You are on page 1of 166

Introduction to

Introduction to Mathematical Analysis I
Mathematical Analysis I

Second Edition

Beatriz Lafferriere

MTH 311
Gerardo Lafferriere
Mau Nam Nguyen

Lafferriere, Lafferriere, Nguyen

 
 
Introduction to Mathematical Analysis I 
Second Edition 
 
 
 
 
 
Beatriz Lafferriere 
Gerardo Lafferriere 
Nguyen Mau Nam 
 
 
 
 
 
PDXOpen Textbooks 
2016   

 

 

 

 

 

 
© 2016 Portland State University 
ISBN: 978‐1‐365‐60552‐9 
 
 
 
 

 
This work is licensed under a 
Creative Commons Attribution‐NonCommercial 4.0 International License 
You are free to: 
 Share — copy and redistribute the material in any medium or format 
 Adapt — remix, transform, and build upon the material 
 
The licensor cannot revoke these freedoms as long as you follow the license terms. 
Under the following terms: 
 Attribution — You must give appropriate credit, provide a link to the license, and indicate if 
changes were made. You may do so in any reasonable manner, but not in any way that suggests 
the licensor endorses you or your use. 
 NonCommercial — You may not use the material for commercial purposes. 
 
 
 
 
 
This publication was made possible by PDXOpen publishing initiative 
 
Published by Portland State University Library  
Portland, OR 97207‐1151 
 
Cover design by Matt Eich 
 

. . 31 2.2 FUNCTIONS 11 1. . .6 OPEN SETS. . .3 MONOTONE SEQUENCES 41 2. . . . . . . . . . AND LIMIT POINTS 55 3 LIMITS AND CONTINUITY .4 THE BOLZANO-WEIERSTRASS THEOREM 47 2. . . . . . . . . . . . . . . . . . . . . .2 LIMIT THEOREMS 37 2. . . . . . . . . . . . .1 CONVERGENCE 31 2. . . . .5 THE COMPLETENESS AXIOM FOR THE REAL NUMBERS 22 1. . . . . . . . . . .6 APPLICATIONS OF THE COMPLETENESS AXIOM 26 2 SEQUENCES . . . .3 CONTINUITY 74 . .4 ORDERED FIELD AXIOMS 18 1. . . . . . . . . . .5 LIMIT SUPERIOR AND LIMIT INFERIOR 50 2. . . . . .1 BASIC CONCEPTS OF SET THEORY 7 1. . . . . . . . . . 63 3. . . . . . . . . . . . 7 1.2 LIMIT THEOREMS 67 3. Contents 1 TOOLS FOR ANALYSIS . . . . . . . . . . . . . . . . . .3 THE NATURAL NUMBERS AND MATHEMATICAL INDUCTION 15 1. . . COMPACT SETS. . . . CLOSED SETS. . . . . . . . . .1 LIMITS OF FUNCTIONS 63 3.

.6 LIMIT SUPERIOR AND LIMIT INFERIOR OF FUNCTIONS 89 3. . . .5 UNIFORM CONTINUITY 83 3. . . .6 CONVEX FUNCTIONS AND DERIVATIVES 126 4. . . . . . . . .7 LOWER SEMICONTINUITY AND UPPER SEMICONTINUITY 94 4 DIFFERENTIATION . . . . . . . . . . .4 L’HOSPITAL’S RULE 117 4. . . . .7 NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBDIFFERENTIALS 132 5 Solutions and Hints for Selected Exercises . . . . . . . .3 SOME APPLICATIONS OF THE MEAN VALUE THEOREM 114 4. . . . . 101 4. . . . . . . . .4 PROPERTIES OF CONTINUOUS FUNCTIONS 77 3. . . . . . . .1 DEFINITION AND BASIC PROPERTIES OF THE DERIVATIVE 101 4. . .5 TAYLOR’S THEOREM 123 4. . 143 . . .3. . . .2 THE MEAN VALUE THEOREM 107 4. . . .

For those exercises. Students should be familiar with most of the concepts presented here after completing the calculus sequence. Such a foundation is crucial for future study of deeper topics of analysis. teachers can choose to simplify them depending on the background of the students. The lecture notes contain topics of real analysis usually covered in a 10-week course: the completeness axiom. related topological concepts can be replaced by more familiar concepts such as open and closed intervals. to make it easier for students to navigate the text. theorem. sequences and convergence. there are explanations about the process itself and examples of more general problems where the same technique may be used. These hyperlinks can be easily recognized because the text or number is on a different color and the mouse pointer changes shape when going over them. continuity. and generalized differentiation of non-differentiable convex functions can be used as optional mathematical projects. differentiation of convex functions. For instance. However. The lecture notes also contain many well-selected exercises of various levels. Changes in the Second Edition The second edition includes a number of improvements based on recommendations from students and colleagues and on our own experience teaching the course over the last several years. Hints and solutions to selected exercises are collected in Chapter 5. Although these topics are written in a more abstract way compared with those available in some textbooks. Preface Our goal in this set of lecture notes is to provide students with a strong foundation in mathematical analysis. the electronic version of these notes contains many hyperlinks that students can click on to go to a definition. example. In this edition we streamlined the narrative in several sections. the lecture notes are suitable for teaching students of different backgrounds. in addition to the solutions. there is at least one exercise fully solved. added more proofs. Finally. rather than introducing the topology of the real line to students. these concepts will be reinforced through rigorous proofs. Some other topics such as lower and upper semicontinuity. Exercises with solutions are indicated by a I and those with hints are indicated by a B. For each section. or exercise at a different place in the notes. many examples . and differentiation. In this way.

and numerous new exercises. Chapter 2 We added the proof that compactness is equivalent to closed and bounded in the main text. As we are now preparing a companion text for the second term (Introduction to Mathematical Analysis II) we now added the roman numeral I to the title.6 worked out in detail. Chapter 4 We clarified the statement and the proof of the second version of L’Hospital’s rule. We defined compactness as what is more commonly termed sequential compactness. this definition is easier to apply and reinforce. We included more prominently the notion of compact set. Moreover. We created a separate section for limit superior/inferior of functions. We have used these notes several times to teach the one-quarter course Introduction to Mathe- matical Analysis I at Portland State University. as the emphasis of the whole text is on sequences. Chapter 3 We added the theorem on extension of uniformly continuous functions and moved the discussion of Lipschiptz and Hölder continuous functions to the section on uniform continuity. Students find this definition easier to absorb than the general one in terms of open covers. . Chapter 1 Added the proofs of several properties of the real numbers as an ordered field. In all we added over 50 examples in the main text and 100 exercises (counting parts). The following are the more significant changes.

we can list all of its elements. The reader should consult the bibliographical references for more details.1 Two sets A and B are equal if and only if A ⊂ B and B ⊂ A. The objects in a set are called the elements or members of the set. We usually use uppercase letters to denote sets and lowercase letters to denote elements of sets. If a is an element of a set A. For instance.1. and write A ⊂ B or B ⊃ A. we write a ∈ A. we write a 6∈ A. we write E = {x : x = 2k for some integer k}. and write A ( B. BASIC CONCEPTS OF SET THEORY FUNCTIONS THE NATURAL NUMBERS AND MATHEMATICAL INDUC- TION ORDERED FIELD AXIOMS THE COMPLETENESS AXIOM FOR THE REAL NUMBERS APPLICATIONS OF THE COMPLETENESS AXIOM 1. 1. a set is a collection of objects with certain properties. To specify a set. c. Two sets are equal if they contain the same elements. or we can use a defining rule. d. If a is not an element of a set A. b. Generalities are kept to a minimum in order to move quickly to the heart of analysis: the structure of the real number system and the notion of limit.1 BASIC CONCEPTS OF SET THEORY Intuitively. c. The following result is straightforward and very convenient for proving equality between sets. To describe the set E containing all even integers. If A ⊂ B and A does not equal B. We say that a set A is a subset of a set B if every element of A is also an element of B. b. if possible. d}. we write A = B. Theorem 1. . TOOLS FOR ANALYSIS This chapter discusses various mathematical concepts and constructions which are central to the study of the many fundamental results in analysis. If A and B are equal. to specify the fact that a set A contains four elements a. we write A = {a. we say that A is a proper subset of B.

8 1.1 BASIC CONCEPTS OF SET THEORY

The set 0/ = {x : x 6= x} is called the empty set. This set clearly has no elements. Using
Theorem 1.1.1, it is easy to show that all sets with no elements are equal. Thus, we refer to the empty
set.
Throughout this book, we will discuss several sets of numbers which should be familiar to the
reader:

• N = {1, 2, 3, . . .}, the set of natural numbers or positive integers.

• Z = {0, 1, −1, 2, −2, . . .}, the set of integers (that is, the natural numbers together with zero
and the negative of each natural number).

• Q = {m/n : m, n ∈ Z, n 6= 0}, the set of rational numbers.

• R, the set of real numbers.

• Intervals. For a, b ∈ R, we have

[a, b] = {x ∈ R : a ≤ x ≤ b},
(a, b] = {x ∈ R : a < x ≤ b},
[a, ∞) = {x ∈ R : a ≤ x},
(a, ∞) = {x ∈ R : a < x},

and similar definitions for (a, b), [a, b), (−∞, b], and (−∞, b). We will say more about the
symbols ∞ and −∞ in Section 1.5.

Since the real numbers are central to the study of analysis, we will discuss them in great detail in
Sections 1.4, 1.5, and 1.6.
For two sets A and B, the union, intersection, difference, and symmetric difference of A and B are
given respectively by

A ∪ B = {x : x ∈ A or x ∈ B},
A ∩ B = {x : x ∈ A and x ∈ B},
A \ B = {x : x ∈ A and x ∈
/ B}, and
A ∆ B = (A \ B) ∪ (B \ A).

If A ∩ B = 0,
/ we say that A and B are disjoint.
The difference of A and B is also called the complement of B in A. If X is a universal set, that is,
a set containing all the objects under consideration, then the complement of A in X is denoted simply
by Ac .

Theorem 1.1.2 Let A, B, and C be subsets of a universal set X. Then the following hold:

(a) A ∪ Ac = X;
(b) A ∩ Ac = 0;
/
(c) c c
(A ) = A;
(d) (Distributive law) A ∩ (B ∪C) = (A ∩ B) ∪ (A ∩C);

9

(e) (Distributive law) A ∪ (B ∩C) = (A ∪ B) ∩ (A ∪C);
(f) (DeMorgan’s law) A \ (B ∪C) = (A \ B) ∩ (A \C);
(g) (DeMorgan’s law) A \ (B ∩C) = (A \ B) ∪ (A \C);
(h) A \ B = A ∩ Bc .

Proof: We prove some of the results and leave the rest for the exercises.
(a) Clearly, A ∪ Ac ⊂ X since both A and Ac are subsets of X. Now let x ∈ X. Then either x is an
element of A or it is not an element of A. In the first case, x ∈ A and, so, x ∈ A ∪ Ac . In the second
case, x ∈ Ac and, so, x ∈ A ∪ Ac . Thus, X ⊂ A ∪ Ac .
(b) No element of x can be simultaneously in A and not in A. Thus, A ∩ Ac = 0. /
(d) Let x ∈ A ∩ (B ∪ C). Then x ∈ A and x ∈ B ∪ C. Therefore, x ∈ B or x ∈ C. In the first
case, since x is also in A we get x ∈ A ∩ B and, hence, x ∈ (A ∩ B) ∪ (A ∩ C). In the second case,
x ∈ A ∩C and, hence, x ∈ (A ∩ B) ∪ (A ∩C). Thus, in all cases, x ∈ (A ∩ B) ∪ (A ∩C). This shows
A ∩ (B ∪C) ⊂ (A ∩ B) ∪ (A ∩C).
Now we prove the other inclusion. Let x ∈ (A ∩ B) ∪ (A ∩ C). Then x ∈ A ∩ B or x ∈ A ∩ C.
In either case, x ∈ A. In the first case, x ∈ B and, hence, x ∈ B ∪ C. It follows in this case that
x ∈ A ∩ (B ∪C). In the second case, x ∈ C and, hence, x ∈ B ∪C. Again, we conclude x ∈ A ∩ (B ∪C).
Therefore, (A ∩ B) ∪ (A ∩C) ⊂ A ∩ (B ∪C) as desired. 
A set whose elements are sets is often called a collection/family of sets and is often denoted by
script letters such as A or B.
Let I be a nonempty set such that to each i ∈ I corresponds a set Ai . Then the family of all sets
Ai as i ranges over I is denoted by

{Ai : i ∈ I}.

Such a family of sets is called an indexed family and the set I is called the index set. Consider the
indexed family of sets {Ai : i ∈ I}. The union and intersection of this family as i ranges over I is
defined respectively by
[
Ai = {x : x ∈ Ai for some i ∈ I}
i∈I

and
\
Ai = {x : x ∈ Ai for every i ∈ I}.
i∈I 

Example 1.1.1 The following examples illustrate the notation.

(a) Let the index set be I = N and for each n ∈ N we have An = [−n, n]. Then
[ \
An = R An = [−1, 1].
n∈N n∈N

(b) Here we let the index set be J = (0, 1] and for each α ∈ J we have Aα = (−α, α). Then
[ \
Aα = (−1, 1) Aα = {0}.
α∈J α∈J

10 1.1 BASIC CONCEPTS OF SET THEORY

The proofs of the following properties are similar to those in Theorem 1.1.2. We include the
proof of part (a) and leave the rest as an exercise.

Theorem 1.1.3 Let {Ai : i ∈ I} be an indexed family of subsets of a universal set X and let B be a
subset of X. Then the following hold:
 T
(a) B ∪ i∈I Ai = i∈I B ∪ Ai ;
T
 S
(b) B ∩ i∈I Ai = i∈I B ∩ Ai ;
S
 S
(c) B \ i∈I Ai = i∈I B \ Ai ;
T
 T
(d) B \ i∈I Ai = i∈I B \ Ai ;
S
c S
Ac ;
T
(e) i∈I Ai =
S c Ti∈I c
(f) i∈I Ai = i∈I A .

Proof of (a): Let x ∈ B ∪ i∈I Ai . Then x ∈ B or x ∈ i∈I Ai . If x ∈ B, then x ∈ B ∪ Ai for all i ∈ I
T T

and, thus, x ∈ i∈I B ∪ Ai . If x ∈ i∈I Ai , then x ∈ Ai for all i ∈I. Therefore, x ∈ B ∪ Ai for all i ∈ I
T T

and, hence, x ∈ i∈I B ∪ Ai . We have thus showed B ∪ i∈I Ai ⊂ i∈I B ∪ Ai .
T T T

Now let x ∈ i∈I B ∪ Ai . Then x ∈ B ∪ Ai for all i ∈ I. If x ∈ B, then x ∈ B ∪ i∈I Ai . Ifx 6∈ B,
T T

then we must have that x ∈ Ai for all i ∈ I. Therefore, x ∈ i∈I Ai and, hence, x ∈ B ∪ i∈I Ai . This
T T

proves the other inclusion and, so, the equality. 
We want to consider pairs of objects in which the order matters. Given objects a and b, we will
denote by (a, b) the ordered pair where a is the first element and b is the second element. The main
characteristic of ordered pairs is that (a, b) = (c, d) if and only if a = c and b = d. Thus, the ordered
pair (0, 1) represents a different object than the pair (1, 0) (while the set {0, 1} is the same as the set
{1, 0})1 .
Given two sets A and B, the Cartesian product of A and B is the set defined by

A × B := {(a, b) : a ∈ A and b ∈ B}. 

Example 1.1.2 If A = {1, 2} and B = {−2, 0, 1}, then

A × B = {(1, −2), (1, 0), (1, 1), (2, −2), (2, 0), (2, 1)}. 

Example 1.1.3 If A and B are the intervals [−1, 2] and [0, 7] respectively, then A × B is the
rectangle

[−1, 2] × [0, 7] = {(x, y) : − 1 ≤ x ≤ 2, 0 ≤ y ≤ 7}.

We will make use of cartesian products in the next section when we discuss functions.

Exercises
1.1.1 Prove the remaining items in Theorem 1.1.2.

