an injection. As in Theorem 1.4, we define a

function g: A⟶ f(A)

g(x) = f(x) for each x ∊ A.

As we saw in Theorem 1.4, the function g is a bijection. But then

A≈ f(A) and f(A) ⊆ B. Hence,

card (A) = card ( f(A) ) and card ( f(A) ) ≤ card (B).

Hence, card (A) ≤ card (B), and this proves the contrapositive.

Hence, if card (A) > card (B), then any function f : A⟶ B is not an

injection. ∎

Definition 3:

(a) The empty set ∅ is said to have 0 elements.

(b) If n ∈ ℕ, a set S is said to have n elements if there exists a bijection

from the set ℕn: = {1,2,...,n} onto S.

(c) A set S is sad to be finite if it is either empty or it has n elements for

some n ∈ ℕ.

(d) A set S is said to be infinite it is not finite.

Since the inverse of a bijection is a bijection, it is easy to see that a set

S has n elements if and only if there is a bijection from S onto the set

{1,2,…,n}. Also, since the composition of two bijections is a bijection, we

see that a set S1 has n elements if and only if there is a bijection from S1

onto another set S2 that has n elements. Further, a set T1 is finite if and

only if there is a bijection from T1 onto another set T2 that is finite.

It is now necessary to establish some basic properties of finite sets to be

sure that the definitions do not lead to conclusions that conflict with our

experience of counting. From the definitions, it is not entirely clear that a

finite set might not have n elements for more than one value of n. Also

it is conceivable possible that the set ℕ: {1,2,3, …} might be a finite set

according to this definition.

Theorem 3.1:

(a) If A is a set with m elements and B is a set with n elements and if A

⋂ B= ∅, then A ⋃ B has m + n elements.

Proof.

(a) Let f be a bijection of ℕm onto A, and let g be a bijection of ℕn onto

B. We define h on ℕm+n by h(i)= f (i) for i=1,…, m and h(i)= g(i−m)

for i= m+1, … , m+n. ∎

Example 3:

52

Suppose that S and T are sets and that T ⊆ S.

(a) If S is a finite set, then T is a finite set.

Proof.

(a) If T = 0, we already know that T is a finite set. Thus, we may

suppose that T≠ ∅. The proof is by induction on the number of

elements in S.

If S has 1 element, then the only nonempty subset T of S must coincide

with S, so T is a finite set.

Suppose that every nonempty subset of a set with k elements is finite.

Now let S be a set having k + 1 elements (so there exists a bijection f of

ℕk+1 onto S), and let T ⊆ S. If f (k + 1) ∉ T, we can consider T to be a

subset of S1 = S\ {f (k + 1)}, which has k elements. Hence, by the

induction hypothesis, T is a finite set.

On the other hand, if f (k + 1) ∈ T, then T1 = T\ {f (k + 1)} is a subset of

S1. Since S1 has k elements, the induction hypothesis implies that T1 is

a finite set. But this implies that T = T1 U {f (k + 1)} is also a finite set. ∎

Definition 4:

For any positive integer n, let Jn be the set whose elements are

the integers 1, 2, ..., n; let J be the set consisting of all positive integers.

For any set A, we say:

be finite).

2. A is infinite if A is not finite.

3. A is countable if A ∼ J.

4. A is uncountable if A is neither finite nor countable.

5. A is at most countable if A is finite or countable.

do not say ‘countably infinite’, although you will hear this phrase from time

to time).

For two finite sets, we have A ∼ B if and only if A and B ’have the same

number of elements’. But for infinite sets this notion becomes vague, while

the idea of 1-1 correspondence (under which, given a mapping from A to

B, the image in B of x1 ∈ A is distinct from the image in B of x2 ∈ A whenever

x1 is distinct from x2) retains its clarity.

Example 4:

Let A be the set of all integers. Then A is countable. For consider the

following arrangement of the sets A and J:

53

J: 1, 2, 3, 4, 5, 6, 7, ...

We can, in this example, even give an explicit formula for a function f from

J to A which sets up a 1-1 correspondence:

𝑛

(𝑛 𝑒𝑣𝑒𝑛)

2

f(n) = { 𝑛−1

− (𝑛 𝑜𝑑𝑑)

2

Example 5:

Let R = {whole numbers between 5 and 45}

Then, R is a finite set and n(R) = 38.

Theorem 4.1:

Every infinite subset of a countable set A is countable.

Proof.

Suppose E ⊂ A, and E is infinite. Arrange the elements x of A in a

sequence {xn} of distinct elements. Construct a sequence {nk} as follows:

Let n1 be the smallest postive integer such that 𝑥𝑛 𝑘 ∈ E. Having chosen

n1, ..., nk−1 (k = 2, 3, 4, ...), let nk be the smallest integer greater than nk−1

such that xnk ∈ E.

Then, letting f(k) = 𝑥𝑛𝑘 (k = 1, 2, 3, ...), we obtain a 1-1 correspondence

between E and J. ∎

One interpretation of the theorem is that countability represents the

’smallest’ kind of infinity, in that no uncountable set can be a subset of a

countable set.

Theorem 4.2:

Let {En} be a sequence of countable sets, and put

S = ⋃∞

𝑛=1 𝐸𝑛

Then S is countable.

Proof.

Let every set En be arranged in a sequence {𝑥𝑛 𝑘 }, k = 1, 2, 3, ... and

consider the infinite array

x21 x22 x23 ...

x31 x32 x33 ...

... ... ... ...

54

in which the elements of En form the nth row. The array contains all

elements of S. We can arrange these elements in a sequence as follows:

x11; x21, x12 ; x31, x22, x13; x41, x32, x23, x14, ...

If any of the sets En have elements in common, these will appear more

than once in the above sequence. Hence there is a subset T of the set

of all positive integers such that S ∼ T, which shows that S is at most

countable. Since E1 ⊂ S, and E1 is infinite, S is also infinite, and thus

countable. ∎

Example 6:

Let S = {x : x ∈ ℤ and x2 – 81 = 0}.

Then, S = {-9, 9} is a finite set and n(S) = 2.

Definition 5:

A set S is said to be infinite if it is not finite.

Theorem 5.1:

Let A and B be sets. If A is infinite and A ≈ B, then B is infinite.

Proof.

We use a proof by contradiction and assume that A is an infinite set, A

≈ B, and B is not infinite. That is, B is a finite set. Since A ≈ B and B is

finite, Theorem 1.1 implies that A is a finite set. This is a contradiction to

the assumption that A is infinite. We have therefore proved that if A is

infinite and A ≈ B, then B is infinite. ∎

Example 7:

Set of all points in a line segment is an infinite set.

Example 8:

Set of all positive integers which is multiple of 3 is an infinite set.

Example 9:

𝕎 = {0, 1, 2, 3, ……..} i.e. set of all whole numbers is an infinite set.

Example 10:

ℕ = {1, 2, 3, ……….} i.e. set of all natural numbers is an infinite set.

55

EXERCISES

(ii){Multiples of 5}

(iii) {Fractions between 1 and 2}

(iv) {Numbers of people in India}

(v) Set of trees in the world

(vi) Set of prime numbers

(vii) Set of leaves on a tree

(viii) Set of children in all the schools of Delhi}

(ix) Set of seven natural numbers

(x) { ……, -4, -, 2, 0, 2, 4, 6, 8}

(xi) {-12, -9, -6, -3, 0, 3, 6, …… }

(xii) {Number of points in a line segment 4 cm long}.

56

ANSWER KEY

where

𝑛

if 𝑛 is even

2

f(n) = {1−𝑛

if 𝑛 is odd.

2

number, then f(n) ≤ 0. So it seems reasonable to use cases to prove

that f is a surjection and that f is an injection.

2𝑦

f(2y)= 2 = y.

If y ≤ 0, then -2y ≥ 0 and 1−2y is an odd natural number. Hence,

1−(1−2𝑦) 2𝑦

f (1−2y) = = 2 = y.

2

These two cases prove that if y ∈ ℤ, then there exists an n ∈ ℕ such that

f( n)=y. Hence, f is a surjection.

First note that if one of m and n is odd and the other is even, then one of f(m)

and f(n) is positive and the other is less than or equal to 0. So if f(m)= f(n), then

both m and n must be even or both m and n must be odd.

𝑚 𝑛

f(m)= f(n) implies that =

2 2

1−𝑚 1−𝑛

f(m)= f(n) implies that = .

2 2

proves that f(m)= f(n), then m=n and hence that f is an injection.

therefore, ℕ ≈ ℤ. Hence, ℤ is countably infinite and card (ℤ)= ℵ0. ∎

57

2. Let A be a set and b ∉ A. Then A is infinite if and only if A bijects A ⋃ {b}.

Proof. Since A is not the same size as A ⋃ {b} when A is finite, we only have

to show that A ⋃ {b} is the same size as A when A is infinite.

infinite.Here’s how: since A is infinite, it certainly has at least one

element; call it a0. But since A is infinite, it has at least two elements, and

one of them must not equal to a0; call this new element a1. But since A

is infinite, it has at least three elements, one of which must not equal

both a0 and a1; call this new element a2. Continuing in this way, we

conclude that there is an infinite sequence a0, a1, a2 , …, an , … of different

elements of A. Now it’s easy to define a bijection e : A ⋃ {b} → A:

e(b)= a0,

e(an )= an+1 for some n ∈ ℕ,

e(a)= a for a ∈ A− {b, a0, a1, …}. ∎

Proof. We can write all the positive rational numbers in a two-dimensional array

as shown in Figure 9.2.

The top row in Figure 9.2 represents the numerator of the rational

number, and the left column represents the denominator. We follow the

arrows in Figure 9.2 to define f : ℕ → ℚ+ . The idea is to start in the upper

left corner of the table and move to successive diagonals as follows:

We start with all fractions in which the sum of the numerator and

1 1

denominator is 2 (𝑜𝑛𝑙𝑦 1). So f(1)= 1.

We next use those fractions in which the sum of the numerator and

2 1

denominator is 3. So f(2)= 1 and f(3)= 2.

58

We next use those fractions in which the sum of the numerator and

1 3 2 2 1

denominator is 4. So f(4)= 3, f(5)= 1. We skipped 2 since 2 = 1. In this

way, we will ensure that the function f is a one-to-one function.

not in lowest terms. This process guarantees that the function f will be

an injection and a surjection. Therefore, ℕ ≈ ℚ+ and card (ℚ+) = ℵ0. ∎

Since x is either in A or not in A, we can consider two cases.

If x ∉ A, define g: ℕ → A ⋃ {x} by

𝑥 𝑖𝑓 𝑛 = 1

g(n)= {

𝑓(𝑛 − 1) 𝑖𝑓 𝑛 > 1.

A ⋃ {x} is countably infinite. ∎

(ii){Multiples of 5} - Infinite

59

(xi) {-12, -9, -6, -3, 0, 3, 6, …… }- Infinite

(xii) {Number of points in a line segment 4 cm long}. - Infinite

60

COUNTABLE and UNCOUNTABLE SETS

MIKAELLA G. PEDARSE

61

Countable and Uncountable Sets

Definition:

(a) A set S is said to be denumerable (or countably infinite) if there

exists a bijection of N onto S.

(b) A set S is said to be countable if it is either finite or denumerable.

(c) A set S is said to be uncountable if it is not countable.

Countable Sets

Definition:

Countable set is a set with the same cardinality (number of elements)

as some subset of the set of natural number. A countable set is either a

finite set or a countably infinite set.

A map f between sets S1 and S2 is called a bijection if f is one-to-one

and onto. In other words

If f(a) = f(b) then a = b. This holds for all a, b ∈ S1.

For each b ∈ S2, there is some a in S1 such that f(a) = b.

We write S1 ∼ S2 if there is a bijection f : S1 → S2. We say that S1 and

S2 are equivalent or have the same cardinality if S1 ∼ S2. This notion of

equivalence has several basic properties:

itself.

2. If S1 ∼ S2 then S2 ∼ S1. If f : S1 → S2 is a bijection then the inverse.

map f−1 is a bijection from S2 to S1.

3. If S1 ∼ S2 and S2 ∼ S3 then S1 ∼ S3. This boils down to the fact that

the composition of two bijections is also a bijection.

Definition:

Let N = {1, 2, 3...} denote the natural numbers. A set S is called count-

Theorem:

62

If S is both countable and infinite, then there is a bijection between S

and N itself.

Proof.

For any s ∈ S, we let f(s) denote the value of k such that s is the kth

smallest element of S. This map is well defined for any s, because

there are only finitely many natural numbers between 1 and s. It is

impossible for two different elements of S to both be the kth smallest

element of S.

Hence f is one-to-one. Also, since S is infinite, f is onto. ∎

Theorem:

Every infinite subset of a countable set A is countable.

Proof.

Suppose E ⊂ A, and E is infinite. Arrange the elements x of A in a

sequence {xn} of distinct elements. Construct a sequence {nk} as

follows:

Let n1 be the smallest positive integer such that xnk ∈ E. Having chosen

n1, ..., nk−1(k = 2, 3, 4, ...), let nk be the smallest integer greater than

nk−1 such that xnk ∈ E.

Then, letting f (k) = xnk (k = 1, 2, 3, ...), we obtain a 1-1 correspondence

between E and J.

One interpretation of the theorem is that countability represents the

‘smallest’ kind of infinity, in that no uncountable set can be a subset of

a countable set.

Lemma:

If S is countable and S′ ⊂ S, then S is countable.

Proof.

Since S is countable, there is a bijection f: S → N. But then f(S′) = N′ is

a subset of N, and f is a bijection between S′ and N′. ∎

Theorem:

Let {En} be a sequence of countable sets, and put

S= ∪ En

n=1

Then S is countable.

Proof.

Let every set En be arranged in a sequence {xnk}, k = 1, 2, 3, ... and

consider the infinite array

63

x11 x12 x13 ...

x21 x22 x23 ...

x31 x32 x33 ...

... ... ... ...

in which the elements of En form the nth row. The array contains all

elements of S. We can arrange these elements in a sequence as follows:

x11; x21, x12; x31, x22, x13; x41, x32, x23, x14, ...

If any of the sets En have elements in common, these will appear more than

once in the above sequence. Hence there is a subset T of the set of all positive

integers such that S ∼ T , which shows that S is at most countable. Since E1 ⊂

S, and E1 is infinite, S is also infinite, and thus countable. ∎

Theorem:

Let A be a countable set, and let Bn be the set of all n-tuples (a1, ..., an)

where ak ∈ A (k = 1, 2, ..., n) and the elements a1, ..., an need not be

distinct. Then Bn is countable.

Proof.

That B1 is countable is evident, since B1 = A. Suppose Bn−1 is countable

(n = 2, 3, 4, ...). The elements of Bn are of the form

(b, a) (b ∈ Bn−1, a ∈ A)

For every fixed b, the set of pairs (b, a) is equivalent to A, and thus

countable. Thus, Bn is the union of a countable set of countable sets;

thus, Bn is countable, and the proof follows by induction on n.

Theorem:

Suppose that S and T are sets and that T ⊆ S.

If S is a countable set, then T is a countable set.\

Theorem:

The following statements are equivalent:

a. S is a countable set.

c. There exists an injection of S into N.

Proof.

(a) ⟹ (b)

If S is finite, there exists a bijection h of some set Nn onto S and we

define H on N by

ℎ(𝑘) 𝑓𝑜𝑟 𝑘 = 1,· · · , 𝑛,

H(k) = {

ℎ(𝑛) 𝑓𝑜𝑟 𝑘 > 𝑛.

Then H is a surjection of N onto S.

64

If S is denumerable, there exists a bijection H of N onto S, which is also

a surjection of N onto S.

(b) ⟹ (c)

If H is a surjection of N onto S, we define H1: S → N by letting H1 (s)

be the least element in the set H-1(s) := {n ∈ N: H(n) = s}.

To see that H1 is an injection of S into N, note that if s, t ∈S and nst = H1

(s) = H1(t), then s = H(nst) = t.

(c) ⟹ (a)

If H1 is an injection of S into N, then it is a bijection of S onto H1 (S) ~

N.

By Theorem (a), H1 (S) is countable, whence the set S is countable.

Example 1:

Let A be the set of all integers. Then A is countable. For consider the

following arrangement of the sets A and J:

A : 0, 1, −1, 2, −2, 3, −3, ...

J : 1, 2, 3, 4, 5, 6, 7, ...

We can, in this example, even give an explicit formula for a function f

from J to A which sets up a 1-1 correspondence:

𝑛

f (n) = { 2 (𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛)

𝑛−1

− (𝑛 𝑜𝑑𝑑)

2

Example 2:

The set of all rational numbers.

We apply the last theorem with n = 2, noting that every rational number

can be written as b/a, where b and a are integers. Since the set of pairs

(b, a) is countable, the set of quotients b/a, and thus the set of rational

numbers, is countable.

Example 3:

The Algebraic Numbers

A real number x is called algebraic if x is the root of a polynomial

equation c0 + c1x + ... + cnxn where all the c’s are integers.

For instance, √2 is an algebraic integer because it is a root of the

equation x2−2 = 0.

To show that the set of algebraic numbers is countable, let Lk denote

the set of algebraic numbers that satisfy polynomials of the form

c0+c1x+...+cnxn where n < k and max (|cj|) < k.

Note that there are at most kk polynomials of this form, and each one

has at most k roots.

Hence Lk is a finite set having at most kk+1 elements.

As above, we make our list L1, L2, L3 of all algebraic numbers and

weed out repeaters.

65

Example 4:

The set E := {2n : n ∈ N } of even natural numbers is denumerable,

since the mapping f : N → E defined by f (n) := 2n for n ∈ N, is a

bijection of N onto E. Similarly, the set O := {2n - 1 : n ∈ N} of odd

natural numbers is denumerable.

Example 5:

Countable Unions of Countable Sets

Lemma:

Suppose that S1, S2, ... ⊂ T are disjoint countable sets. Then S = ∪i Si

is a countable set.

Proof.

There are bijections fi : Si → N for each i. Let Lk denote the set of

elements s ∈ S such that s lies in some Si for i < k, and fi(s) < k.

Note that Lk is a finite set. It has at most k2 members.

The list L1, L2, L3... contains every element of S. Weeding out

repeaters, as above, we see that we have listed all the elements of S.

Hence S is countable. ∎

Indeed, if A = {a1, a2, a3, . . .} and B = {b1, b2, b3, . . .}, we can

enumerate the elements of A U B as

a1, b1, a2, b2, a3, b3, . . .

Uncountable Sets

Definition:

Uncountable Set (or uncountably infinite set) is an infinite set that

contains too many elements to be countable. The uncountability of a

set is closely related to its cardinal number: a set is uncountable if its

cardinal number is larger than that of the set of all natural numbers.

A set is called uncountable if it is not countable.

Lemma:

If S′ ⊂ S and S′ is uncountable, then so is S.

Proof.

This is an immediate consequence of the previous lemma. If S is

countable, then so is S′. But S′ is uncountable. So, S is uncountable as

well.

Cantor's Theorem:

66

If A is any set, then there is no surjection of A onto the set P (A) of all

subsets of A.

Proof.

Suppose that ᵩ : A → P(A) is a surjection. Since ᵩ (a) is a subset of A,

either a belongs to ᵩ (a) or it does not belong to this set. We let

D := {a ∈ A : a ∉ ᵩ (a)}.

Since D is a subset of A, if ᵩ is a surjection, then D = ᵩ (a0) for some

a0 ∈ A.

We must have either a0 ∈ D or a0 ∉ D. If a0 ∈ D, then since D = ᵩ (a0),

we must have a0 ∈ ᵩ (a0), contrary to the definition of D. Similarly, if a0

∉ D, then a0 ∉ ᵩ (a0) so that a0 ∈ D, which is also a contradiction.

Therefore, ᵩ cannot be a surjection.

larger and larger sets. In particular, it implies that the collection P(N) of

all subsets of the natural numbers N is uncountable.

If T is an uncountable set, then S is an uncountable set.

Example 1:

The Set of Binary Sequences

Let S denote the set of infinite binary sequences. Here is Cantor’s

famous proof that S is an uncountable set. Suppose that f : S → N is a

bijection. We form a new binary sequence A by declaring that the nth

digit of A is the opposite of the nth digit of f−1(n). The idea here is that

f−1(n) is some binary sequence and we can look at its nth digit and

reverse it.

Supposedly, there is some N such that f(A) = N. But then the

Nth digit of A = f−1(N) is the opposite of the Nth digit of A, and this is a

contradiction.

Example 2:

The Real Numbers

Let R denote the reals. Let R′ denote the set of real numbers, between

0 and 1, having decimal expansions that only involve 3s and 7s. (This

set R′ is an example of what is called a Cantor set .) There is a bijection

67

between R′ and the set S of infinite binary sequences. For instance, the

sequence 0101001... is mapped to .3737337.... Hence R′ is uncounta-

ble.

Example 3:

The Transcendental Numbers

A real number x is called transcendental if x is not an algebraic number.

Let A denote the set of algebraic numbers and let T denote the set of

transcendental numbers. Note that R = A ∪ T and A is countable. If T

were countable then R would be the union of two countable sets. Since

R is uncountable, R is not the union of two countable sets. Hence T is

uncountable.

The upshot of this argument is that there are many more

transcendental numbers than algebraic numbers.

Example 4:

The Penrose Tiles

To each Penrose tiling P we can associate a tail end 𝜏 (P). Recall that

There is an infinite sequence P = P0, P1, P2, ... where Pn is the parent

of Pn−1. In other words, Pn is obtained from Pn−1 by the grouping

process discussed in class.

We say that the nth digit of 𝜏 (P) is a 0 if x is contained in a kite of Pn

and a 1 if x is contained in a dart of Pn. We might need to move x

slightly to avoid choosing a point that lies right on a crack. If we replace

x by x′, then only the initial part of the sequence changes. So, 𝜏(P) is

well defined.

By using the subdivision operation, we can produce a Penrose tiling P

that has any 𝜏 (P) we like. Hence, there are uncountably many different

Penrose tilings.

68

Exercises:

Part A

Prove or disprove that the following sets are countable:

1. A = {(m; n) ∈ N × N : m ≤ n}

2. A = Q100

3. The set of irrational numbers.

Part B

Prove that the following sets are countable or uncountable:

1. The set of all functions f : N → {0, 1}.

2. The set of all functions f : {0, 1} →N.

3. The set of all infinite sequences of integers.

69

ANSWER KEY

Part A

1. A = {(m; n) ∈ N × N : m ≤ n}

Solution:

A ⊆ N × N. Therefore |A| ≤ |N|. But since A is also infinite

(indeed, n is unbounded in the set), then |A| = N. Therefore A is

countable.

2. A = Q100

Solution:

Q100 = Q × Q × . . . × Q is a cross product of countable sets.

Since the finite cross product of countable sets is countable.

The set is countable.

3. The set of irrational numbers.

Solution:

Let A be the set of irrationals. Since every real number is

either rational or irrational, then A ∪ Q = R. Suppose for the

sake of contradiction that A is countable. Then since Q is

countable, A ∪ Q is also countable. Therefore R is countable.

This contradicts the fact that the set of real numbers is

uncountable. Therefore A is uncountable.

Part B

1. The set of all functions f : N → {0, 1}.

Solution:

UNCOUNTABLE.

Proof. Mapping a function f : N → {0, 1} to the

sequence (a1, a2, ) defined by ai = f(i) for each i yields a

bijection between functions of the given type and infinite binary

set of all functions f : N → {0, 1}.

2. The set of all functions f : {0, 1} →N.

Solution:

COUNTABLE.

Proof. The functions f : {0, 1} →N are in one-to-

one correspondence with N × N (map f to the tuple (a1, a2)

with a1 = f(1), a2 = f(2)). Since the latter set is countable, as a

Cartesian product of countable sets, the given set is countable

as well.

70

3. The set of all infinite sequences of integers.

Solution:

UNCOUNTABLE.

Proof: The set of all infinite integer sequences is a superset of

the set of all infinite binary (i.e., 01) sequences. But an infinite

binary sequence can be interpreted as the binary expansion

of a real number in the interval [0, 1), and since the set of

such real numbers is uncountable, so is the set of all infinite

binary sequences.

71

THE ALGEBRAIC & ORDER PROPERTY OF

REAL NUMBER

MYCHA REYES

72

THE ALGEBRAIC AND ORDER PROPERTIES OF ℝ

Definition:

Algebraic Properties of ℝ on the set ℝ of a real number there are two

binary operations, denoted by + and called addition and multiplication,

respectively.

These operations satisfy the following properties

Field Axioms of ℝ: The real numbers are a field (as are the rational numbers ℚ and

the complex numbers ℂ). That is, there are binary operations +

and defined on ℝ ∋

addition);

all a

in ℝ (existence of a zero element);

=

0 and (−a) + a = 0 (existence of negative elements);

multiplication);

multiplication);

and

73

a ⋅ 1 = a for all a in ℝ (existence of a unit element);

(11⁄𝑎 ) = a and (1⁄𝑎 ) ⋅ a = 1 (existence of a reciprocals);

a,

b, c in ℝ (distributive property of multiplication over addition).

Theorem:

(a) If z and a are elements in ℝ with z + a = a, then z = 0.

(b) If u and b ≠ 0 are elements in ℝ with u ⋅ b =b, then u = 1.

(c) If a ∈ ℝ, then a ⋅ 0 = 0.

Proof.

(a) Using (A3), (A4), (A2), the hypothesis z + a = a, and (A4), we get

z = z + 0 = z + (a +(−a)) = (z +a) + (−a) = a +(−a) = 0.

(b) Using (M3), (M4), (M2), the assumed equality u ⋅ b = b, and (M4) again,

we get u = u ⋅ 1 = u ⋅ (b ⋅ (1⁄𝑏 )) = (u⋅ b) ⋅ (1⁄𝑏 ) = b ⋅ (1⁄𝑏 ) = 1.

(c) We have why?

a + a ⋅ 0 = a ⋅ 1 ⋅ (1 + 0) = a ⋅ 1 = a.

reciprocals and the fact that the product of two numbers in zero only when one of the

factors is zero.

(b) If a ⋅ b = 0, then either a = 0 or b = 0.

Proof.

74

(a) Using (M3), (M4), (M2), the hypothesis a ⋅ b = 1, and (M3), we have

b = 1⋅ b = ((1⁄𝑎 )⋅ a) ⋅ b = (1⁄𝑎 )⋅ (a ⋅ b) = (1⁄𝑎 ) ⋅ 1 = 1⁄𝑎 .

(b) It suffices to assume a ≠ 0 and prove that b = 0. (Why?) We multiply

a ⋅ b by 1⁄𝑎 a and apply (M2), (M3) and (M3) to get

(1⁄𝑎 a) ⋅ (a ⋅ b) = ⋅ b = 1⋅b = b.

Since a ⋅ b = 0, by 2.1.2 (c) this also equals

(11⁄𝑎 ) ⋅ (a ⋅ b) = (1⁄𝑎 ) ⋅ 0 = 0.

Thus we have b = 0.

subset ℝ, by identifying the natural numbers n 𝜖 ℕ with the n- fold sum of the

unit element 1 𝜖 ℝ. Similarly, we identify 0 𝜖 ℤ with the zero element of 0 𝜖 ℝ and we

identify the n- fold sum of −1 with the integer −𝑛.Thus we consider ℕ and ℤ to be

subset of ℝ.

Proof.

Suppose, on the contrary, that p and q are integers such that (𝑝/ 𝑞)2 = 2.

We may assume that p and q are positive and no common integers factors

other than 1. (Why?) Since 𝑝2 = 2𝑞 2 we see that 𝑝2 is even. This implies

that p is also even (because if p = 2n−1 is odd, then its square 𝑝2 = 2 (2𝑛2

−2𝑛 + 1) −1is also odd). Therefore, since p and q do not have 2 as a

common factor, then q must be n odd natural numbers.

satisfied.

a < b, a = b, or a > b

75

(3) For a, b, c ∈ ℝ, if a < b, then a + c < b + c.

Proof. a < b ⟺

a + [(−𝑎) + (−𝑏)] < b + [(−𝑎) + (−𝑏)] ⟺

[𝑎 + (−𝑎) + (−𝑏)] < b + [(−𝑏) + (−𝑎)] ⟺

0 + (−𝑏) < [𝑏 + (−𝑎)] ⟺

−b < 0 + (− a) ⟺ − b < − a > −b

Proof.

c>0 ⇒ – c > 0.

𝑇𝐻4

𝑇𝐻4

Proof.

Suppose a > 0.

1 1

Since 0 < 2 < 1, 0 < 2a < a.

1

Let 𝜖0 = 2a.

Then 0 < 𝜖0 < a, contradicting our hypothesis. Thus a = 0.

1

Let 𝜖0 = 2(𝑎 − b). then

1 1 1 1 1

𝑎 − 𝜖0 = a− 2(𝑎 − b) = 𝑎 + 2 𝑏 > 2 𝑏 + 2 𝑏 = b,

2

76

Examples (1)

Determine the set A of real numbers x such that 2x + 3 ≤ 6.

2

× ∈ 𝐴 ⟺ 2x + 3 ≤ 6 ⟺ 2x ≤ 3 ⟺ x ≤ 3.

2

Therefore A = × ∈ ℝ :{ × ≤ 3}.

× ∈ 𝐴 ⟺ × 𝟐 + × − 2 > 0 ⟺ (× − 1) (× + 2) > 0.

(ii)

× − 1 > 0 and × + 2 > 0. In case (i) we must have both × >

1 𝑎𝑛𝑑 × > − 2, which is satisfied if and only The symbol ⟺ should be read “If

and only If ”

if × > 1. In case (ii) we must have to × > − 2, which is satisfied if

2x + 1

C: = {×∈ ℝ: < 1}.

x+2

We note that

2x + 1 x−1

×∈ 𝐶 ⟺ – 1 < 0 ⟺ x + 2 < 0.

x+2

0 and × + 2 > 0. (Why?) In case (i) we must have both × > 1 𝑎𝑛𝑑 × > −

77

2, which is satisfied if and only if – 2 < x < 1. In case (ii), we must

have both x > 1 and × < − 2, which is never satisfied. We conclude that C =

{x ∈ ℝ: − 2 < x < 1}.

Examples (2)

(a) Let a ≥ 0 and b ≥ 0 Then

We consider the case where a > 0 and b > 0, leaving the case a = 0 to

the reader. It follows from (i) that a + b > 0. Since 𝑏 2 – 𝑎2 = (b − a) (b + a), it

follows from

(b) that b – a > 0 implies that 𝑏 2 − 𝑎2 > 0. Also, it follows that b – a > 0

implies that 𝑏 2 − 𝑎2 > 0. reader. It follows from (i) that a + b > 0. Since

𝑏 2 – 𝑎2 = (b − a) (b + a), it follows from (c) that b – a > 0 implies that 𝑏 2 − 𝑎2

> 0. Also, it follows that b – a > 0 implies that 𝑏 2 − 𝑎2 > 0.

If a > 0 and b > 0, then √𝑎 > 0 and √𝑏 > 0. Since a = (√𝑎 )2 and b =

(√𝑏 )2, We also leave it to the reader to show that if a ≥ 0 and b ≥ 0, then

a ≤ b ⟺ 𝑎2 ≤ 𝑏 2 ⟺ √𝑎 < √𝑏

1

(c) If a and b are positive real numbers, then their arithmetic mean is 2

(a + b) and their geometric mean is √𝑎𝑏. The Arithmetic-Geometric Mean

Inequality for a, b is

1

√𝑎𝑏 ≤ 2

(a + b)

To prove this, note that if a > 0, b > 0, and a ≠ b, then √𝑎 > 0, √𝑏 > 0

and √𝑎 ≠ √𝑏. (Why?) Therefore it follows from (a) that (√𝑎 − √𝑏)2 >. Expanding

this square, we obtain

78

a−2√𝑎𝑏 + b > 0,

1

√𝑎𝑏 < 2 (a + b).

Examples (3).

𝑎 −𝑎 𝑎

Prove that − ( ) = = if b ≠ 0.