1.1.2 I Let Y and Z be subsets of X. Prove that

(X \Y ) ∩ Z = Z \ (Y ∩ Z).
1 For a precise definition of ordered pair in terms of sets see [Lay13]

If x ∈ X.3 Prove the remaining items in Theorem 1. A function from X into Y is a subset f ⊂ X × Y with the following properties (a) For all x ∈ X there is y ∈ Y such that (x. B. 1. In fact. y) ∈ f . and D be sets. is a new function from A into Y given by f|A (a) = f (a) for all a ∈ A. z) ∈ f . f is one-to-one if and only if for all x and x0 in X. (b) (X ×Y ) \ (A × B) = [(X \ A) ×Y ] ∪ [X × (Y \ B)].2 FUNCTIONS Definition 1. Then the two functions are equal if they are equal as subsets of X ×Y . The function f is called injective (or one-to-one) if for each pair of distinct elements of X.5 Let A ⊂ X and B ⊂ Y . It follows from the definition that two equal functions must have the same domain.1 Let X and Y be sets. their images under f are also distinct.1. thus. Both refer to the set {(x. then y = z. Prove the following. f (x)) : x ∈ X}. Determine if the following equalities are true and justify your answer: (a) (X ×Y ) \ (A × B) = (X \ A) × (Y \ B). Note that. (a) (A ∩ B) ×C = (A ×C) ∩ (B ×C).1. The set X is called the domain of f . This element x is then denoted by f −1 (y). corresponds to the geometric interpretation of the graph of a function given in calculus. y) ∈ f }. (c) (A × B) ∩ (C × D) = (A ∩C) × (B ∩ D).2 A function f : X → Y is called surjective (or is said to map X onto Y ) if for every element y ∈ Y . we already built a function from Y to X called the inverse of f . denoted by f|A . the set Y is called the codomain of f . Let f : X → Y be a function and let A be a subset of X. the element f (x) is called the value of f at x or the image of x under f . (b) (A ∪ B) ×C = (A ×C) ∪ (B ×C). 1. y) ∈ f .2. a function is a collection of ordered pairs and. y) ∈ f and (x. C. In this way.2. Let f : X → Y and g : X → Y be two functions.1.1. in this definition. there is exactly one element y ∈ Y such that (x. we will refer indistinctly to the function f or to the graph of f .3. the following implication holds: [ f (x) = f (x0 )] ⇒ [x = x0 ].4 Let A. for any y ∈ Y . If f is both surjective and injective. In this case. It is easy to see that f equals g if and only if f (x) = g(x) for all x ∈ X. Thus. 11 1. The range of f is the subset of Y defined by {y ∈ Y : there is x ∈ X such that (x. 1. (b) If (x. and we write f : X → Y . for each x ∈ X. . We will write y = f (x). there exists a unique element x ∈ X such that f (x) = y. there exists an element x ∈ X such that f (x) = y. It follows from the definition that. it is called bijective or a one-to-one correspondence. Definition 1. The restriction of f on A.

Then (x − 1)2 = (y − 1)2 . from the assumption on h.3 Let f : X → Y be a function. 2) and B = [1. Thus. Let us set up 4 − (x − 1)2 = y and √ solve for x. and let B be a subset of Y .12 1. Theorem 1. −4. 3]. For a subset B of Y .2. f is injective.3 Let f : X → Y be a function and let A be a subset of X. 2] such that f (x) = y. 2] be such that f (x) = f (y).2. Then f (A) = [0. Definition 1.2. so. Therefore. g(y) = g( f (x)) = x. 4) given by f (x) = 4 − (x − 1)2 .  Example 1. Moreover.2. y − 4 has a square root. Then. we have 1 ≥ 4 − y > 0 and. It follows from the definition that f (A) = {b ∈ Y : b = f (a) for some a ∈ A}. This proves f is surjective. we obtain g(y) = f −1 (y). Then x = g( f (x)) = g( f (x0 )) = x0 . −1] ∪ [1. 2]. 4). x = 4 − y + 1. Next we prove that f is injective. 5}. If there are two functions g : Y → X and h : Y → X such that g( f (x)) = x for every x ∈ X and f (h(y)) = y for every y ∈ Y . ∞).2 Note that. 2) and B = {1.2 Let f : R → R be given by f (x) = 3x − 1. Let y ∈ Y and set x = h(y). First let x. despite the notation. Also note that √ since 3 ≤ y < 4. We get. Let x. We have shown that for each y ∈ Y . Next let y ∈ [3. The examples below illustrate these concepts. Let A = [0. This proves f is injective. y ∈ (1. x ∈ (1.2. the preimage of B under f is defined by f −1 (B) = {x ∈ X : f (x) ∈ B}. 7] and f −1 (B) = [4.2 FUNCTIONS Theorem 1. It does not even require the function to be injective. −1. 4). let A be a subset of X. Let A = [0.  Example 1. Since f (h(y)) = y. We want x ∈ (1. hence. 2) and B = (−∞.2. x = y.   Example 1.  Example 1. 2] → [3. we conclude that x − 1 = y − 1 and. Remark 1. The following hold: (a) A ⊂ f −1 ( f (A)). 5) and f −1 (B) = { 23 . Let A = (−1.4 Let f : R → R be given by f (x) = x2 . Since for such a y. there is a unique x ∈ X.2. then f is bijective and g = h = f −1 . x0 ∈ X be such that f (x) = f (x0 ). . Since both x > 1 and y > 1. Proof: First we prove that f is surjective. That is. 2 ≥ 4 − y + 1 > 1. 4) and f −1 (B) = (−2. the definition of preimage does not require the function to have an inverse. Then f (A) = (5. We show that f is bijective. Note that since y < 4.2. which we denote f −1 (y) such that f (x) = y. 2}. 2). Then f (A) = [−1. This shows that f is surjective. we also conclude that h(y) = x = f −1 (y).1 Consider the function f : (1.3 Let f : R → R be given by f (x) = −x + 7.1 Let f : X → Y . 4 − (x − 1)2 = 4 − (y − 1)2 . we have f (x) = f (h(y)) = y. Then the image of A under f is given by f (A) = { f (a) : a ∈ A}. f is surjective if and only if f (X) = Y .

Proof: We prove (b) and leave the other parts as an exercise.5 Let f : X → Y be a function. y ∈ α∈I f (Aα ). (d) f (A \ B) ⊃ f (A) \ f (B). there is α0 ∈ I such that x ∈ Aα0 . Hence x ∈ f −1 (D) and x 6∈ f −1 (C). there is x ∈ α∈I Aα such that S S y = f (x). x ∈ f −1 (D \C).4 Let f : X → Y be a function. Next we prove f −1 (D) \ f −1 (C) ⊂ f −1 (D \C). D ⊂ Y . then f (A) ⊂ f (B). then g is surjective. Theorem 1. .2. (b) We prove first that f −1 (D \ C) ⊂ f −1 (D) \ f −1 (C). then g ◦ f is injective. S S (d) f −1 ( f −1 (Bβ ). we define the composition function g ◦ f of f and g as the function g ◦ f : X → Z given by (g ◦ f )(x) = g( f (x)) for all x ∈ X.  Theorem 1. This means f (x) ∈ D \C and. Thus. Therefore.2. The following hold: (a) If C ⊂ D. (b) If f and g are injective. (c) If A ⊂ B. By the definition of image. 13 (b) f ( f −1 (B)) ⊂ B. The following hold: (a) (g ◦ f )−1 (B) = f −1 (g−1 (B)). We conclude that x ∈ f −1 (D) \ f −1 (C). so. by the definition of preimage. (a) Let x ∈ A. Now. S Given functions f : X → Y and g : Y → Z. Let x ∈ f −1 (D \ C). Thus. T T (c) f −1 ( β ∈J Bβ ) = β ∈J f −1 (Bβ ). Let x ∈ f −1 (D) \ f −1 (C). f (x) ∈ D and f (x) 6∈ C. then f is injective. f (x) ∈ D and f (x) 6∈ C. (c) If f and g are surjective. then f −1 (C) ⊂ f −1 (D). from the definition of inverse image. so. T T β ∈J Bβ ) = β ∈J Proof: We prove (a) and leave the other parts as an exercise.6 Let f : X → Y and g : Y → Z be two functions and let B ⊂ Z. (a) Let y ∈ f ( α∈I Aα ). B ⊂ X. and let C. y = f (x) ∈ f (Aα0 ) and. let {Aα }α∈I be an indexed family of subsets of X. x ∈ f −1 ( f (A)). let A.  Theorem 1. The following hold: S S (a) f ( α∈I Aα ) = α∈I f (Aα ). we get f (x) ∈ D \C. (b) f ( α∈I Aα ) ⊂ α∈I f (Aα ). f (x) ∈ f (A). Therefore. Proof: We prove (a) and leave (b) as an exercise. Then. From the definition of image of a set. (d) If g ◦ f is injective. and let {Bβ }β ∈J be an indexed family of subsets of Y . From the definition of union of a family of sets. x ∈ f −1 (D) and x 6∈ f −1 (C). (b) f −1 (D \C) = f −1 (D) \ f −1 (C). . (e) If g ◦ f is surjective. then g ◦ f is surjective.2.

See [Lay13] for more details.1 I Let f : X → Y be a function.2. 1. . then the following hold: (a) f (A ∩ B) = f (A) ∩ f (B) for A. (d) Suppose g ◦ f is injective and let x. (b) If f is onto. .2. . .2. .5 We say that set A is finite if it is empty or if there exists a natural number n and a one-to-one correspondence f : A → {1. ak } and define f (k + 1) = ak+1 (such an ak+1 exists because A \ {a1 . so.2. . B ⊂ X. x−1 1. n}. . that N is infinite. . . We define f as follows. 3] → [−2.3.  Definition 1.7 Suppose A is an infinite set. We discuss sequences in detail in Chapter 2. 3 (b) g : (1.3 Prove that each of the following functions is bijective. (b) f (A \ B) = f (A) \ f (B) for A. then A = f −1 ( f (A)) for every subset A of X.  To paraphrase. Exercises 1. .5 Prove part (2) of Theorem 1. f (k). the previous theorem says that in every infinite set we can find a sequence made up of all distinct points. . B ⊂ X. It is also easy to show.2. A set is infinite if it is not finite. nonempty). We conclude that f is injective.2. Now the set A \ {a1 } is again infinite (otherwise A = {a} ∪ (A \ {a1 }) would be the union of two finite sets). Definition 1. x0 ∈ X be such that f (x) = f (x0 ). . Proof: Let A be an infinite set. .2.2. 2) → (3. 1) and B = (−1. Choose any element a1 ∈ A and set f (1) = a1 .2 FUNCTIONS Proof: We prove (d) and leave the other parts as an exercise. Then (g ◦ f )(x) = g( f (x)) = g( f (x0 )) = (g ◦ f )(x0 ). Prove that: (a) If f is one-to-one. . 2. ak } is infinite and. 1. by contradiction.2 Let f : R → R be given by f (x) = x2 − 3 and let A = [−2. (a) f : (−∞. 2 This fact relies on a basic axiom of set theory called the Axiom of Choice. then f ( f −1 (B)) = B for every subset B of Y . Then there exists a one-to-one function f : N → A.2. Find f (A) and f −1 (B). . Theorem 1.4 A sequence of elements of a set A is a function with domain N and codomain A. The function f so defined clearly has the desired properties. We leave it as an exercise to prove that the union of two finite sets is finite.14 1. 6). The following result will be useful when studying sequences and accumulation points. So we may choose a2 ∈ A with a2 6= a1 and we define f (2) = a2 2 . we choose ak+1 ∈ A such that ak+1 ∈ A \ {a1 . . it follows that x = x0 .4 Prove that if f : X → Y is injective. 1. Since g ◦ f is injective. . ∞) given by g(x) = . ∞) given by f (x) = |x − 3| − 2. Having defined f (1).

2. so the conclusion follows. Hint: it is easier to show when the sets are disjoint. Well-Ordering Property of the Natural Numbers: If A is a nonempty subset of N.2. ` ≥ 2.1 — Principle of Mathematical Induction. By the Well- Ordering Property of the natural numbers. Set B = N \A. Then A = N.  Example 1. When proving (b).3.2. (b) For each k ∈ N.2. But k + 1 = `.4. By condition (a).7 Prove parts (2). B = {n ∈ N : P(n) is false}. the principle says that. and (5) of Theorem 1. then all propositions are true. 1. 1. The principle of mathematical induction is a useful tool for proving facts about sequences.8 Prove parts (1).2. P(k + 1) is true whenever P(k) is true. It follows that k = ` − 1 is a natural number.3 THE NATURAL NUMBERS AND MATHEMATICAL INDUCTION We will assume familiarity with the set N of natural numbers. (3).6 Prove parts (1). by way of contradiction. the statement P(k) is called the inductive hypothesis. Suppose the following conditions hold: (a) 1 ∈ A. Let A = {n ∈ N : P(n) is true}.2. we have that P(k) is true.1 Prove using induction that for all n ∈ N n(n + 1) 1+2+···+n = . Theorem 1. Hence. 1 6∈ B. In addition. (3). with the usual arithmetic operations of addition and multiplication on N. (3).  To paraphrase.6. 1. Assume.5. which is clearly true. moreover. To paraphrase the previous property. and with the notion of what it means for one natural number to be less than another. Proof: Suppose conditions (a) and (b) hold.3. For each natural number n ∈ N. and P(`) is false. This is a contradiction. that is.9 Prove that the union of two finite sets is finite. Since k < `. and (4) of Theorem 1. k 6∈ B and. 2 1(1+1) The statement P(n) is the equality 1+2+· · ·+n = n(n+1) 2 . if P(1) is true and. since ` ∈ B. 1. We will refer to this principle as mathematical induction or simply induction. hence. given a list of propositions P(n). Condition (a) above is called the base case and condition (b) the inductive step.2. every nonempty subset of positive integers has a smallest element. and (4) of Theorem 1. then there exists an element ` ∈ A such that ` ≤ x for all x ∈ A. suppose that P(n) denotes a proposition which is either true or false. 15 1. we will also assume the following property of the natural numbers. Then B is a nonempty subset of N. there exists a smallest element ` ∈ B. Now the base case says that 1 = 2 . we obtain that P(k + 1) is true. . (2). then k + 1 ∈ A. if k ∈ A. one for each n ∈ N. By condition (b). that A 6= N.

we have 1 + 1 = 2 = 21 . we have 7 − 2 = 5. 2 2 2 This shows that P(k + 1) is true. We conclude by the principle of mathematical induction that n + 1 ≤ 2n for all n ∈ N. 7n − 2n is divisible by 5. k ≥ n0 . We have now proved conditions (a) and (b) of Theorem 1.2 — Generalized Principle of Mathematical Induction. by the principle of mathematical induction we conclude that n(n + 1) 1+2+···+n = for all n ∈ N . Suppose next that k + 1 ≤ 2k for some k ∈ N. so the base case is true. Then B is a nonempty subset of N.3. suppose that 1 + 2 + · · · + k = 2 (this is the inductive hypothesis). which is clearly a multiple of 5. Set B = {n ∈ N : n ≥ n0 . Therefore. The following result is known as the Generalized Principle of Mathematical Induction. This is a contradiction. That is. Hence. By condition (a). By condition (b). suppose that P(n) denotes a proposition which is either true or false. Therefore. Assume. Now.  .3. By the Well- Ordering Property of the natural numbers. if k ∈ A. Since k < `.3 THE NATURAL NUMBERS AND MATHEMATICAL INDUCTION k(k+1) Suppose P(k) is true for some k ∈ N. Proof: Suppose conditions (a) and (b) hold. substituting this expression below. so.  Example 1.3. (b) For each k ∈ N. we obtain that P(k + 1) is true. It follows that 7k+1 − 2k+1 is a multiple of 5.3 Prove using induction that for all n ∈ N n + 1 ≤ 2n For n = 1. Now we have k(k + 1) k(k + 1) + 2(k + 1) (k + 1)(k + 2) 1 + 2 + · · · + k + (k + 1) = + (k + 1) = = . n0 6∈ B.16 1. we have that P(k) is true. This proves the inductive step. It simply states that we can start the induction process at any integer n0 . For n = 1. That is. and P(`) is false. P(n) is false}. Let n0 ∈ N and for each natural n ≥ n0 .2 Prove using induction that for all n ∈ N. Then k + 1 + 1 ≤ 2k + 1. that A 6⊇ {k ∈ N : k ≥ n0 }. Then {k ∈ N : k ≥ n0 } ⊂ A. k 6∈ B and. Suppose that 7k − 2k is a multiple of 5 for some k ∈ N. and then we obtain the truth of all statements P(n) for n ≥ n0 . Theorem 1. Let A = {n ∈ N : P(n) is true}. Let us write 7k = 2k + 5 j. there exists a smallest element ` ∈ B. But k + 1 = `. we also have 1 ≤ 2k . (k + 1) + 1 ≤ 2k + 1 ≤ 2k + 2k = 2 · 2k = 2k+1 . since ` ∈ B. there is an integer j such that 7k − 2k = 5 j.1. Since 2k is a positive integer. by way of contradiction. We conclude by induction that 7n − 2n is divisible by 5 for all n ∈ N. 2  Example 1. It follows that k = ` − 1 ≥ n0 . so the conclusion follows.3. then k + 1 ∈ A. ` ≥ n0 + 1. Suppose the following two conditions hold: (a) n0 ∈ A. we have 7k+1 −2k+1 = 7·7k −2·2k = 7(2k +5 j)−2·2k = 7·2k −2·2k +7·5 j = 2k (7−2)+5·7 j = 5(2k +7 j).

. q > 1 such that k + 1 = pq. Let A = {n ∈ N : P(n) is true}. k}. so. . .3.5 Prove by induction that every positive integer greater than 1 is either a prime number or a product of prime numbers.4 Note that the inductive step above says that. where k ∈ N. (b) For each k ∈ N. There is also a generalized version of this theorem where the base case is for some integer n0 > 1. Theorem 1. 1.. Thus. If k + 1 is prime. For each natural n ∈ N. the statement is true for n = 2. Next we present another variant of the induction principle which makes it easier to prove the inductive step. by the generalized principle of induction.  Example 1. n(n+1)(2n+1) (a) 12 + 22 + · · · + n2 = 6 for all n ∈ N. Suppose next that 3k < 2k for some k ∈ N.3. . 3(k + 1) = 3k + 3 < 2k + 3 < 2k + 2k = 2k+1 . then k + 1 ∈ A. k ≥ 4. .3. . . Since p.2) Thus..1) where the second inequality follows since k ≥ 4 and. . . but also that P(1). 2. Despite its name. The statement is true for n = 4 since 12 < 16..4 Prove by induction that 3n < 2n for all n ≥ 4. we may assume not only that P(k) is true. (c) 1 + 3 + · · · + (2n − 1) = n2 for all n ∈ N. . . Clearly.2 Prove that for all n ∈ N.3. suppose that P(n) denotes a proposition which is either true or false. the inequality holds for all n ≥ 4. Then A = N. Remark 1. q ≤ k. . by the inductive assumption applied to both p and q we can find prime numbers r1 .sm such that p = r1 · · · r` and q = s1 · · · sm (note that ` and m may both equal 1). But then k + 1 = r1 · · · r` s1 · · · sm .3 Prove that for all n ∈ N. 1. . .1 Prove the following using Mathematical Induction. 17  Example 1. Suppose the statement holds for any positive integer m ∈ {2.3.3.3 — Principle of Strong Induction. . there are positive integers p. k ≥ 2.r` and s1 . . (1. This shows that P(k + 1) is true. (1.3. this principle is equivalent to the standard one. 2k ≥ 16 > 3. Suppose the following two conditions hold: (a) 1 ∈ A. k ∈ A. if 1. Exercises 1. P(2). . .P(k − 1) are true. The conclusion now follows by the Principle of Strong Induction. 9n − 5n is divisible by 4. in order to prove P(k + 1) is true. Now. Otherwise. n2 (n+1)2 (b) 13 + 23 + · · · + n3 = 4 for all n ∈ N. the statement holds true for k + 1. the statement holds for k + 1. 7n − 1 is divisible by 3.

1−a 1. Prove by induction that (1 + a)n ≥ 1 + na for all n ∈ N . we will assume that there exists a set. y ∈ R. we will start from an axiomatic presentation of the real numbers.) (c) n3 ≤ 3n for all n ∈ N.4 ORDERED FIELD AXIOMS In this book. unique real numbers x + y and x · y and satisfying the following properties: (1a) (x + y) + z = x + (y + z) for all x. together with the completeness axiom.3. 18 1. We conclude the section with the definition of absolute value of a real number and with several results about it that will be used often later in the text.4 Prove the following using induction. 5 2 2 1. (1b) x + y = y + x for all x.3. y. prove that 1 − an+1 1 + a + a2 + · · · + an = for all n ∈ N . After listing the ordered field axioms we derive from them additional familiar properties of the real numbers.8 B Let a. 1. Use Mathematical Induction to prove the binomial theorem n   n n k n−k (a + b) = ∑ ab . . In this way we identify the basic properties that characterize the real numbers. denoted by R. presented in the next section. This does not require induction.3. Prove that √ √ 1 h 1 + 5 n  1 − 5 n i an = √ − . We assume the existence of a set R (the set of real numbers) and two operations + and · (addition and multiplication) assigning to each pair of real numbers x.4 ORDERED FIELD AXIOMS 1. 2n ≤ n2 + 1. stated below. (b) n2 ≤ 3n for all n ∈ N.7 Let a ≥ −1. b ∈ R and n ∈ N.3. y. (a) 2n + 1 ≤ 2n for n ≥ 3 (n ∈ N). (Hint: show first that for all n ∈ N. z ∈ R. satisfying the ordered field axioms. (Hint: Check the cases n = 1 and n = 2 directly and then use induction for n ≥ 3.6 I The Fibonacci sequence is defined by a1 = a2 = 1 and an+2 = an+1 + an for n ≥ 1.5 Given a real number a 6= 1. That is. (1c) There exists a unique element 0 ∈ R such that x + 0 = x for all x ∈ R.3. k=0 k   n n! where = .) 1. k k!(n − k)! 1.

z ∈ R. y ∈ R. (2d) For each x ∈ R \ {0}. We may also use the notation x > y to represent y < x and the notation x ≥ y to mean x > y or x = y. (We also write 1/x instead of x−1 . y. Some examples are given in the next proposition. exactly one of the three relations holds: x = y. Proposition 1. or x < y. z ∈ R. (3b) For all x. (c) If x 6= 0 and xy = xz. the set of positive integers N does not form a field either. then −x < 0. The proof illustrates how the given axioms are used at each step of the derivation. (d) If x 6= 0. A set F together with two operations + and · and a relation < satifying the 13 axioms above is called an ordered field. The integers Z do not form a field since for an integer m other than 1 or −1. then −x > 0. y. In addition to the algebraic axioms above. Another example of an ordered field is the set of rational numbers Q with the familiar operations and order. axiom 2(d) above does not hold. many familiar properties of R can be derived. if x < y and 0 < z. z ∈ R. then y = z. then 1/(1/x) = x. z ∈ R. Thus. As mentioned above the real numbers R will be defined as the ordered field which satisfies one additional property described in the next section: the completeness axiom. y ∈ R. then xz < yz. y < x. z ∈ R. (f) −x = (−1)x. then x + z < y + z. (i) If x < y and z < 0. z ∈ R. then x < z. (b) −(−x) = x. (2c) There exists a unique element 1 ∈ R such that 1 6= 0 and x · 1 = x for all x ∈ R. (2a) (x · y) · z = x · (y · z) for all x.) (2e) x · (y + z) = x · y + x · z for all x. (2b) x · y = y · x for all x. y. From these axioms. (e) 0x = 0 = x0. there is a relation < on R that satisfies the order axioms below: (3a) For all x.1 For x. then xz > yz. We will use the notation x ≤ y to mean x < y or x = y. if x < y. y. (j) 0 < 1. the real numbers are an example of an ordered field. there exists a unique element x−1 ∈ R such that x · (x−1 ) = 1. there exists a unique element −x ∈ R such that x + (−x) = 0. In particular. if x < y and y < z. (3c) For all x. if x < 0. the following hold: (a) If x + y = x + z. (h) If x > 0. (3d) For all x. thus. then y = z. 19 (1d) For each x ∈ R. its reciprocal 1/m is not an integer and. (g) x(−z) = (−x)z = −(xz). We often write xy instead of x · y. .4. y. y.

we get 1 > 0 which is a contradiction. by axiom (2c) and 0 · 1 = 0 by part (e). Definition 1. that is. The other case follows in a similar way. . that 1 < 0. |x| = −x. (i) Since z < 0. since 0 < 1 (from part (j) of the previous proposition). Thus. if it is not rational. 1 < 2 < 3. −z > 0. etc (in particular all these numbers are distinct). (3c). It follows that 1 > 0. by way of contradiction. etc. x(−z) < y(−z). by part (h). In a similar way. 3 with 1 + 1 + 1. we get 0 = −(0x) + 0x = −(0x) + (0x + 0x) = (−(0x) + 0x) + 0x = 0 + 0x = 0x. and (1c) we get xy = (−xz + xz) + xy = −xz + (xz + xy) < −xy + (xz + xy) = −xy + (xy + xz) = (−xy + xy) + xz = xz. Then by part (i). we have (by uniqueness in axiom (2d)) −(−x) = x. (f) Using axioms (2c) and (2e) we get x + (−1)x = 1x + (−1)x = (1 + (−1))x. Combining this with part (g) we get −xz < −yz. define the absolute value of x by ( x. Adding −x (which exists by axiom (2d)) to both sides. again by axiom (2d). Then axiom (1a) gives [(−x) + x] + y = [(−x) + x] + z. Furthermore. Thus. The proofs of (c) and (d) are similar. Adding −(0x) to both sides (axiom (2d)) and using axioms (1a) and (1c). the number 2 with 1 + 1. in fact) by identifying the 1 in N with the 1 in axiom (2c) above.20 1. We say that a real number x is irrational if x ∈ R \ Q. That 0x = x0 follows from axiom (2b). can include Z and Q as subsets. using axiom (2d) we get that x(−z) = −(xz). axiom (3c) gives. (1b).4. The other equality follows similarly.4 ORDERED FIELD AXIOMS Proof: (a) Suppose x + y = x + z.  Note that we can assume that the set of all natural numbers is a subset of R (and of any ordered field. (j) Axiom (2c) gives that 1 6= 0. we get 0 > −x. Thus. we have (−x) + (x + y) = (−x) + (x + z).1 Given x ∈ R. by axiom (1c). Since 1 · 1 = 1. Then. if x < 0. 0 + y = 0 + z and. if x ≥ 0. Adding xz + yz to both sides and using axioms (1a). From axiom (2d) we get 1 + (−1) = 0 and from part (e) we get x + (−1)x = 0x = 0. 1 · 1 > 0 · 1. using axioms (3c) and (1c) we have x + (−x) > 0 + (−x) = −x. Suppose. (b) Since (−x) + x = 0. y = z. (h) From x > 0. using axiom (2d). From the uniqueness in axiom (2d) we get (−1)x = −x as desired. by axiom (3d). (e) Using axiom (2e) we have 0x = (0 + 0)x = 0x + 0x. (g) Using axioms (2e) and (1c) we have xz + x(−z) = x(z + (−z)) = x0 = 0.

It follows that −M < x < 0 < M. then |x| = x < M as desired. The following properties hold: (a) |x| ≥ 0. Theorem 1. 21 Figure 1. (The same holds if < is replaced with ≤. The following properties of absolute value follow directly from the definition. For the converse. Proof: From the observation above. The next theorem will play an important role in the study of limits. In particular.3 — Triangle Inequality. Proposition 1.4. (d) |x| < M if and only if −M < x < M. we consider different cases. so x > −M. . −M < 0 ≤ x = |x| < M.4. (d) Suppose |x| < M. −x < M and. |x + y| ≤ |x| + |y|. We consider the two cases separately: x ≥ 0 and x < 0. −M < x implies M > −x. Then |x| = −x < M.2 Let x. Therefore. Suppose first x ≥ 0. this implies M > 0. Given x. y. Next suppose x < 0.) Proof: We prove (d) and leave the other parts as an exercise. Now. (b) | − x| = |x|. If x ≥ 0. suppose −M < x < M. Then |x| = −x. Again. since |x| ≤ |x| we get −|x| ≤ x ≤ |x|.1: The absolute value function. Adding up the inequalities gives −|x| − |y| ≤ x + y ≤ |x| + |y|. Now suppose x < 0. hence. M ∈ R and suppose M > 0. (c) |xy| = |x||y|. Then |x| = x and. y ∈ R. we have −|x| ≤ x ≤ |x| −|y| ≤ y ≤ |y|.  Note that as a consequence of part (d) above.

y} = . we are interested in the field of real numbers. Moreover. and (c) of Proposition 1. Prove the following (a) |x|2 = x2 .1 1.5 THE COMPLETENESS AXIOM FOR THE REAL NUMBERS There are many examples of ordered fields.1 Prove that n is an even integer if and only if n2 is an even integer. Definition 1. then n2 is odd.2. y). y.2 Prove parts (c) and (d) of Proposition 1. ||x| − |y|| ≤ |x − y|. A number M is called an upper bound of A if x ≤ M for all x ∈ A. 1.4. (c) |x + y| = |x| + |y| if and only if xy ≥ 0.4 For any x.5.4. d ∈ R. and d(x. y} = and min{x. then A is said to be bounded above.4. (b). Remark 1. There is an additional axiom that will distinguish this ordered field from all others. If A has an upper bound. In order to introduce our last axiom for the real numbers. .4.3 Let a. the number d(a. 2 2 1. More generally. that is. y) ≥ 0. (Hint: prove the “if” part by contraposition. Suppose 0 < a < b and 0 < c < d. 1. 1.5 I Prove Corollary 1.4.4. z. Prove that ac < bd.) 1. z) + d(z.1 Let A be a subset of R.4.4. y ∈ R.  Corollary 1.4 Prove parts (a).5 The absolute value has a geometric interpretation when considering the numbers in an ordered field as points on a line. y) ≤ d(x.4.7 Let x. The number |a| denotes the distance from the number a to 0. Exercises 1. y. (b) |x| < M if and only if x < M and −x < M.4. b.4.2 that d(x.5 THE COMPLETENESS AXIOM FOR THE REAL NUMBERS Since −|x| − |y| = −(|x| + |y|).6 Given two real numbers x and y. y) = 0 if and only if x = y. for all numbers x. b) = |a − b| is the distance between the points a and b. 22 1.4. the triangle inequality implies that d(x.4.4. 1. we first need some definitions. However. c. M ∈ R. It follows easily from Proposition 1. prove that x + y + |x − y| x + y − |x − y| max{x.4.2 (d). prove that if n is odd. the conclusion follows from Proposition 1.

(b) sup{3. Consider next the set [0. This verifies condition (1) in the definition of supremum. it follows as in the previous example that 12 is the supremum. then α ≤ M (this means α is smallest among all upper bounds).1). hence. n n 2 If n ∈ N is odd. We also say that A is bounded if it is both bounded above and bounded below. x ≤ 3. then n ≥ 2 and (−1)n 1 1 = ≤ . Thus. Therefore.5. This verifies condition (2) in the definition of supremum. 3). Since 3 ∈ [0. if (1) x ≤ α for all x ∈ A (that is. then (−1)n −1 1 = <0< . set x = M+3 2 . then M is not an upper bound of [0. 3].1 (a) sup[0. Suppose M is . If y ∈ A. a number L is a lower bound of A if L ≤ x for all x ∈ A. 3 is the supremum of [0. x ∈ R}. 3) as 0 is an element of [0. any upper bound M must satisfy 10 ≤ M as 10 is an element of the set. y = x2 = |x|2 < 4. 3] = 3.5. 3). Since 2 is an element of the set. then it has only one (see Exercise 1. Definition 1. 3) and assume by way of contradiction that 3 > M. 23 Similarly. 3). which shows M is not an upper bound of [0. n n 2 1 1 This shows that 2 is an upper bound of the set. Since we get a contradiction in both cases. In this case. 3]. Next suppose M is an upper bound of [0. Set A = {x2 : −2 < x < 1. 3). We call a number α a least upper bound or supremum of A.  Example 1. α is an upper bound of A). It follows that a set A is bounded if and only if there exist M ∈ R such that |x| ≤ M for all x ∈ A (see Exercise 1. (2) If M is an upper bound of A.2 Let A be a nonempty set that is bounded above. (−1)n   1 (c) sup :n∈N = . Thus 10 is the supremum. 3] = {x ∈ R : 0 ≤ x ≤ 3}. It is easy to see that if A has a supremum. Thus 3 is an upper bound. (d) sup{x2 : −2 < x < 1. 4 is an upper bound of A. First consider the set [0. If 0 > M. we denote such a number by sup A. Then 0 ≤ x < 3 and x > M. 10} = 10. we conclude that 3 ≤ M and. n 2 Note that if n ∈ N is even. It follows as before that 3 is an upper bound of [0. 3) = sup[0. and A is said to be bounded below if it has a lower bound.5. 3]. we get 3 ≤ M. hence. Moreover. 7. 3]. By its very definition we see that for all x ∈ [0.2). Now suppose that M is an upper bound of [0. 3) = {x ∈ R : 0 ≤ x < 3}.5. |x| < 2. 8. then y = x2 for some x satisfying −2 < x < 1 and. Clearly 10 is an upper bound of the set. It follows that 3 is indeed the supremum of [0. 5. If 0 ≤ M. x ∈ R} = 4.

Thus 4 ≤ M. we can prove that if a nonempty set is bounded below.  The following is an axiom of the real numbers and is called the completeness axiom. Choose a number y ∈ R such that M < y < 4 and 0 < y. (2) If N is a lower bound of A. 10} = 3. (b) inf{3. there is a ∈ A such that a > α − ε = M. This contradicts the fact that M is an upper bound. so. suppose conditions (1’) and (2’) hold. The following proposition is convenient in working with suprema. there must exist an element a of A such that α − ε < a as desired. 3] = 0. that M < α.3 Let A be a nonempty subset of R that is bounded below.5). (−1)n   (c) inf : n ∈ N = −1. 5. Let M be an upper bound of A and assume. Then clearly (1’) holds (since this is identical to condition (1) in the definition of supremum). By condition (2) in the statement. The Completeness Axiom. There is a corresponding definition for the infimum of a set. by way of con- tradiction. Now let ε > 0. That is.1 Let A be a nonempty subset of R that is bounded above. condition (2) in the definition of supremum implies that α − ε is not an upper bound of A.2 (a) inf(0. Definition 1. 8. then its infimum exists (see Exercise 1. 7.  Example 1. β is a lower bound of A). sup A exists and is a real number. Proposition 1. denoted by β = inf A.5. However. 3] = inf[0. Then all we need to show is that condition (2) in the definition of supremum holds. This axiom distinguishes the real numbers from all other ordered fields and it is crucial in the proofs of the central theorems of analysis.5. Since α − ε < α. Conversely.5.5 THE COMPLETENESS AXIOM FOR THE REAL NUMBERS an upper bound of A but M < 4. This proves that 4 = sup A. The conclusion now follows. We call a number β a greatest lower bound or infimum of A. Set √ x = − y. n (e) inf{x2 : −2 < x < 1.24 1. y > M which contradicts the fact that M is an upper bound. then β ≥ N (this means β is largest among all lower bounds). (2’) For any ε > 0. Using the completeness axiom. Therefore. y = x2 ∈ A.5. . if (1) x ≥ β for all x ∈ A (that is. Every nonempty subset A of R that is bounded above has a least upper bound. Then −2 < x < 0 < 1 and. n 1 (d) inf{1 + : n ∈ N} = 1. x ∈ R} = 0. there exists a ∈ A such that α − ε < a. Then α = sup A if and only if (1’) x ≤ α for all x ∈ A. Set ε = α − M. Proof: Suppose first that α = sup A.