𝑏 𝑏 −𝑏

𝑎

Proof. Since b ≠ 0 we have 𝑏 ab−1 . So

𝑎

− (𝑏) = −(𝑎𝑏−1 )

= − (𝑏 −1 a)

= ((−𝑏 −1 )𝑎)

= 𝑎(−𝑏 −1 )

𝑎

= −𝑏

Similarly,

−𝑎

−(𝑎𝑏−1 ) = (−𝑎) 𝑏 −1 = . ∎

𝑏

Proof.

Uses Mathematical Induction. The case n = 1 yields equality, so that

assertion is valid in this case. Next, we assume the validity of the

inequality

79

for k ∈ ℕ and will deduce it if k + 1. Indeed, the assumptions that

(1 + 𝑥) 𝑘 ≥ 1 + k× and that 1 + × > 0 imply (why?) that.

(1 + 𝑥) 𝑘+1 = (1 + 𝑥) 𝑘 ⋅ ( 1 + ×)

≥ (1+ k×) ⋅ ( 1 + ×) = (k+1) × + k𝑥 2

≥1+(k+1)×.

all n ∈ ℕ Thus, inequality (4) holds for for n = 𝑘 + 1. Therefore, (4) holds .

Example (4).

Given x > −1, then

(b) (1 + x) 𝑟 ≥ 1 + 𝑟𝑥 for r < 0 or r< 1

Proof.

Use A.M. ≥ G.M.

𝑝

Since r ∈ Q, r=𝑞

(1+ x) 𝑟 = (1+ x) 𝑝/𝑞 = √(1 + x) 𝑝 1 𝑝−𝑞 ≤ 𝑞

𝑝𝑥 + 𝑞 𝑝

= = 1 + 𝑞 𝑥 = 1 + 𝑟𝑥

𝑞

(ii) If 1 + 𝑟𝑥 > 0, rx > −1.

80

1

Since r > 1, we have 0 < 𝑟 < 1. By (a) we get

1

(1+ x) 1/𝑟 ≤ 1 + 𝑟𝑥 = +𝑥

𝑟

∴ (1+ x) 𝑟 ≥ 1 + rx.

Let r > 0, then −𝑟 > 0.

Choose a natural number n sufficiently large such that 0 < − 𝑟/𝑛 < 1 and 1> rx/n >

−1.

1

By (a,) 0 < (1+ x) −𝑟/𝑛 ≤ 1 + 𝑥 > 0.

𝑟

𝑟 𝑟 𝑟 𝑟 𝑟

Since 1≥ 1 − ( 𝑥) 2 = (1 𝑥)(1 + 𝑥) ⇒ (1 𝑥) −1 ≥ 1+ 𝑥

𝑛 𝑛 𝑛 𝑛 𝑛

Hence by (11),

𝑟 𝑟

(1+ x) 𝑟 ≥ = (1 𝑥) 𝑛

≥ 1 + 𝑛 ( 𝑛 𝑥) = 1 + 𝑟𝑥.

𝑛

Example (5).

Prove that f(x) = (1+ x) 𝑛 on [0,x]

(1 + x) 𝑛 −1 𝑛−1

= n (1+ c)

𝑥

𝑛−1

Because n. (1+ c) > n hence

(1 + x) 𝑛 −1 𝑛−1 𝑛−1 𝑛

= n (1+ c) > n (1+ c) > n ⇒ (1+ x) −

𝑥

𝑛

>n⋅ 𝑥 ⇒ (1+ x) − 1 + 𝑛𝑥

Which complete the proof.

The case for −1 ≤ 𝑥 < 0 is similar. The case for x = 0 is simple.

Example (6).

Prove that 3𝑛 ≥ 2𝑛2 + for n = 1, 2… using binomial

𝑛

Theorem, applied (1+ x) with x =2.

81

𝑛

(1+ 2) =

𝑛 𝑛 𝑛 𝑛

( ) + ( )2 + ( ) 22 + ⋯ ( ) ≥ 1 + 2𝑛! + 2𝑛

0 0 0 𝑛

𝑛 4𝑛!

( ) 22 = (−2)!2! = 2𝑛 (n+1).

2

Then you have, since all terms are positive,

𝑛

(1+ 2) = 1 + 2n (n−1) + … = 1

+ 2𝑛2 + ⋯ ≥ 1+ 2𝑛2

EXERCISES

1

Exercise 1. Find all × ∈ ℝ ⋺ 𝑥 < 𝑥 2 .

Exercise 5: Prove that. If a ∈ ℝ is such that 0 ≤ a < 𝜀 for every 𝜀 > 0, then a = 0.

82

Answer Key:

Solution 1.

1

Proof. < 𝑥2 ⟺

𝑥

1

𝑥2 − 𝑥 > 0 ⟺

1 1

{𝑥 > 0 and 𝑥 3 − 1 > 0} 𝒐𝒓 {𝑥 < 0 and 𝑥 3 − 1 < 0} ⟺

{𝑥 > 0 and 𝑥 > 1} or {𝑥 < 0 and 𝑥 < 1} ⟺

x > 1 or x < 0.

Solution 2.

Proof 2.

a<b⟺

a + [(−𝑎) + (−𝑏)] < b + [(−𝑎) + (−𝑏)] ⟺

[𝑎 + (−𝑎)] + (−𝑏)< b + [(−𝑏) + (−𝑎)] ⟺

0 +(−𝑏) < [𝑏 + (−𝑏)] + (−𝑎) ⟺

−𝑏 < 0 + (−𝑎) ⟺ −𝑏 < −𝑎 ⟺ −𝑎 > −𝑏

Solution 3.

Proof 3. By way of contradiction, suppose b > a.

[ Need to find 𝜖 that give a contradiction.]

1

Let 𝜖0 = 2 ( 𝑎 − 𝑏). Then

1 1 1 1 1

𝑎 − 𝜖0 = − 2 ( 𝑎 − 𝑏) = 𝑎 + 2𝑏 > 𝑏 + 𝑏 = 𝑏.

2 2 2

Solution 4.

Proof 4.

̿̿̿̿ [ a + (−𝑎) ] + b 𝐴4

a + [ (−𝑎) + 𝑏 ] 𝐴2 ̿̿̿̿ 0 + b 𝐴3

̿̿̿̿ 𝑏,

so (−𝑎) + b is a solution.

For uniqueness, suppose y is any solution of the equation, i,e., a + y = b. Then

y

̿̿̿̿

𝐴3 0 + 𝑦

̿̿̿̿

𝐴4 [ (−𝑎) + 𝑎 ] + y

83

̿̿̿̿

𝐴2 (−𝑎) + (a + y)

= (−𝑎) + b

Solution 5.

1

Proof 5. Suppose to the contrary that a > 0. Then if we take 𝜀0 : 2 𝑎, we

have

0 <𝜖0 < 𝑎.

Therefore, it is false that a < 𝜀 𝑓𝑜𝑟 𝑒𝑣𝑒𝑟𝑦 𝜀 > 0 and we conclude that a = 0.

84

ABSOLUTE VALUE and REAL LINE

ANIE PADILLA

85

Absolute Value of Real Line

Definition

The absolute value or magnitude of real number a is denoted by /a/ and

defined by

𝑎 𝑖𝑓 𝑎 ≥ 0

/a/ = { }

−𝑎 𝑖𝑓 𝑎 ≥ 0

Example 1:

5/ = 5 /0/ = 0

Since 5 > 0 since 0 ≥ 0

sign if the number is negative and to leave the number unchanged if it is

nonnegative

Example 2:

Solve / x-3 / = 4

Solution.

Depending on whether x-3 is positive or negative, the equation / x-3 / =

4 can be written as

x – 3 = 4 or x – 3 = -4

solving these two equations given x = 7 and x = -1.

Example 3:

Solve / 3x – 2 / = /5x + 4/

Solution. Because two numbers with the same absolute value are either

equal or differ in sign, the given equation will be satisfied if either

3x – 2 = 5x + 4 or 3x – 2 = -(5x + 4)

Solving the first equation yields x = -3 and solving the second yields x = -

¼; thus, the given equation has the solutions x = -3 and -1/4.

Definition

The distance is the same as the absolute value of their difference.

Recall from algebra that a number is called a square root of a if its square

is a.

Recall also that every positive real number has two square roots, one

positive and one negative; the positive square root is denoted by √𝑎 and

86

the negative square root by - √𝑎. For example, the positive square root of

9 is √9 = and the negative square root of 9 is -√9 = - 3.

is nonnegative, it is false for negative a. For example, if a = -4. Then

A result that is correct for all a is given in the following theorem.

Theorem.

For any real number a,

√𝑎 = /a/

Proof.

Since 𝑎2 = (+𝑎)2 = (−𝑎)2, the numbers +a and -a are square roots of

𝑎2 .

If a ≥ 0, then +a is the nonnegative square root of 𝑎2 , and if

a < 0, 𝑡ℎ𝑒𝑛 − 𝑎 is the nonnegative square root of 𝑎2 .

√𝑎2 = +a if a ≥ 0

√𝑎2 = -a if a ≥ 0

Theorem.

If a and b are real numbers, then

(a) /-a/ = /a/

A number and its negative have the same absolute value.

The absolute value of a product is the product of the values.

The absolute value of the ratio is the ratio is the ratio of the absolute

values.

Proof

(a). From Theorem E.2,

87

/-a/ = √(−𝑎)2 = √𝑎2 = /a/

Proof

(b). From Theorem E.2 and a basic property of square roots,

The result in part (b) of Theorem E.3 can be extended to three or more

factors.

More precisely, for any n real,𝑎1 , 𝑎2 , … , 𝑎𝑛, it follows that

/𝑎1 𝑎2 … 𝑎𝑛 / = /𝑎1 //𝑎2 / …/𝑎𝑛 / (1)

In the special case where 𝑎1 , 𝑎2 … , 𝑎𝑛 have the same value, a, it follows

from that

/𝑎𝑛 / = /𝑎/ (2)

The motion of absolute value arises naturally in distances problems. For

example, suppose that A and B are points on a coordinate line that have

coordinates a and b, respectively. Depending on the relative positions of

the points, the distance d between them will be b – a or a-b (Figure E.1).

In either case the distance can be written as d = /b -a/, so we have the

following result.

A B

a b

b-a

(a)

B A

b a

a-b

(b)

Theorem (Distance Formula).

If A and B are points on a coordinate line with coordinate.

A and b, respectively, then the distance d between A and B is d = /b – a

/.

This theorem provides useful geometric interpretation of some common

mathematical expression.

88

Expression: Geometry Interpretation on a Coordinate Line

/x – a/ The distance between x and a

/x + a/ The distance between x and -a (since/x + a/ = /x – (-a)/)

/x/ The distance between x and the origin (since /x/ = /x – 0/)

expression.

The equation |x|=a | x |=a has two solutions x = a and x = -a because both

numbers are at the distance a from 0.

|x+7|=14|x+7|=14

You begin by making it into two separate equations and then solving them

separately.

x+7=14x+7=14

x+7−7=14−7x+7−7=14−7

x=7x=7

or

x+7=−14x+7=−14

x+7−7=−14−7x+7−7=−14−7

x=−21x=−21

equals a negative number since absolute number can never be negative.

The inequality

|x|<2|x|<2

89

Whereas the inequality

|x|>2|x|>2

|ax +b|>c, where c>0|ax+b|>c, where c>0

=ax + b <−c or ax + b >c=ax + b <−c or ax + b>c

When solving an absolute value inequality, it's necessary to first isolate the

absolute value expression on one side of inequality before solving the

inequality.

Example.

2|3x+9|<362|3x+9|<36

2|3x+9|2<3622|3x+9|2<362

|3x+9|<18|3x+9|<18

−18<3x+9<18−18<3x+9<18

−18−9<3x+9−9<18−9−18−9<3x+9−9<18−9

−27<3x<9−27<3x<9

90

−273<3x3<93−273<3x3<93

−9<x<3

One of the most important inequalities in calculus is

0 < │x - a│ and │x - a│< 𝛿 (3)

where 𝛿 (𝐺𝑟𝑒𝑒𝑘 dela) is a positive real number. This is equivalent to the two

inequalities

0 < │x-a│ and │x - a│<𝛿

The first of which is satisfied by all x except x = a, and the second of which

is satisfied by all x that are within 𝛿 units of a on a coordinate line.

Combining these two restrictions, we conclude that the solution set of (3)

consists of all x in the interval (a - 𝛿, a + 𝛿 ) except x = a (Figured E.5).

Stated another way, the solution set of (3) is

(a – 𝛿, a) ∪ (a, a + 𝛿) (4)

It is not generally true that │a + b│ +│a│+│b│

Example

If a = 1 and b = -1, then │a + b│ = 0, here │a│+│b│= 2. It is true, however,

that the absolute value of a sum is always less than or equal to the sum of

the absolute values. This is the content of the following useful theorem,

called triangle inequality.

E.5 Theorem (triangle Inequality)

If a and b are any real numbers, then

Proof.

Observe first that a satisfies the inequality

-│a│≤ 𝑎 ≤ │𝑎│

because either = │a│or a = -│a│, depending on the sign of a. the

corresponding inequality for b is

-│b│≤ b ≤│b│

Adding the two inequalities we obtain

-(│a│+ │b│) ≤ a + b ≤ a (│a│+│b│) (6)

91

Let us now consider the cases a + b ≥ 0 and a + b < 0 separately. In the

case, a + b =│a + b, so the right in (6) yields the triangle inequality (5).

In the second case, a + b = -│a + b│., so the left-handed inequality in (6)

can be written as

- (│a│+ │b│) ≤ --│a│+ │b│

which yields the triangle inequality (5) on multiplying by -1. ∎

Real Number

Definition:

If a is a real number, then we define the absolute value of the

number a

denoted ∣a∣ or abs(a) as:

∣a∣=⎧⎩⎨a0−aifa>0,ifa=0,ifa<0.

Example:

Suppose we want to find the absolute value of 5. Well since 5>0, we note

that ∣5∣=5.

If we wanted to find the absolute value of −5 then since −5<0 we note

that ∣−5∣=−(−5) = 5.

We will now look at some important properties of the absolute values of

real numbers utilizing the The Order Properties of Real Numbers.

Theorem 1:

If a is a real number then ∣a∣=∣−a∣.

Proof:

We will split this proof up into three cases.

Case 1:

Suppose that a>0. Then −a<0.

Therefore, by the definition of the absolute value of a number, ∣a∣=a,

and ∣−a∣=−(−a) = a, and so ∣a∣=∣−a∣

Case 2:

Now suppose that a=0.

Therefore −a=0 and clearly ∣a∣=0 and ∣−a∣=0, and so ∣a∣=∣−a∣.

Case 3:

Lastly suppose that a<0.

Then −a>0.

We obtain that ∣a∣=−a and ∣−a∣=−a, and so ∣a∣=∣−a∣.

In all three cases we get that ∣a∣=∣a∣. ■

Theorem 2:

92

If a and b are real numbers then ∣ab∣=∣a∣∣b∣.

Proof:

Case 1:

Suppose that a=0 or b=0 or both a, b=0. Then a⋅ b=0, and so ∣ab∣=0.

Similari ty ∣a∣∣b∣ is either 0⋅∣b∣ or ∣a∣⋅0 or 0⋅0, all of which equal 0,

so ∣ab∣=∣a∣∣b∣.

Case 2:

so ∣ab∣=∣a∣∣b∣.

Case 3:

and ∣a∣∣b∣=a⋅−b=−ab, so ∣ab∣=∣a∣∣b∣.

Case 4:

Suppose that a, b<0. Then ab>0 and so ∣ab∣=ab and ∣a∣∣b∣=−a⋅−b= (−1)

(−1) ab=ab. So ∣ab∣=∣a∣∣b∣.

Proof:

that a2=∣a2∣=∣a⋅ a∣=∣a∣∣a∣=∣a∣2. ■

Proof:

is equivalent to saying that − c≤ a ≤ c.

then ∣a∣≤c. ■

93

Exercise

1. │3x + 12│+ 7 =7

2. │3x – 7 │+ 7 = 9

3. │𝑥 2 + 1 │= 5

4. │𝑥 2 + 5x + 4│= 0

5. │x + 3│= 𝑥 2 – 4x – 3

94

Answers Key

1.

│3x + 12│+ 7 = 7

│3x + 12│= 0

3x + 12 = 0

3x = -12

x = -4

2.

│3x – 7 │+ 7 = 9

│3x – 7 │ = 2

3x -7 = 2 or 3x – 7 = -2

3x = 9 or 3x = 5

5

x=3 or x = 3

3.

│𝑥 2 + 1 │= 5

𝑥 2 + 1 = 5 or 𝑥 2 + 1 = −5

𝑥 2 = 4 or 𝑥 2 = -6

√𝑥 2 = √4 or √𝑥 2 = √−6

x = ± 2 or x = imaginary

4.

│𝑥 2 + 5x + 4│= 0

𝑥 2 + 5x + 4 = 0

(x + 1) (x + 4) = 0

x + 1 = 0 and x + 4 = 0

x = -1 and x = -4

5.

│x + 3│= 𝑥 2 – 4x – 3

x + 3 = 𝑥 2 – 4x – 3

𝑥 2 – 5x -6

(x – 6) (x + 1) = 0

95

x – 6 = 0 and x + 1 = 0

x = 6 and x = -1

NUMBER

ZOCHEL BASLOT

96

COMPLETENESS PROPERTY OF REAL NUMBERS

The completeness property is also known as the least upper bound property.

Definition: Let S be a nonempty subset of R.

The set S is said to be bounded above if there exists a number u ∈ R

such that s < u for all s E S. Each such number u is called an upper

bound of S.

The set S is said to be bounded below if there exists a number w ∈ R

such that s > w for all s E S. Each such number w is called an lower

bound of S.

A set is said to be bounded if it is both bounded above and below. A

set is said to be unbounded if it is not bounded.

S is said to be bounded above if there exists a number M in R such

that X < M for every x in S. M is called an upper bound for S.

S is said to be bounded below if there exists a number m in R such

that x > m for every x in S. m is called a lower bound for S.

S is said to be bounded if it is bounded above and below.

S is said to be unbounded if it lacks either an upper bound or a lower

bound.

sup S, if and only if for each y < Mo, there exists an x in S for which

y < x < Mo

1. Let us assume that Mo = sup S . We need to prove that for each y < Mo,

there exists an x in S for which y < x < Mo. We do a proof by

contradiction .Let y < Mo be given and assume that there is no element x

of S such that y < x. Then, for every x in S, x in S, x < y. Thus, y is an

97

upper bound of S which is smaller than Mo which contradicts the fact that

Mo is the supremum.

2. Let Mo by an upper bound of S with the property that for each y < Mo,

there exists an x in S for which y < x < Mo. We need to show that Mo =

sup S. Since Mo is already an upper bound, it is enough to show it is the

smallest. If Y were an upper bound strictly smaller than Mo, then by

assumption, there would exists an x in S for which y < x < Mo. But then Y

would not be an upper bound of S, which contradicts our assumption.

Thus, there cannot be an upper bound of S smaller than Mo. It follows

that Mo = sup S.

Theorem: Every nonempty subset of R that is bounded below has an infimum in R.

Proof.

We do a direct proof. We will prove the infimum exists by finding it. Let S

be a nonempty subset of R which is bounded below. Define L to be the set

of lower bounds of S. Since S is bounded below, L = ∅. Furthermore, L is

bounded above by elements of S. By the supremum property, L has a

supremum. Call it 𝛼 that is 𝛼 = sup L. We will show that 𝛼 = inf S. To

prove that 𝛼 = inf S, we first prove that 𝛼 is a lower bound of S. We then

prove that no lower bound greater than 𝛼 can exist, making 𝛼 the greatest

lower bound of S.

that every element of S is larger than 𝛼. Let s ∈ S. Then s is an upper

bound of L. Since 𝛼 = sup L, that is 𝛼 is the least upper bound of L, it

follows that 𝛼 ≤ s. We have proven that if s is an arbitrary element of S,

then we had s ≥ 𝛼. It follows that 𝛼 is a lower bound of S.

straightforward. If 𝛾 is another lower bound of S, then 𝛾 is an element of

L and therefore 𝛾 ≤ 𝛼 since 𝛼 is the least upper bound of L hence an upper

bound of L. Therefore 𝛼 is the greatest lower bound (or the infimum) of S.

Remark: In the first part of the proof, where we proved that 𝛼 is a lower bound of S,

it would have been wrong to say 𝛼 is a lower bound because 𝛼 = sup L

and L is the set of lower bounds of L. It is wrong because the supremum or

the infimum of a set do not necessarily below to the set. Thus 𝛼 is not

necessarily a lower bound of S. It turns out that it is. But we know this after

the proof we gave.

S and 𝛽 is an upper bound of S. Then, we must have 𝛼 ≤ 𝛽.

we have 𝛼 ≤ x. Because 𝛽 is an upper bound of S, we have x ≤ 𝛽. Using

the transitivity property of ≤, we obtain 𝛼 ≤ 𝛽.

98

Proposition: Let S be a subset of R.

2. If S has a largest element, then max S = sup S.

definition, m ≤ s for any s ∈ S. Thus m is also a lower bound of S. If 𝛾 is

another lower bound of S, then 𝛾 ≤ m since m ∈ S. Thus m is the greatest

lower bound of S or m = inf S.

Bounded Sets

other x in S. In this case, we say that x0 is the largest element of S

and we write x0 = max S.

other x in S. In this case, we say that x1 is the smallest element of S

and we write x1 = min S.

Completeness of R

Completeness

is complete. To understand this notion, we first need a couple of

definitions:

bound of A if x ≥ a, ∀a ∈ A.

if

(ii) if x ∈ X is an upper bound of A, then x ≥ s.

99

Lemma: The least upper bound of a set A, if it exists, is unique.

Proof. Let s1 and s2 be two least upper bounds of A. Now since s1 is an upper

bound of A by (i) of definition and s2 is a least upper bound, (ii) of

definition shows that s1 ≥ s2. Similarly since s2 is an upper bound

of A and s1 is a least upper bound, we have s2 ≥ s1. Thus s1 = s2.

Theorem: N is unbounded.

Since b − 1 < b there is an integer n ∈ N so that n > b − 1 (otherwise

b-1 would be an upper bound which is impossible). But then n + 1 > b, a

contradiction.

for all a ∈ A.

maximum (or largest element) of E, and we write

b = max E

minimum (or least element) of E, denoted by min E.

A set may have upper or lower bounds, or it may have neither. Note that a

set may have many upper and lower bounds, but if it has a maximum or a

minimum, those values are unique. Thus, we speak of an upper bound and

the maximum.

there is 𝛼 ∈ S with the following properties:

i) 𝛼 is an upper bound of E

ii) Ɐ𝛼 ∈S, 𝛾 < 𝛼 ⇒ 𝛾 is not an upper bound of E.

we write 𝛼 = sup E. Similarly, if there exists 𝛽 ∈ S such that

i) 𝛽 is a lower bound of E

ii) 𝛾 ∈ S, 𝛾 > 𝛽 ⇒ 𝛾 is not a lower bound of E.

100

Then 𝛽 is called the greatest lower bound of E or the infimum of E, and

we write 𝛽 = inf E.

and the least upper bound is that the least upper bound is not

necessarily an element of set E. The same is true between the

minimum and the greatest lower bound.

of R, known as the completeness axiom

(or the completeness axiom) if for every nonempty subset E ⊆ S that is

bounded above has a least upper bound. That is sup E exists in S.

While the completeness axiom refers only to sets that are bounded above,

the corresponding property for sets bounded below follows readily. The

following theorem tells you that every ordered set with the least-upper-

bound property also has the greatest-lower-bound property.

B be a nonempty and bounded below subset of S. Let L be the set of all lower

bounds of B. Then ∃ ⍺ ∈ S : ⍺ = sup L = inf B.

S : y ≤ x for all x ∈ B} ⊆ S implies L is bounded from above. Finally,

since S has the least upper bound property, then ∃ ⍺ ∈ S : ⍺ = sup L.

Now, if 𝛾 < ⍺ then 𝛾 is not an upper bound of L; hence 𝛾 ∉ B and it

follows (by contrapositive) that ⍺ ≤ x for all x ∈ B. Thus ⍺ ∈ L.

Moreover, if ⍺ < 𝛽 then 𝛽 ∉ L since ⍺ = sup L. We have shown that ⍺

∈ L but 𝛽 ∉ L if 𝛽 > ⍺ which means that ⍺ is a lower bound of B and 𝛽

is not if 𝛽 > ⍺. Hence ⍺ = inf B.∎

Property

a + b is a real number.

101

ab is a real number.

a+b=b+a

ab = ba

(a + b) + c = a + (b + c)

(ab)c = a(bc)

Distributive Properties

a (b + c) = ab + ac

(a + b) c = ac + ab

a+0=0+a=a

Multiplicative Identity Property

a.1 = 1.a = a

a + (-a) = 0

a.1/a = 1, a ≠ 0.

element (or maximum) is 1. More generally, if a and b are two real

numbers such that a ≤ b then min [a, b] = a and max [a, b] = b.

maximum, a number 𝛼 would have to be in [0, 1). Thus, we would have

𝛼+1 𝛼+1

𝛼 < 1. But then 𝛼 < 2 < 1, so 2 is also an element of [0, 1) which

Is larger than 𝛼. This contradicts the fact that 𝛼 = max ([0, 1)).

Theorem: Any nonempty set of real numbers which is bounded above has

a supremum.

102

Proof.

We need a good notation for a real number given by its decimal

representation. A real number has the form

a = a0.a1a2a3a4... where a0 is an integer and a1, a2, a3, ... 2 {0, 1, 2, ...9}

representation, let us decide that if the sequence of decimals ends up with

nines: a = a0.a1a2...an9999... (where an < 9) then we choose this number’s

decimal representation as a = a0.a1a2...(an + 1)0000.... (For example,

instead of 0.4999999.. we write 0.5.)

the least upper bound of S. Consider first all the approximations by

integers of the numbers a of S: if a = a0.a1a2... collect the a0’s. This is a

collection of integer numbers. It is bounded above (by assumption). Then

there is a largest one among them, call it B0.

Next collect only the numbers in S which begin with B0. (There are some!)

Call their collection S0. Any number in S \S0 (number of S not in S0) is

smaller than any number in S0. Look at the first decimal a1 of the numbers in

S0. Let B1 be the largest among them. Let S1 be the set of all numbers in S0

whose first decimal is B1. Note that the numbers in S1 begin with B0.B1

Also note that any number in S \ S1 is smaller than any number in S1. Next

look at the second decimal of the numbers in S1. Find the largest, B2 etc.

Repeating the procedure we construct a sequence of smaller and smaller

sets S0, S1, S2, ...Sn, ...

Note that every set Sn contains al least one element (it is not empty). At

each

step n we have constructed the set Sn of numbers of S which start with

B0.B1B2...Bn; the rest of the decimals can be anything. Also all numbers

S \ Sn are smaller than all numbers of Sn. (The construction is by induction!)

B

is the least upper bound. To show it is an upper bound, let a ∈ S. If a0 < B0

then a < B. Otherwise a0 = B0 and we go on to compare the first decimals.

Either a1 < B1 therefore a < B or, otherwise, a1 = B1. Etc. So either a < B or

a = B. So B is an upper bound.

To show it is the least (upper bound), take any smaller number t < B. Then t

differs from B at some first decimal, say at the nth decimal:

t = B0.B1B2...Bn−1tntn+1... and tn < Bn. But then t is not in Sn and Sn

103

contains numbers bigger than t.

(−1)𝑛

Example 3: Consider the set A = f{ : n ∈ N}.

𝑛

a. Show that A is bounded from above. Find the supremum. Is this supre-

mum a maximum of A?

SOLUTION:

A: We will show that 1/2 ≤ M: Suppose the contrary. That is, suppose

that

(−1)𝑛

M < 1/2: Since M is an upper bound of A, we have ≤ M for all n ∈

𝑛

N: In particular, letting n = 2 we obtain 1/2 ≤ M < 1/2 which is

impossible. Thus, 1/2 ≤ M so that sup {A} = 1/2: Since the supremum

is an element of A we conclude that 1/2 is also the maximum of A:

b. Show that A is bounded from below. Find the infimum. Is this infimum

a minimum of A?

SOLUTION:

show that m ≤ -1: Suppose the contrary. That is, suppose that m > -1:

(−1)

Letting n = 1 we find that -1 = 1 ≥ m > -1; which is impossible.

Therefore, we must have m ≤ -1: This establishes that inf {A} = -1: Since

-1 is in A; it is the minimum of A.

104

EXERCISES

(a) Show that A is bounded from above. Find the supremum. Is this

supremum a maximum of A?

(b) Show that A is bounded from below. Find the infimum. Is this infimum a

minimum of A?

𝐿+2

(b) Let L be a lower bound of A such that L > 2. Let y = 2 Show that

2<y<L.

(c) Show that y ∈ A and L ≤ y: Show that this leads to a contradiction.

Hence, we must have L ≤ 2 which means that 2 is the infimum of A:

3. Suppose that α = sup A < ∞. Let 𝜖> 0 be given. Prove that there is an

x ∈ A such that 𝛼 − 𝜖 < x.

4. Suppose that 𝛽 = inf A < ∞: Let 𝜖 > 0 be given. Prove that there is an

x ∈ A such that 𝛽 +𝜖 > x.

Show that each of these numbers is irrational.

105

ANSWER KEY

(a) Show that A is bounded from above. Find the supremum. Is this

supremum a maximum of A?

(b) Show that A is bounded from below. Find the infimum. Is this infimum a

minimum of A?

Solution:

A: We will show that 2 ≤ M: Suppose the contrary. That is, suppose

that 1 < M < 2: Let r be a rational number such that M < r < 2: Then

r ∈ A and M < r which contradicts the fact that M is an upper bound

of A. Hence, we must have 2 ≤ M so that sup {A} = 2. Since the

supremum is not an element of A we conclude that 2 is not a

maximum of A.

(b) Clearly, 1 is a lower bound of A. Let m be a lower bound of A. We

will show that m ≤ 1. That is, suppose that 1 < m < 2.

Let r be rational number such that 1 < r < m. Then r ∈ A and r < m

which contradicts the fact that m is a lower bound of A. Thus, we

must have 1 ≤ m so that inf {A} = 1. Since 1 is not in A, it is not a

minimum of A. ∎

𝐿+2

(b) Let L be a lower bound of A such that L > 2. Let y = 2 Show that

2<y<L.

(c) Show that y ∈ A and L ≤ y: Show that this leads to a contradiction.

Hence, we must have L ≤ 2 which means that 2 is the infimum of A:

Solution:

𝐿+2

(b) Since L > 2 we have L + 2 > 4 and this implies y = 2 > 2. Also,

𝐿+2 𝐿+𝐿

y = 2 < 2 = L.

(c) Since y > 2 we have y2 > 4 so that y ∈ A: But L is a lower bound of

A so we must have L ≤ y: But this contradicts y < L from (b). It follows

that 2 is the greatest lower bound of A. ∎

3. Suppose that α = sup A < ∞. Let 𝜖 > 0 be given. Prove that there is an

x ∈ A such that 𝛼 − 𝜖 < x.

106

Solution:

𝛼 − 𝜖 is an upper bound of A. Thus, we must have 𝛼 ≤ 𝛼 − 𝜖 which

` is impossible. ∎

4. Suppose that 𝛽 = inf A < ∞: Let 𝜖 > 0 be given. Prove that there is an

x ∈ A such that 𝛽 +𝜖 > x.

Solution:

𝛽 + 𝜖 is a lower bound of A: Thus, we must have 𝛽 + 𝜖 ≤ 𝛽 which is

Impossible.

Show that each of these numbers is irrational.

Solution:

which is an irrational number. Suppose that sk is irrational for

k = 1, 2, … n: We want to show that sn+1 is irrational. Suppose the

contrary, then sn = s2n+1 − 2 is rational which contradicts the

assumption that sn is irrational. Hence, sn+1 must be irrational

107

APPLICATIONS OF SUPREMUM PROPERTY

JEASON GALINDEZ

108

The Supremum Property

Definition:

Every nonempty set of real numbers that is bounded above has a

supremum which is a real number.

Definition:

Every nonempty set of real numbers that is bounded below has an

infimum which is a real number.

Theorem:

The Supremum Property and the Completeness Axiom are equivalent.

This is an if and only if statement.

Proof.