1 Prove that a subset A of R is bounded if and only if there is M ∈ R such that |x| ≤ M for all x ∈ A.5. We denote the extended real number set by R.4 The extended real number system consists of the real field R and the two symbols ∞ and −∞. The following is a basic property of suprema. we will write inf A = −∞. Definition 1. then for x ∈ B. −∞ + (−∞) = −∞. since sup B is an upper bound for B. (d) ∞ + ∞ = ∞. Thus.5. A is also bounded above. then x + ∞ = ∞. The expressions 0 · ∞.5. (b) If x > 0. Then sup A ≤ sup B. ∞ · ∞ = (−∞) · (−∞) = ∞. Then β = inf A if and only if (1’) x ≥ β for all x ∈ A. then x · ∞ = −∞.5. (2’) For any ε > 0. Proposition 1. sup A ≤ sup B as desired. by the second condition in the definition of supremum. The set R with the above conventions will be convenient to describe results about limits in later chapters. x · (−∞) = ∞. there exists a ∈ A such that a < β + ε. x · (−∞) = −∞. x ≤ M.5. (c) If x < 0. Theorem 1. and (−∞) + ∞ are left undefined.2 Let A be a nonempty subset of R that is bounded below. We preserve the original order in R and define −∞ < x < ∞ for every x ∈ R The extended real number system does not form an ordered field. Definition 1. every nonempty subset of R has a supremum and an infimum in R. 25 The following proposition is useful when dealing with infima and its proof is completely analogous to that of Proposition 1. we will write sup A = ∞. With this definition. If A is not bounded below in R. To complete the picture we adopt the following conventions for the empty set: sup 0/ = −∞ and inf 0/ = ∞.5 If A 6= 0/ is not bounded above in R. Suppose B is bounded above. Exercises 1. .  It will be convenient for the study of limits of sequences and functions to introduce two additional symbols. x + (−∞) = −∞. Now. it is also an upper bound for A. but it is customary to make the following conventions: (a) If x is a real number. Additional ones are described in the exercises. Then. and (−∞) · ∞ = ∞ · (−∞) = −∞.1.5. ∞ + (−∞). then x · ∞ = ∞. it is also true that x ≤ M for x ∈ A.3 Let A and B be nonempty sets and A ⊂ B. Proof: Let M be an upper bound for B. In particular.

17} (b) [0. determine if it is bounded above. then αA is bounded below and inf αA = α sup A. bounded below.2 Let A be a nonempty set and suppose α and β satisfy conditions (1) and (2) in Definition 1. then inf A ≥ inf B. Define αA = {αa : a ∈ A}. Define A + B = {a + b : a ∈ A and b ∈ B}. (b) If α < 0 and A is bounded above. Prove that A + B is bounded above and sup(A + B) = sup A + sup B. Define −A = {−a : a ∈ A}.3 For each subset of R below. then A has an infimum in R and inf A = − sup(−A). 26 1.6 APPLICATIONS OF THE COMPLETENESS AXIOM We prove here several fundamental properties of the real numbers that are direct consequences of the Completeness Axiom. 1. Prove the following statements: (a) If α > 0 and A is bounded above. .5 Let A be a nonempty subset of R. B be nonempty subsets of R that are bounded below. then αA is bounded above and sup αA = α sup A. both are suprema of A). Prove that A + B is bounded below and inf(A + B) = inf A + inf B. Justify all your conclusions.5. Prove that if A ⊂ B. 1.5. 1. 1. or both.5.6 Let A be a nonempty subset of R and α ∈ R. Prove that α = β .4 I Suppose A and B are nonempty subsets of R that are bounded above.5. If it is bounded above (below) find the supremum (infimum).5. then −A is bounded above. (b) Prove that if A is bounded below. 5. ∞) (e) {x ∈ R : x2 − 3x + 2 = 0} (f) {x2 − 3x + 2 : x ∈ R} (g) {x ∈ R : x3 − 4x < 0} (h) {x ∈ R : 1 ≤ |x| < 3} 1.2 (that is.5. (a) Prove that if A is bounded below. 1.7 Suppose A and B are nonempty subsets of R that are bounded below.6 APPLICATIONS OF THE COMPLETENESS AXIOM 1.5. 5) (−1)n   (c) 1 + :n∈N n (d) (−3.5. (a) {1. 1.8 Let A.

A has a smallest element `. In particular. This implies k − N − 1 ≤ x < k − N. If ` − 1 ∈ N. Then. we have ` − 1 = 0 < x.2 The following hold: (a) For any x ∈ R. The proof is now complete. there exists n ∈ N such that x < n. (c) We only need to apply (a) for the real number y/x. But then n + 1 > α. Proof: Let us assume by contradiction that N is bounded above. A is nonempty. either ` − 1 ∈ N or ` − 1 = 0. This is a contradiction since n + 1 is a natural number. Define the set A = {n ∈ N : x < n}. there exists n ∈ N such that x < n. Therefore. This implies 1/n < ε. Since ` ∈ N. there exists n ∈ N such that 1/ε < n. Proof: (a) Fix any x ∈ R. there exists N ∈ N such that |x| < N.6.  The following theorem presents several immediate consequences. there exists a natural number k such that k − 1 ≤ x + N < k. there exists n ∈ N such that y < nx. there exists m ∈ Z such that m − 1 ≤ x < m. since ` − 1 6∈ A we get ` − 1 ≤ x.  . the conclusion follows. x cannot be an upper bound of N. By Proposition 1.5. In this case. Theorem 1. 27 Theorem 1. The set of natural numbers is unbounded above. there exists n ∈ N such that α − 1 < n ≤ α. there exists n ∈ N such that 1/n < ε.6. by the Well-Ordering Property of the natural numbers. (d) For any x ∈ R. Since A is a subset of N. (b) Fix any ε > 0. (d) First we consider the case where x > 0. x < ` and ` − 1 is not in A. Thus. (c) For any x > 0 and for any y ∈ R. Setting m = k − N. α = sup N exists and is a real number. By (1). If ` − 1 = 0.1 (with ε = 1). (b) For any ε > 0. Since N is not bounded above. Then 1/ε is a real number. Since N is nonempty. by part (1). −N < x < N. in both cases we have ` − 1 ≤ x < ` and the conclusion follows with m = `. by the result just obtained for positive numbers. so x + N > 0. In the case x ≤ 0. From part (a).1 — The Archimedean Property.

√ Proposition 1. Then there are integers r and s with s 6= 0. Therefore. Since 1 − 1/n < 1 for all n ∈ N.6. by Theorem 1. one can choose m ∈ Z such that m − 1 ≤ nx < m. Then ny = nx + n(y − x) > nx + 1. x < m/n < y. there exists n ∈ N such that 1 < n(y − x). Then nx < m ≤ nx + 1 < ny. that 2 ∈ Q. we obtain condition (1’). Theorem 1.1) √ that there exists a (unique) 2 positive real number x such that x = 2.1. The proof is now complete. such that √ r 2= .2 (3). let ε > 0. Proof: We are going to prove that there exist an integer m and a positive integer n such that x < m/n < y. By Theorem 1.4 The number 2 is irrational.6. we may assume that r and s have no common factors. we get r2 2= . From Theorem 1. in particular. equivalently.4.6.6. that R 6= Q.5. or. by way of contradiction.3. √ Proof: Suppose. nx < m < ny = nx + n(y − x).  We will prove in a later section (see Examples 3. Then 1 1−ε < 1− .2 and 4. by squaring both sides of the equation above. then there exists a rational number r such that x < r < y. We denote that number by 2. We use Proposition 1.28 1. Now. We claim that sup A = 1.2 (b) we can find n ∈ N such that 1n < ε.3 — The Density Property of Q.6.2 (4). The following result shows. s By canceling out the common factors of r and s. If x and y are two real numbers such that x < y.6 APPLICATIONS OF THE COMPLETENESS AXIOM  Example 1.1 Let A = sup{1 − 1n : n ∈ N}. s2 .6. Next. n This proves condition (2’) with a = 1 − 1n and completes the proof. Since y − x > 0.

Thus. Prove that for every n ∈ N.6. hence. Hence r2 = 4 j2 .   3n (a) :n∈N n+4   n 1 (b) (−1) + : n ∈ N n n   n (−1) (c) (−1) − :n∈N n 1. Proof: Since x < y. 29 and. Justify all your conclusions.6. s2 is even.3 Let x ∈ R. Theorem 1. Therefore. n+1 n 1 1. n .6. 2s2 = r2 .5 Let x and y be two real numbers such that x < y.3.1).  The next theorem shows that irrational numbers are as ubiquitous as rational numbers. there exists a rational number r such that √ √ x− 2 < r < y− 2 This implies √ x < r + 2 < y.1 For each sets below determine if it is bounded above. (1.4.6. Therefore.3).2 I Let r be a rational number such that 0 < r < 1. or both. one has √ √ x− 2 < y− 2 By Theorem 1. the number t = r + 2 is irrational (see Exercise 1.6. If it is bounded above (below) find the supremum (infimum). Then there exists an irrational number t such that x < t < y. We conclude as before that s is even.6. Substituting in (1. √ Since r is rational. we get s2 = 2 j2 . there is r ∈ Q such that |x − r| < . both r and s have a common factor. which is a contradiction. Prove that there is n ∈ N such that 1 1 <r≤ . r is an even integer (see Exercise 1. We can then write r = 2 j for some integer j. bounded below.4) and x < t < y.3) It follows that r2 is an even integer.  Exercises 1.

6 APPLICATIONS OF THE COMPLETENESS AXIOM 1. What can you say about xy? 1.6.6 Prove that in between two real numbers a and b with a < b. .4 Prove that if x is a rational number and y is an irrational number. 1. there are infinitely many rational numbers.6. there are infinitely many irrational numbers.6. then x + y is irrational.5 Prove that in between two real numbers a and b with a < b.30 1.

Each value an is called a term of the sequence. there exists a real number N such that for any n ∈ N with n > N. . we call a the limit of the sequence (see Theorem 2.1.0.  Example 2. namely. SEQUENCES We introduce the notion of limit first through sequences. a sequence of elements of a set A is a function f : N → A. −1. 2. or even {an }.1. . . an = f (n). −1}. one has |an − a| < ε. that a sequence {an } converges to a if and only if for any ε > 0. AND LIMIT POINTS 2. we may list the terms explicitly as in 1 1 1 1 1.1 Let {an } be a sequence of real numbers.0. we call the sequence divergent. 1. . using the Archimedean property. .1. one has |an − a| < ε (or equivalently. . . On occasion. CLOSED SETS. We will denote the image of n under the function with subscripted variables. there exists a positive integer N such that for any n ∈ N with n ≥ N. This should not be confused with the two-element set {1. −1. the n-th term of the sequence. 2 3 4 5 n  Example 2.1 It follows directly from the definition. More precisely. This is a sequence of rational numbers. {an }n . . . a − ε < an < a + ε). . a sequence is just a function with domain N. CONVERGENCE LIMIT THEOREMS MONOTONE SEQUENCES THE BOLZANO-WEIERSTRASS THEOREM LIMIT SUPERIOR AND LIMIT INFERIOR OPEN SETS. 1. Remark 2. We say that the sequence {an } converges to a ∈ R if. .2 Let an = (−1) for n ∈ N. This is a sequence of integers. for any ε > 0.3 below) and write limn→∞ an = a. 1. for example. −1. We will also denote sequences by {an }∞ n=1 . when the pattern is clear. If the sequence {an } does not converge. more precisely. As mentioned in Chapter 1.1 CONVERGENCE Definition 2. COMPACT SETS. Note that the sequence takes on only two values. In this case.1 Consider the sequence an = 1n for n ∈ N.

We claim that limn→∞ an = 0.1 Let an = 1n for n ∈ N. we get .) Then.32 2. Let ε > 0.1 CONVERGENCE  Example 2.1. if n ≥ N. (Note that such an integer N exists due to the Archimidean Property. Choose an integer N > 1/ε. We verify it using the definition.

.

.

1.

1 1 1 |an − 0| = .

.

.

.

1. hence. one has n > ( ε1 )1/α and. n n N 1/ε  Example 2. nα Let ε > 0. = ≤ < = ε.2 We now generalize the previous example as follows. This implies . Choose an integer N such that N > ( ε1 )1/α . nα > ε1 . For every n ≥ N. Let α > 0 and consider the sequence given by 1 an = for n ∈ N .

.

.

1 .

− 0.

= 1 < 1 = ε. .

.

nα .

we simplify and estimate the expression |an − a|. Notice that 2 2 2 .3 Consider the sequence {an } where 3n2 + 4 an = . nα 1/ε We conclude that limn→∞ an = 0.1. To that end. We first search for a suitable N. Let ε > 0. 2n2 + n + 5 We will prove directly from the definition that this sequence converges to a = 32 .  Example 2.

.

.

.

.

.

.

.

.

an − 3 .

= .

3n + 4 − 3 .

=

2(3n + 4) − 3(2n + n + 5)

=

−7 − 3n .

.

.

.

.

.

.

.

.

.

2 .

.

2n2 + n + 5 2 .

.

2(2n2 + n + 5) .

.

2(2n2 + n + 5) .

2(2n + n + 5) 4n 4n 10 To guarantee that the last expression is less than ε. Indeed. First notice that . 3n + 7 10n 10 = 2 < 2 = . 4ε we get 10 10 10 |an − a| ≤ ≤ < 10 = ε. 4n 4N 4 4ε  Example 2.1. if n ≥ N. it will suffice to choose N > .4 Let {an } be given by 4n2 − 1 an = . Let ε > 0. 3n2 − n We claim limn→∞] an = 43 . We search for a suitable N.

2 .

4n − 1 4 .

.

12n2 − 3 − 12n2 + 4n .

.

4n − 3 .

.

.

.

.

.

.

3n2 − n − 3 .

= .

.

=

3(3n2 − n)

.

.

.

.

.

.

3(3n2 − n) .

33

Since n ≥ 1, we have n2 ≥ n and 4n > 3. Therefore we have

2

4n − 1 4 .

4n − 3 4n − 3 4n 4 .

3n2 − n − 3 .

. = 3(3n2 − n) ≤ 3(3n2 − n2 ) < 6n2 = 6n .

.

for n ≥ N . we have. 4 Thus. if N > 6ε .

2 .

.

4n − 1 4 .

4 4 .

3n2 − n − 3 .

. ≤ 6n ≤ 6N < ε.

.

1. 4n2 + n We prove directly from the definition that {an } converges to 14 .  Example 2. Now. Let ε > 0.5 Consider the sequence given by n2 + 5 an = . .

2 1 .

.

.

.

4n2 + 20 − 4n2 − n .

.

.

.

.

.

n +5 |20 − n| .

− .

4n2 + n 4 .

.

= .

4(4n2 + n) If n ≥ 20. then |20 − n| = n − 20. Therefore. for such n we have . = 4(4n2 + n) .

2 .

.

n +5 1 .

.

n − 20 n 1 .

4n2 + n − 4 .

= 4(4n2 + n) ≤ 16n2 = 16n . .

1 Choose N > max{ 16ε . 20}. for n ≥ N we get . Then.

2 .

.

n +5 1 .

1 1 .

4n2 + n − 4 .

≤ 16n ≤ 16N < ε. .

.

Proof: Suppose {an } converges to a and b. If ` > 0. .  The following lemma is a simple generalization of (2. Proof: This is easily proved by contraposition. Let N = max{N1 .2).1. N2 }. hence. then there is a positive number. such that ε < `.2 Let ` ≥ 0. a = b. If ` < ε for all ε > 0. Then |a − b| ≤ |a − aN | + |aN − b| < ε/2 + ε/2 = ε. The following result is quite useful in proving certain inequalities between numbers. Lemma 2.1.  Theorem 2. |a − b| = 0 and.2. there exist positive integers N1 and N2 such that |an − a| < ε/2 for all n ≥ N1 and |an − b| < ε/2 for all n ≥ N2 . by Lemma 2. for example ε = `/2.1. then ` = 0. Since ε > 0 is arbitrary. Then given ε > 0.3 A convergent sequence {an } has at most one limit.1.

1. And suppose. The other direction follows immediately from the order axioms. limn→∞ bn = `. the sequence {an } is bounded below if the set {an : n ∈ N} is bounded below. Therefore.  Theorem 2. Then ε ε a− < aN ≤ bN < b + . that a > b.4 Given real numbers a. for n ≥ N.  Definition 2.5 — Comparison Theorem. then set ε0 = a − b.  The following comparison theorem shows that (non-strict) inequalities are preserved “in the limit”.6 — The Squeeze Theorem.1. Using Lemma 2. which is a contradiction. It follows that a ≤ b. Proof: For any ε > 0. By assumption. Similarly. 2 2 Choose N > max{N1 . . respec- tively. we have ` − ε < an ≤ bn ≤ cn < ` + ε. and an ≤ bn for all n ∈ N. and {cn } satisfy an ≤ bn ≤ cn for all n ∈ N.5. and limn→∞ an = limn→∞ cn = `. 2 2 ε ε b − < bn < b + . that is.34 2.4 we conclude a ≤ b.1 that the sequence {an } is bounded if and only if there is M ∈ R such that |an | ≤ M for all n ∈ N. there exists N1 ∈ N such that ` − ε < an < ` + ε for all n ≥ N1 . Proof: Suppose a < b + ε for all ε > 0. by way of contradiction.1. N2 ∈ N such that ε ε a− < an < a + . Since limn→∞ an = `. We say that the sequence {an } is bounded if the set {an : n ∈ N} is bounded. Then. Let N = max{N1 . N2 }. Similarly. Theorem 2. then a ≤ b if and only if a < b + ε for all ε > 0. since limn→∞ cn = `. for n ≥ N2 . there exists N2 ∈ N such that ` − ε < cn < ` + ε for all n ≥ N2 . b. there exist N1 . we should have a < b + ε0 = b + a − b = a. Suppose {an } and {bn } converge to a and b. Proof: Fix any ε > 0.1.2 A sequence {an } is bounded above if the set {an : n ∈ N} is bounded above. Then a ≤ b.1. if it is both bounded above and bounded below. a < b + ε for any ε > 0. Then ε0 > 0. 2 2 Thus. Suppose the sequences {an }. It follows from the observation after Definition 1. N2 }. for n ≥ N1 . {bn }.1 CONVERGENCE Lemma 2. Then limn→∞ bn = `. which implies |bn − `| < ε.

such that bk = ank for each k ∈ N.6 Consider the sequence an = (−1)n for n ∈ N. 35 Theorem 2. it is clear that n1 ≥ 1 since n1 is a positive integer. {ank } converges to a as k → ∞. .3 Let {an }∞ n=1 be a sequence of real numbers. {an } is bounded. nk+1 ≥ nk + 1. Proof: Suppose {an } converges to a and let ε > 0 be given. .  Example 2. For any k ≥ N. we also have |ank − a| < ε. n→∞ .   Example 2. Similarly. Now nk+1 > nk and. {a2k+1 } is also a subsequence of {an } and a2k+1 = −1 for all k (here nk = 2k + 1 for all k). Proof: We use mathematical induction. then any subsequence {ank } of {an } also con- verges to a. . When k = 1.1. suppose by contradiction that lim an = `. Therefore.  Theorem 2. Therefore.8 Let {nk }k be a sequence of positive integers with n1 < n2 < n3 < · · · Then nk ≥ k for all k ∈ N. Thus. there exists N ∈ N such that |an − a| < 1 for all n ≥ N. Indeed. Then there exists N such that |an − a| < ε for all n ≥ N. Then {a2k } is a subsequence of {an } and a2k = 1 for all k (here nk = 2k for all k).1. this implies. Proof: Suppose the sequence {an } converges to a. since nk and nk+1 are integers.1. . |aN−1 |. Assume nk ≥ k for some k.1.1. Lemma 2. Then |an | ≤ M for all n ∈ N. nk+1 ≥ k + 1 by the inductive hypothesis. |a| + 1}. The conclusion now follows by the principle of mathematical induction. Set M = max{|a1 |.7 A convergent sequence is bounded. for ε = 1. Then the sequence {an } is divergent.9 If a sequence {an } converges to a. Then. this implies |an | < 1 + |a| for all n ≥ N.7 Let an = (−1)n for n ∈ N. since nk ≥ k.1.  Definition 2. Since |an | − |a| ≤ ||an | − |a|| ≤ |an − a|. The sequence {bn }n=1 is called a ∞ subsequence of {an }∞ n=1 if there exists a sequence of increasing positive integers n1 < n2 < n3 < · · · .

1. that is. n→∞ 2n2 + 1 . (n − 1)(n − 2) although a1 and a2 are not defined. For simplicity of notation. a function defined only for the integers greater than or equal to k0 . 3n3 + 1 n2 + 1 (b) limn→∞ 2 = 1.6 Prove that the following limits are 0.36 2.1.1 CONVERGENCE Then every subsequence of {an } converges to a number ` ∈ R. this example illustrates the fact that the converse of theorem 2.7 is not true.1.6.1 Prove the following directly from the definition of limit. we talk of the sequence {an } given by n+1 an = . we may also denote this sequence by {an } whenever the integer k0 is clear from the context.1. Is the converse true? 2. In all cases. in particular. it follows. Prove that the sequence {xn yn } converges to 0. k→∞ k→∞ This contradiction shows that the sequence is divergent.5 Let {xn } be a bounded sequence and let {yn } be a sequence that converges to 0. Exercises 2. the sequence must be defined from some integer onwards.1. Prove that if the sequence {|an |} converges to 0.1. Since the sequence {an } is bounded but not convergent. 2. For instance. it will be convenient to also talk about the sequence {an }n≥k0 .3 Let {an } be a sequence.) n + cos(n2 − 3) (a) lim . n −n 2. n 2.2 Prove that if {an } is a convergent sequence. 2n2 + 2 (a) limn→∞ = 0.1. then {an } also converges to 0. sin n 2. 4n − n 2 3n2 + 5 1 (d) limn→∞ 2 = . Remark 2.1.10 Given a positive integer k0 . 5n + n + 1 5 2n3 + 1 1 (c) limn→∞ 3 = .4 Prove that limn→∞ = 0. then {|an |} is a convergent sequence. 6n + n 2 4n2 − 1 (e) limn→∞ 2 = 4.1. From the previous theorem. (Hint: use Theorem 2. that ` = lim a2k = 1 and ` = lim a2k+1 = −1.

n (a) Prove that if limn→∞ an = `. 2. . limk→∞ a3k+1 = `. Prove that if limn→∞ an = `.   an an a (d) If in addition b 6= 0 and bn 6= 0 for n ∈ N. . then limn→∞ an = `.2 LIMIT THEOREMS We now prove several theorems that facilitate the computation of limits of some sequences in terms of those of other simpler sequences.8 I Prove that if limn→∞ an = ` 6= 0.12 I Consider a sequence {an }.1 Let {an } and {bn } be sequences of real numbers and let k be a real number. 2. then limn→∞ bn = `.1.7 Prove that if limn→∞ an = ` > 0.13 Given a sequence {an }. (c) limn→∞ (an bn ) = ab.3.1. Is the conclusion still true if ` = 0? 2. n→∞ 3 2.) n→∞ an 3 √ √ 2. n→∞ n! n! (c) lim n . (Hint: prove first that there is N such that an > 1 forn ≥ N. + an bn = .11 Prove that the sequence {an } with an = sin(nπ/2) is divergent. {kan }.1. then there exists N ∈ N such that an > 0 for all n ≥ N. Theorem 2. define a new sequence {bn } by a1 + a2 + .1. bn bn b . 2.1. n→∞ an + 1 4 (b) lim = . Then the sequences {an + bn }. (b) Find a counterexample to show that the converse does not hold in general.1. and limk→∞ a3k+2 = `. a 2. (b) Prove that limn→∞ an = ` if and only if limk→∞ a3k = `. 2.9 Let {an } be a sequence of real numbers such that limn→∞ an = 3. and {an bn } converge and (a) limn→∞ (an + bn ) = a + b. n→∞ n n2 (d) lim n .4(c)). Use Definition 2. then converges and limn→∞ = .1. (a) Prove that limn→∞ an = ` if and only if limk→∞ a2k = ` and limk→∞ a2k+1 = `.10 Let an ≥ 0 for all n ∈ N. (b) limn→∞ (kan ) = ka. 37 3n (b) lim . Suppose {an } converges to a and {bn } converges to b.1 to prove the following (a) lim 3an − 7 = 2.2. (Hint: see Exercise 1.1. then limn→∞ an+1 n = 1.

2 Let N = max{N1 . Since {an } converges to a. For any n ≥ N.1. This proves (c). n→∞ bn b . we may choose N2 ∈ N such that ε |bn − b| < for all n ≥ N2 . since {bn } converges to b. 2M 2(|b| + 1) Therefore.1) Let ε > 0. it follows from (2. (c) Since {an } is convergent. This proves (a). let N ∈ N be such that |an − a| < for n ≥ N. 2(|b| + 1) Similarly.38 2. (d) Let us first show that 1 1 lim = . there exists M > 0 such that |an | ≤ M for all n ∈ N . N2 }. 2M Let N = max{N1 . Since {an } converges to a. |k| |kan − ka| = |k||an − a| < ε. it follows from Theorem 2. The conclusion follows immediately. then ka = 0 and kan = 0 for all n. 2 Similarly. limn→∞ (an + bn ) = a + b. there exists N1 ∈ N such that ε |an − a| < for all n ≥ N1 . one has ε ε |(an + bn ) − (a + b)| ≤ |an − a| + |bn − b| < + = ε. It follows that limn→∞ (kan ) = ka as desired. (2.2 LIMIT THEOREMS Proof: (a) Fix any ε > 0. we may choose N1 ∈ N such that ε |an − a| < for all n ≥ N1 . 2 2 Therefore. This proves (b). limn→∞ an bn = ab. Suppose ε next that k 6= 0. Given ε > 0. Thus. for n ≥ N. we have the following estimate: |an bn − ab| = |an bn − an b + an b − ab| ≤ |an ||bn − b| + |b||an − a|. N2 }. Then for n ≥ N. Then.7 that it is bounded. For every n ∈ N. (b) If k = 0.1) that ε ε |an bn − ab| < M + |b| < ε for all n ≥ N. there exists N2 ∈ N such that ε |bn − b| < for all n ≥ N2 .

2 2 2 For each n ≥ N1 . 2 |b| |b| |b| It follows (using a triangle inequality) that. there is N1 ∈ N such that |b| |bn − b| < for n ≥ N1 . 39 Since {bn } converges to b. < |bn |. we have the following estimate . − < |bn | − |b| < and. hence. for such n.

.

.

1 1 .

|bn − b| 2|bn − b| .

− .

2) .= ≤ . (2.

bn b .

one has . By (2. N2 }. Since limn→∞ bn = b. |bn ||b| b2 Now let ε > 0.2). 2 Let N = max{N1 . there exists N2 ∈ N such that b2 ε |bn − b| < for all n ≥ N2 .

.

.

1 1 .

2|bn − b| .

− .

≤ < ε for all n ≥ N. .

bn b .

2. n→∞ bn n→∞ bn b The proof is now complete. In this case.4) 1/n2 − 4/n + 7 Therefore.3) 1 − 4n + 7n2 Dividing numerator and denominator by n2 . where b > 0. By the binomial theorem. (2. Consider the case where b > 1. we can apply part (c) and have an 1 a lim = lim an = . by the limit theorems above.2. bn b Finally. 3 − 2/n + 5/n2 limn→∞ 3 − limn→∞ 2/n + limn→∞ 5/n2 3 lim an = lim = = .   Example 2.5) n→∞ n→∞ 1/n2 − 4/n + 7 limn→∞ 1/n2 − limn→∞ 4/n + limn→∞ 7 7 √n  Example 2.2 Let an = b. n . we can write 3 − 2/n + 5/n2 an = (2. (2. b2 1 1 It follows that limn→∞ = .1 Consider the sequence {an } given by 3n2 − 2n + 5 an = . This implies b−1 0 < an − 1 ≤ . an > 1 for every n. b = ann = (an − 1 + 1)n ≥ 1 + n(an − 1).

It follows that for n ≥ N.40 2. it has been shown that limn→∞ xn = 1. √ √ (c) limn→∞ ( 3 n3 + 3n2 − n2 + n).4 Find the following limits.2. choose N > .   2 2 2 (b) limn→∞ + +···+ n . √ √ (d) limn→∞ ( n + 1 − n). n→∞ 3n − 2n2 + 1 2. n N Thus.5 Prove or disprove the following statements: . √ (b) limn→∞ ( 3 n3 + 3n2 − n). hence. limn→∞ an = 1.2 Find the following limits: √ 3n + 1 (a) lim √ √ . In the case where b = 1. limn→∞ (b + br + br2 + · · · + brn ).2. n→∞ n→∞ xn Exercises 2. it is obvious that an = 1 for all n and. ε b−1 b−1 |an − 1| = an − 1 < ≤ < ε. limn→∞ an = 1. 1 If 0 < b < 1. 10 102 10 2.2 LIMIT THEOREMS b−1 For each ε > 0. let c = and define b √ n 1 xn = c= . n→∞ 4n2 − 3 1 + 3n − n3 (b) lim 3 . This implies 1 lim an = lim = 1.1 Find the following limits: 3n2 − 6n + 7 (a) lim .2.2.2. n→∞ n 2. 2.3 I Find the following limits if they exist: √ (a) limn→∞ ( n2 + n − n). n→∞ n + 3 r n 2n + 1 (b) lim . √ √ (e) limn→∞ ( n + 1 − n)/n. an Since c > 1. (a) For |r| < 1 and b ∈ R.

` − ε < an < ` + ε for all n ≥ N. then {an + bn } is divergent sequence. then {an bn } is a convergent sequence.1 A sequence {an } is called increasing if an ≤ an+1 for all n ∈ N. 41 (a) If {an } and {bn } are convergent sequences. (e) If {an } and {an + bn } are convergent sequences. Proof: (a) Let {an } be an increasing sequence that is bounded above. then an ≤ am whenever n ≤ m.3. since ` is an upper bound for A.1. Thus. Let ` = sup A and let ε > 0. Define bn = −an . If {an } is increasing or decreasing. Let {an } be a sequence of real numbers. It is easy to show by induction that if {an } is an increasing sequence. ` − ε < aN ≤ an for all n ≥ N. then {an + bn } is a convergent sequence. then it is convergent. Theorem 2. hence. By Proposition 1. then it is called a monotone sequence. we have an ≤ ` for all n. Therefore. (c) If {an } and {bn } are convergent sequences. (b) Let {an } be a decreasing sequence that is bounded below. limn→∞ an = `. On the other hand. (d) If {an } and {bn } are divergent sequences. Then A is a subset of R that is nonempty and bounded above and. (f) If {an } and {an + bn } are divergent sequences. there exists N ∈ N such that ` − ε < aN ≤ `.3 MONOTONE SEQUENCES Definition 2. an > an+1 for all n ∈ N). then it is convergent. It is called decreasing if an ≥ an+1 for all n ∈ N. strictly decreasing) if an < an+1 for all n ∈ N (resp. Define A = {an : n ∈ N}. 2. Since {an } is increasing.1 — Monotone Convergence Theorem.3. The sequence is called strictly increasing (resp. . then {an bn } is a divergent sequence. (b) If {an } is decreasing and bounded below. sup A exists.5. The following hold: (a) If {an } is increasing and bounded above. (b) If {an } and {bn } are divergent sequences. then {bn } is a convergent sequence. then {bn } is a divergent sequence.

3 MONOTONE SEQUENCES Then {bn } is increasing and bounded above (if M is a lower bound for {an }. then lim an = inf{an : n ∈ N}.3. then lim an = sup{an : n ∈ N}. Then 0 ≤ an+1 = ran ≤ an for all n. Then ak + 5 3 + 5 8 ak+1 = ≤ = ≤ 3.1. suppose ak < ak+1 for some k ∈ N.2. 3 3 3 It follows that an ≤ 3 for all n ∈ N. In the general case. Next we prove that the sequence is bounded by 3. Suppose that ak ≤ 3 for some k ∈ N.7) 3 First we will show that the sequence is increasing. an < an+1 . ak + 5 ak+1 + 5 ak+1 = < = ak+2 . Again.  Remark 2. (2. the sequence converges. n→∞ n→∞ Then {an } converges to −` by Theorem 2. Let ` = lim bn = lim (−an ).1.42 2. {an } is decreasing and bounded below. The statement is clearly true for n = 1.3. see Exercise 2. if {an } is decreasing and bounded below. the statement is true for n = 1. n→∞ Since an+1 = ran for all n. We prove by induction that for all n ∈ N.3.3. n→∞ This is clear if r = 0.1 that if {an } is increasing and bounded above. Thus. define an = rn for n ∈ N.1.2 Consider the sequence {an } defined as follows: a1 = 2 (2. (1 − r)` = 0 and. then −M is an upper bound for {bn }). ` = 0. Then ak + 5 < ak+1 + 5 and (ak + 5)/3 < (ak+1 + 5)/3. Since a2 = a13+5 = 73 > 2 = a1 . Let us first consider the case where 0 < r < 1.6) an + 5 an+1 = for n ≥ 1. n→∞  Example 2. By Theorem 2. we proceed by induction. .3. Next. Let ` = lim an . Therefore.2 It follows from the proof of Theorem 2.3. 3 3 It follows by induction that the sequence is increasing. Therefore. taking limits on both sides gives ` = r`.1 Given r ∈ R with |r| < 1. we only need to consider the sequence defined by bn = |an | for n ∈ N.  Example 2. n→∞ Similarly. Then lim an = 0. hence.