Assume the Completeness Axiom and show that sup X and inf X exist

and are a real numbers.

Let X ⊆ R be a nonempty set that is bounded above.

Let U be the set of all upper bounds for X.

Since X is bounded above, U ≠ ∅. If x X and u U, x ≤ u, since u is

an upper bound for X.

So, x ≤ u ∀x X; u U

By the Completeness Axiom,

Ǝα R such that x ≤ α ≤ u ∀x X, u U

α is an upper bound for X, and it is less than or equal than every other

upper bound for X, so it is the least upper bound for X, so sup X = α

R.

The case in which X is bounded below is similar. (show it) Thus, the

Supremum Property holds.

Assume the Supremum Property and show that the completeness axiom

holds.

Suppose L, H ⊆ R, L ≠ Ø H, and,

l ≤ h for al l L, h H

exist and it is real number, so let α = supL

l ≤ α for all l L

of supremum, α ≤ h.

109

l ≤ α ≤ h for al l L, h H

so the Completeness Axiom holds.

Every nonempty set of real numbers that is bounded above has a

supremum, which is a real number.

Every nonempty set of real numbers that is bounded below has an

inÖmum, which is a real number.

The supremum property is useful to prove other properties of real

numbers.

Definition:

For all x,y R, y > 0, Ǝn N such that x < ny = (y + . . . + y) → n times.

Theorem:

The set of real numbers (an ordered field with the Least Upper Bound

property) has the Archimedean Property.

This is the proof I presented in class. It is one of the standard proofs.

The key is the following Lemma.

Lemma:

The set N of positive integers N = {0, 1, 2, . . .} is not bounded from above.

Proof Reasoning by contradiction, assume N is bounded from above.

Since N ⊂ R and R has the least upper bound property, then N has a

least upper bound α ∈ R.

Consequently α − 1 is not an upper bound for N (if it were, since α − 1 <

α, then α would not be the least upper bound).

Therefore, there is some integer k with α − 1 < k. But then α < k + 1.

This contradicts that α is an upper bound for N.

Proof:

equivalent to finding an n with n > y/x for some n, But if there is no such

n then n < y/x for all integers n.

That is, y/x would be an upper bound for the integers. This contradicts

the Lemma.

110

And we say b is a minimum for X if,

Maximum and minimum do not always exist if the set is bounded, but the

sup and the inf do always exist if the set is bounded.

If the sup and inf are also elements if the set, then they coincide with the

max and the min.

Simple Result

Theorem:

Given a set X ⊆ R, if max X exist it is equal to sup X.

Proof:

TO do this we must show: that (1) a is an upper bound, and (2) for

every other upper bound a’. we have a’ ≥ a.

a is an upper bound on X since a ≥ x for all x X by definition of max X.

suppose a’ < a.

(a’ cannot be an upper bound).

Definition:

An essential property of the natural numbers is the following induction

principle, which expresses the idea that we can reach every natural

number by counting upwards from one.

1 ∈ A; (b) n ∈ A implies (n + 1) ∈ A. Then A = N. Before stating the next

theorem, we give a formal definition of the maximal and minimal

elements of a set.

111

Suppose that A ⊂ R is a set of real numbers. A maximal element of A is

a number M = max A such that M ∈ A and M ≥ x for every x ∈ A. A

minimal element of A is a number m = min A such that m ∈ A and m ≤ x

for every x ∈ A.

minimal element, then sup A = max A or inf A = min A. However, sup A

or inf A may exist, even when max A or min A don’t exist, if they don’t

belong to A. The following result states a fundamental property of the

natural numbers called the well-ordering property.

Theorem:

Every nonempty subset of natural numbers contains a minimal

element.

Proof:

element, then S is the empty set, which proves the result.

Let

A = {n ∈ N : 1, 2, . . . , n /∈ S} .

n for every n ∈ N, which contradicts our assumption on S.

Hence, 1 ∈ A.

n /∈ S. If n+ 1 ∈/ A, then n+ 1 ∈ S, so n+ 1 is a minimal

element of S. This contradiction shows that n + 1 ∈ A. It

follows by induction that A = N, which implies that

S = ∅.

implies the induction axiom, so the well-ordering of N is

equivalent to induction. The integers Z are not well-ordered

(for example, Z itself has no minimal element), but a similar

property holds for sets of integers that are bounded from

below.

Corollary:

If a nonempty set S ⊂ Z is bounded from below in R, then S has a

minimal element

Proof.

implies that S is bounded from below by some integer a ∈ Z. Then

B = {n − a + 1 : n ∈ S} ⊂ N,

112

so it implies that B has a minimal element b ∈ B. It follows that m

= b + a − 1 ∈ S is a minimal element of S.

We can now prove the following (intuitively obvious) lemma,

which we will use to prove the density of Q in R.

Lemma:

If x, y ∈ R satisfy y − x > 1, then there exists m ∈ Z such that x < m < y.

Proof:

S = {n ∈ Z : x < n} .

Then S ⊂ Z is nonempty (by the Archimedean property) and

bounded from below (by x), so it has a minimal element m ∈ S

with m − 1 ∈/ S. It follows that x < m and x ≥ m − 1, so x < m ≤ x

+ 1 < y, which proves the result.

Theorem:

If x, y ∈ R and x < y, then there exists r ∈ Q such that x < r < y

Proof:

Let = y−x > 0. By the Archimedean property, there exists n ∈ N such

that 0 < 1/n < , which implies that ny − nx > 1. By Lemma 5, there

exists m ∈ Z such that nx < m < ny, which proves the result with r = m/n.

infinitely many rational numbers between any pair of distinct real

numbers. One can also use the result to prove that the irrational

numbers R \ Q are dense in R.

numbers in the real numbers, namely that between any two real

numbers there exists a rational numbers.

Theorem 1:

Then there exists a rational number r ∈ Q such that x < r < y.

Proof:

Suppose that x > 0. Since x < y we have that y > 0 and furthermore we

have that y – x > 0.

Now we know by the Archimedean properties that since y – x > 0, then

there exists a natural number n ∈ N such that 1n < y − x.

If we multiply this out we get that 1 < ny − nx or rather nx + 1 < ny. Now

we know that since n > 0 and since x > 0, and by the Archimedean

properties that since nx > 0 then there exists a natural number, call it A

∈ N such that A – 1 ≤ nx < A or equivalently A ≤ nx + 1 ≤ A + 1.

113

Therefore nx ≤ A ≤ nx + 1 ≤ ny and so nx < A < ny and so the rational

number r = An works for x < r < y.

Theorem 2:

Let x,y ∈ R be any two real numbers where x < y. Then there exists an

irrational number q ∈ R ∖Q such that x < q < y.

Proof:

Consider the real numbers x / √2 and y / √2. By the theorem above there

exists a rational number r such that:

(1)

X / √2 < r < y / √2

again to get an r ∗ ∈ Q such that:

(2)

X / √2 < r∗ < r < y / √2 ⇒ x / √2 <r∗ < y / √2

And we can choose the r∗ instead of r. So we assume that r≠0.

We multiply the inequality at (∗) by √2 to get x < r √2 < y, and let q

= r √2. Since r ≠ 0 and r ∈ supQ we have that q is an irrational

number.

114

Exercises:

(a) Show that A is bounded from above. Find the supremum. Is this supremum

a maximum of A?

(b) Show that A is bounded from below. Find the infimum. Is this infimum a

minimum of A?

(a) Show that A is bounded from above. Find the supremum. Is this supremum

a maximum of A?

(b) Show that A is bounded from below. Find the infimum. Is this infimum a

minimum of A?

(a) What is a lower bound of A?

(b) Let L be a lower bound of A such that L > 2. Let y = L+2 2 . Show that 2 < y

< L.

(c) Show that y ∈ A and L ≤ y. Show that this leads to a contradiction. Hence,

we must have L ≤ 2 which means that 2 is the infimum of A.

4.) Show that for any real number x there is a positive integer n such that n > x.

5.) Let a and b be any two real numbers such that a < b.

(a) Let w be a fixed positive irrational number. Show that there is a rational

number r such that a < wr < b.

(b) Show that wr is irrational. Hence, between any two distinct real numbers

there is an irrational number.

6.) Suppose that α = sup A < ∞. Let > 0 be given. Prove that there is an x ∈ A such

that α − < x.

7.) Suppose that β = inf A < ∞. Let > 0 be given. Prove that there is an x ∈ A such

that β + > x

8.) For each of the following sets S find sup {S} and inf {S} if they exist. You do not

need to justify your answer.

(a) S = {x ∈ R : x 2 < 5}.

(b) S = {x ∈ R : x 2 > 7}.

(c) S = {−1 n : n ∈ N}.

9.) (a) Show that for any positive numbers a and b we have a+b / 2 ≥ √ab.

(b) Let ai > 0 for i = 1, 2, · · · , n. Suppose that n√ a1a2 · · · an = 1. Use (a) to

show that (1 + a1)(1 + a2)· · ·(1 + an) ≥ 2 n .

10.) Consider the numbers s1, s2, · · · where s1 = √ 2 and sn+1 = √ 2 + sn for n ∈ N.

Show that each of these numbers is irrational.

115

Answers

1.) (a) Clearly, 1 / 2 is an upper bound of A. Let M > 0 be an upper bound of A. We will

show that 1 / 2 ≤ M. Suppose the contrary.

That is, suppose that M < 1 2 . Since M is an upper bound of A, we have (−1)n

n ≤ M for all n ∈ N. In particular, letting n = 2 we obtain 1 / 2 ≤ M < 1 / 2 which is

imposssibe. Thus, 1 / 2 ≤ M so that sup{A} = 1 / 2 .

maximum of A. (b) Clearly, −1 is a lower bound of A. Let m be a lower bound of A. We

will show that m ≤ −1. Suppose the contrary. That is, suppose that m > −1.

we must have m ≤ −1. This establishes that inf{A} = −1. Since −1 is in A, it is the

minimum of A

2.) (a) Clearly, 2 is an upper bound of A. Let M > 1 be an upper bound of A. We will

show that 2 ≤ M.

Suppose the contrary. That is, suppose that 1 < M < 2. Let r be a rational

number such that M < r < 2. Then r ∈ A and M < r which contradicts the fact that M is

an upper bound of A.

element of A we conclude that 2 is not a maximum of A.

that m ≤ 1.

Let r be a rational number such that 1 < r < m. Then r ∈ A and r < m which

contradicts the fact that m is a lower bound of A. Thus, we must have 1 ≤ m so

that inf{A} = 1. Since 1 is not in A, it is not a minimum of A

(b) Since L > 2 we have L + 2 > 4 and this implies y = L+2 / 2 > 2.

Also, y = L+2 / 2 < L+L / 2 = L.

(c) Since y > 2 we have y 2 > 4 so that y ∈ A. But L is a lower bound of A so we

must have L ≤ y. But this contradicts y < L from (b). It follows that 2 is the

greatest lower bound of A.

5.) (a) Since a < b, we have a w < b w ., there is a rational number r such that a w < r

< b w or a < rw < b.

(e) If rw = s with s rational then w = s r which is a rational, a contradiction. Hence,

rw is irrational

116

6.) Suppose the contrary. That is, α − ≥ x for all x ∈ A. In this case, α − is an

upper bound of A. Thus, we must have α ≤ α − which is impossible.

7.) (a) Note that S = {x ∈ R : − √ 5 < x < √ 5}. So √ 5 is an upper bound of S. Let M be

an upper bound of S.

Suppose that M < √ 5. Let r be a rational number such that M < r < √ 5.

Then r ∈ S and M < r.

But this contradicts the fact that M is an upper bound of S. Thus, √ 5 ≤ M so

that sup{S} = √ 5.

Likewise one can show that inf{S} = − √ 5.

(c) sup{S} = 0 and inf{S} = −1

we find

( 1 + a1 / 2) (1 + a2 / 2) · · · (1 + an / 2) ≥ √ a1a2 · · · an = 1

…Hence, (1 + a1)(1 + a2)· · ·(1 + an) ≥ 2 n

For n = 1 we have s1 = √ 2 which is an irrational number. Suppose that sk is

irrational for k = 1, 2, · · · , n. We want to show that sn + 1 is irrational.

the assumption that sn is irrational. Hence, sn + 1 must be irrational.

117

INTERVALS

SALDIE G. INCINADA

118

INTERVALS

Interval Notation

Interval notation uses parentheses and brackets to describe sets of real numbers

and their endpoints.

Learning Objectives

Use interval notation to represent sets of numbers

Key Takeaways

Key Points

A real interval is a set of real numbers with the property that any number that lies

between two numbers included in the set is also included in the set.

The interval of numbers between a and b, including a and b, is denoted [a,b]. The

two numbers a and b are called the endpoints of the interval.

To indicate that an endpoint of a set is not included in the set, the square bracket

enclosing the endpoint can be replaced with a parenthesis.

An open interval does not include its endpoints, and is enclosed in parentheses.

A closed interval includes its endpoints, and is enclosed in square brackets.

An interval is considered bounded if both endpoints are real numbers. An

interval is unbounded if both endpoints are not real numbers.

Replacing an endpoint with positive or negative infinity—e.g., (−∞,b ]

— indicates that a set is unbounded in one direction, or half-bounded.

Key Terms

Interval: A distance in space.

Bounded interval: A set for which both endpoints are real numbers.

Open interval: A set of real numbers that does not include its endpoints.

Endpoint: Either of the two points at the ends of a line segment.

Half-bounded interval: A set for which one endpoint is a real number and the

other is not.

Closed interval: A set of real numbers that includes both of its endpoints.

Unbounded interval: A set for which neither endpoint is a real number.

FORMAL DEFINITION

A “real interval” is a set of real numbers such that any number that lies

between two numbers in the set is also included in the set. For example, the

set of all numbers x satisfying 0≤x≤1 is an interval that contains 0 and 1, as

well as all the numbers between them. Other examples of intervals include the

set of all real numbers and the set of all negative real numbers.

The interval of numbers between a and b, including a and b, is often denoted

[a,b]. The two numbers are called the endpoints of the interval.

An open interval does not include its endpoints and is indicated with parentheses.

For example, (0,1) describes an interval greater than 0 and less than 1.

A closed interval includes its endpoints and is denoted with square brackets

rather than parentheses.

For example, [0,1]

119

- describes an interval greater than or equal to 0 and less than or equal

to 1.

To indicate that only one endpoint of an interval is included in that set,

both symbols will be used. For example, the interval of numbers between 1

and 5, including 1 but excluding 5, is written as [1,5)

.

The image below illustrates open and closed intervals on a number line.

Intervals: Representations of open and closed intervals on the real number line.

An interval is said to be bounded if both of its endpoints are real numbers.

Bounded intervals are also commonly known as finite intervals. Conversely, if

neither endpoint is a real number, the interval is said to be unbounded.

For example,

The interval (1,10) is considered bounded; the interval (−∞,+∞) is considered

unbounded.

The set of all real numbers is the only interval that is unbounded at both ends;

the empty set (the set containing no elements) is bounded.

or more descriptively, left-bounded or right-bounded. For example, the interval

(1,+∞) is half-bounded; specifically, it is left-bounded.

We have just looked at Open and Closed Intervals. The following theorem will

verify that if we have an interval, then this interval contains all points in

between its endpoints, for example, the interval (2,4) must contain all values

of x such that 2<x<4, e.g., 3∈(2,4).

Let S⊆R that contains at least two points. Then if S has the property such that

if x<y and x,y]⊆S, then S is an interval.

Proof:

120

We will consider four cases to this theorem.

Case 1: Suppose that the set S is bounded. Since this set is nonempty and

bounded, by the completeness property, this set contains an supremum (and

similarly, an infimum).

Let a=infS and let b=supS. Therefore, for all s∈S we have that a≤s≤b and so

S⊆[a,b]. Now we want to show that (a,b)⊆S. Let z∈(a,b). Therefore a<z<b.

Therefore, z is not a lower bound of the set S, so there exists an x∈S such

that x<z. Similarly, z is not an upper bound of the set S so there exists a y∈S

such that z<y. Combining these two inequalities we get that x<z<y. Now since

x<y and [x,y]⊆S, we thus have that z∈[x,y]. Therefore (a,b)⊆S. So S is an

interval. Either S=[a,b], S=[a,b), S=(a,b] or S=(a,b) depending on whether the

end points a,b are contained in S or not.

Case 2: Suppose that S is bounded above only. Then let b=supS. Therefore,

∀s∈S we have that s≤b and so S⊆(−∞,b]. We now want to show that

(−∞,b)⊆S. Let z∈(−∞,b). Then z<b, so z is not an upper bound to this set. So

there exists a y∈S such that z<y.

Furthermore, since this set is not bounded below, then there exists an x∈S

such that x<z. Combining these inequalities we have that x<z<y, and so x<y

and [x,y]⊆S so z∈[x,y], and once again, since z is arbitrary this implies that

(−∞,b)⊆S. Therefore S=(−∞,b] or S=(−∞,b) depending on whether b∈S..

Case 3: Suppose that S is bounded below only. Then let a=infS. Therefore

∀s∈S we have that a≤s and so S⊆[a,∞). We now want to show that (a,∞)⊆S.

Let z∈(a,∞). Then a<z, so z is not a lower bound to this set. So there exists an

x∈S such that x<z.

Furthermore, since this set is not bounded above, then there exists a y∈S

such that z<y. Combining these inequalities we have that x<z<y, and so x<y

and [x,y]⊆S so z∈[x,y], and since z is arbitrary this implies that (a,∞)⊆S.

Therefore S=[a,∞) or S=(a,∞) depending on whether a∈S.

Case 4: Suppose that S is not bounded above and not bounded below. Then

clearly S⊆(−∞,∞). We now want to show that (−∞,∞)⊆S. Let z∈(−∞,∞). Then

since this set is unbounded there exists x,y∈S such that x<z<y. We have that

x<y and [x,y]⊂S, so then z∈[x,y]. Since z is arbitrary this implies that

(−∞,∞)⊆S. Therefore S=(−∞,∞)

We have just looked at what exactly a Nested Interval is, and we are about to

look at a critically important theorem in Real Analysis. Before we look at the

Nested Intervals Theorem let's first look at the following important lemma that

will be used to prove the Nested Intervals Theorem.

Lemma 1:

Let a<b and let c<d and let I=[a,b] and J=[c,d]. Then I⊆J if and only if

c≤a<b≤d.

121

Theorem 1:

If the interval In=[an,bn] for n∈N is a sequence of closed bounded nested

intervals then there exists a real number ξ=sup{an:n∈N} such that ξ∈⋂∞n=1In.

Proof of Theorem:

We note that by the definition of nested intervals that In⊆I1 for all n∈N so then

an≤b1.

Now consider the nonempty set A={an:n∈N} that is bounded above by b1.

Thus, this set has a supremum in the real numbers and denote it supA=ξ so

that an≤ξ for all n∈N.

We now want to show that ξ≤bn for all n∈N. We will do this by showing that

an≤bk for all n,k∈N.

First consider the case where n≤k. We thus have that In⊇Ik by the definition of

nested intervals and so by lemma 1 we get that an≤ak≤bk≤bn. Here we see

that an≤bk.

Now consider the case where n>k. We thus have that Ik⊇In by the definition

of nested intervals and so by lemma 1 once again we have that ak≤an≤bn≤bk.

Once again we have that an≤bk.

So then an≤bk for all n,k∈N, and so bk is an upper bound to the set A and so

supA=ξ≤bk for all k∈N. Furthermore we have that ak≤ξ for all k∈N, and so

ξ∈Ik for every k∈N and thus ξ∈⋂∞n=1In and so the set theoretic union is

nonempty. ■

Theorem 2:

If the interval In=[an,bn] for n∈N is a sequence of closed bounded nested intervals

then there exists a real number η=inf{bn:n∈N} such that η∈⋂∞n=1In.

Proof:

We note that by the definition of nested intervals that In⊆I1 for all n∈N so then

a1≤bn.

Now consider the nonempty set B={bn:n∈N} that is bounded below by a1.

Thus this set has an infimum in the real numbers and denote it infB=η so that

η≤bn for all n∈N.

We now want to show that an≤η for all n∈N. We will do this by showing that

ak≤bn for all n,k∈N.

First consider the case where n≤k. We thus have that In⊇Ik by the definition of

nested intervals and so by lemma 1 we get that an≤ak≤bk≤bn. Here we see

that ak≤bn.

Now consider the case where n>k. We thus have that Ik⊇In by the definition

of nested intervals and so by lemma 1 once again we have that ak≤an≤bn≤bk.

Once again we have that ak≤bn.

122

So then ak≤bn for all n,k∈N, and so ak is an upper bound to the set A and so

ak≤η=infB for all k∈N. Furthermore we have that η≤bk for all k∈N, and so η∈Ik

for every k∈N and thus η∈⋂∞n=1In. ■

Theorem 3:

Let A:={an:n∈N} and B:={bn:n∈N}. If supA=ξ and infB=η then if the

interval In=[an,bn] is a sequence of closed bounded nested intervals then

[ξ,η]=⋂∞n=1In.

Proof:

ξ≤x≤η. But we know that an≤ξ for all n∈N and we know that η≤bn for all n∈N

and so an≤ξ≤x≤η≤bn for all n∈N. Therefore x∈[an,bn] for all n∈N or in other

words x∈⋂∞n=1In. Therefore [ξ,η]⊆⋂∞n=1In.

We will now prove that ⋂∞n=1In⊆[ξ,η]. Let x∈⋂∞n=1In. Then an≤x≤bn for all

n∈N. We also know that an≤ξ≤η≤bn for all n∈N. Suppose that x is such that

an≤x≤ξ. Since ξ is the supremum of the set A then there exists an element

ax∈{an:n∈N} such that x<ax and so x∉[ax,bx] and therefore x∉⋂∞n=1In, a

contradiction. Now suppose that x is such that η≤x≤bn for all n∈N. Since η is

the infimum of the set B then there exists an element bx∈{bn:n∈N} such that

bx<x and so x∉[ax,bx] and so x∉⋂∞n=1In,

Once again, a contradiction. We must therefore have that an≤ξ≤x≤η≤bn and

so x∈[ξ,η] and so ⋂∞n=1In⊆[ξ,η].

Since [ξ,η]⊆⋂∞n=1In and ⋂∞n=1In⊆[ξ,η] we have that [ξ,η]=⋂∞n=1In

Set-Builder Notation

In this text, we use interval notation. However, other resources that you are likely

to encounter use an alternate method for describing sets called set-builder

notation. We have used set notation to list the elements such as the integers

The braces group the elements of the set and the ellipsis marks indicate that the

integers continue forever. In this section, we wish to describe intervals of real

numbers—for example, the real numbers greater than or equal to 2.

Since the set is too large to list, set-builder notation allows us to describe it using

familiar mathematical notation. An example of set-builder notation follows:

123

Here x∈R describes the type of number, where the symbol (∈) is read “element of.”

This implies that the variable x represents a real number.

The vertical bar (|) is read “such that.”

Finally, the statement x≥2 - is the condition that describes the set using

mathematical notation.

At this point in our study of algebra, it is assumed that all variables represent real

numbers.

For this reason, you can omit the “∈R” and write {x|x≥2}, which is read “the set of all

real numbers x such that x is greater than or equal to 2.”

6 }, which is read “ the set of all real numbers x such that x is less than

or x is greater than or equal to 6”.

set of all real numbers x such that x is greater than or equal to -1 and less than 3”.

Key Takeaways

Inequalities usually have infinitely many solutions, so rather than presenting an

impossibly large list, we present such solutions sets either graphically on a number

line or textually using interval notation.

Inclusive inequalities with the “or equal to” component are indicated with a closed

dot on the number line and with a square bracket using interval notation.

Strict inequalities without the “or equal to” component are indicated with an open dot

on the number line and a parenthesis using interval notation.

124

Compound inequalities that make use of the logical “or” are solved by solutions of

either inequality. The solution set is the union of each individual solution set.

Compound inequalities that make use of the logical “and” require that all inequalities

are solved by a single solution. The solution set is the intersection of each individual

solution set.

Compound inequalities of the form n<A<m can be decomposed into two inequalities

using the logical “and.” However, it is just as valid to consider the argument A to be

bounded between the values n and m.

125

Topic Exercises

Graph all solutions on a number line and provide the corresponding interval notation.

1. x≤10

2. x>0

3. x≤−3

4. −4<x

5. x<−12

6. x≥−134

Graph all solutions on a number line and give the corresponding interval notation

7. −2<x<5

8. −5<x≤20

9. 10<x≤40

11. −58<x<18

12. −1≤x<112

126

ANSWERS

1: (−∞,10]

2: (0,∞)

3: (−∞,−3]

4: (−4,∞)

5: (−∞,−12)

6: [−134,∞)

7: (−2,5)

8: (−5,20]

9: (10,40]

127

10: (0,50]

11: (−58,18)

128

SEQUENCES and Their LIMITS

YASDAN DIAZ

129

Sequences and Their Limits

Definition:

A sequence of real numbers (or a sequence in R) is a function from N

into R.

Notation:

(1) The values of X : N ! R are denoted as X(n) or xn, where X is the

sequence.

(2) (xn : n ∈ N) or simply (xn) may denote a sequence — this is not the

same as {xn : n ∈ N}.

(3) (x1, x2, . . . , xn, . . .).

Examples:

(1) (3n) = (3n : n ∈ N) = (3, 6, 9, . . . , 3n, . . .).

(2) (1) = (1 : n ∈ N) = (1, 1, 1, . . . , 1, . . .).

(3) ((−2)𝑛 ) =(−2)𝑛 : n ∈ N =( -2, 4, -8, . . . , (−2)𝑛 , . . . .)

1 1 1 1

(4) (2 + 2 (−1)𝑛 )= (2 + 2 (−1)𝑛 : n ∈ N) = (0, 1, 0, 1, . . . , 0, 1, . . .).

𝑛 1 1(−1)𝑛 𝑛 1 1(−1)𝑛 𝑛

(5)( ( 2)2+2 )= ( ( 2)2+2 : n ∈ N )= ( 1, 1, 1, 2, 1, 3, 1, 4, . . , 1, 2 ,

...).

Definition:

Sequences may also be defined inductively or recursively.

Example:

(1) 𝑥1 = 5, 𝑥2 +1 = 2𝑥𝑛 −3 (n≥ 1) gives (5, 7, 11, 19, 35, . . .).

(2) Fibonacci sequence: 𝑥1 = 𝑥2 = 1, 𝑥𝑛+1 = 𝑥𝑛−1 + 𝑥𝑛 (n ≥ 2) gives (1, 1,

2, 3, 5, 8, 13, . . .).

Definition:

A sequence of real numbers converges to a real number a if, for every

positive number ∈ , there exists an N ∈ N such that for all n ≥ N, |an -

a| < ∈ . We call such an a the limit of the sequence and write lim 𝑎𝑛

𝑛→∞

an = a.

Proposition 1:

The sequence 1 n converges to zero.

Proof.

1

Let ∈ > 0. We choose N ∈ N such that N > 1 ∈. Such a choice is

always possible by the Archimedean property. To verify that this choice

1

of N is appropriate, let n ∈ N satisfy n ≥ N. Then, n ≥ N implies n > ∈

1 1

which is equal to 𝑛 = | 1 n −0 | < ∈ , proving that converges to zero

𝑛

by the definition of convergence.

130

Proposition 2.

An example of a sequence that does not converge is the

following:(1, −1, 1, −1, ...) If a sequence does not converge, it is said to

diverge, which we will explain later in the paper, along with the

explanation of why the above sequence does not converge.

Proposition 3.

If 𝑥𝑛 ≤ 𝑦𝑛 ≤ 𝑧𝑛 for all n ∈ N and lim 𝑥𝑛 = lim 𝑧𝑛 =l, then lim 𝑦𝑛 =l too

𝑛→∞ 𝑛→∞ 𝑛→∞

Proof.

Let ∈ > 0. We want to show there exists an N such that for all n > N, | yn

− l |< ∈. We know that xn → l. Therefore, there exists an N1 such that

for all n > N1, | xn − l |< ∈. Also, we know that zn → l. Therefore, there

exists an N2 such that for all n > N2, | zn − l |< ∈. Let N = max(N1, N2)

and n > N. Then, n > N1 so | xn − l |< ∈. Also, n > N2 so | zn − l |< ∈. We

want to show that | yn − l |< ∈. This is equivalent to showing that both yn

− l < ∈ and l − yn < ∈. We know that yn ≤ zn, so yn − l ≤ zn − l < ∈. Also,

we know that yn ≥ xn, so l − yn ≤ l − xn < ∈.

Theorem:

(Algebraic Limit Theorem).

Let lim 𝑎𝑛 = a and lim 𝑏𝑛 = b. Then,

𝑛→∞ 𝑛→∞

(i) lim 𝑐𝑛 = ca for all c ∈ R

𝑛→∞

(ii) lim ( 𝑎𝑛 +𝑏𝑛 ) = a+b

𝑛→∞

(iii) lim ( 𝑎𝑛 𝑏𝑛 ) = ab

𝑛→∞

(iv) lim ( 𝑎𝑛 /𝑏𝑛 = a/b provided b ≠ = 0

𝑛→∞

Example:

If (xn) → 2, then ((2xn - 1)/3) → 1.

Proof.

First, we will start with the information given in the example:

xn → 2.

4 3

Next, we simply use the fact that (3 )(2 ) = 2.

4 3

xn → (3 )(2 )

4 3 2

Now, let an = xn, and let a = (3 )(2 ), and let c = (3 ). From the Algebraic

2 2 4 3

Limit Theorem, we know that 𝑐𝑎𝑛 → ca. Then, (3 )(xn) → (3) (3 )(2 ), which

is equal to the following:

2𝑥𝑛 4

( ) →(3 )

3

4 1

The next step follows from the fact that (3 ) = 1 + (3 ) .

2𝑥𝑛 1

( ) →1 + (3 ) .

3

2𝑥 1 −1 −1 −1

Let ( 3𝑛 ) = an, let (1 + 3 ) = a, let bn = ( 3 , 3 , ...), and let b = 3 . Then,

by the Algebraic Limit Theorem, we know that an + bn → a + b.

131

2𝑥𝑛 −1 1 −1

Therefore, we know that ( ) + → (1 + 3 ) + , which is equal to the

3 3 3

following:

2𝑥𝑛 1

( )- →1

3 3

2𝑥𝑛 1 2𝑥𝑛−1

This last step follows because ( )- =( )=

3 3 3

2𝑥𝑛 1

( )- → 1.

3 3

Therefore, using the Algebraic Limit Theorem, we have shown that if

2𝑥 1

(xn) → 2, then ( 3𝑛 ) - 3 → 1.

Example:

2 +1 2

The following sequence converges to the proposed limit lim (5𝑛+4)= 5

𝑛

Proof.

2 +1 1 2 +1

Let 5𝑛+4 be an, let 5 be bn and let 5𝑛+4be cn, and cn = an + bn. By

𝑛 𝑛 +4 𝑛

Theorem 2.3, we know that lim(cn) = lim (an+bn) = lim(an)+lim(bn). We

must therefore determine what lim(an) and lim(bn) are. First, we will

1

show that lim(5 +4 ) = 0. Let ∈ > 0. By the Archimedean principle, there

𝑛

exists an N ∈ N such that N > 1/∈. Then, for n > N, 1 5n+4 < 1 5N+4 <

1/N < ∈. Therefore, the limit of 1 5n+4 is zero. Then, because lim(cn) =

lim(an+bn), lim(cn) = lim(an + 0) = lim(an). We will therefore find the limit

2 +1 2 2 +1

of an in order to prove lim(5𝑛+4 ) = 5. We now want to show that lim(5𝑛+4

𝑛 𝑛

2

) = 5. Let ∈ > 0. By the Archimedean Principle, there exists an N such

that 1/∈ < N. Let n > N. We then want to show the following:

2 +1 2 2 +1 2 −8

|5𝑛+4 − 5| < ∈, then |5𝑛+4 − 5|=|5(5 |

𝑛 𝑛 𝑛 +4)

−8

<∈

5(5𝑛 +4)

8

<∈

5(5𝑛 +4)

−8

We know that the inequality −5(5 < ∈ is true for every value of n

𝑛 +4)

because n > N > 1/∈ and ∈. Therefore we only need to show that the

8

inequality 5(5 +4)< ∈

𝑛

is true. Using the fact that N > 1/∈, we can say the following:

8 8 8∈

< 5(5(1/∈+4)= 25+20∈

5(5𝑛 +4)

8∈ 8∈ 8

Then 25+20∈ < 25< ∈. Therefore 5(5 < ∈.