Theorem 2.7). e is an irrational number and e ≈ 2. taking limits on both sides of the equation in (2. T / (b) If. This shows that {an } is an increasing sequence bounded above by b1 and {bn } is a decreasing sequence bounded below by a1 . By the . Moreover. Then the following hold: (a) ∞ n=1 In 6= 0.2 2. 43 From the Monotone Convergence Theorem.3 —The number e. Since the subsequence {ak+1 }∞ k=1 also converges to `. In particular. bn ].3 — Nested Intervals Theorem. n By the binomial theorem. Thus. n    k n 1 an = ∑ k=0 k n n(n − 1) 1 n(n − 1)(n − 2) 1 n(n − 1) · · · (n − (n − 1)) 1 = 1+1+ 2 + 3 +···+ n 2! n   3!   n   n!   n   1 1 1 1 2 1 1 2 n−1 = 1+1+ 1− + 1− 1− +···+ 1− 1− ··· 1− . n ∈ N. 3` = ` + 5 and. It is then clear that an < an+1 for all n ∈ N. Consider the sequence {an } given by 1 n   an = 1 + . In fact. The following fundamental result is an application of the Monotone Convergence Theorem. we deduce that there is ` ∈ R such that limn→∞ an = `.  Example 2. hence. in addition. 2! n 3! n n n! n n n The corresponding expression for an+1 has one more term and each factor (1 − nk ) is replaced by the k larger factor (1 − n+1 ). 3 Therefore. limn→∞ an exists and is denoted by e. By the Monotone Convergence Theorem.71828. Given that In+1 ⊂ In . an ≤ bn for all n ∈ N. 1 1 1 an ≤ 1 + 1 + + +···+ 2! 3! n! 1 1 1 < 2+ + +···+ 1. the lengths of the intervals In converge to zero. we obtain `+5 `= .3 (n − 1) · n n−1   1 1 1 = 2+ ∑ − = 3 − < 3.3. ` = 5/2. Proof: Let {In } be as in the statement with In = [an . k=1 k k+1 n Hence the sequence is bounded above. the sequence is increasing. we have an ≤ an+1 and bn+1 ≤ bn for all n ∈ N. Let {In }∞ n=1 be a sequence of nonempty closed bounded intervals satisfying In+1 ⊂ In for all n ∈ N. then ∞ T n=1 In consists of a single point.3.

2 A sequence {an } is said to diverge to ∞ if for every M ∈ R. that limn→∞ an = ∞. Thus. This proves the desired inclusion and. Therefore. we get a ≤ b.4 Consider the sequence an = n 5n . n→∞ Similarly. Then b = limn→∞ bn = limn→∞ [(bn − an ) + an ] = a.  .44 2. This means bn − an → 0 as n → ∞. Indeed. Then an ≥ aN ≥ M for all n ≥ N because {an } is increasing. T an ≤ x ≤ bn for all n. it does not converge. we write limn→∞ an = ∞. 5n 5 5n 5 Choose N > 5M. We now prove part (b). Let M ∈ R. the behavior follows a clear pattern. using Definition 2. ∞ T n=1 In = [a. b ∈ R such that limn→∞ an = a and limn→∞ bn = b. [a.3.5.3. there exists N ∈ N such that an > M for all n ≥ N. Definition 2. Thus. To make this precise we provide the following definition. then x ∈ In for all n. if n ≥ N. b]. It follows that ∞ T n=1 In = {a} as desired. n→∞ Proof: Fix any real number M. we have n N an ≥ ≥ > M. Since an ≤ bn for all n. there exists N ∈ N such that aN ≥ M.3. In this case. This shows that if a ≤ x ≤ b.3. Now. Using Theorem 2. Similarly. there exist a. b].  When a monotone sequence is not bounded. one that does not converge). The proof for the second case is similar. if x ∈ n=1 In .1.1). b] ⊂ ∞ T∞ n=1 In .5 If a sequence {an } is increasing and not bounded above.1. one that satisfies the previous definition). by Theorem 2.3. Note that n2 + 1 n 1 n = + ≥ . limn→∞ an = ∞. hence. if {an } is decreasing and not bounded below. Then. we conclude a ≤ x ≤ b. T / This proves part (a).3 MONOTONE SEQUENCES Monotone Convergence Theorem (Theorem 2. Now note also that ∞ T∞ n=1 In ⊂ [a. x ∈ [a.5. then x ∈ In for all n ∈ N. However. Since {an } is not bounded above. then lim an = ∞. +1 2  Example 2. Theorem 2.3. with a divergent sequence (that is. Therefore. We will show.4 We should not confuse a sequence that diverges to ∞ (that is. b]. we also have an ≤ a and b ≤ bn for all n ∈ N (since {an } is increasing and {bn } is decreasing). 5 5 Th following result completes the description of the behavior of monotone sequences. then lim an = −∞. Suppose the lengths of the intervals In converge to zero. there exists N ∈ N such that an < M for all n ≥ N. Remark 2. we say that {an } diverges to −∞ and write limn→∞ an = −∞ if for every M ∈ R. It follows that n=1 In 6= 0.2.

(b) limn→∞ (an bn ) = ∞. . Since limn→∞ an = ∞. ε This implies that for n ≥ N.6 Let {an }. 2 Let N = max{N1 . lim bn = ∞. (a) Fix any M ∈ R.3. Then it is clear that an + bn ≥ M for all n ≥ N. Since limn→∞ an = ∞. {bn }. and limn→∞ kan = −∞ if k < 0. This implies (a). and lim cn = −∞ n→∞ n→∞ n→∞ Then (a) limn→∞ (an + bn ) = ∞. there exists N1 ∈ N such that M an ≥ for all n ≥ N1 . N2 }. (c) limn→∞ (an cn ) = −∞. there exists N2 ∈ N such that M bn ≥ for all n ≥ N1 . 1 (e) For any ε > 0. 2 Similarly. Suppose lim an = ∞.) an Proof: We provide proofs for (a) and (e) and leave the others as exercises. and {cn } be sequences of real numbers and let k be a constant. there exists N ∈ N such that ε 1 an > for all n ≥ N. let M = . (d) limn→∞ kan = ∞ if k > 0. (Here we assume an 6= 0 for all n. 45 Theorem 2. 1 (e) limn→∞ = 0.

.

.

1 .

− 0.

. = 1 < ε.

.

an .

1. . (a) If limn→∞ an = ∞. then limn→∞ an = −∞.  The proof of the comparison theorem below follows directly from Definition 2.2 (see also Theorem 2.3.7 Suppose an ≤ bn for all n ∈ N. an Thus.5). (b) If limn→∞ bn = −∞.3. (e) holds. Theorem 2. then limn→∞ bn = ∞.

· · · 1 1 (b) 1/2.3 B Prove that each of the following sequences is convergent and find its limit. (b) Prove that {an } is an increasing sequence.3.   1 1 (c) an+1 = 2an + 2 .3.3.2. 2 2.3. (a) Prove that an < 2 for all n ∈ N.3.··· 2 + 1/2 1 2+ 2 + 1/2 2. . b1 = b. . an + 3 (a) a1 = 1 and an+1 = for n ≥ 1.6 Prove the following using Definition 2.46 2.3. 3 an   1 b (d) an+1 = an + . 2! 3! n! 2.1 I Let a1 = 2. 2. 2 an 2. n→∞ n + 2 2.6. and (d) of Theorem 2.3.7.3 MONOTONE SEQUENCES Exercises √ 2. 2. 2n2 + n + 1 (a) lim = ∞. n ∈ N. √ √2 (b) a1 = 6 and an+1 = an + 6 for n ≥ 1.2 B Prove that each of the following sequences is convergent and find its limit.8 Prove Theorem 2. (c). 2 2 2.3.4 Prove that the following sequence is convergent: 1 1 1 an = 1 + + +···+ . Define a1 = a.3. a1 > 0. n→∞ n−2 1 − 3n2 (b) lim = −∞.3. . 2. Define p an+1 = an + 2 for n ≥ 1. √ p √ √ q p (a) 2.3. b > 0.5 B Let a and b be two positive real numbers with a < b. n ≥ 1. and p an + bn an+1 = an bn and bn+1 = for n ≥ 1. (c) Prove that limn→∞ an = 2. 2 Show that {an } and {bn } are convergent to the same limit.7 Prove parts (b).

Proof: Let {an } be a convergent sequence and let lim an = a. one has |am − an | < ε. Let I2 = [c. This is possible since I2 ∩ A is infinite. Every bounded sequence {an } of real numbers has a convergent subsequence. d] is infinite. c+d c+d c+d 2 ] and [ 2 .3. d]. Now. that {cn } converges. Set I1 = [c. dn ]. Choose n2 > n1 such that an2 ∈ I2 . {an } is a Cauchy sequence. Set In = [cn . n ≥ N. d] otherwise. d]. Define A = {an : n ∈ N} (the set of values of the sequence {an }). Pick n1 such that an1 = x.2 A convergent sequence is a Cauchy sequence. Suppose now that A is infinite. We can then define a convergent subsequence as follows. limn→∞ dn = limn→∞ [(dn − cn ) + cn ] = ` as well. 47 2. This completes the proof. Then limn→∞ (dn − cn ) = 0.4. Continuing in this way. it follows from Theorem 2.1 (Cauchy sequence). A sequence {an } of real numbers is called a Cauchy sequence if for any ε > 0. n ≥ N. 2 ] is infinite and I2 = [ 2 . Thus.4. we can define a subsequence {ank } which is constant. In this way. then at least one of the elements of A. For any m.  .4. We define a sequence of nonempty nested closed bounded intervals as follows. d]. one has |am − an | ≤ |am − a| + |an − a| < ε/2 + ε/2 = ε. Theorem 2. We also know from the proof of the Monotone Convergence Theorem (Theorem 2. at least one of A ∩ [c. we obtain a subsequence {ank } such that ank ∈ Ik for all k ∈ N. Thus. converges to x. n→∞ Then for any ε > 0. Next choose n3 > n2 such that an3 ∈ I3 . We now construct the desired subsequence of {an } as follows. thus. Let n1 = 1. More precisely. Since A is infinite. we can choose n2 > n1 such that an2 = x. in fact. Say ` = limn→∞ cn . First observe there exist c.1). Theorem 2. say x. A ⊂ [c. If A is finite. 2 ] if A ∩ [c. we construct a nested sequence of nonempty closed bounded intervals {In } such that In ∩ A is infinite and the length of In tends to 0 as n → ∞.1. Continuing in this way. d ∈ R such that c ≤ an ≤ d for all n ∈ N.4 THE BOLZANO-WEIERSTRASS THEOREM The Bolzano-Weierstrass Theorem is at the foundation of many results in analysis. 2 ] c+d c+d c+d c+d or A ∩ [ 2 . Bx = {n ∈ N : an = x} is infinite. Since ck ≤ ank ≤ dk for all k ∈ N. Proof: Suppose {an } is a bounded sequence.5 that limk→∞ ank = `. equivalent to the completeness axiom of the real numbers. Next consider the two subintervals [c. there exists a positive integer N such that |an − a| < ε/2 for all n ≥ N. there exists a positive integer N such that for any m. must be equal to an for infinitely many choices of n. since Bx is infinite.1 — Bolzano-Weierstrass. that is. It is.  Definition 2. equal to x and.

Therefore. .6 It follows from Definition 2. |an | ≤ M for all n ∈ N and.48 2. thus. |an − aN | < 1 for all n ≥ N. Then. |aN−1 |.4. {an } converges to a. there exists a positive integer N such that |am − an | ≤ ε/2 for all m. .4 THE BOLZANO-WEIERSTRASS THEOREM Theorem 2.4. it is convergent by Lemma 2. For the above ε. . Therefore. {an } has a convergent subsequence.4. we have |an − a| ≤ |an − an` | + |an` − a| < ε.3 A Cauchy sequence is bounded. Let {ank } be a subsequence of {an } that converges to some point a. Proof: Let {an } be a Cauchy sequence. n ≥ N. . Then it is bounded by Theorem 2.4. Therefore. Thus.4. By the Bolzano- Weierstrass theorem. n ≥ N. In particular. 1 + |aN |}. Let M = max{|a1 |. . Proof: Let {an } be a Cauchy sequence. there exists a positive number K such that |ank − a| < ε/2 for all k ≥ K. |an | = |an − aN + aN | ≤ |an − aN | + |aN | ≤ 1 + |aN | ≤ M. . for n ≥ N. . Definition 2.  Theorem 2.4. .4 A Cauchy sequence that has a convergent subsequence is convergent.3. for n = 1. . N − 1. we clearly have |an | ≤ M. Proof: Let {an } be a Cauchy sequence that has a convergent subsequence. we can find a positive integer n` > N such that |an` − a| < ε/2.  Remark 2.4.4. 1) such that |an+2 − an+1 | ≤ k|an+1 − an | for all n ∈ N . Moreover. there exists a positive integer N such that |am − an | < 1 for all m.5 Any Cauchy sequence of real numbers is convergent.1 that {an } is a Cauchy sequence if and only if for every ε > 0. Then for ε = 1. there exists N ∈ N such that |an+p − an | < ε for all n ≥ N and for all p ∈ N . For any ε > 0. {an } is bounded.  Lemma 2.4. Then for any n ≥ N.2 A sequence {an } is called contractive if there exists k ∈ [0.

1 The condition k < 1 in the previous theorem is crucial.4. Consider the following example. |an+p − an | ≤ |an+1 − an | + |an+2 − an+1 | + · · · + |an+p − an+p−1 | ≤ (kn−1 + kn + · · · + kn+p−2 )|a2 − a1 | ≤ kn−1 (1 + k + k2 + · · · + k p−1 )|a2 − a1 | kn−1 ≤ |a2 − a1 |.4. Since 1 < n+2 n+1 n+1 < n for all n ∈ N and the natural logarithm is an increasing function. Proof: By induction. 49 Theorem 2. 1−k for all n. p ∈ N. hence.   Example 2. Since kn−1 → 0 as n → ∞ (independently of p). we have . Thus. it is convergent. this implies {an } is a Cauchy sequence and.7 Every contractive sequence is convergent. Let an = ln n for all n ∈ N. one has |an+1 − an | ≤ kn−1 |a2 − a1 | for all n ∈ N .

 .

  .

n + 2 .

.

n+2 |an+2 − an+1 | = | ln(n + 2) − ln(n + 1)| = .

ln .

= ln n+1 .

4. (c) an = n/(n + 1). (a) an = (−1)n . the inequality in Definition 2.4. (d) an = (cos n)/n. . Exercises 2.4.1 I Determine which of the following are Cauchy sequences. n Therefore. (b) an = (−1)n /n.2 is satisfied with k = 1. yet the sequence {ln n} does not converge.2 Prove that the sequence n cos(3n2 + 2n + 1) an = n+1 has a convergent subsequence. 2. n+1   n+1 < ln = | ln(n + 1) − ln n| = |an+1 − an |.