𝑛 +4)

Example:

Let xn ≥ 0. If (xn) → 0, then (√𝑥𝑛 ) → 0.

Proof.

First, we have to prove that lim( √𝑥𝑛 ) exists. We know that xn is

decreasing but is greater than or equal to 0 for all values of n. The square

132

root of a positive number is also positive. Therefore, √𝑥𝑛 ≥ 0. Also, note

that if 0< a < b, then 0< √𝑎 < √𝑏. So if xn is decreasing, then so is √𝑥𝑛 .

Therefore, lim(√𝑥𝑛 ) exists. Next, we must prove that ( √𝑥𝑛 ) → 0. Let

lim(xn) = lim√𝑥𝑛 √𝑥𝑛 )) = 0. By the Algebraic Limit Theorem, we know that

if lim(an) = a and lim(bn) = b then lim((an)(bn)) = ab. By this theorem,

lim((√𝑥𝑛 )(√𝑥𝑛 )) = lim(√𝑥𝑛 )lim√𝑥𝑛 ) = 0. Thus, (lim(√𝑥𝑛 ))2=0. This implies

that lim(√𝑥𝑛 )=0.

Definition:

A sequence that does not have a limit or in other words, does not

converge, is said to be divergent.

Example:

Recall proposition 2, which says that the following sequence does not

converge: (3.3) (1, −1, 1, −1...).

1 1 1 1 1 1

(1, , , , , , … )

2 3 4 5 6 7

This converges to zero, as we proved earlier in this paper. However,

these sequences do have something in common. They are both

bounded.

Definition:

A sequence (xn) is bounded if there exists a number M>0 such that | xn

| ≤ M for all n ∈ N. Geometrically, this means we can find an interval

[−M,M] that contains every term in the sequence (xn).

Example:

Given the sequence xn = (1, 2, 1, 2, 1, 2 ...), we can see that the interval

[1, 2] contains every term in xn. This sequence is therefore a bounded

sequence.

Example:

Given the sequence xn = (10, 100, 1000, 10000, ...), we can see that

there is no real number that serves as an upper bound because lim(xn)

is infinity. Therefore, there does not exist any interval that contains every

term in the sequence xn, and xn is not a bounded sequence.

Theorem:

Every convergent sequence is bounded.

Example:

𝑛+1

Theorem being illustrated: Let xn = , which is the following sequence:

𝑛

2 3 4 5

( . , , )

1 2 3 4

We know this converges to 1 and can verify this using the same logic

used in the proof under the definition of convergence showing that 1 n

converges to zero. Therefore, as n becomes very large, xn approaches

1, but is never equal to 1. By the above theorem, we know that this

sequence is bounded because it is convergent. We can see that xn is a

133

decreasing sequence, so the x1 is the largest value of the sequence and

is the “upper bound.” The limit of the sequence, 1, is the lower bound.

An interval that contains every term in the sequence xn is (1,2].

Continuity

Theorem:

If f: R → R is continuous, xn → x implies f(xn) → f(x)

Example:

Theorem being applied: Let f(x) = 3x. This function is continuous. Let

lim(xn) = 5. In order words, xn → 5. By the above theorem, this implies

that f(xn) → f(5). This is equal to 3xn →(3)(5) which is also equal to 3xn

→ 15. Therefore, we are able to see what the limit of f(xn) is using this

theorem.

Example:

Theorem failing when function is non-continuous: Let f(x) be 1 x , a non-

continuous function. We know this is non-continuous because there is

an asymptote at x=0. Let xn be 1 n. We know this converges to zero

based on a previous proof. Let’s see if the continuity theorem fails for a

noncontinuous function f. The theorem states that f(xn) converges to f(x)

if xn → x. We know that xn → 0, so if the theorem works, then f(xn) →

f(0). But f(0) = 1 0 which does not exist. Therefore, f(xn) cannot converge

to 1/0, and the theorem fails for this non-continuous function.

Definition:

Let (an) be a sequence of real numbers, and let n1 < n2 < n3 < n4... be

an increasing sequence of natural numbers. Then, the sequence (an1 ,

an2 , an3 , an4 ...) is called a subsequence of (an) and is denoted by

(𝑎𝑛𝑘 ), where k ∈ N indexes the subsequence.

Example:

1 1 1

Let xn = 1 n = (1, 2 , 3 , 4...). Below are two examples of valid

subsequences:

1 1 1 1

( , , , )

3 6 9 12

1 1 1

( , , …)

20 200 2000

Bolzano-Weierstrass Theorem:

Every bounded sequence contains a convergent subsequence.

Example:

Given a sequence xn = (1,2,3,4,1,2,3,4...), a convergent subsequence

can be found.

134

Proof.

We know that this sequence is bounded by the interval [1, 4]. By the

Bolzano-Weierstrass Theorem, we can say that there indeed exists a

convergent subsequence of xn. Just by looking at this sequence, we can

see four convergent subsequences: (1,1,1...), (2,2,2...), (3,3,3...), and

(4,4,4...). These subsequences converge to 1, 2, 3, and 4 respectively.

Example:

Given an unbounded sequence xn = (1,2,3,4,5...), a convergent

subsequence of xn does not exist Proof. A convergent subsequence

does not necessarily exist because this sequence does not satisfy the

Bolzano-Weierstrass Theorem. Recall that any subsequence of a

sequence is non-repeating and in the order of the original entries of xn.

Notice that xn is increasing for all values of n and is divergent,

considering the sequence continues until infinity. Therefore, for any

subsequence an, the values will be increasing toward infinity as well, and

the subsequence will also be divergent.

Theorem:

Subsequences of a convergent sequence converge to the same limit as

the original sequence.

Example:

Let us return to the example of a divergent sequence that was given

under the definition of divergence. Recall that this sequence, xn, was (1,

-1, 1, -1, 1, -1...). One subsequence of xn is (1, 1, 1, ...). This

subsequence converges to 1. Another subsequence of xn is (-1, -1, -1,

...). This subsequence converges to -1. Now, we will prove that xn is

divergent by contradiction. Assume xn is convergent. Then, by the above

theorem, all its subsequences converge to lim(xn), implying that all its

subsequences converge to the same value. The two subsequences of

xn stated above converge to different values. Therefore, this contradicts

our original hypothesis that xn is convergent. We are then able to

conclude that xn is divergent. Acknowledgements. I would like to thank

my two mentors, Sean Howe and Abby Ward for meeting with me each

week for the duration of my stay at the program. Thank you so much for

suggesting this topic, explaining everything to me that I was confused

about, editing and helping with my paper, and calming me down when I

got overwhelmed. I would also like to thank Peter May for organizing this

wonderful REU and for giving me the opportunity to participate.

Theorem:

Let X = (xn) be a sequence in R, and let x ∈ R. The following are

equivalent:

(a) X converges to x.

(b) ∀ ∈ > 0, ∃ K ∈ N ∋ ∀ n ≥ K. |𝑥𝑛 − 𝑥| < ∈.

(c) ∀ ∈ > 0, ∃ K ∈ N ∋ ∀ n ≥ K , x- ∈< 𝑥𝑛 < x+ ∈.

(d) ∀ ∈-nbhd. 𝑉𝜖 (x) of x, ∃ K ∈ N ∋ ∀ n ≥ K , 𝑥𝑛 ∈ 𝑉𝜖 (x) .

135

The K-ǫ principle

Definition:

Can a sequence have more than one limit? Common sense says no: if

there were two different limits L and L ′ , the an could not be arbitrarily

close to both, since L and L ′ themselves are at a fixed distance from

each other. This is the idea behind the proof of our first theorem about

limits. The theorem shows that if {an} is convergent, the notation lim an

makes sense; there’s no ambiguity about the value of the limit. The proof

is a good exercise in using the definition of limit in a theoretical argument.

Try proving it yourself first.

Theorem:

A Uniqueness theorem for limits.

A sequence an has at most one limit: an → L and an → L ′ ⇒ L = L ′ .

Proof.

By hypothesis, given 𝜖 > 0, an ≈ ǫ L for n ≫ 1, and an ≈ ǫ L ′ for n ≫ 1

Therefore, given ǫ > 0, we can choose some large number k such that L

≈ ǫ ak ≈ ǫ L ′ . By the transitive law of approximation (2.5 (8)), it follows

that (4) given ǫ > 0, L ≈ 2ǫ L ′ . To conclude that L = L ′ , we reason

indirectly (cf. Appendix A.2). Suppose L 6= L ′ ; choose ǫ = 1 2 |L − L ′ |.

We then have |L − L ′ | < 2ǫ, by (4); i.e., |L − L ′ | < |L − L ′ |, a contradiction.

Remarks:

The line (4) says that the two numbers L and L ′ are arbitrarily close. The

rest of the argument says that this is nonsense if L 6= L ′ , since they

cannot be closer than |L − L ′ |.

Theorem:

{an} increasing, L = lim an ⇒ an ≤ L for all n; {an} decreasing, L = lim an

⇒ an ≥ L for all n. Proof. Both cases are handled similarly; we do the

first. Reasoning indirectly, suppose there were a term aN of the

sequence such that aN > L. Choose ǫ = 1 2 (aN − L). Then since {an} is

increasing, an − L ≥ aN − L > ǫ, for all n ≥ N, contradicting the Definition

3.1 of L = lim an.

Definition:

In the proof of Theorem 3 , note the appearance of 2ǫ in line (4). It often

happens in analysis that arguments turn out to involve not just ǫ but a

constant multiple of it. This may occur for instance when the limit

involves a sum or several arithmetic processes. Here is a typical

example. Example 3.2 Let an = 1 n + sin n n + 1 . Show an → 0, from

the definition. Solution To show an is small in size, use the triangle

inequality: ¯ ¯ ¯ ¯ 1 n + sin n n + 1 ¯ ¯ ¯ ¯ ≤ ¯ ¯ ¯ ¯ 1 n ¯ ¯ ¯ ¯ + ¯ ¯ ¯

¯ sin n n + 1 ¯ ¯ ¯ ¯ . At this point, the natural thing to do is to make the

separate estimations ¯ ¯ ¯ ¯ 1 n ¯ ¯ ¯ ¯ < ǫ, for n > 1 ǫ ; ¯ ¯ ¯ ¯ sin n n

+ 1 ¯ ¯ ¯ ¯ < ǫ, for n > 1 ǫ − 1 ; so that, given ǫ > 0, ¯ ¯ ¯ ¯ 1 n + sin n n

136

+ 1 ¯ ¯ ¯ ¯ < 2ǫ , for n > 1 ǫ . This is close, but we were supposed to

show |an| < ǫ. Is 2ǫ just as good?

The usual way of handling this would be to start with the given ǫ, then

put ǫ ′ = ǫ/2, and give the same proof, but working always with ǫ ′ instead

of ǫ. At the end, the proof shows ¯ ¯ ¯ ¯ 1 n + sin n n + 1 ¯ ¯ ¯ ¯ < 2ǫ ′ ,

for n > 1 ǫ ′ ; and since 2ǫ ′ = ǫ, the limit definition is satisfied. Instead of

doing this, let’s once and for all agree that if you come out in the end with

2ǫ, or 22ǫ, that’s just as good as coming out with ǫ. If ǫ is an arbitrary

small number, so is 22ǫ. Therefore, if you can prove something is less

than 22ǫ, you have shown that it can be made as small as desired. We

formulate this as a general principle, the “K-ǫ principle”. This isn’t a

standard term in analysis, so don’t use it when you go to your next

mathematics congress, but it is useful to name an idea that will recur

often.

Principle 3.2:

The K-ǫ principle. Suppose that {an} is a given sequence, and you can

prove that (5) given any ǫ > 0, an ≈ Kǫ L for n ≫ 1 , where K > 0 is a

fixed constant, i.e., a number not depending on n or ǫ. Then limn→∞ an

= L . The K-ǫ principle is here formulated for sequences, but we will use

it for a variety of other limits as well. In all of these uses, the essential

point is that K must truly be a constant, and not depend on any of the

variables or parameters.

Infinite limits

Definition:

Even though ∞ is not a number, it is convenient to allow it as a sort of

“limit” in describing sequences which become and remain arbitrarily

large as n increases. The definition is like the one for the ordinary limit.

Definition:

We say the sequence {an} tends to infinity if (6) given any M ≥ 0, an > M

for n ≫ 1 . In symbols: limn→∞ {an} = ∞, or an → ∞ as n → ∞.As for

regular limits, to establish that lim{an} = ∞, what you have to do is give

an explicit value for the N concealed in “for n ≫ 1”, and prove that it does

the job, i.e., prove that an > M when n ≥ N. In general, this N will depend

on M: the bigger the M, the further out in the sequence you will have to

go for the inequality an > M to hold. As before, it is not you who chooses

the M; the limit demon does that, and you have to prove the inequality in

(6) for whatever positive M it gives you. Note also that even though we

are dealing with size, we do not need absolute values, since an > M

means the an are all positive for n ≫ 1. One should not think that infinite

limits are associated only with increasing sequences. Consider these

examples, neither of which is an increasing sequence.

An important limit

137

As a good opportunity to practice with inequalities and the limit definition,

we prove an important limit that will be used constantly later on.

Theorem:

The limit of a n.

Proof.

We consider the case a > 1 first. Since a > 1, we can write a = 1 + k, k >

0. a n = (1 + k) n Thus , which by the binomial theorem = 1 + nk + n(n −

1) 2! k 2 + n(n − 1)(n − 2) 3! k 3 + ... + k n . Since all the terms on the

right are positive, a n (5) > 1 + nk ; > M, for any given M > 0, if n > M/k,

say. This proves that lim a n = ∞ if a > 1, according to Definition 3.3. ¤

The second case a = 1 is obvious. For the third, in outline the proof is:

|a| < 1 ⇒ 1 |a| > 1 ⇒ ³ 1 |a| ´n → ∞ ⇒ a n → 0 . Here the middle implication

follows from the first case of the theorem. The last implication uses the

definition of limit; namely, by hypothesis, given ǫ > 0, ³ 1 |a| ´n > 1 ǫ for

n large; by the reciprocal law of inequalities (2.1) and the multiplication

law for | |, |a n | < ǫ for n large.

Why did we begin by writing a = 1 + k? Experimentally, you can see that

when a > 1, but very close to 1 (like a = 1.001), a increases very slowly

at first when raised to powers. This is the worst case, therefore, and it

suggests writing a in a form which shows how far it deviates from 1. The

case a ≤ −1 is not included in the theorem; here the a n alternate in sign

without getting smaller, and the sequence has no limit. A formal proof of

this directly from the definition of limit is awkward; instead we will prove

it at the end of Chapter 5, when we have more technique.

THEOREM:

Suppose that an ≤ bn ≤ cn for all n > N, for some N. If limn→∞ an =

limn→∞ cn = L, then limn→∞ bn = L. And a final useful fact:

THEOREM:

limn→∞ |an| = 0 if and only if limn→∞ an = 0.This says simply that the

size of an gets close to zero if and only if an gets close to zero.

EXAMPLE:

Determine whether n n + 1∞ n=0 converges or diverges. If it converges,

compute the limit. Since this makes sense for real numbers we consider

limx→∞ x x + 1 = limx→∞ 1 − 1 x + 1 = 1 − 0 = 1. Thus the sequence

converges to 1.

EXAMPLE:

Determine whether ln n n ∞ n=1 converges or diverges. If it converges,

compute the limit. We compute limx→∞ ln x x = limx→∞ 1/x 1 = 0, using

L’Hˆopital’s Rule. Thus the sequence converges to 0.

EXAMPLE:

Determine whether {(−1)n } ∞ n=0 converges or diverges. If it converges,

compute the limit. This does not make sense for all real exponents, but

the sequence is easy to understand: it is 1, −1, 1, −1, 1 . . . and clearly

diverges.

EXAMPLE:

138

Determine whether {(−1/2)n } ∞ n=0 converges or diverges. If it

converges, compute the limit. We consider the sequence {|(−1/2)n |}∞

n=0 = {(1/2)n } ∞ n=0. Then limx→∞ 1 2 x = limx→∞ 1 2 x = 0, so by

theorem 11.1.4 the sequence converges to 0.

Sometimes we will not be able to determine the limit of a sequence, but

we still would like to know whether it converges. In some cases we can

determine this even without being able to compute the limit. A sequence

is called increasing or sometimes strictly increasing if ai < ai+1 for all i.

It is called non-decreasing or sometimes (unfortunately) increasing if ai

≤ ai+1 for all i. Similarly a sequence is decreasing if ai > ai+1 for all i and

non-increasing if ai ≥ ai+1 for all i. If a sequence has any of these

properties it is called monotonic.

EXAMPLE:

The sequence 2 i − 1 2 i ∞ i=1 = 1 2 , 3 4 , 7 8 , 15 16 , . . . , is increasing,

and n + 1 n ∞ i=1 = 2 1 , 3 2 , 4 3 , 5 4 , . . . is decreasing.

A sequence is bounded above if there is some number N such that an ≤

N for every n, and bounded below if there is some number N such that

an ≥ N for every n. If a sequence is bounded above and bounded below

it is bounded. If a sequence {an} ∞ n=0 is increasing or non-decreasing

it is bounded below (by a0), and if it is decreasing or nonincreasing it is

bounded above (by a0). Finally, with all this new terminology we can

state an important theorem.

THEOREM:

If a sequence is bounded and monotonic then it converges. We will not

prove this; the proof appears in many calculus books. It is not hard to

believe: suppose that a sequence is increasing and bounded, so each

term is larger than the one before, yet never larger than some fixed value

N. The terms must then get closer and closer to some value between a0

and N. It need not be N, since N may be a “too-generous” upper bound;

the limit will be the smallest number that is above all of the terms ai .

EXAMPLE:

All of the terms (2i − 1)/2 i are less than 2, and the sequence is

increasing. As we have seen, the limit of the sequence is 1—1 is the

smallest number that is bigger than all the terms in the sequence.

Similarly, all of the terms (n+ 1)/n are bigger than 1/2, and the limit is 1—

1 is the largest number that is smaller than the terms of the sequence.

We don’t actually need to know that a sequence is monotonic to apply

this theorem— it is enough to know that the sequence is “eventually”

monotonic, that is, that at some point it becomes increasing or

decreasing. For example, the sequence 10, 9, 8, 15, 3, 21, 4, 3/4, 7/8,

15/16, 31/32, . . . is not increasing, because among the first few terms it

is not. But starting with the term 3/4 it is increasing, so the theorem tells

us that the sequence 3/4, 7/8, 15/16, 31/32, . . . converges. Since

convergence depends only on what happens as n gets large, adding a

few terms at the beginning can’t turn a convergent sequence into a

divergent one.

EXAMPLE:

139

Show that {n 1/n} converges. We first show that this sequence is

decreasing, that is, that n 1/n > (n+1)1/(n+1). Consider the real function

f(x) = x 1/x when x ≥ 1. We can compute the derivative, f ′ (x) = x 1/x(1−ln

x)/x2 , and note that when x ≥ 3 this is negative. Since the function has

negative slope, n 1/n > (n + 1)1/(n+1) when n ≥ 3. Since all terms of the

sequence are positive, the sequence is decreasing and bounded when

n ≥ 3, and so the sequence converges. (As it happens, we can compute

the limit in this case, but we know it converges even without knowing the

limit; see exercise 1.)

EXAMPLE:

Show that {n!/nn } converges.

Again we show that the sequence is decreasing, and since each term is

positive the sequence converges. We can’t take the derivative this time,

as x! doesn’t make sense for x real. But we note that if an+1/an < 1 then

an+1 < an, which is what we want to know. So we look at an+1/an: an+1

an = (n + 1)! (n + 1)n+1 n n n! = (n + 1)! n! n n (n + 1)n+1 = n + 1 n + 1

n n + 1n = n n + 1n < 1.

which look like infinite polynomials. Since calculus, that is, computation

of derivatives and antiderivatives, is easy for polynomials, the obvious

question is whether the same is true for infinite series. The answer is

yes:

THEOREM:

Suppose the power series f(x) = X∞ n=0 an(x − a) n has radius of

convergence R. Then f ′ (x) = X∞ n=0 nan(x − a) n−1 , Z f(x) dx = C +

X∞ n=0 an n + 1 (x − a) n+1 , and these two series have radius of

convergence R as well.

EXAMPLE:

Starting with the geometric series: 1 1 − x = X∞ n=0 x n Z 1 1 − x dx = −

ln |1 − x| = X∞ n=0 1 n + 1 x n+1 ln |1 − x| = X∞ n=0 − 1 n + 1 x n+1.

when |x| < 1. The series does not converge when x = 1 but does

converge when x = −1 or 1 − x = 2. The interval of convergence is [−1,

1), or 0 < 1 − x ≤ 2, so we can use the series to represent ln(x) when 0

< x ≤ 2.

140

Sequences and Limits Exercises

2𝑛+4

1.Prove that the sequence 𝑎𝑛 = has a limit of 2. Also, calculate the terms whose

𝑛

distance from 2 is less than 0.1.

4𝑛+1

2. Prove that the sequence 𝑎𝑛 = 𝑛 has a limit of 4 and calculate how many terms

of the succession are not within (4 − 0.001, 4 + 0.001).

𝑛3

3. Prove that the sequence 𝑎𝑛 = 𝑛2 +3 has a limit of 1 and calculate how many terms

of the succession are not within (1 − 0.001, 1 + 0.001).

(−1)𝑛

5. an = (−1)n n + 1

𝑛+1

6. Determine whether or not the sequence, xn defined for each integer n, by the finite

series xn = arctan 1 + arctan 2 + arctan 3 + ... + arctan n, for n ≥ 1, converges as n →

∞?

141

Answer Keys:

2𝑛+4 1

1.| − 2| < 10

𝑛

2𝑛+4−2𝑛 1

.| | < 10

𝑛

4 1 4 1

|𝑛| < 10 ; < 10 ; n> 40.

𝑛

4𝑛+1 1

2. | − 4| < 1000

𝑛

4𝑛+1−4𝑛 1

| | < 1000

𝑛

1 1 1 1

|𝑛| < 1000 ; < 1000 ; n> 1000.

𝑛

𝑛3 1

3. |𝑛2 +3 − 1| < 1000

−3 1

|𝑛2 +3| < 1000

3 1

< 1000 ; 𝑛3 + 3> 3000 ; 𝑛2 > 2997 ; n> 54. The first 54 terms are out.

𝑛2 +3

5 8 65 163 1957

4. 1, 2, 2 , 3 , 24 , , …

60 720

5. lim sup 𝑎𝑛 =+∞ ; lim inf 𝑎𝑛 = −∞, lim 𝑎𝑛

𝑛→∞ 𝑛→∞ 𝑛→∞

6. No, the limit cannot exist since arctan n → π/2 as n → ∞ and so the Divergence

Test for infinite series applies and this result follows.

142

LIMIT THEOREM

FERLY CANONAYON

143

Limit Theorem

Definition:

A sequence (𝑥𝑛 ) is bounded if there is a number 𝑀 > 0 such that

|𝑥𝑛 | ≤ 𝑀 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑛 ∈ 𝑁.

Theorem:

Every convergent sequence is bounded.

Proof :

Assume that ( 𝑥𝑛 ) is a convergent sequence: there is a unique

𝑙𝑖𝑚

number 𝑙 such that 𝑥 =1

𝑛→∞ 𝑛

Theorem:

If 𝑎 = 𝑙𝑖𝑚 𝑎𝑛 and 𝑏 = 𝑙𝑖𝑚 𝑏𝑛 , then

(i) 𝑙𝑖𝑚(𝑐𝑎𝑛 ) = 𝑐𝑎 for all 𝑐 ∈ 𝑅,

(ii) 𝑙𝑖𝑚(𝑎𝑛 + 𝑏𝑛 ) = 𝑎 + 𝑏,

(iii) 𝑙𝑖𝑚(𝑎𝑛 𝑏𝑛 ) = 𝑎𝑏,

(iv) 𝑙𝑖𝑚(𝑎𝑛 /𝑏𝑛 ) = 𝑎/𝑏 provided 𝑏 ≠ 0.

Proof:

a. We consider two cases.

1. c ≠ 0

We are trying to show that 𝑐𝑎𝑛 → 𝑐𝑎. Hence we are interested in the

inequality

|𝑐𝑎𝑛 − 𝑐𝑎| < 𝜖

⇔ |𝑐||𝑎𝑛 − 𝑎| < 𝜖

𝜖

⇔ |𝑎𝑛 − 𝑎| <

|𝑐|

𝜖

Let 𝜖 > 0 be arbitrary. Since (𝑎𝑛 ) converges to a, given |𝑐| ≥ 0, there exists

𝜖

𝑁 ∈ ℕ such that if 𝑛 ≥ 𝑁, then|𝑎𝑛 − 𝑎| < |𝑐| . Then we compute

𝜖

|𝑎𝑛 − 𝑎| <

|𝑐|

⟹ |𝑐||𝑎𝑛 − 𝑎| < 𝜖

⟹ |𝑐𝑎𝑛 − 𝑐𝑎| < 𝜖

|𝑐𝑎𝑛 − 𝑐𝑎| < 𝜖

Hence, 𝑐𝑎𝑛 converges to 𝑐𝑎.

b. 𝑐 = 0. If 𝑐 = 0 then the sequence 𝑐𝑎𝑛 is simply the sequence

(0,0,0…)which consists only of 0’s. by the theorem, this sequence

converges to 0 = 𝑐𝑎, as desired.

Theorem:

144

Suppose that a = lim 𝑎𝑛 and b = lim𝑏𝑛 .

(i) If 𝑎𝑛 ≥ 0 for all 𝑛 ∈ 𝑁, then 𝑎 ≥ 0.

(ii) If 𝑎𝑛 ≤ 𝑏𝑛 for all 𝑛 ∈ 𝑁, then 𝑎 ≤ 𝑏.

(iii) If there exists 𝑐 ∈ 𝑅 for which c ≤ 𝑏𝑛 for all 𝑛 ∈ 𝑁, then

𝑐 ≤ 𝑏. Similiary, if 𝑎𝑛 ≤ c for all 𝑛 ∈ 𝑁, then 𝑎 ≤ 𝑐.

Proof.

(i) This will be proved by contradiction: we suppose that 𝑎 <

0 and we will show that 𝑎𝑛 < 0 for some n, a contradiction.

For 𝜖= |a| there exists 𝑁 ∈ 𝑁 such that |𝑎𝑛 − 𝑎| < = |𝑎| when

𝑛 ≥ 𝑁. Taking n = N, we have |𝑎𝑁 − a| < |a|, or unwrapping the

absolute value in the middle,

−|𝑎| + 𝑎 < 𝑎𝑁 < |𝑎| + 𝑎 = 0.

Because (𝑏𝑛 − 𝑎𝑛 )≥ 0 for all 𝑛 ∈ 𝑁, we apply part (i) to conclude

that 𝑏 − 𝑎 ≥ 0.

(ii) Take 𝑎𝑛 = c (or 𝑏𝑛 = c) and apply part (ii).

Theorem:

Two Fundamental Limits

𝑙𝑖𝑚

i. 𝑐 = 𝑐, where c is constant

𝑥→𝑎

𝑙𝑖𝑚

ii. 𝑥=𝑎

𝑥→𝑎

Examples:

𝑙𝑖𝑚

a. 10 = 10

𝑥→2

𝑙𝑖𝑚

b. 𝝅=𝝅

𝑥→6

𝑙𝑖𝑚

c. 𝑥=2

𝑥→2

𝑙𝑖𝑚

d. 𝑥=0

𝑥→0

Theorem:

Limit of a Constant Multiple

If 𝒄 is a constant, then

𝑙𝑖𝑚 𝑙𝑖𝑚

𝑐 𝑓(𝑥) = 𝑐 𝑓(𝑥)

𝑥→𝑎 𝑥→𝑎

𝑙𝑖𝑚 𝑙𝑖𝑚

a. 5𝑥 = 5 𝑥 = 5 · 8 = 40

𝑥→8 𝑥→8

𝑙𝑖𝑚 3 3 𝑙𝑖𝑚 3

b. (− 2) 𝑥 = (− 2) 𝑥 = (− 2) (−2) = 3

𝑥 → −2 𝑥 → −2

Theorem:

145

Limits of a Sum, Product, and Quotient

𝑙𝑖𝑚 𝑙𝑖𝑚

Suppose 𝑎 is a real number and 𝑓(𝑥) and 𝑔(𝑥) exist. If

𝑥→𝑎 𝑥→𝑎

𝑙𝑖𝑚 𝑙𝑖𝑚

𝑓(𝑥) = 𝐿1 and 𝑔(𝑥) = 𝐿2 , then

𝑥→𝑎 𝑥→𝑎

𝑙𝑖𝑚 [𝑓(𝑥) 𝑙𝑖𝑚 𝑙𝑖𝑚

i. ± 𝑔(𝑥)] = 𝑓(𝑥) ± 𝑔(𝑥) = 𝐿1 ± 𝐿2

𝑥→𝑎 𝑥→𝑎 𝑥→𝑎

𝑙𝑖𝑚 [𝑓(𝑥)𝑔(𝑥)] 𝑙𝑖𝑚 𝑙𝑖𝑚

ii. =( 𝑓(𝑥)) ( 𝑔(𝑥)) = 𝐿1 𝐿2 , and

𝑥→𝑎 𝑥→𝑎 𝑥→𝑎

𝑙𝑖𝑚

𝑙𝑖𝑚 𝑓(𝑥) 𝑥→𝑎𝑓(𝑥) 𝐿1

iii. = 𝑙𝑖𝑚 = 𝐿 , 𝐿2 ≠ 0

𝑥 → 𝑎 𝑔(𝑥) 𝑥→𝑎 𝑔(𝑥) 2

i. The limit of a sum is the sum of the limits

ii. The limit of a product is the product of the limits, and

iii. The limit of a quotient is the quotient of the limits

Theorem:

Limit of a Power

𝑙𝑖𝑚

Let 𝑥→𝑎 𝑓(𝑥) = 𝐿 and 𝑛 be a positive integer. Then

𝑛

𝑙𝑖𝑚 𝑙𝑖𝑚

[𝑓(𝑥)]𝑛 = [ 𝑓(𝑥)] = 𝐿𝑛

𝑥→𝑎 𝑥→𝑎

For the special case for 𝑓(𝑥) = 𝑥, the result given in theorem 2.2.4

yields

𝑙𝑖𝑚 𝑛

𝑥 = 𝑎𝑛

𝑥→𝑎

Theorem:

The Limit That Does Not Exist

𝑙𝑖𝑚 𝑙𝑖𝑚

Let 𝑓(𝑥) = 𝐿1 ≠ 0 and 𝑔(𝑥) = 0. Then

𝑥→𝑎 𝑥→𝑎

𝑙𝑖𝑚 𝑓(𝑥)

𝑥 → 𝑎 𝑔(𝑥)

does not exist.

Proof:

We will give an indirect proof of this result by using Theorem 2.2.3.

𝑙𝑖𝑚 𝑙𝑖𝑚

Suppose that 𝑓(𝑥) = 𝐿1 and 𝑔(𝑥) = 0 and suppose

𝑥→𝑎 𝑥→𝑎

𝑙𝑖𝑚 𝑓(𝑥)

further that exists and equal 𝐿2 . Then

𝑥 → 𝑎 𝑔(𝑥)

𝑙𝑖𝑚 𝑙𝑖𝑚 𝑓(𝑥)

𝐿1 = 𝑓(𝑥) = (𝑔(𝑥) · ) , 𝑔(𝑥) ≠ 0

𝑥→𝑎 𝑥→𝑎 𝑔(𝑥)

146

𝑙𝑖𝑚 𝑙𝑖𝑚 𝑓(𝑥)

=( 𝑔(𝑥)) = 0 · 𝐿2 = 0

𝑥→𝑎 𝑥 → 𝑎 𝑔(𝑥)

By contradicting the assumption that 𝐿1 ≠ 0, we have proved the

theorem.

Theorem:

Limit of a Root

𝑙𝑖𝑚

Let 𝑓(𝑥) = 𝐿 and 𝑛 be a positive integer. Then

𝑥→𝑎

𝑙𝑖𝑚 𝑛 𝑛 𝑙𝑖𝑚 𝑛

√𝑓(𝑥) = √ 𝑓(𝑥) = √𝐿

𝑥→𝑎 𝑥→𝑎

provided that 𝐿 ≥ 0 when 𝑛 iseven.

Theorem:

Existence Implies Uniqueness

𝑙𝑖𝑚

If 𝑓(𝑥) exists, then it is unique.

𝑥→𝑎

Theorem:

Suppose that f and g are functions such that f(x) = g(x) for all x in

some open interval interval containing a except possibly for a, then

Theorem:

Suppose that f and g are functions such that the two limits

integer. Then

1.

2.

3.

4.

5.

6.

147

7.

8.

10.

Theorem C.

The limit

if and only if the right-hand limits and left-hand limits exist and are equal to

M:

Theorem D.

(Squeeze Theorem)

Suppose that f, g and h are three functions such that f(x) < g(x) < h(x)

for all x. If

then

Theorem E.

Suppose that f and g are two functions such that

and

Examples

148

𝑙𝑖𝑚 (4𝑥 2 𝑙𝑖𝑚 𝑙𝑖𝑚 𝑙𝑖𝑚

1. − 2𝑥 + 1) = 4𝑥 2 − 2𝑥 + 1

𝑥→2 𝑥→2 𝑥→2 𝑥→2

𝑙𝑖𝑚 2 𝑙𝑖𝑚 𝑙𝑖𝑚

=4 𝑥 −2 𝑥+ 1

𝑥→2 𝑥→2 𝑥→2

2

= 4(2) − 2(2) + 1

=9

𝑙𝑖𝑚 2 𝑙𝑖𝑚 2 𝑙𝑖𝑚

𝑙𝑖𝑚 √𝑥 2 −1 √𝑥→−3(𝑥 −1) √𝑥→−3𝑥 −𝑥→−31

2. = 𝑙𝑖𝑚 = 𝑙𝑖𝑚

𝑥 → −3 2𝑥 𝑥→−3

2𝑥

𝑥→−3

2𝑥

(−32 )−1 √8 √2

= = −6 = −

2(−3) 3

149

Exercises

𝑙𝑖𝑚

1. Evaluate (10𝑥 + 7)

𝑥→5

2. Evaluate

𝑙𝑖𝑚 3

a. 𝑥

𝑥 → 10

𝑙𝑖𝑚 5

b.

𝑥 → 10 𝑥 2

3.Using Theorem 2.2.3. Evaluate

𝑙𝑖𝑚

a. (𝑥 2 − 5𝑥+6)

𝑥→3

4. Using Theorem 2.2.4. Evaluate

𝑙𝑖𝑚

a. (3𝑥 − 1)10

𝑥→1

𝑙𝑖𝑚 3𝑥−4

5. Evaluate 2

𝑥 → −1 8𝑥 =2𝑥−2

𝑙𝑖𝑚 𝑥−1

6. Evaluate

𝑥 → 1 𝑥 2 +𝑥−2

150

Answers Key

𝑙𝑖𝑚 (10𝑥 𝑙𝑖𝑚 𝑙𝑖𝑚

1. + 7) = 10𝑥 + 7

𝑥→5 𝑥→5 𝑥→5

𝑙𝑖𝑚 𝑙𝑖𝑚

= 10 𝑥+ 7

𝑥→5 𝑥→5

= 10 · 5 + 7

= 57

𝑙𝑖𝑚 3

2. a. 𝑥 = 103 = 1000

𝑥 → 10

𝑙𝑖𝑚

𝑙𝑖𝑚 5 5

𝑥→4

b. 2 = 𝑙𝑖𝑚

𝑥 → 4𝑥 𝑥→4

𝑥2

5

= 42

5

= 16

3. − 5𝑥 + 6) = 𝑥 − 5𝑥 + 6

𝑥→3 𝑥→3 𝑥→3 𝑥→3

= 32 − 5(3) + 6

= 9 − 15 + 6

=0

𝑙𝑖𝑚

4. (3𝑥 − 1)10

𝑥→1

First, we see from Theorem 2.2.3 that

𝑙𝑖𝑚 𝑙𝑖𝑚 𝑙𝑖𝑚

(3𝑥 − 1)10 = 3𝑥 − 1=2

𝑥→1 𝑥→1 𝑥→1

𝑙𝑖𝑚 𝑙𝑖𝑚

(3𝑥 − 1)10 = [ (3𝑥 − 1)10 ]

𝑥→1 𝑥→1

= 210

= 1024

3𝑥−4

5. Solution: f(x)= 8𝑥 2 =2𝑥−2 is a rational function and so if we

identify the polynomials

𝑝(𝑥) = 3𝑥 − 4 and (𝑥) = 8𝑥 2 = 2𝑥 − 2 , then

𝑙𝑖𝑚 𝑙𝑖𝑚

𝑝(𝑥) = 𝑝(−1) = −7 and 𝑞(𝑥) = 𝑞(−1) = 4

𝑥→1 𝑥→1

Since, 𝑞(−1) ≠ 0 it follows that

𝑙𝑖𝑚 3𝑥 − 4 𝑝(−1) −7 7

= = = −

𝑥 → −1 8𝑥 2 = 2𝑥 − 2 𝑞(−1) 4 4

𝑙𝑖𝑚 𝑥−1 𝑙𝑖𝑚 𝑥−1

6. =

𝑥 → 1 𝑥 2 +𝑥−2 𝑥 → 1 (𝑥−1)(𝑥+1)

151

𝑙𝑖𝑚 1

=

𝑥 → 1 𝑥+1

𝑙𝑖𝑚

1

𝑥→1

= 𝑙𝑖𝑚

(𝑥+2)

𝑥→1

1

=3

152

MONOTONE SEQUENCES

CHRISTY IGNALIG

153

Monotone Sequence

Definition:

We begin by a definition. We say that a real sequence (an) is

monotone increasing if n1 < n2 =) an1 < an2 monotone decreasing if

n1 < n2 =) an1 > an2 monotone non-decreasing if n1 < n2 =) an1 6 an2

monotone non-increasing if n1 < n2 =) an1 > an2

Example:

Let an = n. Then (an) is monotone increasing. So is an = (2n + 1)2.

Definition:

Remarks:

definition.

Monotonic decreasing sequences are defined similarly.

Then the big result is

Theorem:

Proof:

We will prove that the sequence converges to its least upper bound

(whose existence is guaranteed by the Completeness axiom).

So let α be the least upper bound of the sequence. Given ε > 0, we'll

show that all the terms of the sequence (except the first few) are in the

interval (α - ε, α + ε). Now since α + ε is an upper bound of the

sequence, all the terms certainly satisfy an< α + ε. Also

since α - ε is not an upper bound of the sequence, we must

have aN> α - ε for some N. But then all the later terms will be > α - ε also

and so (for n > this N) we have our condition for convergence.

Remarks:

(which converges to its greatest lower bound).

For example, the sequences (-1 , 1/2 , -1/3 , 1/4 , ... ) and

(1/2 , 1/22 , 1/3 , 1/32 , ... ) both converge to 0.

what their limits are.

154

Examples:

have an= α (approx) and an+1= α (approx) and so we must have α =

√(2α -1) and hence α2 = 2α - 1 and we get α = 1. To show that it does

indeed have a limit, we'll prove that it is monotonic decreasing and

bounded below.

Since the terms of the sequence are positive, the sequence is clearly

bounded below by 0.

Claim:

2 2

an+1 - an= (an - an-1 )/2 = (an + an-1)(an - an-1)/2 and so (since an > 0) the

result follows by induction.

The limit α satisfies α = (α2+ 1)/2 and hence α = 1.

But since the sequence starts at 2 and increases, it cannot converge to

1 and hence it has no limit.

One may show that if one takes the starting value a1 = 1/2 then the

sequence is bounded above and does converge to 1.

Definition:

𝑠1 ≤ 𝑠2 ≤ 𝑠3

The sequence is𝑠𝑛 said to be decreasing if 𝑠𝑛 ≥+1 𝑠𝑛 for all n ≥ 1,

i.e., 𝑠1 ≥ 𝑠2 ≥ 𝑠3

A sequence is said to be monotone if it is either increasing or

decreasing.

Example:

The sequence n2 : 1, 4, 9, 16, 25, 36, 49, ... is increasing. The sequence

1/2n : 1/2, 1/4, 1/8, 1/16, 1/32, ... is decreasing. The sequence

(−1)n_1/n : 1,−1/2, 1/3,−1/4, 1/5,−1/6, ... is not monotone.

Example:

155

Let r 6= 0. Consider the geometric sequence sn = r n: 1, r , r 2, r 3, r 4, ....

Boundedness

Definition:

The sequence 𝑠𝑛 is bounded above if there exists a number M such that

𝑠𝑛 ≤ M for all n. M is called an upper bound for 𝑠𝑛 .

The sequence 𝑠𝑛 is bounded below if there exists a number K such that K

≤ 𝑠𝑛 for all n. K is called a lower bound for𝑠𝑛 .

A sequence is bounded if it is both bounded above and bounded below.

Example:

1

𝑠𝑛 𝑠𝑛 =𝑛 is bounded, since it is both bounded above (by 1, for example) and

bounded below

𝑠𝑛 = −n is bounded above (by 0, for example) but not bounded below.

𝑠𝑛 = 2𝑛 is bounded below (by 0, for example) but not bounded above.

𝑠𝑛 = (−1𝑛−1 )n is neither bounded above nor bounded below.

Remarks:

If a sequence is bounded above, then there are infinitely many upper

bounds, not just one. Any number greater than some upper bound is itself

an upper bound. However, among all the upper bounds, there is at least

one, the least upper bound, also known as the supremum.

If a sequence is bounded below, then there are infinitely many lower

bounds, not just one. Any number less than some lower bound is itself a

lower bound. However, among all the lower bounds, there is a greatest

one, the greatest lower bound, also known as the infimum.

Definition:

A sequence 𝑎𝑛 𝑛 ≥ 1of real numbers is called an increasing sequence if 𝑎𝑛 ≤

𝑎𝑛+1 for all n ≥ 1, and (𝑎𝑛 )𝑛 ≥ 1 is called a decreasing sequence if 𝑎𝑛 ≥

𝑎𝑛+1 for all n ≥ 1. A sequence that is increasing or decreasing is said to be a

monotone sequence.

Example:

The sequence (1, 1, 2, 2, 3, 3, 4, 4, . . .) is increasing, but (−1, 1, −1, 1,

. . .) is not monotone.

Exercise:

a) Show that any bounded above increasing sequence is bounded.

b) Show that any bounded below decreasing sequence is bounded.

Exercise:

a) Prove that (𝑛2−𝑛 )𝑛≥2 is a decreasing sequence.

b) Let (𝑎𝑛 )𝑛≥1 be an increasing sequence of positive numbers and

define (𝜎𝑛 )𝑛≥1 . Prove that (𝜎𝑛 )𝑛≥1 is also an increasing sequence.

156

Theorem:

Every bounded monotone sequence converges.

Proof.

We will prove the theorem for increasing sequences. The case of

decreasing sequences is left to Exercise 4.3. So, let a sequence

(𝑎𝑛 )𝑛≥1 increase. By the assumption of the theorem, (𝑎𝑛 )𝑛≥1 is

bounded, that is, there exists C ∈ R such that ⃓𝑎𝑛 ⃓ ≤ 𝐶 for all 𝑛 ≥ 1.

This implies that the set 𝐴 ≔ 𝑎𝑛 : 𝑛 ≥ 1is also bounded. Thus, by

Theorem 2.2 (i) there exists sup 𝐴 = : 𝑠𝑢𝑝 𝑛 ≥ 1 an denoted by a. Let

us prove that 𝑎𝑛 → 𝑎, 𝑛 → ∞. We first note that 𝑎𝑛 ≤ a for all 𝑛 ≥ 1,

since the supremum of A is also its upper bound (see Definition 2.6).

Next, we take an arbitrary ε > 0 and use Theorem 2.1 (i). So, there

exists a number m such that〖 a〗_m > a − ε. By the monotonicity, a −

ε < 𝑎𝑚 ≤ 𝑎𝑛 for all 𝑛 ≥ 𝑚. Thus, setting 𝑁 ∶= 𝑚, one has

𝑎 − 𝜀 < 𝑎𝑛 ≤ 𝑎 for all 𝑛 ≥ 𝑁 which implies |𝑎𝑛 − 𝑎| < 𝜀.

Exercise:

Prove Theorem above for decreasing sequences.

Theorem above remains valid if one requires the monotonicity of

(𝑎𝑛 )𝑛 ≥ 1 starting from some number m, that is, the monotonicity of

(𝑎𝑛 )n≥m = (𝑎𝑚 , 𝑎𝑚+1 , . . .).

Example:

𝑙𝑖𝑚 10𝑛

Prove that n! = 0, where 𝑛! ∶= 1 · 2 · 3 · . . .· 𝑛. Solution.

𝑛 → ∞ 𝑛!𝑛+1

10 10𝑛

First we note that(𝑛+1)! < 𝑛! ⇔ 10𝑛 + 1 ⇔ 𝑛 > 9. Hence, the sequence

10𝑛

( 𝑛! )𝑛≥10 is decreasing. Moreover, it is bounded below by zero. Thus,

10𝑛

( 𝑛! )𝑛≥10 is bounded, by Exercise 4.1 b). Using Theorem 4.1, one gets

𝑙𝑖𝑚 10𝑛 10𝑛+1 10𝑛

that there exists a ∈ R such that = 𝑎. But we can = •

𝑛 → ∞ 𝑛! (𝑛+1)! 𝑛!

10

. So,

𝑛+1

10𝑛+1 10𝑛 10

𝑎 = lim = lim · lim = 𝑎 ·

𝑛→∞ (𝑛 + 1)! 𝑛→∞ 𝑛! 𝑛→∞ 𝑛 + 1

This implies 𝑎 = 0

Exercise:

Show that

𝑛!

a) lim 2 =0

𝑛→∞ 2𝑛

𝑛

b) lim =0

𝑛→∞ 2√𝑛

157

Exercise 4.5.

Exercise 4.6.

1

Let 𝑎1 = 1 and 𝑎𝑛+1 = 3 (𝑎𝑛+1 ) for all n ≥ 1.

b) Use induction to show that 𝑎𝑛 > 1 2 for all 𝑛 ≥ 1.

c) Show that (𝑎𝑛 )𝑛 is a decreasing sequence. d) Show that limn→∞ an

exists and find it.

Exercise 4.7.

Let c > 0, a1 > 0 and let an+1 = 1 2 an + c an for all n ≥ 1.

a) Show that an ≥ √ c for all n ≥ 2.

b) Show that (an)n≥2 is a decreasing sequence.

c) Show that limn→∞ an exists and find it.

Theorem 4.2.

(i) If (an)n≥1 is an unbounded increasing sequence, then limn→∞ an

= +∞.

(ii) If (an)n≥1 is an unbounded decreasing sequence, then limn→∞ an

= −∞.

Proof.

We will prove only Part (i) of the theorem. The proof of Part (ii) is

similar. If (an)n≥1 is an unbounded increasing sequence, then it must

be unbounded above, since it is bounded below by a1. Taking any C

and using the unboundedness of (an)n≥1, one can find a number m ∈

N such that am ≥ C. Next, by the monotonicity of (an)n≥1, the

inequality an ≥ am ≥ C trivially holds for all n ≥ N := m. This proves

limn→∞ an = +∞ (see Definition 3.4).

Corollary 4.1.

If (an)n≥1 is a monotone sequence, then the sequence either converges,

diverges to +∞, or diverges to −∞. Thus limn→∞ an is always meaningful

for monotone sequences. Proof. The proof immediately follows from

theorems 4.1 and 4.2.

Exercise 4.8.

Let A be a bounded nonempty subset of R such that sup A is not in A.

Prove that there is an increasing sequence (an)n≥1 of points from A such

that limn→∞ an = sup A.

158

4.2 The number e In this section, we will consider two sequences of

positive numbers an := 1 + 1 n n n≥1 and bn := 1 + 1 n n+1! n≥1 (1) and

study their properties.

Theorem 4.3.

The sequences defined in (1) satisfy the following properties: 1) an < bn

for all n ≥ 1; 2) the sequence (an)n≥1 increases; 3) the sequence (bn)n≥1

decreases.

Proof.

Since bn = an 1 + 1 n = an + an n > an for all n ≥ 1, Property 1) is

proved. To prove 2), we are going to use Bernoulli’s inequality (see

Theorem 2.6). So, one has an an−1 = n + 1 n n n − 1 n n−1 = n n − 1 1

− 1 n2 n > n n − 1 1 − n n2 = 1, for all n ≥ 2. Thus, an > an−1 for all n ≥

2. For the prove of 3) we use the same argument. We consider bn−1 bn

= n n − 1 n n n + 1n+1 = n − 1 n n 2 n2 − 1 n+1 = n − 1 n 1 + 1 n2 − 1

n+1 > n − 1 n 1 + n + 1 n2 − 1 = 1 for all n ≥ 2. Hence, bn−1 > bn for all n

≥2

Theorem 4.3

yields the following inequalities a1 < a2 < . . . < an < . . . < bn < . . . < b2 <

b1. (2) Consequently, the sequences (an)n≥1 and (bn)n≥1 are monotone

and bounded. By Theorem 4.1, they converge. We set e := limn→∞ an =

limn→∞ 1 + 1 n n = 2, 718281828459045... It is known that e is an

irrational number. The number e is one of the most important constants in

mathematics. Since bn = an 1 + 1 n for all n ≥ 1, one has bn → e, n → ∞.

We also note that 1 + 1 n n < e < 1 + 1 n n+1 , (3) by inequalities (2).

Definition 4.2.

The logarithm to base e is called the natural logarithm and is denoted by

ln := loge , that is, for each a > 0 ln a is a (unique!) real number such that

e ln a = a. The inequality (3) immediately implies 1 n + 1 < ln 1 + 1 n < 1

n for all n ≥ 1.

Exercise 4.9. Show that limn→∞ n ln 1 + 1 n = 1.

Exercise 4.10. Prove that for each x > 0 the sequence 1 + x n n n≥1 is

increasing and bounded.

Subsequences

Definition4.3.

1 Subsequences and Subsequential Limits Let (an)n≥1 be a sequence.

We consider any subsequence (nk)k≥1 of natural numbers such that 1 ≤

n1 < n2 < . . . < nk < nk+1 < . . .. We note that nk ≥ k and nk → +∞, k →

∞.

Example 4.3.

1) nk = k, k ≥ 1; then (nk)k≥1 = (1, 2, 3, . . . , k, . . .);

2) nk = 2k, k ≥ 1; then (nk)k≥1 = (2, 4, 6, . . . , 2k, . . .);

159

3) nk = k 2 , k ≥ 1; then (nk)k≥1 = (1, 2, 9, . . . , k2 , . . .);

4) nk = 2k , k ≥ 1; then (nk)k≥1 = (2, 4, 8, . . . , 2 k , . . .).

Definition 4.3.

A sequence (ank )k≥1 = (an1 , an2 , an3 , . . . , ank , . . .) is said to be a

subsequence of (an)n≥1. Thus, (ank )k≥1 is just a selection of some

(possibly all) of the an’s taken in order.

Remark 4.2.

The following properties follows from the definition of subsequence. 1. If

a sequence is bounded, then every its subsequence is bounded. 2. If a

sequence converges to a (that could be +∞ or −∞), then every its

subsequence also converges to a.

Exercise 4.11.

Prove that a monotone sequences which contains a bounded

subsequence is bounded.

Exercise 4.12.

Prove that a sequence (an)n≥1 converges iff (a2k)k≥1, (a2k−1)k≥1 and

(a3k)k≥1 converge.

Definition 4.4.

A subsequential limit of a sequence (an)n≥1 is any real number or the

symbol +∞ or −∞ that is the limit of some subsequence of (an)n≥1. Let A

denotes the set of all subsequential limit of (an)n≥1.

Example 4.4.

a) For the sequence (1, 2, 3, . . . , n, . . .) the set of all subsequential limit

A = {+∞}.

b) For the sequence (−1, 1, −1, . . . ,(−1)n , . . .) the set of all

subsequential limit A = {−1, 1}.

c) If an → a, then A = {a}, by Remark 4.2.

Exercise 4.13. Prove the following statements. a) −∞ ∈ A ⇔ (an)n≥1 is

unbounded below. b) +∞ ∈ A ⇔ (an)n≥1 is unbounded above.

Exercise 4.14. Find the set A of all subsequential limits of the following

sequences.

a) (sin 3πn)n≥1;

b) (sin απn)n≥1 for α ∈ Q;

c) (an)n≥1, where an = ( (−1) n+1 2 + n, if n is odd, (−1) n 2 + 1 n , if n is

even.

4.3.2 Existence of Monotone Subsequence

Theorem 4.4.

160

A number a ∈ R is a subsequential limit of a sequence (an)n≥1 iff ∀ε > 0

∀N ∈ N ∃n˜ ∈ N : ˜n ≥ N, |an˜ − a| < ε. (4)

Proof.

We first prove the necessity. Let a ∈ A. Then there exists a subsequence

(ank )k≥1 such that ank → a, k → ∞. We fix an arbitrary ε > 0 and N ∈ N.

By the definition of the limit, ∃K1 ∈ N ∀k ≥ K1 : |ank − a| < ε. Similarly,

∃K2 ∈ N ∀k ≥ K2 : nk ≥ N. Thus, taking ˜k := max{K1, K2}, ˜n := nk˜, one

has ˜n ≥ N and |an˜ − a| < ε. To prove the sufficiency, we are going to

construct a subsequence of (an)n≥1 converging to a. Let (4) holds. Then,

by (4), for ε = 1 and N = 1 there exists n1 ≥ 1 such that |an1 − a| < 1.

Similarly, for ε = 1 2 and N = n1 + 1 there exists n2 ≥ n1 + 1 such that

|an2 − a| < 1 2 and so on. Consequently, we obtain a subsequence (ank

)k≥1 satisfying |ank − a| < 1 k for all k ≥ 1. Using Theorem 3.7 and

Exercise 3.5 a), one can see that ank → a, k → ∞

Exercise 4.15.

Show that +∞ ∈ A (−∞ ∈ A) provided ∀C ∈ R ∀N ∈ N ∃n˜ ∈ N : ˜n ≥ N

and an˜ ≥ C(an˜ ≤ C).

Theorem 4.5.

Every sequence of real numbers contains a monotone subsequence.

Proof.

We consider the set M := {n ∈ N : ∀m > n am > an}. If M is infinite, then

M can be written as M = {n1, n2, . . . , nk, . . .}, where n1 < n2 < . . . <

nk < . . .. By the definition of M, we have an1 < an2 < . . . < ank < . . ..

So, the subsequence (ank )k≥1 increases. If M is finite, then let n1 be

the smallest natural number such that ∀m ≥ n1 : m 6∈ M. Since n1 6∈

M, one can find n2 > n1 such that an1 ≥ an2 . Similarly, since n2 6∈ M,

one can find n3 > n2 such that an2 ≥ an3 and so on. Thus, the

constructed subsequence (ank )k≥1 decreases.

Corollary 4.2

For every sequence the set of its sub sequential limits is not empty.

Proof

The corollary immediately follows from Theorem 4.5 and Corollary 4.1.

Theorem 4.6

(Bolzano-Weierstrass theorem). Every bounded sequence has a

convergent subsequence

Proof.

The theorem is a direct consequence of theorems 4.5 and 4.1.

161

Exercise

1

1.{4𝑛}−∞

𝑛=1

2𝑛−3

2.{3𝑛−2 }+∞

𝑛=1

𝑎𝑛+1

Determine if the sequence is increasing or decreasing by calculating .

𝑎𝑛

1

3.{.4𝑛}+∞

𝑛=1

𝑒 𝑛 −𝑒 −𝑛

4. {𝑒 𝑛+𝑒 −𝑛}+∞

𝑛=1

Determine if the sequence is increasing or decreasing by calculating the derivative

𝑎′𝑛 .

1

5. { 𝑛}−∞

𝑛=1

4

𝐼𝑛(2𝑛) +∞

6{ }

𝐼𝑛(6) 𝑛=1

Use an appropriate test for monotonicity to determine if

the sequence increases, decreases, eventually increases, or eventually decreases.

3𝑛 +∞

7.{ }𝑛=1

2𝑛+1

1

8.{𝑛 − 𝑛}+∞ 𝑛=1

𝑛2 +∞

9.{ 𝑛! }𝑛=1

2𝑛=1

10{. (2𝑛)! }+∞

𝑛=1

𝑒 √𝑛

11{. 𝑛 }+∞

𝑛=1

12.{𝑒 𝑛 𝜋 −𝑛 }+∞

𝑛=1

𝑛! +∞

13 {𝑛!}𝑛=1

2)

3(𝑛

14.{1000𝑛}+∞𝑛=1

15.{𝑛3 𝑒 _𝑛 }+∞

𝑛=1

162

Answers key

2. The sequence is (strictly) increasing.

3. The sequence is (strictly) decreasing.

4. The sequence is (strictly) increasing.;

5. The sequence is (strictly) decreasing.

6. The sequence is (strictly) increasing.

7. The sequence is (strictly) increasing.

8. The sequence is (strictly) increasing.

9. The sequence is eventually (strictly) decreasing.

10. The sequence is (strictly) decreasing.;

11. The sequence is eventually (strictly) increasing.

12. The sequence is (strictly) decreasing.

13. The sequence is (strictly) decreasing.

14. The sequence is eventually (strictly) increasing.

15. The sequence is eventually (strictly) decreasing.

163

SUBSEQUENCES

JOAN DIGOLTO

164

Sub- Sequences

Definition:

Let (an) be a sequence of real numbers, and let n1 < n2 < n3 < · · · be

a strictly increasing sequence of natural numbers. Then the sequence

an1 , an2 , an3 , . . . , is called a subsequence of (an) and is denoted

by (anj ), where j ∈ N indexes the subsequence. Notice that the order

of the terms in a subsequence (anj ) is the same as in the original

sequence

Example:

If an = 1/n2 , then 1, 1 9 , 1 25 , 1 49 , . . .

1 1 1

(1, 9,,25,49,…)

And

1 1 1 1

(4,16,64 , 256,…)

Are subsequences of (𝑎𝑛 ).

For the first of these we have (𝑎𝑛 ) where

𝑛𝑗 =2𝑗 -1

And for the second of these we have

𝑛𝑗 =2𝑗 .

Both of these forms of 𝑛𝑗 give strictly increasing sequences of positive

integers.

Theorem:

All subsequences of a convergent sequence converge to the same limit

as the original sequence.

Proof:

Let (𝑎𝑛 ) be a convergence sequence with limit l.

Suppose (𝑎𝑛 ) is a subsequence of (𝑎𝑛 ).

For 𝜖 > 0 we must find J ∈ N such that |𝐴𝑛̇ 𝑗 − l| < 𝜖 for all j ≥ J.

Since 𝑎𝑛 → l as n → ∞, there is N ∈ N such that |𝑎𝑛 − l| < 𝜖 for all n ≥ N.

By the nature of 𝑛𝑗 , there is a J ∈ N such that 𝑛𝑗 ≥ N for all j ≥ J.

Then because |𝑎𝑛 − l| < 𝜖 for all n ≥ N, and because 𝑛𝑗 ≥ N for all j ≥ J,

we have that |𝑎𝑛𝑗 − l| < 𝜖 for all j ≥ J.

Example:

For 0 < b < 1 we have b > b2 > b3 > b4 > · · · > bn > · · · > 0.

Thus the sequence (b n ) is decreasing and bounded below, and so it

converges by the Monotone Convergence Theorem.

A reasonable guess for the limit is 0, but we can confirm that by the

Algebraic Limit Theorem and a

If l is the limit of (𝑏 𝑛 ), then l is the limit of the subsequence (𝑏 2𝑛 ).

But 𝑏 2𝑛 = 𝑏 𝑛 𝑏𝑛 and so we have l = 12 , and thus l = 0 (why not 1?).

Can you extend this to −1 < b < 0? It is true.

165

Divergence Criterion for Sequences

Definition:

Since all subsequences of a convergence sequence converge to the

same limit as the original, then we can detect a divergence sequence

if we can produce two subsequences that converge to different limits.

The sequence ( −1)𝑛 is not convergent because it has two

subsequences (−1)2𝑛 and ( −1)2𝑛+1 which converge to 1 and -1

respectively.

But as we have seen, a bounded sequence might have a convergent

subsequence, like (−1)𝑛 does.

It is an amazing result that every bounded sequence has a convergent

subsequence.

Example:

Show that the sequence ((−1)n) is divergent.

To show this sequence is divergent, consider the subsequence of even

terms which is (1,1,1,...) which converges to the real number 1.

Now consider the subsequence of odd terms which

is (−1,−1,−1,...) which converges to the real number −1.

Since these two subsequences converge to different limits, we

conclude that ((−1)n) is divergent.

Example:

Show that the sequence (an) defined by an={1/nnifnisevenifnisodd.

Let's first look at a few terms of this sequence. We have

that (an)=(1,12,3,14,5,16,...). We can see this sequence is not bounded

above and hence not bounded, which we will prove.

Suppose that there exists an M∈N such that ∣an∣<M for all n∈N. By the

Archimedean property since M∈R there exists a natural

number nM∈N such that M≤nM. We also note that M≤nM<nM+1.

So, either nM or nM+1 is a term in the sequence (an) which contradicts

the sequence (an) from being bounded.

Example:

Show that the sequence (n) is divergent.

Once again, this sequence is unbounded. Suppose that instead (n) is

bounded, that is ∣n∣=n<M for some M∈R. But this contradicts the

Archimedean property which says that for any M∈R there exists

an nM∈N such that M≤nM, and so in fact (n) is not bounded and by the

divergence criteria, is divergent as well.

Theorem:

Every bounded sequence contains a convergent subsequence.

166

Proof.

Let (𝑎𝑛 ) be a bounded sequence.

Then there is ℳ Є ℝ such that |𝑎𝑛 |≤

м for all n Є N.

We will construct a convergent subsequence of (an) through a

bisection technique. Bisect the closed interval [−M, M] into the closed

subintervals [−M, 0] and [0, M].

Notice the midpoint is included in both subintervals, but as we shall

see, this does not complicate things.

Since there are infinitely many an, one of the two subintervals must

contain infinitely many of them; label this closed interval I1 and choose

n1 so that 𝑎𝑛1 ∈ 𝐼1 .

Now bisect the closed interval 𝐼1 into two closed subintervals that

overlap at the midpoint.

Since there are infinitely many 𝑎𝑛 for n > 𝑛1 , one of the two closed

subintervals must contain infinitely many of them; label this closed

interval 𝐼2 and choose 𝑛2 > 𝑛2 so that 𝑎𝑛2 Є 𝐼2 .

Notice that 𝐼1 ⊇ 𝐼2 .

We can repeat this step countably many times to obtain a nested

sequence of closed intervals

𝐼2 ⊇ 𝐼2 ⊇ 𝐼3 ⊇ 𝐼4 ⊇…

and positive integers 𝑛1 < 𝑛2 < 𝑛3 < 𝑛4 < · · · such that anj ∈ Ij for all j ∈

N.

By the Nested Interval Property

Theorem:

There is at least one x ∈ R contained in every 𝐼𝐽 .

Now the suspicion is that this x is the limit of the subsequence( 𝑎𝑛𝑗 ).

Let 𝜖 > 0.

1 𝐽−1

By the bisection technique, the length of 𝐼𝐽 is M (2) which converges

to 0. Choose J so that j ≥ J implies that the length of 𝐼𝐽 is less than 𝜖.

Then as 𝑎𝑛𝑗 and x are both in the closed interval 𝐼𝐽 of length less than

𝜖, we have |𝑎𝑛 − 𝑥|< 𝜖.