5. n→∞ n→∞ Note that lim supn→∞ an = limn→∞ sn . Therefore. the limit inferior of {an }. both sequences are convergent by Theorem 2. is defined by lim sup an = lim sup{ak : k ≥ n}. denoted by lim infn→∞ an . then so is {sn }. ∞) → R be such that f (x) > 0 for all x.1.5 LIMIT SUPERIOR AND LIMIT INFERIOR 2.9) Then {sn } and {tn } are convergent. Then the limit superior of {an }. 2. 50 2. 2n Prove that the sequence an is contractive. the sequence {sn } is decreasing.1 Let {an } be a sequence.8). Define an = rn for n ∈ N.5 LIMIT SUPERIOR AND LIMIT INFERIOR We begin this section with a proposition which follows from Theorem 2. {tn } is increasing and bounded above. In particular.9). Similarly. is defined by lim inf an = lim inf{ak : k ≥ n}. where sn is defined in (2. (2.4 Define 1 + 2n an = for n ∈ N .  Definition 2. n→∞ n→∞ Note that lim infn→∞ an = limn→∞ tn . Similarly. Proof: If n ≤ m. denoted by lim supn→∞ an . .5.8) and tn = inf{ak : k ≥ n}.4.6 Prove that the sequence is not contractive. Prove that the sequence {an } is contractive. 2.  ∞ 1 2.5 Let r ∈ R be such that |r| < 1. {sn } is bounded below.3 Let f : [0.4. n n=1 2.4. so. Define f (n) an = .1 Let {an } be a bounded sequence. All sequences in this section are assumed to be of real numbers.4. Proposition 2.3 that sn ≥ sm and.3. Since {an } is bounded. f (n) + 1 Prove that the sequence an has a convergent subsequence. where tn is defined in (2.3. Define sn = sup{ak : k ≥ n} (2.5. it follows from Theorem 1. Therefore. then {ak : k ≥ m} ⊂ {ak : k ≥ n}.1.

n→∞ Similarly. This implies sN = sup{an : n ≥ N} < ` + ε. then lim sup an = ∞.9).5. Thus. then lim tn = −∞. for ε = 1.2. there exists N ∈ N such that an < ` + ε for all n ≥ N. then lim sn = ∞. the set {ai : i ≥ k} is also not bounded above.  Remark 2. and there exists a subsequence of {ank } of {an } such that lim ank = `. Therefore. limk→∞ sk = ∞.4 Let {an } be a sequence and ` ∈ R.5.8). .5. Similarly. Thus. Proof: Suppose {an } is not bounded above. where sn is defined as in (2. sk = sup{ai : i ≥ k} = ∞ for all k.3 By Theorem 2. n→∞ where {sn } is defined in (2. there exists N ∈ N such that ` − ε < sn < ` + ε for all n ≥ N. an < ` + ε for all n ≥ N. n→∞ Theorem 2.8). Thus. k→∞ Proof: Suppose lim supn→∞ an = `. Moreover. there exists n1 ∈ N such that ` − 1 < an1 < ` + 1. if {an } is not bounded below. (b) For any ε > 0. we see that if {an } is not bounded above. The following are equivalent: (a) lim supn→∞ an = `. The proof for the second case is similar. Then limn→∞ sn = `. For any ε > 0.5. Then for any k ∈ N. there exists N1 ∈ N such that ` − 1 < sN1 = sup{an : n ≥ N1 } < ` + 1. n→∞ where {tn } is defined in (2. then lim inf an = −∞. if {an } is not bounded below. 51 Theorem 2.2 If {an } is not bounded above.

8. there exists N ∈ N such that an < ` + ε and ` − ε < ank < ` + ε for all n ≥ N and k ≥ N.6 Let {an } be a sequence. Let any m ≥ N. so we also have sm = sup{ak : k ≥ m} ≥ anm > ` − ε. Given any ε > 0.52 2. nm ≥ m.5.5 LIMIT SUPERIOR AND LIMIT INFERIOR 1 For ε = . we have sm = sup{ak : k ≥ m} ≤ ` + ε. .5.5. The following are equivalent: (a) lim infn→∞ an = `. there exists N ∈ N such that an > ` − ε for all n ≥ N. limk→∞ ank = `.1. there exists N2 ∈ N and N2 > n1 such that 2 1 1 `− < sN2 = sup{an : n ≥ N2 } < ` + .  The following result is proved in a similar way.7 Let {an } be a sequence. there exists n2 > n1 such that 1 1 `− < an2 < ` + . Theorem 2. We now prove the converse.5. (b) For any ε > 0.5. By Lemma 2. Corollary 2.5 Let {an } be a sequence and ` ∈ R. Then lim an = ` if and only if lim sup an = lim inf an = `.5. k→∞ The following corollary follows directly from Theorems 2. k k Therefore. limm→∞ sm = lim supm→∞ an = `. 2 2 In this way. and there exists a subsequence of {ank } of {an } such that lim ank = `. Therefore.4 and 2. n→∞ n→∞ n→∞ Corollary 2. 2 2 Thus. we can construct a strictly increasing sequence {nk } of positive integers such that 1 1 `− < ank < ` + .

Define A = {x ∈ R : there exists a subsequence {ank } with lim ank = x}. Theorem 2. By induction. there exists N ∈ N such that an < ` + ε and `0 − ε < ank < `0 + ε for all n ≥ N and k ≥ N.5. Each element of the set A called a subsequential limit of the sequence {an }.  By a similar method. Since limn→∞ qn−N aN = 0. It follows from Theorem 2. Thus. k→∞ Then `0 ≥ `. (b) Suppose lim infn→∞ an = ` and {ank } is a subsequence of {an } with lim ank = `0 . 53 (a) Suppose lim supn→∞ an = ` and {ank } is a subsequence of {an } with lim ank = `0 . By Theorem 2.5.5. we obtain the theorem below. .9 Suppose {an } is a sequence such that an > 0 for every n ∈ N and an+1 lim sup = ` < 1.4.5.5. k→∞ Then `0 ≤ `. n→∞ an Then limn→∞ an = 0. n→∞ n→∞ Theorem 2. Then 0 < q < 1.5.4 and the definition of limits. for any ε > 0.8 Let {an } be a bounded sequence.7 that A 6= 0/ and lim sup an = max A and lim inf an = min A. `0 < ` + 2ε and. one has limn→∞ an = 0. Proof: We prove only (a) because the proof of (b) is similar. hence. an Let q = ` + ε. and Corollary 2. this implies `0 − ε < anN < ` + ε. 0 < an ≤ qn−N aN for all n ≥ N.  Remark 2. Then there exists N ∈ N such that an+1 < ` + ε for all n ≥ N. Since nN ≥ N. Proof: Choose ε > 0 such that ` + ε < 1.5. `0 ≤ ` because ε is arbitrary.

(b) lim supn→∞ an = −∞ if and only if limn→∞ an = −∞. Is the conclusion still true in each of parts (a) and (b) if the sequences involved are not necessarily bounded? 2.10 Suppose {an } is a sequence such that an > 0 for every n ∈ N and an+1 lim inf = ` > 1.5. define αn an = . (b) infk≥n (an + bn ) ≥ infk≥n ak + infk≥n bk . Suppose α > 0. it is obvious that limn→∞ an = 0. (b) Prove that lim infn→∞ (an + bn ) ≥ lim infn→∞ an + lim infn→∞ bn .5. n  nπ (d) an = n sin . 2 2. (a) Prove that lim supn→∞ (an + bn ) ≤ lim supn→∞ an + lim supn→∞ bn .5 Let {an } be a convergent sequence and let {bn } be an arbitrary sequence.5.5 LIMIT SUPERIOR AND LIMIT INFERIOR Theorem 2. Prove that (a) lim supn→∞ (an + bn ) = lim supn→∞ an + lim supn→∞ bn = limn→∞ an + lim supn→∞ bn .4 I Let {an } and {bn } be bounded sequences. Exercises All sequences in this set of exercises are assumed to be in R. 2. (b) lim infn→∞ (an + bn ) = lim infn→∞ an + lim infn→∞ bn = limn→∞ an + lim infn→∞ bn .5. n ∈ N.5.5. 2 1 + (−1)n (c) an = . . 2.  nπ  (b) an = sin . 2. (c) Find two counterexamples to show that the equalities may not hold in part (a) and part (b).3 Let {an } and {bn } be bounded sequences. Prove that: (a) supk≥n (an + bn ) ≤ supk≥n ak + supk≥n bk . Then an+1 α lim sup = lim = 0 < 1.2 For a sequence {an }. n→∞ an n→∞ n+1 Thus. (a) an = (−1)n .3.1 Find lim supn→∞ an and lim infn→∞ an for each sequence. n! When α = 0.5. In the general case. n→∞an Then limn→∞ an = ∞.1. we can also show that limn→∞ an = 0 by considering limn→∞ |an | and using Exercise 2. limn→∞ an = 0.54 2.  Example 2. prove that: (a) lim infn→∞ an = ∞ if and only if limn→∞ an = ∞.1 Given a real number α.

δ ) = (a − δ . δ ). Indeed. (b) The union of any collection of open subsets of R is open. Let us show that C is not open. δ ) = (a − δ . Proof: The proof of (a) is straightforward. there exists δ > 0 such that B(c. Therefore. (2) The sets A = (−∞. d) is open.6.6.1 (1) Any open interval A = (c. Definition 2. c + δ ) ⊂ C.6 OPEN SETS. there exists δ > 0 such that B(a.1 The following hold: (a) The subsets 0/ and R are open. ∞) are open. However. CLOSED SETS. Assume by contradiction that C is open. (c) The intersection of a finite number of open subsets of R is open. δ ) ⊂ A. ∞) is not open. a + δ ). A is open. That means Gα is open for every α ∈ I. COMPACT SETS. one has c < a < d. d − a}.1 A subset A of R is said to be open if for each a ∈ A. but the set C = [c. AND LIMIT POINTS The open ball in R with center a ∈ R and radius δ > 0 is the set B(a. δ ) ⊂ G because Gα0 ⊂ G. a + δ ) ⊂ A. Theorem 2. Thus. Then there exists α0 ∈ I such that a ∈ Gα0 . δ ) ⊂ Gα0 .  Example 2. for the element c ∈ C. Let δ = min{a − c. there exists δ > 0 such that B(a. Then B(a. δ ) = (c − δ . Let us show that the set [ G= Gα α∈I is open. c) and B = (c. The reader can easily verify that A and B are open. . G is open. (b) Suppose {Gα : α ∈ I} is an arbitrary collection of open subsets of R. 55 2. Then. Take any a ∈ G. for each a ∈ A. This implies B(a.6. but c − δ /2 ∈ / C. Since Gα0 is open. this is a contradiction because c − δ /2 ∈ B(c.

it follows that a ∈ A. (b). .3 A subset A of R is closed if and only if for any sequence {an } in A that converges to a point a ∈ R. ε) = (a − ε.6 OPEN SETS. there exists δi > 0 such that B(a. ∞). (−∞. CLOSED SETS. Theorem 2. . S is closed. . (−∞. Sc = R \S. b) ∪ (b. Sc is open because it is a union of opens sets in R (Theorem 2.2 The following hold: (a) The sets 0/ and R are closed. is open. Let δ = min{δi : i = 1. b]c = (−∞.2 A subset S of R is called closed if its complement. are open subsets of R. (c) The union of a finite number of closed subsets of R is closed.6. .  Example 2. n. Proof: The proofs for these are simple using the De Morgan’s law. a]. .6. (b) The intersection of any collection of closed subsets of R is closed. . Theorem 2. say.3 It follows from part (c) and Example 2. Take any a ∈ G. AND LIMIT POINTS (c) Suppose Gi . Let {Sα : α ∈ I} be a collection of closed sets. {b}c = (−∞. δ ) ⊂ G. Then a ∈ Gi for i = 1.6. n. Since A is closed. ∞)c = (−∞. . This implies aN ∈ Ac . Indeed. G is open. b]c is open by Theorem 2.6. COMPACT SETS. ∞). b]. there exists N ∈ N such that a − ε < aN < a + ε. Also.2 The sets [a.2 that any finite set is closed. ∞) and [a.6. . Let us prove. and [a. a) ∪ (b. there exists ε > 0 such that B(a.6.6. Thus. i = 1. Let us show that the set n \ G= Gi i=1 is also open. We will prove that the set \ S= Sα α∈I is also closed. a + ε) ⊂ Ac . δi ) ⊂ Gi . Then δ > 0 and B(a. Proof: Suppose A is a closed subset of R and {an } is a sequence in A that converges to a. . a) which are open by Example 2. We have !c \ [ Sc = Sα = Sαc .1. [a. a contradiction. Since [a. for instance. Therefore. α∈I α∈I Thus.  Definition 2. Suppose by contradiction that a 6∈ A. n}. Since {an } converges to a. . single element sets are closed since. a]c = (a.6. Since each Gi is open.56 2.1. .1(b)). .6. .   Example 2. ∞) are closed.

By the Bolzano-Weierstrass theorem (Theorem 2. Since A is compact. then max A exists. 1n ) ∩ A. b ∈ R. Let {an } be a sequence in A that converges to a point a ∈ R. we can obtain a convergent subsequence {ank }. So there exists δ > 0 such that (m − δ . limk→∞ |ank | = ∞. that a ≤ s ≤ b and. By the definition of compactness. Say. it 1 This definition of compactness is more commonly referred to as sequential compactness. It is clear that the sequence {an } is in A and it is convergent to a (because |an − a| < n1 . Therefore. Therefore.5.5 A subset A of R is compact if and only if it is closed and bounded. This is a contradiction since a ∈/ A. {an } has a subsequence {ank } that converges to b ∈ A.1. Then Ac is not open. / In particular. This contradicts the fact that m is the least upper bound of A (see Proposition 1.6. b] is compact. there exists an ∈ B(a. hence. A is closed by Theorem 2. if A is a nonempty subset of R that is closed and bounded below. Then sup A exists. Since Ac is not open. Theorem 2. We conclude that A is compact.5. This means there exists no a ∈ A with m − δ < a ≤ m. we will show that m ∈ A. there exists a subsequence {ank } that converges to a point a ∈ A. max A exists. one has B(a.4. Let us now show that A is closed. Then a = b ∈ A and. then the sequence is bounded. We now must show that s ∈ A. Since A is bounded.1  Example 2. To complete the proof. the sequence is bounded and.3 A subset A of R is called compact if for every sequence {an } in A. Since a ≤ an ≤ b for all n. 57 Let us now prove the converse. Similarly. Proof: Suppose A is a compact subset of R. Suppose by contradiction that A is not closed.  Theorem 2. . Let us first show that A is bounded. for all n ∈ N).3. This is a contradiction because {|ank |} converges to |a|. limk→∞ ank = s.6. Thus A is bounded. Then m ∈ Ac . We show that the set A = [a. hence.6. Therefore.6.4 If A is a nonempty subset of R that is closed and bounded above.1).4 Let a. A is closed. suppose A is closed and bounded and let {an } be a sequence in A. a ≤ b. m + δ ) ⊂ Ac . by the Bolzano-Weierstrass theorem (Theorem 2. Then for every n ∈ N.1). there exists an ∈ A such that |an | ≥ n. Suppose. Assume by contradiction that m ∈ / A.1).  Definition 2. that A is not bounded.6. for such an a and for each n ∈ N. For the converse.4. Let {an } be a sequence in A. s ∈ A as desired. which is an open set. Then |ank | ≥ nk ≥ k for all k. Let m = sup A. then min A exists Proof: Let A be a nonempty closed set that is bounded above. there exists a subsequence {ank } that converges to some a ∈ A. by contradiction. it follows from Theorem 2. ε) ∩ A 6= 0. Since a ≤ ank ≤ b for all k. there exists a ∈ Ac such that for any ε > 0.

δ ) contains an infinite number of points of G \ {a}. δ ) ∩ D ⊂ V. any point of the interval [0. CLOSED SETS. Then A does not have any limit points. δ ) for k ≥ K. Definition 2. by the Bolzano-Weierstrass theorem (Theorem 2. a) ∪ (a. let δ > 0 be such that B(a. Since the set {ank : k ≥ K} is infinite.5 Let D be a subset of R. We say that a subset V of D is open in D if for every a ∈ V . there exists K ∈ N such that ank ∈ B(b.6. Proof: Let A be an infinite subset of R and let {an } be a sequence of A such that am 6= an for m 6= n (see Theorem 1.6.6 Any infinite bounded subset of R has at least one limit point. 1] is a limit point of A. AND LIMIT POINTS has a convergent subsequence. Proof: Suppose V is open in D. it has a convergent subsequence {ank }. Say.6. The following theorem is a variation of the Bolzano-Weierstrass theorem.58 2.4.  The following definitions and results provide the framework for discussing convergence within subsets of R.4 (cluster/limit/accumulation point). COMPACT SETS. Then a = 0 is a limit point of A and b = 1 is also a limit point of A. there exists δa > 0 such that B(a. In fact.6. 1) has no isolated points. Then a = 0 is the only limit point of A.7). All elements of A are isolated points. Let A be a subset of R. If G is open and a ∈ G. thus B(a. Every element of Z is an isolated point of Z. {ank }. then a is a limit point of G \ {a}.6 If G is an open subset of R then every point of G is a limit point of G.3 that a ∈ A.6. Theorem 2. δ ) ⊂ G. By definition. Indeed. a + δ ) and. Then (G \ {a}) ∩ B(a. A subset V of D is open in D if and only if there exists an open subset G of R such that V = D ∩ G. δ ) = (a − δ .  Example 2.  Definition 2. it follows that b is a limit point of A.6 OPEN SETS. (3) Let A = {1/n : n ∈ N}. δ ) contains an infinite number of points of A. It now follows from Theorem 2. limk→∞ ank = a. A point a ∈ A which is not an accumulation point of A is called an isolated point of A.2. This shows that A is compact as desired. 1). Given δ > 0. for every a ∈ V .6. Theorem 2. Since {an } is bounded.7 Let D be a subset of R. δa ) ∩ D ⊂ V.5 (1) Let A = [0.1). (2) Let A = Z. The set [0. Set b = limk→∞ ank . more is true. there exists δ > 0 such that B(a. A point a ∈ R (not necessarily in A) is called a limit point of A if for any δ > 0. δa ) .  Example 2. the open ball B(a.6. In fact. Define G = ∪a∈V B(a.