For all j ≥ 𝐽

This holds for all j ≥ 𝐽 because of the nested property of 𝐼𝐽 and

because 𝑎𝑛𝑗 ∈ 𝐼𝐽 .

Thus, we have that (𝑎𝑛𝑗 ) converges to x.

Definition:

Let (𝑥𝑛 ) be a sequence of points in X. Suppose that (𝑛𝑘 ) is a strictly

increasing sequence of natural numbers. Then we say that the

sequence (𝑥𝑛𝑘 ) of points in X indexed by k, is a subsequence of (𝑥𝑛 ).

In the following examples, we take the sequence:

167

Example:

1 1 1 1 1 1

The sequence 2 , 4 , 6 , 8 , 10 , 12 , … is a subsequence of (𝑥𝑛 ). For we take

(𝑛𝑘 ) = 2𝑘 then the sequence is (𝑥𝑛𝑘 ) = 1⁄2 𝑘 for 𝑘 = 1,2,3, …

Example:

1 1 1 1 1 1 1

The sequence 3 , 5 , 7 , 9 , 11 , 13 , 15 , … is a subsequence of (𝑥𝑛 ). For if we

take 𝑛𝑘 = 2𝑘 + 1 then the sequence is (𝑥𝑛𝑘 ) = 1⁄2 𝑘 for 𝑘 ∈ ℕ

Example:

1 1 1 1 1 1 1

The sequence 4 , 3 , 2 , 1 , 5 , 6 , 7 , … is not ha subsequence of (𝑥𝑛 ). For we

take 𝑛𝑘 such that (𝑥𝑛𝑘 ) gives these terms then 𝑛1 = 4 and 𝑛2 = 3,

which is enough to contradict the fact that 𝑛𝑘 must be increasing for a

subsequence.

Example:

1 1 1 1 1 1

The sequence 1, 2 , 3 , 4 , 5 , 6 , 7 , … is not ha subsequence of (𝑥𝑛 ). For we

take 𝑛𝑘 such that 𝑥𝑛𝑘 gives these terms then 𝑛4 = 𝑛5 = 4, which is

contradicts the fact that 𝑛𝑘 must be strictly increasing for a

subsequence.

Example:

Let 𝑝𝑘 denote the kth prime number. Thus 𝑝1 = 2, 𝑝2 = 3, 𝑝3 = 5, 𝑝4 = 7

etc. This is a strictly increasing sequence of natural numbers, so the

1 1 1 1 1 1 1

sequence 𝑥𝑝𝑘 , 2 , 3 , 5 , 7 , 11 , 13 , 17 , … is a subsequence of (𝑥𝑛 ).

One important reason subsequence is useful is because very often

even when a sequence does not converge itself, it does have

subsequences which converge.

Example:

Let 𝑥𝑛 = (−1)𝑛 . Then the sequences 𝑥2𝑘 = 1 and 𝑥2𝑘+1 = −1 both

converge, although the sequence itself does not converge.

On the other hand, if a sequence does converge then all its

subsequences also converge, and to the same thing.

Proposition:

If the sequence (𝑥𝑛 ) converges to x then all its subsequences also

converges to x.

Remark:

Although it seems a pretty unsurprising result, this last proposition does

give us a useful criterion for establishing that a given sequence

does not converge. All one has to do is to identify two subsequences

which converge to different limits. Since we have far more results at

our disposal which do prove convergence than we have which disprove

it, this is oftentimes much easier.

Example:

1

The sequence 𝑥𝑛 = (−1)𝑛 (1 − 𝑛) diverges.

168

Proof.

1

The subsequence 𝑥2𝑘 = 1 − 2𝑘 converges to 1. The subsequence

1

𝑥2𝑘+1 = 2𝑘+1 − 1 converges to -1. Thus, by the proposition above

(𝑥𝑛 ) does not converge. The following theorem shows that, at least in

the real numbers, many sequences have convergent subsequences.

So, this behavior is very common indeed.

Theorem:

Bolzano- Weierstrass Theorem

Every bounded sequence in ℝ hs a convergent subsequence.

Remark:

This proof of the Bolzano-Weierstrass theorem uses ideas based on

Ramsey's Theorem from combinatorics. Indeed the key step of the

proof, showing that every sequence of real numbers has a monotonic

subsequence can be proved using Ramsey's Theorem.

We can use Ramsey's theorem to prove again that every sequence of

real numbers has a monotonic subsequence:

Proof:

Let a sequence (𝑥𝑛 ) of real numbers be given. We define a collection

of 2-element subsets of ℕ as follows:

Let {𝑚, 𝑛} be a 2-element subset of ℕ. Then 𝑚 ≠ 𝑛, so one of 𝑚, 𝑛 is

smaller than the other. Assume that 𝑚 < 𝑛. We include {𝑚, 𝑛} in S if

𝑥𝑚 < 𝑥𝑛 , otherwise we exclude {𝑚, 𝑛} from S and use Ramsey’s

theorem to obtain an infinite set L with one or other of the two

properties described.

If the first case holds, then whenever 𝑚 < 𝑛 and 𝑚, 𝑛 ∈ 𝐿 then {𝑚, 𝑛} is

in S, and so 𝑥𝑚 < 𝑥𝑛 . Thus, if we list the elements of L in increasing

order as 𝑙𝑘 then (𝑥𝑙𝑘 ) is an increasing subsequence.

If the second case holds, then whenever 𝑚 < 𝑛 and 𝑚, 𝑛 ∈ 𝐿, then

{𝑚. 𝑛} ∉ 𝑆 and so 𝑥𝑚 ≥ 𝑥𝑛 . Thus, in this case, if we list the elements of

L in increasing order as 𝑙𝑘 then (𝑥𝑙𝑘 ) is decreasing subsequence.

Following on the Bolzano-Weierstrass Theorem, we can use obtain a

useful property of closed and bounded subsets of the real numbers:

Definition:

We say that a sequence of real numbers {an} is a Cauchy sequence

provided that for every > 0, there is a natural number N so that when

n, m ≥ N, we have that |an − am| ≤ .

Example:

Let x be a real number and 𝑡𝑛 (x) be the nth truncation of its decimal

expansion as in Lectures 2 and 3. Then if n, m ≥ N, we have that | 𝑡𝑛

(x)−𝑡𝑚 (x)| ≤ 10−N , since they share at least the first N places of their

decimal expansion. Given any real number 𝜖 > 0, there is an N(𝜖) so

that 10−N() < .Thus, we have shown that the sequencese {𝑡𝑛 (x)} is

a Cauchy sequence.

169

Example 1 was central in our construction of the real numbers. We got

the least upper bound property by associating to each sequence as in

Example 1, the real number x which is its limit. The class of Cauchy

sequences should be viewed as minor generalization of Example 1 as

the proof of the following theorem will indicate.

Theorem:

Every Cauchy sequence of real numbers converges to a limit.

Proof:

Let {𝑎𝑛 } be a Cauchy sequence.

For any j, there is a natural number Nj so that whenever n, m ≥ 𝑁𝐽 , we

have that |𝑎𝑛 − 𝑎𝑚 | ≤ 2 −j . We now consider the sequence {𝑏𝑗 } given

by 𝑏𝑗 = 𝑎𝑁𝑗 − 2−𝑗 .

Notice that for every n larger than 𝑁𝑗 , we have that 𝑎𝑛 > 𝑏𝑗 . Thus, each

bj serves as a lower bound for elements of the Cauchy sequence {𝑎𝑛 }

occuring later than 𝑁 𝑗 .

Each element of the sequence {𝑏𝑗 } is bounded above by 𝑏1 + 1, for the

same reason. Thus, the sequence {𝑏𝑗 } has a least upper bound which

we denote by L. We will show that L is the limit of the sequence e {𝑎𝑛 }.

Suppose that n > 𝑁𝑗 .

Then |𝑎𝑛 − L| < 2 −j + |𝑎𝑛 − 𝑏𝑗 | = 2−j + 𝑎𝑛 − 𝑏𝑗 ≤ 3(2−𝑗 ).

For every > 0 there is j(𝜖) so that 21−𝑗 < and we simply take N(𝜖) to

𝑁𝑗 (𝜖) .

The idea of the proof of Theorem 1 is that we recover the limit of the

Cauchy sequence by taking a related least upper bound. So, we can

think of the process of finding the limit of the Cauchy sequence as

specifying the decimal expansion of the limit, one digit at a time, as this

how the least upper bound property worked.

The converse of Theorem 1 is also true.

Theorem:

Let {𝑎𝑛 } be a sequence of real numbers converging to a linit L. Then

the sequence {𝑎𝑛 } is a Cauchy sequence.

Proof:

Since {𝑎𝑛 } converges to L, for every 𝜖 > 0, there is an N > 0 so that

𝜖

when j > N, we have |𝑎𝑗 − L| ≤ 2.

𝜖 𝜖

(The reason we can get 2on the right hand side is that we put 2in the

role of 𝜖 in the definition of the limit.) Now if j and k are both more than

𝜖 𝜖

N, we have |𝑎𝑗 − L| ≤ 2 and |𝑎𝑗 − L| ≤ 2 . Combining these using the

triangle inequality, we get

|𝑎𝑗 − 𝑎𝑘 | ≤ 𝜖,

so that the sequence {𝑎𝑗 } is a Cauchy sequence as desired.

Combining Theorems above, we see that what we have learned is that

Cauchy sequences of real numbers and convergent sequences of real

numbers are the same thing. But the advantage of the Cauchy criterion

170

is that to check whether a sequence is Cauchy, we don’t need to know

the limit in advance.

Example:

Consider the series (that is, infinite sum)

S = ∑∞𝑛=1 1𝑛2 .

We may view this as the limit of the sequence of partial sums

𝑗

𝑎𝑗 = ∑𝑛=1 1𝑛2.

We can show that the limit converges using Theorem 1 by showing that

{𝑎𝑗 } is a Cauchy sequence. Observe that if j, k > N, we definitely have

𝑗

|𝑎𝑗 − 𝑎𝑘 | ≤ ∑𝑛=𝑁 1𝑛2.

It may be difficult to get an exact expression for the sum on the right,

but it is easy to get an upper bound.

∞ ∞ ∞

1 1 1

∑ 1𝑛2 ≤ ∑ = ∑ − .

𝑛(𝑛 − 1) 𝑛(𝑛 − 1) 𝑛

𝑛=𝑁 𝑛=𝑁 𝑛=𝑁

1

The reason we used the slightly wasteful inequality, replacing by

𝑛2

1

that now the sum on the right telescopes, and we know it is exactly

𝑛2−𝑛

1

equal to 𝑁−1 . To sum up, we have shown that when j, k > 𝑁, we have

1

|𝑎𝑗 − 𝑎𝑘 | ≤ 𝑁−1 .

Since we can make the right hand side arbitrarily small by taking N

sufficiently large, we see that {𝑎𝑗 } is a Cauchy sequence. This example

gives an indication of the power of the Cauchy criterion. You would not

have found it easier to prove that the limit exists if I had told you in

𝜏2

advance that the series converges to 6 .

Let {𝑎𝑛 } be a sequence of real numbers. Let {𝑛𝑘 } be a strictly increasing

sequence of natural numbers. We say that {𝑎𝑛𝑘 } is a subsequence of

{𝑎𝑛 } . We will now prove an important result which helps us discover

convergent sequences in the wild.

Theorem 3:

(Bolzano-Weierstrass)

Let {𝑎𝑛 be a bounded sequence of real numbers. (That is, suppose

there is a positive real number B, so that |𝑎𝑗 | ≤ B for all j.) Then {𝑎𝑛 }

has a convergent subsequence.

Proof of Bolzano-Weierstrass

All the terms of the sequence live in the interval

𝐼0 = [−B, B].

We can make the right-hand side arbitrarily small by making N

sufficiently large. Thus, we have shown that the subsequence is a

Cauchy sequence and hence convergent.

A question you might ask yourselves is: How is the proof of the

Bolzano Weierstrass theorem related to decimal expansions?

171

Our final topic for today’s lecture is the Squeeze theorem. It is a result

that allows us to show that limits converge by comparing them to limits

that we already know converge.

Theorem 4:

(Squeeze theorem)

Given three sequences of real numbers {𝑎𝑛 }, {𝑏𝑛 }, and {𝑐𝑛 }. If we know

{𝑎𝑛 } and {𝑏𝑛 }, that both converge to the same limit L and we know that

for each n we have

𝑎𝑛 ≤ 𝑐𝑛 ≤ 𝑏𝑛 ,

then the sequence {𝑐𝑛 } also converges to the limit L.

172

Exercises:

(a) what is the value of a3?

(b) Find the value of ∑5𝑛=1 𝑎𝑛

3. List the first four terms of the sequence {an} = {n2}, starting with n = 1.

4. List the first four terms of the following sequence, beginning with n = 0.

5. Find the sum of the first six terms of An, where an = 2an–1 + an–2, a1 = 1, and a2 =1.

173

Answer Key: Subsequences

1. Given the sequence An = {1, 3, 5, 7, 9},

(a) what is the value of a3?

(b) Find the value of ∑5𝑛=1 𝑎𝑛

Solution:

(a) The index of a3 is n = 3 so they're asking me for the third term, which is "5".

(b) The funky symbol is the Greek capital letter "sigma", indicating a series. That

means that they're asking me here to do the addition of the terms of the sequence.

The "value" they're asking me to find is the total, the sum, of all the

terms an from a1 to a5; in other words:

a1 + a2 + a3 + a4 + a5

= 1 + 3 + 5 + 7 + 9 = 25

Then my answers are:

value of a3: 5

value of sum: 25

They've given me a rule for each term of this series; the rule is to multiply the index by

two. So, to find each term, I'll plug the value of n into the formula; namely, I'll take the

index and multiply by two. I'll be starting with n = 0 and ending with n = 4. To find the

series sum, I'll be adding all the terms, like this:

2(0) + 2(1) + 2(2) + 2(3) + 2(4)

= 0 + 2 + 4 + 6 + 8 = 20

I'll just plug n into the formula, and simplify:

{a1, a2, a3, a4}

= {12, 22, 32, 42}

= {1, 4, 9, 16}

My answer is the simplified form of the sequence:

{1, 4, 9, 16}

4. List the first four terms of the following sequence, beginning with n = 0

An=(n+1)!(−1)n

Sequences and series are often the first place students encounter this exclamation-

mark notation. The notation doesn't indicate that the series is "emphatic" in some

manner; instead, this is technical mathematical notation. It indicates that the terms of

this summation involve factorials. (If you're not familiar with factorials, brush up now.)

A factorial symbol, k!, indicates that I need to find the product of all the whole numbers

from 1 through k. The first few factorial values are: 1! = 1

174

2! = 1 × 2 = 2

3! = 1 × 2 × 3 = 6

4! = 1 × 2 × 3 × 4 = 24

(Your graphing calculator can probably find factorials for you. Look for an appropriate

command, probably somewhere in a "Prob" or "Probability" submenu.)

I'll use these factorial values in my computations:

I'll use these factorial values in my computations:

1 1

𝑎0= (−1)0 =1!=1=1

(0+1)!

(−1)1 −1 1

𝑎1 =(1+1)!= 2! =− 2

(−1)2 1 1

𝑎2 = = =

(2 + 1)! 3! 6

(−1)3 −1 1

𝑎3 =((3+1)= 4! =− 24

So the first four terms are:

1 1 1

1, 2, 6,- 24

5. Find the sum of the first six terms of An, where an = 2an–1 + an–2, a1 = 1, and a2 =1.

This formula looks much worse than it really is; I just have to give myself some time,

and dissect the formula carefully.

They gave me the values of the first two terms, and then they gave me a formula that

says that each term (after the first two terms) is a sum formed from the previous two

terms. At each stage, I'll be taking the previous term and multiplying it by two; to this,

I'll be adding the term before that one. For instance, the third term will be twice the

second term, plus the first term.

Plugging into this formula, I get:

a3 = 2a3–1 + a3–2

= 2a2 + a1

= 2(1) + (1)

=2+1=3

a4 = 2a4–1 + a4–2

= 2a3 + a2

= 2(3) + (1)

=6+1=7

a5 = 2a5–1 + a5–2

= 2a4 + a3

= 2(7) + (3)

= 14 + 3 = 17

a6 = 2a6–1 + a6–2

= 2a5 + a4

= 2(17) + (7)

175

= 34 + 7 = 41

Now that I've found the values of the third through the sixth terms, I can find the value

of the series; the sum is:

1 + 1 + 3 + 7 + 17 + 41 = 70

176

THE CAUCHY CRITERION

177

THE CAUCHY CRITERION

Definition:

A sequence {𝑥𝑛 } of real numbers is said to be Cauchy sequence if for

every ε > 0 there exists N ∈N such that if n, m > N ⇒|𝑥𝑛 −𝑥𝑚 |< ε.

A sequence is Cauchy if the terms eventually get arbitrarily close to each

other.

Example:

1

The sequence {𝑛} is Cauchy. To see this let ε > 0 be given. Choose N ∈ N

1 𝜀 1 1 1 1 𝜀 𝜀

such that 𝑁 <2. Now, if n, m > N ⇒|𝑛 − 𝑚| ≤ 𝑛 + 𝑚 < 2 + 2 = ε.

Example:

𝑛

The sequence { } is Cauchy. To see this let ε > 0 be given. Choose N ∈N

𝑛+1

1 𝜀 𝑛 𝑚

such that 𝑁 <2. Now, if n, m > N ⇒| 𝑛+1 − 𝑚+1|=

𝑚+1−𝑛−1 𝑚−𝑛 1 1 𝜀 𝜀

| (𝑛+1)(𝑚+1)|≤| 𝑛𝑚 |< 𝑛 + 𝑚 < 2 + 2= ε.

Lemma:

Let sequence {𝑥𝑛 } be a Cauchy sequence of real numbers. Then {𝑥𝑛 } is

bounded.

Proof:

Since {𝑥𝑛 } is a Cauchy sequence, then there exists N ∈N such that if n, m > N

⇒|𝑥𝑛 −𝑥𝑚 |< 3.

if n, m > N ⇒|𝑥𝑛 −𝑥𝑚 |< 3

Let m = N +1, if n > N ⇒|𝑥𝑛 −𝑥𝑁 +1|< 3 Note: |𝑥𝑛 |−|𝑥𝑁 +1|≤|𝑥𝑛 −𝑥𝑁 +1|

⇒|𝑥𝑛 |−|𝑥𝑁 +1|≤|𝑥𝑛 −𝑥𝑁 +1|< 3

if n > N ⇒|𝑥𝑛 |< 3+|𝑥𝑁 +1|.

Let M = max {|𝑥1 |,|𝑥2 |,···|𝑥𝑁 |,|𝑥𝑁 +1|+3}

Now, if n > N ⇒|𝑥𝑛 |< 3+|𝑥𝑁 +1|≤M

Now, if n ≤ N ⇒|𝑥𝑛 |< max {|𝑥1 |,|𝑥2 |,···|𝑥𝑁 |}≤M

Thus ∀ n ∈ N, |𝑥𝑛 |≤ M. ∎

178

Theorem [Cauchy Convergence Criterion]

Cauchy sequence.

Proof:

Let {𝑥𝑛 } be a sequence of real numbers.

(⇒) Suppose that limn→∞𝑥𝑛 = x ∈R. We want to show that {𝑥𝑛 } is

Cauchy sequence.

Give let ε > 0. Since 𝑙𝑖𝑚𝑛→∞ 𝑥𝑛 = x ∴ ∃ N ∈N 3

𝜀

if n > N ⇒|𝑥𝑛 −x|<2

𝜀

Also, if m > N ⇒|𝑥𝑚 −x|< 2

𝜀 𝜀

Now, if n, m > N ⇒|𝑥𝑛 −𝑥𝑚 |=|𝑥𝑛 –x + x−𝑥𝑚 |≤|𝑥𝑛 −x|+|𝑥𝑚 −x|<2 + 2= ε.

(⇐) Suppose that {𝑥𝑛 } is a Cauchy sequence. We want to show that {𝑥𝑛 } is

convergent.

Let ε > 0 be given. Since {𝑥𝑛 } is a Cauchy sequence, then by Lemma 1

it is bounded.

Hence {𝑥𝑛 } has a converge subsequence {𝑥𝑛𝑘 }.

Suppose lim 𝑥𝑛𝑘 = x∈R.

𝑘→∞

𝜀

There exist 𝑁1 ,𝑁2 ∈N 3 if n, m > 𝑁1 ⇒|𝑥𝑛 − 𝑥𝑚 |< 2

𝜀

And, if k > 𝑁2 ⇒|𝑥𝑛𝑘 −x|<2.

𝜀

Now, ﬁx k > 𝑁2 such that 𝑛𝑘 > 𝑁1 and, if n > 𝑁1 ⇒|𝑥𝑛 −𝑥𝑛𝑘 |< 2 and |𝑥𝑛𝑘

𝜀

−x|<2.

𝜀 𝜀

Now, if n > 𝑁1 ⇒|𝑥𝑛 −x|=|𝑥𝑛 −𝑥𝑛𝑘 +𝑥𝑛𝑘 −x|≤|𝑥𝑛 −𝑥𝑛𝑘 |+|𝑥𝑛𝑘 −x|< 2 + 2= ε.

179

Cauchy’s insight

Our diﬃculty in proving “𝑎𝑛 → l” is this: What is `? Cauchy saw that it was

enough to show that if the terms of the sequence got suﬃciently close to

each other then completeness will guarantee convergence.

Remark. In fact, Cauchy’s insight would let us construct R out of Q if we

had time.

Deﬁnition

Let (𝑎𝑛 ) be a sequence [R or C]. We say that (an) is a Cauchy sequence if,

for all ε > 0 there exists N ∈N such that

m, n > N =⇒|𝑎𝑚 −𝑎𝑛 | < ε.

[Is that all? Yes, it is!

Cauchy ⇒ Bounded

Theorem:

Every Cauchy sequence is bounded [R or C].

Proof:

1 > 0 so there exists N such that m, n > N =⇒ |𝑎𝑚 − 𝑎𝑛 | < 1. So form > N,

|𝑎𝑚 |6 1+|𝑎𝑁 | by the ∆ law. So for all m

|𝑎𝑚 |6 1+|𝑎1 |+|𝑎2 |+···+|𝑎𝑁 |. ∎

Convergent ⇒ Cauchy [R or C]

Theorem:

Every convergent sequence is Cauchy.

Proof:

Let 𝑎𝑛 → l and let ε > 0. Then there exists N such that

k > N =⇒|𝑎𝑘 −l| < ε/2

For m, n > N we have

|𝑎𝑚 −l| < ε/2

|𝑎𝑛 −l| < ε/2

180

So

|𝑎𝑚 −𝑎𝑛 | ≤ |𝑎𝑚 −l| + |𝑎𝑛 −l| by the ∆ law

< ε/2 + ε/2 = ε ∎

Theorem:

Every real Cauchy sequence is convergent.

Proof:

Let the sequence be (𝑎𝑛 ). By the above, (𝑎𝑛 ) is bounded. By Bolzano-

Weierstrass (𝑎𝑛 ) has a convergent subsequence (𝑎𝑛𝑘 ) → l, say. So let ε > 0.

Then

∃N1 such that r ⩾ N1 =⇒ |𝑎𝑛𝑟 −l| < ε/2

∃N2 such that m, n ⩾ N2 =⇒ |𝑎𝑚 −𝑎𝑛 | < ε/2

Put s: = min {r|𝑛𝑟 ⩾ N2} and put N =𝑛𝑠 . Then

M, n⩾ N ⇒ |𝑎𝑚 −𝑎𝑛 |

⩽ |𝑎𝑚 −𝑎𝑛𝑠 | + |𝑎𝑛𝑠 −l|

< ε/2 + ε/2 = ε ∎

Theorem:

Every complex Cauchy sequence is convergent.

Proof:

Put 𝑧𝑛 = x + iy. Then 𝑥𝑛 is Cauchy: |𝑥𝑥 −𝑥𝑚 | 6 |𝑧𝑛 −𝑧𝑚 | (as |Rw| ⩽ |w|).

So 𝑥𝑛 → x, 𝑦𝑛 → y and so 𝑧𝑛 → x+iy. ∎

Cauchy Sequence in R

Deﬁnition:

A sequence 𝑥𝑛 ∈R is said to converge to a limit x if

181

• ∀² > 0, ∃N s. t. n > N ⇒ |𝑥𝑛 −x|< ².

A sequence 𝑥𝑛 ∈R is called Cauchy sequence if

• ∀², ∃N s. t. n > N & m > N ⇒ |𝑥𝑛 −𝑥𝑚 |< ².

Proposition:

Every convergent sequence is a Cauchy sequence.

Proof:

Assume 𝑥𝑘 →x. Let ² > 0 be given.

• ∃N s .t. n > N ⇒ |𝑥𝑛 −x|< ² 2.

• n, m ≥N ⇒ |𝑥𝑛 −𝑥𝑚 |≤ |x −𝑥𝑛 |+|x −𝑥𝑚 |< ² 2 + ² 2 = ².

Theorem:

Every bounded sequence in R has a subsequence that converges to some

point in R.

Proof:

Suppose 𝑥𝑛 is a bounded sequence in R. ∃M such that

−M ≤𝑥𝑛 ≤M, n = 1, 2,···. Select 𝑥𝑛0 =𝑥1 .

• Bisect 𝑙0 := [−M, M] into [−M, 0] and [0,M].

• At least one of these (either [−M, 0] or [0, M]) must contain 𝑥𝑛 for

inﬁnitely many indices n.

• Call it I1 and select n1 > n0 with xn1 ∈I0.

• Continue in this way to get a subsequence 𝑥𝑛𝑘 such that

• 𝑙0 ⊃𝑙1 ⊃𝑙2 ⊃𝑙3 ···

• 𝑙𝑘 = [𝑎𝑘 ,𝑏𝑘 ] with |𝑙𝑘 |= 2−𝑘 M.

• Choose 𝑛0 < 𝑛1 < 𝑛2 <··· with 𝑥𝑛𝑘 ∈ 𝑙𝑘.

• Since 𝑎𝑘 ≤𝑎𝑘 +1 ≤M (monotone and bounded), 𝑎𝑘 →∃ x.

• Since 𝑥𝑛𝑘 ∈𝑙𝑘 and |𝑙𝑘 |= 2−𝑘 M, we have

|𝑥𝑛𝑘 −x|<|𝑥𝑛𝑘 −𝑎𝑘 |+|𝑎𝑘 −x|≤2−𝑘−1 M+|𝑎𝑘 −x|→0 as k →∞. 2

182

Corollary: (Compactness)

Every sequence in the closed interval [a ,b] has a subsequence in R that

converges to some point in R.

Proof:

Assume a≤𝑥𝑛 ≤b for n = 1, 2,···. By Theorem 1.4.3, ∃ a subsequence

𝑥𝑛𝑘 and a≤ ∃𝑥 ≤b such that 𝑥𝑛𝑘 →x.

Lemma. (Boundedness of Cauchy sequence)

If 𝑥𝑛 is a Cauchy sequence, 𝑥𝑛 is bounded.

Proof:

∃N s. t. n ≥N ⇒ |𝑥𝑛 −x|< 1. Then 𝑠𝑢𝑝𝑛 |𝑥𝑛 | ≤1 + max {|x1|, ···,|𝑥𝑁 |} (Why?)

Theorem. (Completeness)

Every Cauchy sequence in R converges to an element in [a, b].

Proof:

Cauchy seq. ⇒ bounded seq. ⇒ convergent sub seq.

Uniform Convergence

Definition:

A sequence of functions {𝑓𝑛 } is said to converge uniformly on an Interval [a,b]

to a function f if for any ε > 0 and for all x ∈ [a, b] there exists an integer N

(independent of x but dependent on ε) such that for all x∈[a, b]

|𝑓𝑛 (x) − f(x)| < ε, ∀n≥N (1)

Remark. Every uniformly convergent sequence is point wise convergent, and

the uniform limit function is same as the pointwise limit function. However, the

converse is not true. However non-pointwise convergence implies non-

uniform convergence.

Definition:

A series of functions ∑ 𝑓𝑛 is said to converge uniformly on [a, b] if the

sequence {𝑆𝑛 } of its partial sums, defined by

S n(x) =∑𝑛𝑖−! 𝑓𝑖 (x)

183

Converges uniformly on [a, b].

Definition:

A series of functions ∑ 𝑓𝑛 converges uniformly to f on [a, b] if for ∈ > 0 and all

x ∈ [a, b] there exists an integer N (independent of x and dependent on ε)

such that for all x in [a, b]

|𝑓1 (x) + 𝑓2 (x) + … + 𝑓𝑛 (x) − f(x)| < ε, for n ≥ N

Theorem 1:

The sequence of functions {𝑓𝑛 } defined on [a, b] converges uniformly on [a, b]

if and only if for every ε > 0 and for all x ∈ [a, b], there exists an integer N such

that

|𝑓𝑛+𝑝 (x) − 𝑓𝑛 (x) | < ε, ∀ n ≥ N, p ≥ 1 … (1)

Proof:

Let the sequence {𝑓𝑛 } uniformly converge on [a, b] to the limit function f, so

that for a given ε > 0, and for all x ∈ [a, b], there exist integers n1, n2 such that

| 𝑓𝑛 (x) − f(x)| < ε/2, ∀ n ≥ n1

and

|𝑓𝑛+𝑝 (x) − f(x)| < ε/2, ∀ n ≥𝑛2 , p ≥ 1

∴ |𝑓𝑛+𝑝 (x) − 𝑓𝑛 (x)| ≤ |𝑓𝑛+𝑝 (x) − f(x)| + |𝑓𝑛 (x) − f(x)|

Conversely. Let the given condition hold so by Cauchy’s general principle of

convergence, {𝑓𝑛 } converges for each x ∈ [a, b] to a limit, say f and so the

sequence converges pointwise to f.

For a given ε > 0, let us choose an integer N such that (1) holds. Fix n, and let

p→∞ in (1). Since 𝑓𝑛+𝑝 →f as p → ∞, we get

184

Which proves that 𝑓𝑛 (x) → f(x) uniformly on [a, b].

Remark: Other form of this theorem is

The sequence of functions {𝑓𝑛 } defined on [a, b] converges uniformly on [a,

b]if and only if for every ε > 0 and for all x ∈ [a, b], there exists an integer N

such that

|𝑓𝑛 (x) − 𝑓𝑚 (x)| < ε, ∀ n, m ≥ N

Theorem 2:

A series of functions ∑ 𝑓𝑛 defined on [a, b] converges uniformly on [a, b] if

and only if for every ε > 0 and for all x∈ [a, b], there exists an integer N such

that

|𝑓𝑛+1 (x) + 𝑓𝑛+2 (x) +…+ (x)| < ε, ∀ n ≥ N, p ≥ 1 … (2)

Proof:

Taking the sequence {𝑠𝑛 } of partial sums of functions ∑ 𝑓𝑛, defined by

S n(x) = ∑𝑛𝑖−1 𝑓𝑖(𝑥)

In addition, applying above theorem, we get the result.

Example:.

Show that the sequence {𝑓𝑛 }, where

𝑛𝑥

𝑓𝑛 (x) =1+𝑛2𝑥2, for x ∈ [a, b].

Solution:

The sequence converges pointwise to f, where f(x) = 0, ∀ real x. Let {𝑓𝑛 }

converge uniformly in any interval [a, b], so that the pointwise limit is also the

uniform limit. Therefore for given ε>0, there exists an integer N such that for

all x∈ [a, b], we have

𝑛𝑥

| 1+𝑛2𝑥2-0 | < ε, ∀n≥N

185

1

If we take ε =3, and t an integer greater than N such that 1/t ∈ [a, b],

we find on

Taking n = t and x = 1/t, that

𝑛𝑥 1 1

= 2 ∢ 3 = 𝜖.

1+𝑛2𝑥2

the interval [a, b], having the point 1/t. However, since 1/t→0, the interval [a,

b] contains 0. Hence the sequence is not uniformly convergent on any interval

[a, b] containing 0.

Deﬁnition:

We say that a sequence of real numbers{an}is a Cauchy sequence provided

that for every > 0, there is a natural number N so that when n, m ≥ N, we

have that |an −am|≤ .

Example 1 Let x be a real number and 𝑡𝑛 (x) be the nth truncation of its

decimal expansion as in Lectures 2 and 3. Then if n, m ≥ N, we have that

|𝑡𝑛 (x) −𝑡𝑚 (x) |≤ 10−N, since they share at least the ﬁrst N places of their

decimal expansion. Given any real number > 0, there is an N (𝜖) so that

10−𝑁(𝜖) <𝜖. Thus, we have shown that the sequence {𝑡𝑛 (x)} is a Cauchy

sequence.

Example 1 was central in our construction of the real numbers. We got the

least upper bound property by associating to each sequence as in Example 1,

the real number x which is its limit. The class of Cauchy sequences should be

viewed as minor generalization of Example 1 as the proof of the following

theorem will indicate.

Theorem 1:

Every Cauchy sequence of real numbers converges to a limit.

Proof:

Let {an} be a Cauchy sequence. For any j, there is a natural number 𝑁𝑗 so that

whenever n, m ≥ 𝑁𝑗 , we have that |𝑎𝑛 −𝑎𝑚 |≤ 2−𝑗 j. We now consider the

sequence {𝑏𝑗 } given by

𝑏𝑗 = 𝑎𝑁𝑗 −2−𝑗.

186

Notice that for every n larger than 𝑁𝑗 , we have that an >𝑏𝑗 . Thus, each

𝑏𝑗 serves as a lower bound for elements of the Cauchy sequence {𝑎𝑛 }

occurring later than𝑁𝑗 . Each element of the sequence {bj} is bounded above

by b1 + 1, for the same reason. Thus, the sequence {𝑏𝑗 } has a least upper

bound which we denote by L. We will show that L is the limit of the sequence

{an}. Suppose that n >𝑁𝑗 . Then

For every > 0 there is j (𝜖) so that 21−j < and we simply take N (𝜖) to 𝑁𝑗 (𝜖).

The idea of the proof of Theorem 1 is that we recover the limit of the Cauchy

sequence by taking a related least upper bound. Therefore, we can think of

the process of ﬁnding the limit of the Cauchy sequence as specifying the

decimal expansion of the limit, one digit at a time, as this how the least upper

bound property worked.

The converse of Theorem 1 is also true.

Theorem 2:

Let {𝑎𝑛 } be a sequence of real numbers converging to a limit L. Then the

sequence {𝑎𝑛 } is a Cauchy sequence.

Proof:

Since {𝑎𝑛 } converges to L, for every > 0, there is an N > 0 so that when j > N,

we have

𝜖

|𝑎𝑗 −L|≤ 2

𝜖 𝜖

(The reason we can get 2 on the right hand side is that we put in the role of

2

in the deﬁnition of the limit.) Now if j and k are both more than N, we have |𝑎𝑗

𝜖

−L|≤ 2 and

𝜖

|𝑎𝑘 −L|≤2. Combining these using the triangle inequality, we get

|𝑎𝑗 −𝑎𝑘 |≤ 𝜖.

Cauchy sequences of real numbers and convergent sequences of real

numbers are the same thing. However, the advantage of the Cauchy criterion

is that to check whether a sequence is Cauchy, we do not need to know the

limit in advance.

187

EXERCISES

∞

1. (A) Show that if {𝑎𝑛 } 𝑛=1

is Cauchy then {a𝑛2 }𝑛=1

∞

is also Cauchy.

∞ ∞

such that {𝑎𝑛 }𝑛=1

Is not Cauchy.

∞

2. Let {𝑎𝑛 }𝑛=1 be a Cauchy sequence such that 𝑎𝑛 is an integer for all n ∈ N. Show that

there is a positive integer N such that 𝑎𝑛 = C for all n ≥ N, where C is a constant.

∞

3. What does it mean for a sequence {𝑎𝑛 }𝑛=1 to not be Cauchy?

sequence also converges to L.

n ∈ N. Show that such a sequence is not a Cauchy sequence. Does this sequence

converge?

188

ANSWER KEY

∞

1. (A) since {𝑎𝑛 }𝑛=1 is Cauchy, it is convergent. Since the product of two convergent

sequences is convergent the sequence {a𝑛2}𝑛=1

∞

is convergent and therefore is Cauchy.

∞

(b) Let 𝑎𝑛 = (−1)𝑛 for all n ∈N. The sequence { 𝑎𝑛 } 𝑛=1 is not Cauchy since it is

divergent. However, the sequence {a𝑛2}𝑛=1

∞

= {1,1,···} converges to 1so it is Cauchy∎

1 ∞

2. Let ∈ =2. Since {𝑎𝑛 }𝑛=1 is Cauchy, there is a positive integer N such that if m, n ≥ N

1

we have |𝑎𝑚 −𝑎𝑛 | <2. But am−an is an integer so we must have 𝑎𝑛 = 𝑎𝑁 for all n ≥ N∎

∞

3. A sequence {𝑎𝑛 }𝑛=1 is not a Cauchy sequence if there is a real number > 0 such

that for all positive integers N there exist n,m ∈N such that n, m ≥ N and |𝑎𝑛 −𝑎𝑚 |≥ ∈

∎

∞ ∞ ∞

4. Let {𝑎𝑛 } 𝑛=1 be a Cauchy sequence. Let {𝑎𝑛𝑘 } 𝑘=1 be a subsequence of {𝑎𝑛 } 𝑛=1

∞

converging to L. By Exercise? The sequence {𝑎𝑛 }𝑛=1 is convergent say to a limit L0.

By Exercise? We must have L = L’ ∎

5. We will show that there is an ∈ > 0 such that for all N ∈N there exist m and n such

that m, n ≥ N but |𝑎𝑚 −𝑎𝑛 |≥ . Note that |𝑎𝑛+1 −𝑎𝑛 | = 𝑛3 ≥ 1. Let = 1. Let N ∈N. Choose

m = N + 1 and n = N. In this case, |𝑎𝑚 −𝑎𝑛 | = 𝑁 3 ≥ 1 =. Hence, the given sequence is

not a Cauchy sequence. Since every convergent sequence must be Cauchy, the given

sequence is divergent.∎

189

PROPERLY DIVERGENT SEQUENCE

190

PROPERLY DIVERGENT SEQUENCES

Definition:

Let {𝑥𝑛 } be sequence of real numbers.

i. We say {xn } diverges to ∞ and we write lim𝑛→∞ 𝑥𝑛 = ∞, if for every

𝝀 > 0 there exists 𝑁 ∈ ℕ such that if 𝑛 > 𝑁 ⟹ 𝑥𝑛 > 𝝀.

ii. We say {xn } diverges to −∞ and we write lim𝑛→∞ 𝑥𝑛 = −∞, if for

every 𝝀 > 0 there exists 𝑁 ∈ ℕ such that if 𝑛 > 𝑁 ⟹ 𝑥𝑛 > −𝝀.

lim𝑛→∞ 𝑥𝑛 = −∞.

Example 1:

The sequence {𝑛} is Properly divergent and lim𝑛→∞ 𝑥𝑛 = ∞. To see

this let 𝝀 > 0 be given. Choose 𝑁 ∈ ℕ such that 𝜆 < 𝑁. Now, if

𝑛 > 𝑁 ⟹ 𝑛 > 𝑁 ⟹ 𝝀. Hence, 𝑙𝑖𝑚𝑛→∞ 𝑛 = ∞.

Example 2:

𝑛2 𝑛2

The sequence {𝑛+1} is Properly divergent and lim𝑛→∞ 𝑛+1 = ∞. To see

this let 𝝀 > 0 be given. Choose 𝑁 ∈ ℕ such that 𝑁 > 2𝝀. Now, if

𝑛2 𝑛2 𝑛 𝑁 2𝜆

𝑛 > 𝑁 ⟹ 𝑛+1 > 2𝑛 = > > = 𝝀.

2 2 2

Lemma 1:

Let sequence {𝑥𝑛 } and {𝑦𝑛 } be two sequences of real numbers such

that 𝑥𝑛 ≤ 𝑦𝑛 ∀ 𝑛 ∈ ℕ.

i. If lim𝑛→∞ 𝑥𝑛 = ∞, then lim𝑛→∞ 𝑦𝑛 = ∞

ii. If lim𝑛→∞ 𝑦𝑛 = ∞, then lim𝑛→∞ 𝑥𝑛 = −∞

Proof:

i. Let 𝝀 > 0 be given. Since lim𝑛→∞ 𝑥𝑛 = ∞ ∴ ∃ 𝑁 ∈ ℕ such that if 𝑛 >

𝑁 ⟹ 𝑥𝑛 > 𝝀. Now, 𝑥𝑛 ≤ 𝑦𝑛 , then, if 𝑛 > 𝑁 ⟹ 𝑦𝑛 ≥ 𝑥𝑛 > 𝝀 .

Thus, lim𝑛→∞ 𝑦𝑛 = ∞.

ii. Let 𝝀 > 0 be given. Since lim𝑛→∞ 𝑦𝑛 = −∞ ∴ ∃ 𝑁 ∈ ℕ such that if 𝑛 >

𝑁 ⟹ 𝑦𝑛 < −𝝀. Now, 𝑥𝑛 ≤ 𝑦𝑛 , then, if 𝑛 > 𝑁 ⟹ 𝑥𝑛 ≤ 𝑦𝑛 < −𝝀 .

Thus, lim𝑛→∞ 𝑥𝑛 = −∞.

Notations:

Let {𝑥𝑛 } be sequence of real numbers.

Let 𝐴1 = {𝑥1 , 𝑥2 , … , 𝑥𝑛 , … } = {𝑥𝑘 ⃒𝑘 ≥ 1}

𝐴2 = {𝑥2 , 𝑥3 , … , 𝑥𝑛 , … } = {𝑥𝑘 ⃒𝑘 ≥ 2}

⋮

𝐴𝑛 = {𝑥𝑛 , 𝑥𝑛+1 , … , … } = {𝑥𝑘 ⃒𝑘 ≥ 𝑛}

Note 𝐴1 ⊇ 𝐴2 ⊇ ⋯ ⊇ 𝐴𝑛 ⊇ 𝐴𝑛+1 ⊇ ⋯.

191

Now, 𝑠𝑢𝑝𝐴1 ≥ 𝑠𝑢𝑝𝐴2 ≥ ⋯ ≥ 𝑠𝑢𝑝𝐴𝑛 ≥ 𝑠𝑢𝑝𝐴𝑛+1 ≥ ⋯,

and 𝑖𝑛𝑓𝐴1 ≤ 𝑖𝑛𝑓𝐴2 ≤ ⋯ ≤ 𝑖𝑛𝑓𝐴𝑛 ≤ 𝑖𝑛𝑓𝐴𝑛+1 ≤ ⋯.

{𝑖𝑛𝑓𝐴_𝑛} which is increasing.

We define

𝑖𝑛𝑓 𝑖𝑛𝑓 sup 𝑥𝑘 𝑙𝑖𝑚 sup 𝑥𝑘

lim sup 𝑥𝑛 = sup 𝐴𝑛 = =

𝑛≥1 𝑛 ≥ 1𝑘 ≥ 𝑛 𝑛 → ∞𝑘 ≥ 𝑛

and

lim 𝑖𝑛𝑓 𝑥𝑛 = inf 𝐴𝑛 = =

𝑛≥1 𝑛 ≥ 1𝑘 ≥ 𝑛 𝑛 → ∞𝑘 ≥ 𝑛

Example 3:

Let 𝑥𝑛 = 3 − (−1)𝑛 ∀ 𝑛 ∈ ℕ. Find lim 𝑠𝑢𝑝𝑥𝑛 and lim 𝑖𝑛𝑓𝑥𝑛 .

𝐴1 = {4,2,4,2, … }, sup 𝐴1 = 4 and inf 𝐴1 = 2.

𝐴2 = {2,4,2,4, … }, sup 𝐴2 = 4 and inf 𝐴2 = 2.

⋮

It is clear that sup 𝐴𝑛 = 4 and inf 𝐴𝑛 = 2 ∀ 𝑛 ∈ ℕ.

Hence, lim sup 𝑥𝑛 = lim sup 𝐴𝑛 = lim 4 = 4 and

𝑛→∞ 𝑛→∞

lim inf 𝑥𝑛 = lim inf 𝐴𝑛 = lim 2 = 2.

𝑛→∞ 𝑛→∞

Example 4:

1

Let 𝑥𝑛 = (−1)𝑛 + 𝑛 ∀ 𝑛 ∈ ℕ. Find lim sup 𝑥𝑛 and lim inf 𝑥𝑛 .

1 1 1 1

𝐴1 = {0,1 + 2 , −1 + 3 , 1 + 4 , … } , sup 𝐴1 = 1 + 2 , 𝑎𝑛𝑑 inf 𝐴1 = −1

1 1 1 1

𝐴1 = {1 + 2 , −1 + 3 , 1 + 4 , … } , sup 𝐴2 = 1 + 2 , 𝑎𝑛𝑑 inf 𝐴2 = −1

⋮

1 1 1 1

𝐴2𝑛 = {1 + 2𝑛 , −1 + 2𝑛+1 , 1 + 2𝑛+2 , … } , sup 𝐴2𝑛 = 1 + 2𝑛 , 𝑎𝑛𝑑 inf 𝐴2𝑛 =

−1

1 1 1

𝐴2𝑛+1 = {−1 + 2𝑛+1 , 1 + 2𝑛+2 , −1 + 2𝑛+3 , … } , sup 𝐴2𝑛+1 = 1 +

1

, 𝑎𝑛𝑑 inf 𝐴2𝑛+1 = −1

2𝑛+1

1

Now, sup 𝐴𝑛 = 1 + 𝑛 and inf 𝐴𝑛 = −1 ∀ 𝑛 ∈ ℕ

1

Hence, lim sup 𝑥𝑛 = 𝑙𝑖𝑚𝑛→∞ sup 𝐴𝑛 = 𝑙𝑖𝑚𝑛→∞ (1 + 𝑛) = 1 and

lim inf 𝑥𝑛 = 𝑙𝑖𝑚𝑛→∞ inf 𝐴𝑛 = 𝑙𝑖𝑚𝑛→∞ (−1) = −1

192

Definition:

Let 𝑥𝑛 be a sequence of real numbers.

1. We say that (𝑥𝑛 ) tends to −∞, and write 𝑙𝑖𝑚 𝑥𝑛 = +∞, if for

every 𝛼 ∈ ℝ there exists a natural number 𝐾(𝛼)such that if 𝑛 ≥

𝐾(𝛼), then 𝑥𝑛 > 𝛼.

2. We say that (𝑥𝑛 ) tends to +∞, and write 𝑙𝑖𝑚 𝑥𝑛 = −∞, if for

every 𝛽 ∈ ℝ there exists a natural number 𝐾(𝛽)such that if 𝑛 ≥

𝐾(𝛽), then 𝑥𝑛 < 𝛽.

We say that (𝑥𝑛 ) is properly divergent in case we have either

𝑙𝑖𝑚 𝑥𝑛 = +∞ or 𝑙𝑖𝑚 𝑥𝑛 = −∞.

Examples:

1. lim(𝑛) = +∞

2. lim(𝑛2 ) = +∞

3. If 𝑐 > 1 then lim(𝑐 𝑛 ) = +∞

Theorem:

A monotone sequence of real numbers is properly divergent if and only

if it is unbounded.

1. If (𝑥𝑛 ) is unbounded increasing sequence, then 𝑙𝑖𝑚 (𝑥𝑛 ) = +∞

2. If (𝑥𝑛 ) is unbounded decreasing sequence, then 𝑙𝑖𝑚 (𝑥𝑛 ) = −∞

Theorem:

Let (𝑥𝑛 ) and (𝑦𝑛 ) be two sequences of real numbers and suppose that

𝑥𝑛 ≤ 𝑦𝑛 for all 𝑛 ∈ ℕ.

a. If 𝑙𝑖𝑚(𝑥𝑛 ) = +1, then 𝑙𝑖𝑚(𝑦𝑛 ) = +1.

b. 2 If 𝑙𝑖𝑚(𝑦𝑛 ) = −1, then 𝑙𝑖𝑚(𝑥𝑛 ) = −1

Theorem:

Let (𝑥𝑛 ) and (𝑦𝑛 ) and be two sequences of positive real numbers and

𝑙𝑖𝑚 (𝑥𝑛)

suppose that for some 𝐿 ∈ ℝ, 𝐿 > 0, we have =𝐿

𝑛 → ∞ (𝑦𝑛)

Then 𝑙𝑖𝑚(𝑥𝑛 ) = +1 if and only if 𝑙𝑖𝑚(𝑦𝑛 ).

Definition:

A sequence of real numbers (𝑎𝑛 ) is said to be properly divergent to ∞ if

𝑙𝑖𝑚

𝑎 = ∞ that is ∀𝑀 ∈ ℝ there exists an 𝑁 ∈ ℕ such that if 𝑛 ≥ ℕ

𝑛→∞ 𝑛

then 𝑎𝑛 > 𝑀. Similarly, (𝑎𝑛 ) is said to be properly divergent to −∞ if

𝑙𝑖𝑚

𝑎 = −∞ that is ∀𝑀 ∈ ℝ there exists an 𝑁 ∈ ℕ such that if 𝑛 ≥ ℕ

𝑛→∞ 𝑛

then 𝑎𝑛 < 𝑀.

Theorem 1:

An increasing sequence of real numbers (𝑎𝑛 ) is properly divergent to ∞

if it is unbounded. A decreasing sequence of real numbers (𝑎𝑛 ) is

properly divergent to −∞ if it is unbounded.

Proof:

193

Suppose that (𝑎𝑛 ) is a sequence of real number that is increasing.

Since, (𝑎𝑛 ) is unbounded, then for any 𝑀 ∈ ℝ there exists a term 𝑎𝑀

(dependent on 𝑀) such that 𝑀 < 𝑎𝑀 . Since (𝑎𝑛 ) is an increasing

sequence, then for 𝑛 ≥ 𝑀 we have that 𝑀 < 𝑎𝑛 and since 𝑀 is arbitrary

we have that lim 𝑎𝑛 = ∞.

𝑛→∞

Similarly suppose that (𝑎𝑛 ) is a sequence of real number that is

decreasing. Since, (𝑎𝑛 ) is unbounded, then for any 𝑀 ∈ ℝ there exists

a term 𝑎𝑀 (dependent on 𝑀) such that 𝑀 > 𝑎𝑀 . Since (𝑎𝑛 ) is an

decreasing sequence, then for 𝑛 ≥ 𝑀 we have that 𝑀 < 𝑎𝑛 and since

𝑀 is arbitrary we have that lim 𝑎𝑛 = −∞.

𝑛→∞

Theorem 2:

Let (𝑎𝑛 ) and (𝑏𝑛 ) be sequences of real numbers such that 𝑎𝑛 ≤ 𝑏𝑛 for

all 𝑛 ∈ ℕ. Then if lim 𝑎𝑛 = ∞ then lim 𝑏𝑛 = ∞.

𝑛→∞ 𝑛→∞

Proof :

Let (𝑎𝑛 ) and (𝑏𝑛 )be sequences of real numbers such that 𝑎𝑛 ≤ 𝑏𝑛 for

all 𝑛 ∈ ℕ, and let lim 𝑎𝑛 = ∞. Then it follows that for all 𝑀 ∈ ℝ that

𝑛→∞

there exists an 𝑁(dependent on 𝑀 such that if 𝑛 ≥ ℕ then 𝑎𝑛 ≥ 𝑀). But

we have that 𝑏𝑛 ≥ 𝑎𝑛 for all 𝑛 ∈ ℕ and so for 𝑛 ≥ ℕ we have that

𝑏𝑛 ≥ 𝑀. Since 𝑀 is arbitrary it follows that lim 𝑏𝑛 = ∞.

𝑛→∞

Theorem 3:

Let (𝑎𝑛 ) and (𝑏𝑛 ) be sequences of real numbers such that 𝑎𝑛 ≤ 𝑏𝑛 for

all 𝑛 ∈ ℕ. Then if lim 𝑏𝑛 = −∞ then lim 𝑎𝑛 = ∞.

𝑛→∞ 𝑛→∞

Proof:

Let (𝑎𝑛 ) and (𝑏𝑛 )be sequences of real numbers such that 𝑎𝑛 ≤ 𝑏𝑛 for

all 𝑛 ∈ ℕ, and let lim 𝑏𝑛 = −∞. Then it follows that for all 𝑀 ∈ ℝ tht

𝑛→∞

there exists an 𝑁(dependent on 𝑀 such that if 𝑛 ≥ ℕ then 𝑏𝑛 ≤ 𝑀). But

we have that 𝑎𝑛 ≤ 𝑏𝑛 for all 𝑛 ∈ ℕ and so for 𝑛 ≥ ℕ we have that

𝑎𝑛 ≤ 𝑀. Since 𝑀 is arbitrary it follows that lim 𝑎𝑛 = ∞.

𝑛→∞

Theorem 4:

Let (𝑎𝑛 ) and (𝑏𝑛 ) be sequences of positive real numbers that for some

𝑎

real numbers 𝐿 > 0 that lim 𝑏𝑛 = 𝐿. Then lim 𝑎𝑛 = ∞ if and only if

𝑛→∞ 𝑛 𝑛→∞

lim 𝑏𝑛 = ∞.

𝑛→∞

Proof:

𝑎𝑛

Suppose (𝑎𝑛 ) and (𝑏𝑛 )are convergent sequences and that lim =𝐿

𝑛→∞ 𝑏𝑛

𝐿

for 𝐿 ∈ ℝ and 𝐿 > 0. Then for 𝜖 = 2 > 0 we have that for some 𝑛 ∈ ℕ if

𝑎 𝐿

𝑛 ≥ ℕ then |𝑏𝑛 − 𝐿| < 2 or equivalently

𝑛

𝐿 𝑎𝑛 3𝐿

< <

2 𝑏𝑛 2

𝐿 3𝐿

𝑏 < 𝑎𝑛 < 𝑏𝑛 .

2 𝑛 2

194

3𝐿

If lim 𝑎𝑛 = ∞ then since 𝑎𝑛 < 𝑏𝑛 it follows that lim 𝑏𝑛 = ∞. Similarly

𝑛→∞ 2 𝑛→∞

𝐿

if lim 𝑏𝑛 = ∞ then since 2 𝑏𝑛 < 𝑎𝑛 it follows that lim 𝑎𝑛 = ∞.

𝑛→∞ 𝑛→∞

Theorem 5:

If (𝑎𝑛 ) is a properly divergent subsequence then there exists no

convergent subsequences 𝑎𝑛𝑘 of 𝑎𝑛 .

Proof:

We will first deal with the case where (𝑎𝑛 )is properly divergent to ∞.

Suppose instead that there exists a subsequences 𝑎𝑛𝑘 that converges

to 𝐿. Then ∀𝜖 > 0 ∃𝐾 ∈ ℕ such that if 𝑘 ≥ 𝐾 then 𝑎𝑛𝑘 − 𝐿 < 𝜖, and so

for 𝑘 ≥ 𝐾 then 𝐿 − 𝜀 < 𝑎𝑛𝑘 < 𝐿 + 𝜖 and so 𝑎𝑛𝑘 < 𝐿.

Now if (𝑎𝑛 ) diverges to ∞ then for 𝐿 + 𝜖 ∈ ℝ ∃𝑁 ∈ ℕ such that if 𝑛 ≥ 𝑁

then 𝑎𝑛 > 𝐿 + 𝜖. So for 𝑛𝑘 ≥ 𝑚𝑎𝑥{𝐾, 𝑁},we have that

𝐿 + 𝜀 < 𝑎𝑛𝑘 < 𝐿 + 𝜖 which is a contradiction. So our assumption that

𝑎𝑛𝑘 converges was false, and so there exists no convergent

subsequences 𝑎𝑛𝑘 .

Definition 2:

Let (𝑠𝑛 ) be a sequence of real numbers. We say that (𝑠𝑛 ) diverges to

∞ and write lim 𝑠𝑛 = ∞ or 𝑠𝑛 → ∞ as 𝑛 → ∞ provided that for every

𝑛→∞

number 𝑀 there is an integer 𝑁 so that 𝑠𝑛 ≥ 𝑀 whenever 𝑛 ≥ 𝑁.

Example:

𝑛2 +1

Let us prove that → ∞ using the definition.

𝑛+1

If 𝑀 is any positive number we need to find some point in the sequence

after which all terms exceed 𝑀. Thus, we need to consider the

inequality

𝑛2 + 1

≥𝑀

𝑛+1

After some arithmetic we see that this is equivalent to

1 𝑛

𝑛+ − ≥𝑀

𝑛+1 𝑛+1

𝑛

Since 𝑛+1 < 1 we see that, as long as 𝑛 ≥ 𝑀 + 1 this will be true.

Thus, take any integer 𝑁 ≥ 𝑀 + 1 and it will be true that

𝑛2 + 1

≥𝑀

𝑛+1

for all 𝑛 ≥ 𝑁. (Any larger value of 𝑁 would work too.)

Definition:

Let (𝑥𝑛 ) be a sequence.

∀𝛼 ∈ ℝ ∃𝐾(𝛼) ∈ ℕ ∋ ∀𝑛 ≥ 𝐾(𝛼), 𝑥𝑛 > 𝛼

b. We say that (𝑥𝑛 ) tends to −∞ and write lim 𝑥𝑛 = −∞ if

∀𝛽 ∈ ℝ ∃𝐾(𝛽) ∈ ℕ ∋ ∀𝑛 ≥ 𝐾(𝛽), 𝑥𝑛 < 𝛽

195

We say (𝑥𝑛 ) is properly divergent in either case.

Example:

For 𝐶 > 1, 𝑙𝑖𝑚𝑐 𝑛 = +∞

Proof:

Let 𝛼 ∈ ℝ be given. [How to express 𝐶 > 1.]

𝑐 = 1 + 𝑏 where 𝑏 > 0. By the Archimedian property,

𝛼

∃𝐾(𝛼) ∈ ℕ ∋ 𝐾(𝛼) > 𝑏 . Then ∀𝑛 ≥ 𝐾(𝛼),

𝑐 𝑛 = 1 + 𝑏 𝑛 ≥ 1 + 𝑛𝑏 > 1 + 𝛼 > 𝛼

Thus, 𝑙𝑖𝑚𝑐 𝑛 = +∞

Definition:

Let (𝑥𝑛 ) be a sequence in real numbers

i. We say that 𝑥𝑛 tends to +∞ and 𝑙𝑖𝑚𝑥𝑛 = +∞ if ∀𝛼 ∈ ℝ, ∃𝐾(𝛼) ∈

ℕ such that ∀𝑛 ≥ 𝐾, we have that 𝑥𝑛 > 𝛼.

ii. We say that 𝑥𝑛 tends to −∞ and 𝑙𝑖𝑚𝑥𝑛 = −∞ if ∀𝛽 ∈ ℝ, ∃𝐾(𝛽) ∈

ℕ such that ∀𝑛 ≥ 𝐾, we have that 𝑥𝑛 < 𝛽.

iii. We say that 𝑥𝑛 is properly divergent if either 𝑙𝑖𝑚𝑥𝑛 = +∞ or

𝑙𝑖𝑚𝑥𝑛 = −∞

Theorem:

A monotone sequence in ℝ is properly divergent if and only if it is

unbounded, and

a. If (𝑥𝑛 ) is unbounded and increasing sequence then 𝑙𝑖𝑚𝑥𝑛 = +∞

b. If (𝑥𝑛 ) is unbounded and decreasing sequence then 𝑙𝑖𝑚𝑥𝑛 = −∞

Proof

Suppose (𝑥𝑛 ) is monotone increasing sequence.

⟹ Suppose ( 𝑥𝑛 ) is bounded, then it must be convergent by MCT,

which is a contradiction.

Theorem:

Let 𝑥𝑛 ≤ 𝑦𝑛 ∀𝑛 ∈ ℕ.

a.If 𝑙𝑖𝑚 𝑥𝑛 = +∞ then 𝑙𝑖𝑚 𝑦𝑛 = +∞

b. If 𝑙𝑖𝑚 𝑥𝑛 = −∞ then 𝑙𝑖𝑚 𝑦𝑛 = −∞

Proof

(b) let 𝛽 ∈ ℝ, ∃𝐾(𝛽) ∈ ℕ such that ∀𝑛 ≥ 𝐾, then 𝑦𝑛 < 𝛽 ⟹ 𝑥𝑛 ≤ 𝑦𝑛 <

𝛽.

⟹ 𝑙𝑖𝑚 𝑥𝑛 = −∞

Theorem:

Let (𝑥𝑛 ) , (𝑦𝑛 ) be two sequences of positive real numbers and suppose

that

𝑥𝑛

𝑙𝑖𝑚 ( ) = 𝐿 > 0

𝑦𝑛

Then 𝑙𝑖𝑚𝑥𝑛 = +∞ if and only if 𝑙𝑖𝑚𝑦𝑛 = +∞

Proof:

196

𝑥

Since 𝑙𝑖𝑚 (𝑦𝑛 ) = 𝐿

𝑛

𝐿 𝐿

Let 𝜖 = 2 > 0, then ∃𝐾 , (2) ∈ ℕ such that ∀𝑁 ≥ 𝐾,

𝑥 𝐿

|𝑦𝑛 − 𝐿| < 2⟹

𝑛

𝐿 𝑥 𝐿

− 2 < 𝑦𝑛 − 𝐿 < 2⟹

𝑛

𝐿 𝑥𝑛 3𝐿

( )< <( )

2 𝑦𝑛 2

𝐿 3𝐿

⟹(2) 𝑦𝑛 < 𝑥𝑛 < ( 2 ) 𝑦𝑛

3𝐿

Now, if 𝑙𝑖𝑚𝑥𝑛 = +∞ ⟹ 𝑙𝑖𝑚 ( 2 ) 𝑦𝑛 = +∞

⟹𝑙𝑖𝑚𝑦𝑛 = +∞

𝐿

and if 𝑙𝑖𝑚𝑦𝑛 = +∞ ⟹ 𝑙𝑖𝑚 (2) 𝑦𝑛 = +∞

⟹ 𝑙𝑖𝑚𝑥𝑛 = +∞

197

Exercises

1. Prove that

a. lim 𝑛 = ∞

b. lim(−𝑛2 ) = −∞

c. If 𝑐 > 1, then 𝑙𝑖𝑚𝑐 𝑛 = +∞

2. Determine if the following series is convergent or divergent. If it

converges determine its value.

d. ∑∞𝒏=𝟏 𝒏

3. Determine if the following series converges or diverges. If it

converges determine its sum.

e. ∑∞𝒏=𝟎(−𝟏)

𝒏

4. Let 𝑏𝑟 = √𝑟 + 1 − √𝑟

198

Answers Key

1.

a. lim 𝑛 = ∞

Let 𝛼 ∈ ℝ be given. Then , be given. Then , ∃𝐾(𝛼) ∈ ℕ such

that 𝐾 > 𝛼. So ∀𝑛 ≥ 𝐾, we have 𝑥𝑛 = 𝑛 ≥ 𝐾 > 𝛼

b. lim(−𝑛2 ) = −∞

Let 𝛽 ∈ ℝ be given. Then , ∃𝐾(𝛽) > −𝛽 and so ∀𝑛 ≥ 𝐾, we

have

𝑛2 ≥ 𝑛 ≥ 𝐾 > −𝛽 ⟹ 𝑥𝑛 = −𝑛2 < 𝛽

c. If 𝑐 > 1, then 𝑙𝑖𝑚𝑐 𝑛 = +∞

Let 𝑐 = 1 + 𝑏, 𝑏 > 0 ⟹ 𝑐 𝑛 = (1 + 𝑏)𝑛 ≥ 1 + 𝑛𝑏 > 𝑛𝑏

(Bernoulli’s Inequaliity)

𝛼 𝛼

So, if 𝛼 ∈ ℝ is given . then ∃𝐾 (𝑏 ) such that 𝐾 > 𝑏 . So if 𝑛 ≥

𝐾, then we have

𝛼

𝑥𝑛 = 𝑐 𝑛 ≥ 1 + 𝑛𝑏 > 𝑛𝑏 > 𝐾𝑏 > 𝑏 · 𝑏 = 𝛼.

2. To determine if the series is convergent we first need to get our hands

on a formula for the general term in the sequence of partial sums

𝑛

𝑠𝑛 = ∑ 𝑖

𝑖=1

This is a known series and its value can be shown to be

𝑛

𝑛(𝑛 + 1)

𝑠𝑛 = ∑ 𝑖 =

2

𝑖=1

So, to determine if the series is convergent we will first need to see if

the sequence of partial sums,

𝑛(𝑛 + 1) ∞

{ }

2 𝑛=1

is convergent or divergent. That’s not terribly difficult in this case. The

limit of the sequence terms is,

𝑙𝑖𝑚 𝑛(𝑛 + 1)

=∞

𝑛→∞ 2

Therefore, the sequence of partial sums diverges to ∞ and so the

series also diverges.

3. In this case we really don’t need a general formula for the partial sums

to determine the convergence of this series. Let’s just write down the first

few partial sums.

𝑠0 =1

𝑠1 =1−1=0

𝑠2 =1−1+1=1

𝑠3 =1−1+1−1=0

⋮

So, it looks like the sequence of partial sums is,

199

{𝑠𝑛 }∞

𝑛=0 = {1,0,1,0, . . . }

𝑙𝑖𝑚

and this sequence diverges since 𝑠 doesn’t exist. Therefore,

𝑛→∞ 𝑛

the series also diverges.

4. 𝑏𝑟 = √𝑟 + 1 − √𝑟

√𝑟 + 1 + √𝑟

𝑏𝑟 = (√𝑟 + 1 − √𝑟) ( )

√𝑟 + 1 + √𝑟

𝑟+1−𝑟

𝑏𝑟 =

√𝑟 + 1 + √𝑟

1 1

𝑏𝑟 = ≥

√𝑟 + 1 + √𝑟 2√𝑟 + 1

∞ ∞

1 1

∑ 𝑏𝑟 ≥ ∑

2 √𝑟 + 1

𝑛=2 𝑟=1

∞ ∞

1 1

∑ 𝑏𝑟 ≥ ∑ , {𝑛 = 𝑟 + 1}

2 𝑛

𝑛=2 𝑟=1

1

𝑤ℎ𝑒𝑟𝑒 ∑∞

𝑟=2 is divergent

√𝑛

∑ 𝑏𝑟 is divergent

𝑛=2

𝑛=1

200

INTRODUCTION TO INFINITE SERIES

JOVIE AGUE

201

Introduction to Infinite Series

Definition:

A series or infinite series is the sum of all the terms in a sequence.

Example:

(Examples of infinite series).

1. ∑∞

𝑛=1 𝑛 = 1 + 2 + 3 + ⋯

1 1 1

2. A geometric series ∑∞

𝑛=1 2𝑛=1+2 + 2 + ...

1

by multiplying by the common ratio . This is what geometric means.

2

1 1 1

3. The harmonic series ∑∞

𝑛=1 𝑛=1+ 2 + 3 + ...

4. An alternating series ∑∞

𝑛=0(−1)𝑛=1+(−1)+1+(−1)+ ...

Notes

1. For now these infinite sums are just formal expressions or arrangements

of symbols. Whether it is meaningful to think of them as numbers or not is

something that can be investigated.

confuse these terms. A sequence is just a list of numbers. A series is an

infinite sum.

3. The “sigma” notation for sums : sigma (lower case σ, upper case Σ) is a

letter from the Greek alphabet, the upper case Σ is used to denote sums.

The notation ∑𝑖𝑛=𝑖 𝑎𝑛 means: i and j are integers and i ≤ j. For each n from

i to j the number an is defined; the expression above means the sum of the

numbers an where n runs through all the values from i to j, i.e.

∑ 𝑎𝑛 = 𝑎𝑖 + 𝑎𝑖 + 1 + 𝑎𝑖2 + ⋯ + 𝑎𝑗 − 1 + 𝑎𝑗.

𝑛=1

For example

∑5𝑛=2 𝑛2 = 22 + 32 + 42 + 52 = 54.

For infinite sums it is possible to have −∞ and/or ∞ (instead of fixed integers i

and j) as subscripts and superscripts for the summation.

What does it mean to talk about the sum of infinitely many numbers? We

cannot add infinitely many numbers together in practice, although we can (in

principle) at least, add up any finite collection of numbers. In the examples

above we can start from the beginning, adding terms at the start of the series.

Adding term by term we get the following lists.

202

1. ∑∞

𝑛=1 𝑛 = 1 + 2 + 3 + ...

1, 1 + 2, 1 + 2 + 3, 1 + 2 + 3 + 4, 1 + 2 + 3 + 4 + 5, · · · : 1, 3, 6, 10,

15, . . .

Since the terms being added on at each stage are getting bigger, the

numbers in the list above will keep growing (faster and faster as n

increases) - we can’t associate a numberical value with this infinite

sum.

2. A geometric series

1 1 1

∑∞

𝑛=1 = 1 + 2 + 22 + ⋯

2𝑛

1 1 1 1 1 1

1,1+2 , 1 + 2 + 22 , 1 + 2 + 22 + 23 …

3 7 15 31 63

1,1+2 , 4 , , 16 , 32 …

8

1

In this example the terms that are being added on at each step ( 2𝑛 )

are getting smaller and smaller as n increases, and the numbers in the

list appear to be converging to 2.

1 1 1

∑∞

𝑛=1 𝑛 = 1 + 2 + 3 + ⋯

1 1 1 1 1 1 3 11 25 137

1, 1 + 2 , 1 +2 + 3 , 1 + 2 +3 + 4 . . . : 1, 2, , 12, ,...

6 60

4. An alternating series

∑∞

𝑛=0(−1) n = 1 + (−1) + 1 + (−1) + . . .

1, 1 − 1, 1 − 1 + 1, 1 − 1 + 1 − 1, 1 − 1 + 1 − 1 + 1 . . . : 1, 0, 1, 0, 1, .

.

The terms being “added on” at each step are alternating between 1 and

−1, and as we proceed with the summation the “running total”

alternates between 0 and 1. So there is no numerical value that we can

associate with the infinite sum ∑∞

𝑛=0(−1)n.

Note

203

he series in 2. above converges to 2, the series in 1. and 4. are both

divergent and it is not obvious yet but the series in 3. is divergent as well. Our

next task is to give precise meanings to these terms for series. In order to do

this we need some terminology. Bear in mind that we know what it means for

a sequence to converge, but we don’t yet have a definition of convergence for

series.

Definition:

For a series ∑∞

𝑛=1 an, and for k ≥ 1, let

sk =∑𝑘𝑛=1 an = a1 + a2 + a3 + · · · + ak.

Then sk is called the kth partial sum of the series, and the sequence

(sk)∞ k=1 is called the sequence of partial sums of the series.

to converge and s is called its sum. In this situation we can write

∑∞

𝑛=1 𝑎𝑛 = 𝑠.

Example:

(Convergence of a geometric series). Recall the second example

above :

1 1 1

∑∞

𝑛=0 = 1 + 4 + 22 + ⋯

2𝑛

1 1 1 1

Sk=∑𝑘𝑛=0 2 = 1 + 2 + 4 + ⋯ 2𝑘

1 1 1 1 1 1

𝑠𝑘 = ∑𝑘𝑛=1 2𝑛+1 = + 4 + ⋯ 2𝑘 + 2𝑘+1

2 2

Then

1 1 1 1

Sk-2 𝑠𝑘 = 2 𝑠𝑘 = 1 − 2𝑘+1 ⇒ 𝑠𝑘 = 2 − 2𝑘.

1

So the sequence of partial sums has kth term 2 −2. This sequence

converges to 2 so the series converges to 2; we can write

1

∑𝑛 = 0∞ = 2.

2𝑘

204

General geometric series :

Consider the sequence of partial sums for the geometric

series

∑∞

𝑛=0 𝑎𝑟𝑛 = 𝑎 + 𝑎𝑟 + 𝑎𝑟2 + ⋯

ratio r.) The kth partial sum sk is given

By

𝑎(1−𝑟𝑘+1)

Then (1 − r)sk = a − ark+1 ⇒ sk = 1−𝑟 . If |r| < 1, then

rk+1 → 0 as k → ∞, and the sequence of partial sums

𝑎

(hence the series) converges to 1−𝑟. If |r| ≥ 1 the series is

divergent.

Next we show that the harmonic series is divergent.

Theorem:

1

The harmonic series ∑∞

𝑛=1 𝑛 is divergent. We give two proofs.

Proof 1:

1 1

Think of 𝑛 as the area of a rectangle of height 𝑛 and width 1,

sitting on the interval [n, n + 1] on the x-axis. So the first

1

term 1 corresponds to

a square of area 1 sitting on the interval[1, 2], the term 12

1

corresponds to a rectangle of area 2 sitting on the interval [2, 3] and so

on, as in the following picture.The total area accounted for by these

triangles is the sum of the harmonic series, and this exceeds the area

accounted for by the improper integral

∞1

∫1 𝑋

dx.

From Section 1.5 we know that this area is infinite, hence the series is

divergent.

Proof 2:

205

We show that the sequence of partial sums of the harmonic

series is not bounded above.

1

• The second term is 2

1

• The sum of the 3rd and 4th terms exceeds 2.

1 1 1 1 1 1

+ 4 > 4+4 + 4 = 2.

3

1

• The sum of the 5th, 6th, 7th and 8th terms exceeds2:

1 1 1 1 1 1 1 1 1 1

+ 6 + 7 + 8 > 8 + 8 + 8 + 8 + 8 = 2.Type equation here.

5

the same reason, the sum of the next 8 terms (terms 9 through 16) also

1

exceeds 2

1 1 1

• In general the sum of the 2𝑛−1 terms 𝑛−1 +1 𝑡ℎ𝑟𝑜𝑢𝑔ℎ exceeds2.

2 2𝑛

12n exceeds 12

.

So, as we list terms in the sequence of partial sums of the harmonic

series, we keep accumulating non-overlapping stretches of terms that

1

add up to more than 2. Thus the entire series has infinitely many non-

1

overlapping stretches all individually summing to more than 2. Then the

sum of this series is not finite and the series diverges.

Note:

A necessary condition for the series

∑∞𝑛−1 𝑎𝑛 an to converge is that the sequence {𝑎𝑛 }∞𝑛=1 converges to 0;

i.e. that 𝑎𝑛 → 0 as n → ∞. If this does not happen, then the sequence

of partial sums

has no possibility of converging.

The example of the harmonic series shows that the condition 𝑎𝑛 → 0 as

n → ∞ is not sufficient to guarantee that the series ∑∞

𝑛=1 will

converge.

series to converge;

• Give examples of convergent and divergent series;

• show that the harmonic series is divergent;

• Use the “sigma” notation for sums.

206

Many of our in_nite sequences, for the remainder of the course, will be

de_ned by sums.

For example, the sequence

Sm :=

Xm

n=112n : (1) is defined by a sum. Its terms (partial sums) are 1 2; 1 2

+14=34;12+14+18=78;12+14+18+116=1516;:::

These in_nite sequences de_ned by sums are called in_nite series.

Review of sigma notation

The Greek letter _ used in this notation indicates that we are adding

(\summing") elements of a

certain pattern. (We used this notation back in Calculus 1, when we

_rst looked at integrals.) Here our

sums may be \in_nite"; when this occurs, we are really looking at a

limit.

Resources

An introduction to sequences a standard part of single variable

calculus. It is covered in every calculus

textbook. For example, one might look at

* section 11.3 (Integral test), 11.4, (Comparison tests) , 11.5 (Ratio &

Root tests), 11.6 (Alternating,

abs. conv & cond. conv) in Calculus, Early Transcendentals (11th ed.,

2006) by Thomas, Weir,

Hass, Giordano (Pearson)

* section 11.3 (Integral test), 11.4, (comparison tests), 11.5 (alternating

series), 11.6, (Absolute conv,

ratio and root), 11.7 (summary) in Calculus, Early Transcendentals (6th

ed., 2008) by Stewart

(Cengage)

* sections 8.3 (Integral), 8.4 (Comparison), 8.5 (alternating), 8.6,

Absolute conv, ratio and root, in

Calculus, Early Transcendentals (1st ed., 2011) by Tan (Cengage)

Integral tests, comparison tests,

ratio & root tests.

* section 9.4 (Convergence Tests), 9.5 (Comparison, ratio, root tests),

9.6 (Alternating, abs. conv &

cond. conv) Calculus, Early Transcendentals (11th ed., 2009) by

Anton, Bivens, Davis (John Wiley

& Sons) p. 645 of Anton has a nice list.

* section 10.3 (integral test), 10.4 (alt series), 10.5 (comparison), 10.6

(absolute convergence), 10.7 (ratio and root) in the Whitman College

online textbook: http://www.whitman.edu/mathematics/multivariable/*

Whitman's online textbook:

http://www.whitman.edu/mathematics/multivariable/calculus_

10_Sequences_and_Series.pdf

1 3.2.1 What is a series?

Given an infinite series 1X n=1 an we de_ne the partial sum Sm :=Xm

n=1 an: Thus, in the series 1X n=1 12n we have the partial sums

S1 =1 2

S2 =3 4

207

S3 =7 8

S4 =15 16

We mean, by the expression, 1X n=1, the limit, as n ! 1, of the partial

sums Sn:

In this case, the partial sums appear to have the pattern Sn = 1 12n :

So 1X n=1 1 2n really means limm!1 Xmn=112n = limm!11 12m =

1:Since, in this case, the limit is 1, we say that1Xn=112n = 1:3.2.2 An

easy divergence test

Intuitively, if a series is to converge to a _nite limit L we would expect

that eventually the terms we are

adding up are contributing very little to the series { that at some point

the sum is close to L and each

new term is not changing that. This argument can be made precise (but

we won't do that here.)

This gives us a theorem, the \n-th term divergence test":

Theorem.

(n-th term test) If a series X

an converges then the limit, as n goes to in_nity, of the terms an, must

be zero.

Although stated in terms of convergence, the theorem is really a

statement about divergence, for it is equivalent to the statement:

The n-th term divergence test: If in a series X the terms an do not go to

zero then the series does not converge! (This statement is the

contrapositive of the statement in the theorem.)

The complexity of our investigation into series (initiated long ago by the

Bernoulli brothers and Euler) can be displayed by examining the family

of geometric series { central to our understanding of series {and two

interesting \sporadic" series, the harmonic series and the alternating

harmonic series.

2

In an earlier section, we examined the series 1Xn=11 2n = 2+14+18+1

16+ : : : 12n + : : : :

We concluded, just by observation, that the partial sums had the form

1 1

2n and thus the series converged to 1. Note that this series has this

property: each term added on is exactly 12 of the previous term.A

series which has a \ratio" r such that each new term is exactly r times

the previous term, is said

to be geometric. (See http://en.wikipedia.org/wiki/Geometric_series for

a general discussion of these series, including modern applications.)

The main idea.

There is a nice way to work out a formula for the partial sum of a

geometric series. In general, a geometric series has form a + ar + ar2 +

ar3 + : : : + arn1 + : : : = 1X n=0 arnwhere a is the _rst term and r is

the common ratio between terms.

Let us write

Sm := a + ar + ar2 + ar3 + : : : + arm1 = mX1 n=0 arn:

208

Notice that rSm := ar + ar2 + ar3 + : : : + arm1 + arm = Xm n=1 arn

and so rSm Sm := arm a: (Notice how most terms cancel!)

Therefore Sm = a(rm 1) r 1 = a(1 rm) 1 r :

If jrj > 1 then the expression a(1 rm) 1 r does not converge and so

the geometric series does not converge.

If r = 1 then the expression a(1rm) 1r is unde_ned but it is easy to

check that the partial sums are

Sm = ma and so the series diverges to in_nity.

If r = 1 then the a(1rm)

1r does not converge and so the geometric series does not converge.

The partial sums alternate between a and 0:

But if jrj < 1 then the a(1rm) 1r converges to a 1r : Therefore the

geometric series converges to a 1r :

This is important enough to emphasize as a theorem.

Theorem.

(Geometric series)

But if jrj _ 1 then 1X n=0 arn diverges.3

The geometric series are central to the study of in_nite series. We will

see later in this course that if a

series is not geometric, we will attempt (in a certain way) to \pretend" it

is geometric anyway! Sometimes

this \pretense" gives us very useful information. (This will motivate the

\ratio" and \root" tests.)

3.2.4 The Harmonic Series

The n-th term divergence test says that the terms of a series must go

to zero if there is any hope of the

series converging.

Warning! It is tempting to believe in the converse statement. Is it true

that if the terms go to zero

then the series converges? That would be nice { but it is not true in

general.

Here is an example { a classical one { where the terms go to zero but

the series diverge. The series is called the \harmonic series". The

harmonic series 1Xn1n= 1 +12+13+14+ : : : +1n+ : : : :

(See http://en.wikipedia.org/wiki/Harmonic_series_(mathematics) for

Wikipedia's discussion of

this series, including an explanation for its name.)

The terms 1n go to zero as n goes to in_nity. Yet this series diverges!

We give one argument that this series diverges. This argument is a

\comparison test". Another

argument will be given later.

Compare the series 1Xn1n= 1 +12+13+14+15+16+17+18+19+110+

::n+ : : : and the series 1 + 1 2 + (14+14) + (18+

18+18+18 + (116+116+ : : :116) + : : :In the second series, any

expression of the form 1n , 2s < n _ 2s+1, we replace n by 2s+1: Since

12s+1_1n , we have a series which is \smaller" than the harmonic

series. Each partial sum of the second

series no bigger than the partial sum of the harmonic series.

209

But notice that, since in the second series there are 2s terms of the

form 12s+1 then we can collect them together to form the term12:So

the second series becomes1+12+(14+14)+(18+18+18+18)+(116+116+

: : :116)| {z }8 terms+: : : = 1+12+(12)+(12)+(12)+: : :+(12)+: : :

We continue, forever, to add 12 in the sum. So this second series

diverges (by the \n-th term test”, if you will.) But since the harmonic

series is forever larger than this diverging series, then the harmonic

series diverges! The harmonic series forms a nice counterpoint to the

n-th term test for divergence. If the n-th term of a series goes to zero,

the series still might not converge. The harmonic series is a nice

example of thisphenomenon.43.2.5 The Alternating Harmonic Series

One more interesting series: Consider the alternating harmonic series

1Xn(1)n+1n= 1 12+14+15+ : : : +(1)n+1n+ : : : :

Note the elect of the expression (1)n+1; it forces the signs to

alternate, so that we are adding positive, then negative terms.

It turns out that an alternating series, where the n-th term does to zero,

does converge!

The alternating harmonic series converges { indeed, we will see later

that its sum is ln 2 _ 0:693147180559945.

3.2.6 Telescoping series

There is another type of geometric series we run into from time to time,

where a little trick will give not

only convergence, but the exact value of the limit of the series.

Consider, for example, the series1X11n (n + 1)=12+16+112+ : : : ::::

We can use the algebraic concept of partial fractions (remember that?!)

to rewrite 1n(n + 1)=1n1n + 1: So1X11n(n + 1)=1X 1 1n1n + 1=

(1112) + (1213) + (1314) + : : : ::::

Notice how the terms begin to cancel! We can rewrite the sum

as=1+(12+12) + (13+13) + (14+14) + : : : ::::In this case, the series

collapses (\telescopes") and only the _rst term survives. So1X11n(n +

1)= 1:5

210

EXERCISES

𝒆𝒏

1. 𝒂𝒏 = 𝟑𝒏

2. 𝒂𝒏 = (−𝟏)𝒏 √𝒏

𝟏

3. 𝒂𝒏 = (−𝟏)𝒏

√𝒏

4. 𝒂𝒏 = (−𝟏)𝟐𝒏+𝟏

(−𝟏)𝒏 +𝒏

5. 𝒂𝒏 = (−𝟏)𝒏 −𝒏

𝒏

𝒍𝒏((𝒆𝟒 ) )

6. 𝒂𝒏 = 𝟑𝒏

𝝅

7. 𝒂𝒏 = (−𝟏)𝒏 𝒄𝒐𝒔( 𝟐 (𝒏 + 𝟏))

𝝅

8. (−𝟏)𝒏 𝒄𝒐𝒔( 𝟐 (𝟐𝒏 − 𝟏))

211

ANSWERS KEY

1. Converges to 0

2. Diverges

3. Converges to 0

4. Converges to −1

5. Converges to −1

4

6. Converges to3

7. Diverges

8. Converges to 0

212

LIMITS OF A FUNCTION

BRIX LAPIAD

213

Limits found Graphically

approach particular values of x. Left-hand and right-hand limits are

the idea of looking at what happens to a function as you approach a

particular value of x, from a particular direction.

The limit of f(x) as x approaches the value of a from the left is written

lim f ( x)

x a

and the limit of f(x) as x approaches the value of a from the right is

written

lim f ( x)

x a

Let’s explore these ideas with the graph of f(x) in Figure 3.1 below.

Figure 3.1

Looking at f(x) when x = -2, you notice there is a “break” in the function.

However, if you approach x = -2 “from the left” (Figure 3.2a) you can see

that the function values are getting closer and closer to 1. On the other

hand, if we approach x = -2 “from the right” (Figure 3.2b) you can see

that the function values are getting closer and closer to 3.

214

Figure 3.2a Figure 3.2b

we approach f(x) from the left or from the right (Figure 3.3), you can see

that the function values are getting closer and closer to 2.

Figure 3.3

x 2 x 2 x 1

lim f ( x) 2

x 1

215

Example 1

Using the given graph of g(x), find the following left- and right-hand

limits.

a. lim g ( x)

x 0

b. lim g ( x)

x 0

c. lim g ( x)

x 1

d. lim g ( x)

x 1

Solution

a. This asks us to look at the graph of g(x) as x approaches 0 from the left. You

can see that the function values are getting closer and closer to -1. So,

lim g ( x) 1

x 0

b. This asks us to look at the graph of g(x) as x approaches 0 from the right.

You can see that the function values are getting closer and closer to -1. So,

lim g ( x) 1

x0

216

c. This asks us to look at the graph of g(x) as x approaches 1 from the left. You

can see that the function values are getting closer and closer to -2. So,

lim g ( x) 2

x 1

d. This asks us to look at the graph of g(x) as x approaches 1 from the right.

You can see that the function values are getting closer and closer to -2. So,

lim g ( x) 2

x 1

Notice that in the solutions to parts (c) and (d) above, the function value

g(1)=1 does not play a role in determining the values of the limits. A limit

is strictly the behavior of a function “near” a point.

217

Example 2

Using the graph of h(x) below, find the following left- and right-hand

limits.

a. lim h( x)

x 4

b. lim h( x)

x 4

Solution

a. Looking at the graph of h(x), as x approaches 4 from the left, you can see that

the function values keep getting more and more negative, without end. Thus,

we say that the function values approach negative infinity, written

lim h( x)

x 4

b. Looking at the graph of h(x), as x approaches 4 from the right, you can see

that the function values keep getting more and more positive without end.

Thus, we say that the function values approach positive infinity, written

lim h( x)

x 4

approach a particular value of x, you can determine whether or not the

limit of the function at that point exists.

x a x a x a

Therefore, if the left-hand limit does not equal the right-hand limit as x

approaches a, then the limit as x approaches a does not exist.

218

Example 3

Using the graph of f(x) below, find the following limits.

a. lim f ( x)

x 1

b. lim f ( x)

x 2

Solution

x 1 x 1

x 1 x 1 x 1

It is important to notice that this limit exists even though f(1) does not

exist.

x 2 x 2

x 2 x 2 x 2

Example 4

219

Using the given graph of g(x), find lim g ( x) .

x 1

Solution:

In previous investigations of this function, we found that

x1 x1

x 1 x 1 x 1

Example 5

Using the given graph of h(x), find lim h( x) .

x4

220

Solution

In previous investigations of this function, we found that

x 4 x 4

x 4 x 4 x4

particular values of x by looking at what is happening to the function

values corresponding to values very near to the x value. Let’s now

explore what happens to the function values when we allow x to

approach positive and negative infinity.

Figure 3.4

Figure 3.4a Figure 3.4b Figure 3.4c

In Figure 3.4a, you can see that if you move to the right on the graph

and allow x to continually become larger (approach infinity), the function

values also become larger and larger. If you move to the left on the

graph and allow x to become more and more negative (approach

negative infinity), you can see that the function values are again

becoming larger and larger. Thus, we have

x x

In Figure 3.4b, you can see that if you allow x to approach infinity, the

function values go towards negative infinity. If you allow x to approach

negative infinity, you can see that the function values go towards positive

infinity. Thus, we have

221

lim m( x) and lim m( x)

x x

In Figure 3.4c, you can see that if you allow x to approach either positive

or negative infinity, the function values approach zero. Thus, we have

x x

x , we say that the function has a horizontal asymptote at y = L.

Thus, we have just found that p(x) has a horizontal asymptote at y=0,

while k(x) and m(x) have no horizontal asymptotes.

(Note: You can only approach positive infinity from the left and you can

only approach negative infinity from the right so there is no discussion

of left- and right-hand limits at infinity.)

Example 6

Using the graph of f(x) below, find the following limits.

a. lim f ( x)

x 5

b. lim f ( x )

x

c. lim f ( x)

x 0

d. lim f ( x)

x

Solution

a. We need to find and compare the left- and right-hand limits of f(x) at x= -5. As

x approaches -5 from the left, f(x) approaches 1 and as x approaches -5 from

the right, f(x) also approaches 1. Therefore,

lim f ( x) 1

x 5

b. As x , the function values get more and more positive without end.

Therefore,

lim f ( x)

x

222

lim f ( x) 5

x 0

lim f ( x) 0

x

discussing, it is not always practical or necessary if the function is

defined algebraically.

x2 9

For instance, say we are given that f ( x) . If we are looking for

x 3

lim f ( x) , instead of having to graphically search for the answer, we can

x 3

find both the left- and right-hand limits by using tables. By choosing x

values that get closer and closer to x = 3 from both sides, we can analyze

the behavior of f(x).

Table 3.1

values that were very close to x = 3. It seems that as x approaches 3

from either side, the function values are approaching 6. Therefore, it

seems reasonable to conclude that

x2 9

lim 6

x 3 x 3

as shown below.

223

Example 7

4, x 1

Using tables, find the following limits given that f ( x) 2

x , x 1

a. lim f ( x)

x2

b. lim f ( x)

x 1

Solution

Since all of these x-values are in the domain of x >1, we will use the part of

the function defined by x 2 to determine the function values in our table.

Approaching x=2 from both the left and the right sides shows that the function

values are approaching 4. Thus, lim f ( x) 4 .

x2

b. We need to construct a table with x-values approaching 1 from both sides. All

x-values approaching x = 1 from the left are in the domain x < 1, so we will be

using the part of the function defined by 4 when finding these function values.

All x-values approaching x = 1 from the right are in the domain x > 1, so we

will use x 2 to find these function values in our table.

224

As x 1 from the left, f(x) seems to be approaching 4, while as

x 1 from the right, f(x) seems to be approaching 1. Since these are

not equal, by definition, lim f ( x) does not exist.

x 1

the following rules to algebraically evaluate limits more efficiently. Most

of these rules can intuitively be verified from looking at the previously

worked examples.

xa

x a

3) lim c f ( x) c lim f ( x)

x a x a

constant times the limit of the function.)

x a x a x a

(The limit of the sum or difference of two functions equals the sum

or difference of the limits of the functions.)

xa xa xa

(The limit of the product of two functions is the product

of the limits of the functions.)

f ( x) lim f ( x)

6) lim x a , if lim g ( x) 0

x a g ( x) lim g ( x) x a

x a

(The limit of a quotient is the quotient of the limits of the numerator and

denominator if the limit of the denominator is not zero.)

225

n

7) lim f ( x) lim f ( x) (provided this is defined)

n

x a xa

(The limit of a function raised to a power equals the limit of the function

raised to the power provided the math makes sense.)

Example 8

Evaluate

a. lim x2 2 x 4

x 1

x2 9

b. lim

x 3 x 3

1

c. lim

x 4 x4

Solution

x 1

x 2 9 lim x2 9

b. lim x 3

(Rule 6)

x 3 x 3 lim x 3

x 3

0

(Rule 2)

0

When you get 0/0 you have what is called an indeterminant form and you

must try other techniques to determine the limit. In this case, factor both the

numerator and denominator and cancel common terms to remove the zero in

the denominator. Then, apply the limit rules to the simplified expression.

x2 9 ( x 3)( x 3)

lim lim (Factor)

x 3 x 3 x 3 ( x 3)

x 3

=3+3=6 (Rule 2)

1 lim1

c. lim x 4 (Rule 6)

x 4 x 4 lim x 4

x 4

1

(Rules 1 and 2)

0

226

This is not defined and whenever you get a result of a non-zero number over

zero, there are no common factors in the numerator and denominator which

can be cancelled. Therefore, there is no way to rid the denominator of its zero

term, meaning that the limit does not exist. Looking at the graph of the

function near

Notice this is the same function we analyzed when finding limits graphically.

There, we also found that the limit does not exist.

While these rules also apply when looking for limits at infinity (or negative

infinity), it is almost always necessary to algebraically manipulate the

expression of the function to determine the limit.

Example 9

2x2 7

Evaluate lim

x 6 3 x 2

Solution

2 x 2 7 lim 2 x2 7

lim x

x 6 3 x 2 lim 6 3x 2

x

or or are all also known as indeterminant forms. When this form

occurs when finding limits at infinity (or negative infinity) with rational functions,

divide every term in the numerator and denominator by the highest power of x

in the denominator to determine the limit.

227

2 x 2 7 lim 2 x2 7 x2

lim x

x 6 3 x 2 lim 6 3x 2 x 2

x

7

lim 2 2

x

x

6

lim 2 3

x x

2 0 2

03 3

Continuity

Definition:

2. lim f ( x) exists (A limit value exists as you approach x=a.)

xa

x a

x=a.)

Graphically, this means that a function is continuous wherever the graph

of the function has no holes, gaps, or jumps. A function is said to be

discontinuous at x = a, if a hole, gap, or break occurs in the graph at x =

a, meaning the function violates one of the three items above.

Notice that the third item in the definition of continuity and Rule 2 of the

Limit Rules show that all polynomial functions are continuous for all real

values of x.

Example 10

Using the graph of f(x) below, find all values of x where f(x) is

discontinuous and state why f(x) is discontinuous at these points,

according to the definition of continuity.

228

Solution

x = -2, f(-2) is undefined

x = 1, while the function is defined by f(1)=5 and lim f ( x) 1 , these are not

x 1

x = 4, lim f ( x) does not exist

x4

Example 11

x4

Is f ( x) continuous at x =1? At x =3?

x 3

Solution

For both values of x, we must check each of the three items in the

definition of continuity. If one of the items fails, f(x) not continuous at

that particular x-value.

x = 1:

1 4 5

First, check to see if a function value exists: f (1)

1 3 2

229

x 4 lim x4 5

Next, check to see if a limit value exists: lim x1

x 1 x 3 lim x 3 2

x 1

Last, check that the function value equals the limit value, which it does

in this case. Therefore, f(x) is continuous at x =1.

x = 3:

3 4 7

First, check to see if a function value exists: f (3) is

33 0

undefined

discontinuous at x = 3.

values of x. This is also true for exponential functions. Moreover, it is

true that rational functions are continuous for all real values of x that do

not make the denominator zero and logarithmic functions are

continuous for all x-values in their domains.

230

Exercises

x 1

x

x 2

Exercise 3.5 Fill in the given table and use it to find lim f ( x) .

x2

x2

f ( x) ?

x2

Exercise 3.6 Fill in the given table and use it to find lim g ( x) .

x 0

231

x 1, x 0

g ( x) 3, x0 ?

x 2 1, x 0

x 2 4 x 21

Exercise 3.7 Evaluate lim 2

x 7 x 5 x 14

x 1

Exercise 3.8 Evaluate lim

x 1 x 2x 1

2

x 2 2 x x3

Exercise 3.9 Evaluate lim

x 3x 4 7

x 2, x 1

Exercise 3.10 Is f ( x) continuous at x = -1? At x=1?

x , x 1

232

## Гораздо больше, чем просто документы.

Откройте для себя все, что может предложить Scribd, включая книги и аудиокниги от крупных издательств.

Отменить можно в любой момент.