6. Theorem 2. 59 Then G is a union of open subsets of R. Then D \ K = D \ (D ∩ F) = D \ F = D ∩ F c .   Example 2. The following theorem is now a direct consequence of Theorems 2. It follows that K = D \ (D \ K) = D \ (D ∩ G) = D \ G = D ∩ Gc .7. (c) The intersection of a finite number of open sets in D is open in D. we conclude from Theorem 2. We say that a subset A of D is closed in D if D \ A is open in D. 12 ) is open in R.6 Let D be a subset of R.7 and 2.9 Let D be a subset of R. V ⊂ G ∩ D = ∪a∈V [B(a. Suppose V = G ∩ D. (b) The union of any collection of open sets in D is open in D. Moreover. applying Theorem 2. suppose that there exists a closed subset F of R such that K = D ∩ F. The proof is now complete. 12 ).  . so G is open. Then F is a closed subset of R and K = D ∩ F. Therefore. Therefore.6. V = G ∩ D. Let us now prove the converse.7 Let D = [0. 21 ). where G is an open set.7 again. A subset K of D is closed in D if and only if there exists a closed subset F of R such that K = D ∩ F. Since (−1. Then D \ K is open in D. Let F = Gc . Notice that V itself is not an open subset of R. δ ) ∩ D ⊂ G ∩ D = V.6. We can write V = D ∩ (−1. we have a ∈ G. so there exists δ > 0 such that B(a.6. Conversely. For any a ∈ V .1. there exists an open set G such that D \ K = D ∩ G. K is closed in D by definition. Since F c is an open subset of R. 1) and V = [0.6. δa ) ∩ D] ⊂ V.6. Proof: Suppose K is a closed set in D. The following hold: (a) The subsets 0/ and D are open in D. one has that D \ K is open in D. Definition 2. δ ) ⊂ G. By Theorem 2.7 that V is open in D.6.6. Theorem 2.8 Let D be a subset of R.6. It follows that B(a.

1). Let {xk } be a sequence in K that converges to a point x̄ ∈ D. for every k ∈ N.9 Consider the set D = [0. . Corollary 2. Exercises 2. Proof: Let D be a subset of R. 1) is neither open nor closed.6.8 that A is closed in D. COMPACT SETS.9. We conclude that K is closed in D.6.9 and 2.11 Let D be a subset of R. 1) and K = [ 12 .2 Prove that the interval [0. Notice that K itself is not a closed subset of R.9 that K is closed in D. A subset K of D is closed in D if and only if for every sequence {xk } in K that converges to a point x̄ ∈ D it follows that x̄ ∈ K. Then {xk } is a sequence in K and.1 Prove that a subset A of R is open if and only if for any x ∈ A. Then an ∈ A for all n. (c) The union of a finite number of closed sets in D is closed in D.6. However. one has B(x̄. We can write K = D ∩ [ 12 . then A ∪ B is a compact set. it follows that A is not compact. (b) The intersection of any collection of closed sets in D is closed in D. By Theorem 2. hence.   1 B x̄. {xk } converges to x̄ ∈ D. 1 Indeed. 2. moreover. δ ) ∩ D * D \ K. x̄ ∈ D ∩ F = K. Suppose by contradiction that K is not closed in D or D \ K is not open in D. Let us prove the converse. 2].6. Since [ 12 . In particular. AND LIMIT POINTS  Example 2. 1k ) ∩ D such that xk ∈ / D \ K. Clearly. We showed in Example 2. every subsequence also converges to 1.3 I Prove that if A and B are compact subsets of R.60 2.2. This is a contradiction. CLOSED SETS.6. ∩ D * D \ K. x + 1/n) ⊂ A. {an } converges to 1 and. A is bounded.6.  The following theorem is a direct consequence of Theorems 2.6. 2] is closed in R. x̄ ∈ F.6 OPEN SETS.10 Let D be a subset of R. Since {xk } is also a sequence in F and F is a closed subset of R.6. Suppose K is closed in D. The following hold: (a) The subsets 0/ and D are closed in D. Thus. We show this by finding a sequence {an } in A for which no subsequence converges to a point in A. choose xk ∈ B(x̄.6. k For each k ∈ N. Then x̄ ∈ K.  Example 2. Theorem 2. there exists n ∈ N such that (x − 1/n. consider the sequence an = 1 − 2n for n ∈ N. 1). A is not compact. we conclude from Theorem 2. 1) and the subset A = [ 12 .6. Then there exists x̄ ∈ D \ K such that for every δ > 0.6.8 Let D = [0.6. Since 1 6∈ A. Moreover. 2. there exists a closed subset F of R such that K = D ∩ F.

closed in D. 1).6. (a) A = (0. neither or both. 1).6. 61 2. 2. (b) B = [0. (d) D = {m + 1/n : m. (c) C = Q.6. 1]. Justify your answers. (b) B = N. Classify each subset of D below as open in D.4 Prove that the intersection of any collection of compact subsets of R is compact. . ∞).5 Find all limit points and all isolated points of each of the following sets: (a) A = (0.6 Let D = [0. n ∈ N}. 1). ∞). (d) D = (−1. (c) C = Q ∩ D. (e) E = (−2. 2.

.

we write lim f (x) = `. if |x − 1| < δ and x ∈ [0. We say that f has a limit at x̄ if there exists a real number ` such that for every ε > 0.1. we have ε | f (x) − `| = |x2 + x − 2| = |x − 1||x + 2| < 3δ = 3 = ε. x→x̄ Remark 3.1 Note that the limit point x̄ in the definition of limit may or may not be an element of the domain D.1. |x + 2| ≤ |x| + 2 ≤ 3. 3 This shows that limx→1 f (x) = 2. This suggests the choice δ = ε/5. LIMITS AND CONTINUITY In this chapter. if |x − 2| < δ we have | f (x) − 2| = 5|x − 2| < 5δ = ε. First note that | f (x) − 2| = |5x − 7 − 3| = |5x − 10| = 5|x − 2|. Then. the inequality | f (x) − `| < ε need only be satisfied by elements of D.2 Let f : [0. given ε > 0.1.1 Let f : R → R be given by f (x) = 5x − 7.1 Let f : D → R and let x̄ be a limit point of D. Let x̄ = 1 and ` = 2. First note that | f (x) − `| = |x2 + x − 2| = |x − 1||x + 2| and for x ∈ [0. Let ε > 0. there exists δ > 0 with | f (x) − `| < ε for all x ∈ D for which 0 < |x − x̄| < δ . Then. In this case.1 LIMITS OF FUNCTIONS Definition 3.  Example 3. we extend our analysis of limit processes to functions and give the precise definition of continuous function. . We prove that limx→2 f (x) = 3. We derive rigorously two fundamental theorems about continuous functions: the extreme value theorem and the intermediate value theorem. 1) → R be given by f (x) = x2 + x. LIMITS OF FUNCTIONS LIMIT THEOREMS CONTINUITY PROPERTIES OF CONTINUOUS FUNCTIONS UNIFORM CONTINUITY LIMIT SUPERIOR AND LIMIT INFERIOR OF FUNCTIONS LOWER SEMICONTINUITY AND UPPER SEMICONTINUITY 3. Now.  Example 3. In any case. 1). 3. choose δ = ε/3.1. 1).

2 .2. However. that |x| < 3 and. we impose the condition δ ≤ 1. We show that limx→2 f (x) = 4.1.1. for such x. If |x − 2| < 1. whenver |x − 2| < δ we get | f (x) − 4| = |x − 2|x + 2| ≤ |x − 2|(|x| + 2) < δ 5 ≤ ε. It follows.64 3. We prove that limx→1 f (x) = − . Since the domain is all of R the expression |x + 2| is not bounded and we cannot proceed as in Example 3. Then. then −1 < x − 2 < 1 and. Now. First note that | f (x) − 4| = |x2 − 4| = |(x − 2)(x + 2)| = |x − 2||x + 2|.1 LIMITS OF FUNCTIONS  Example 3. hence |x| + 2 < 5. so.1. ε5 }. thus.3 Let f : R → R be given by f (x) = x2 . we are interested only in values of x close to 2 and. First x +3 2 we look at the expression | f (x) − (− 12 )| and try to identify a factor |x − 1| (because here x̄ = 1). 3x − 5 1  Example 3. given ε > 0 we choose δ = min{1.4 Let f : R → R be given by f (x) = 2 . 1 < x < 3.

 .

.

.

.

.

.

f (x) − − 1 .

= .

3x − 5 + 1 .

= .

6x − 10 + x + 3 .

= |x − 1||x + 7| ≤ 1 |x − 1||x + 7|. .

.

.

.

.

.

.

2 .

.

x2 + 3 2 .

.

x2 + 3 .

It follows that if |x − 1| < δ we get . 2(x2 + 3) 6 Proceeding as in the previous example. Thus |x| < 2 and |x + 7| ≤ |x| + 2 < 9. we choose δ = min{1. 0 < x < 2. Now. 32 ε}. given ε > 0. if |x − 1| < 1 we get −1 < x − 1 < 1 and. so.

.

.

f (x) − (− ).

≤ |x + 7| |x − 1| < 9 δ ≤ ε. 1 .

.

.

2 .

Then there exists N ∈ N with 0 < |xn − x̄| < δ for all n ≥ N. there exists xn ∈ D with 1 0 < |xn − x̄| < and | f (xn ) − `| ≥ ε0 . there exists δ > 0 such that | f (x) − `| < ε whenever x ∈ D and 0 < |x − x̄| < δ . suppose (3.  .2 — Sequential Characterization of Limits. the sequence {xn } n converges to x̄. It follows that (3. Conversely. Then lim f (x) = ` (3. Let {xn } be a sequence in D with xn 6= x̄ for every n and such that {xn } converges to x̄. By the squeeze theorem (Theorem 2. we have | f (xn ) − `| < ε. xn 6= x̄ for every n. Theorem 3. there exists x ∈ D with 0 < |x − x̄| < δ and | f (x) − `| ≥ ε0 .2). Then there exists ε0 > 0 such that for every δ > 0.2) implies (3. Thus. 6 6 The following theorem will let us apply our earlier results on limits of sequences to obtain new results on limits of functions. For such n. Given any ε > 0. Let f : D → R and let x̄ be a limit point of D.1) holds.1.2) is false. Proof: Suppose (3. Moreover. for every n ∈ N.1) is false. This shows that (3.2) n→∞ for every sequence {xn } in D such that xn 6= x̄ for every n and {xn } converges to x̄.1) x→x̄ if and only if lim f (xn ) = ` (3. This implies (3.1.6).1) and the proof is complete.

 Example 3. Then limx→x̄ f (x) does not exist for any x̄ ∈ R. {sn } converging to x̄ such that rn ∈ Q and sn 6∈ Q for all n ∈ N. and { f (xn )} does not converge.1. Let {xn } be a sequence in D \ {x̄} that converges to x̄. fix x̄ ∈ R and choose two sequences {rn }. Then there exists δ > 0 such that f (x) < g(x) for all x ∈ B(x̄.4 Let f : D → R and let x̄ be a limit point of D. Suppose lim f (x) = `1 . n→∞ n→∞ Since f (xn ) ≤ g(xn ) for all n ∈ N. if n = 2k. lim g(x) = `2 . It is clear that {xn } converges to x̄. x 6= x̄.1. g : D → R and let x̄ be a limit point of D.1. Suppose that lim f (x) = `1 . f (x) = 0.1. Corollary 3.1.5 Consider the Dirichlet function f : R → R given by ( 1. g : D → R and let x̄ be a limit point of D. Define a new sequence {xn } by ( rk . x 6= x̄.2. if x ∈ Q.3. Moreover. It follows from the sequential characterization of limits that limx→x̄ f (x) does not exist. x→x̄ x→x̄ and that there exists δ > 0 such that f (x) ≤ g(x) for all x ∈ B(x̄.5. Then `1 ≤ `2 . the sequence { f (xn )} converges to two different limits `1 and `2 .1. x̄ + δ ) ∩ D that converges to x̄ and xn 6= x̄ for all n. . lim f (xn ) = `1 and lim g(xn ) = `2 .9 implies that the sequence { f (xn )} does not converge. Proof: Let {xn } be a sequence in B(x̄. 65 Corollary 3. δ ) ∩ D = (x̄ − δ . By Theorem 3.1. if x ∈ Qc . if n = 2k − 1.1.2. x→x̄ x→x̄ and `1 < `2 . then this limit is unique. δ ) ∩ D.  The following corollary follows directly from Theorem 3. By Theorem 3. lim g(x) = `2 . Then f does not have a limit at x̄ if and only if there exists a sequence {xn } in D such that xn 6= x̄ for every n. This is a contradiction to Theorem 2. Proof: Suppose by contradiction that f has two different limits `1 and `2 . since { f (rn )} converges to 1 and { f (sn )} converges to 0.6 Let f .1. If f has a limit at x̄.3 Let f : D → R and let x̄ be a limit point of D. we obtain `1 ≤ `2 .1. Theorem 2.5 Let f .2. xn = sk . Theorem 3. Indeed.  Theorem 3.1. applying Theorem 2. δ ) ∩ D. {xn } converges to x̄.

If limx→x̄ f (x) = limx→x̄ h(x) = `. Suppose there exists δ > 0 such that f (x) ≤ g(x) ≤ h(x) for all x ∈ B(x̄.66 3. . √x+1 (d) limx→0 x = 0.1.7 Let f . x 6= x̄. f (x) < `1 + ε < `2 − ε < g(x) for all x ∈ B(x̄.1. δ ) ∩ D. 3.1 Use the definition of limit to prove that (a) limx→2 3x − 7 = −1. x 6= x̄.1. (b) limx→3 (x2 + 1) = 10. Proof: The proof is straightforward using Theorem 2.1. Exercises 3. δ ) ∩ D. Prove that if limx→x̄ f (x) = `.1. x→x̄ Give an example to show that the converse is not true in general.8 We will adopt the following convention. ∞).  Theorem 3. then f (x) results in a real number. such that ε < 2 ). (e) limx→2 x3 = 8. |x| (b) limx→0 cos(1/x). The proof is now complete. g. For example.1. Then there exists δ > 0 such that `1 − ε < f (x) < `1 + ε and `2 − ε < g(x) < `2 + ε for all x ∈ B(x̄. x 6= x̄.2 Prove that the following limits do not exist. in 1 lim x→2 x + 3 we assume D = R \{−3} and in √ lim x x→1 we assume D = [0. x+3 (c) limx→1 = 2. When we write limx→x̄ f (x) without speci- fying the domain D of f we will assume that D is the largest subset of R such that if x ∈ D.1. Thus. h : D → R and let x̄ be a limit point of D. 3.6 and Theorem 3.3 Let f : D → R and let x̄ be a limit point of D. then limx→x̄ g(x) = `. x (a) limx→0 . δ ) ∩ D.1 LIMITS OF FUNCTIONS `2 −`1 Proof: Choose ε > 0 such that `1 + ε < `2 − ε (equivalently. then lim | f (x)| = |`|.  Remark 3.2.

Prove that if limx→x̄ f (x) = `. (b) limx→x̄ ( f g)(x) = `m. Suppose f (x) ≥ 0 for all x ∈ D. Definition 3. Suppose x̄ is a limit point of D and lim f (x) = `. (a) limx→1/2 f (x). g : D → R and let c ∈ R.5 Find limx→0 x sin(1/x). e Theorem 3. The functions f + g. lim g(x) = m. The function f is defined as a function from D for x ∈ D.2 LIMIT THEOREMS Here we state and prove various theorems that facilitate the computation of general limits. (c f )(x) = c f (x) e = {x ∈ D : g(x) 6= 0}. and c f are respectively defined as functions from D to R by ( f + g)(x) = f (x) + g(x). 3. 1]. 1].2. x→x̄ 3. 3. if x ∈ Q ∩ [0. if x ∈ Qc ∩ [0.1. (c) limx→x̄ (c f )(x) = c`.1.1.1 Let f . For those that exist find their values. x→x̄ x→x̄ Then (a) limx→x̄ ( f + g)(x) = ` + m. 67 3. f (x) = 1 − x. f g.1 Let f . Determine which of the following limits exist. ( f g)(x) = f (x)g(x). Let D e to R by g   f f (x) (x) = g g(x) for x ∈ D. g : D → R and let c be a constant. (b) limx→0 f (x). then p √ lim f (x) = `.4 Let f : D → R and let x̄ be a limit point of D.   f ` (d) limx→x̄ (x) = provided that m 6= 0.2. (c) limx→1 f (x).6 I Let f be the function given by ( x. g m .

Then.2.1. e Since x̄ is a limit point of D.2. x→0 x2 + 7 limx→0 x2 + limx→0 7 0+7 7  Example 3. we first simplify the expression keeping in mind that in the definition of limit we never need to evaluate the expression at the limit point itself.2. this means we may assume that x 6= −1. n→∞ Applying Theorem 3. The proofs of (b) and (c) are similar. limx→−2 x + limx→−2 7 −2 + 7 5  Example 3.1. we get limx→x̄ ( f + g)(x) = ` + m.1 Consider f : R \ {−7} → R given by f (x) = .2 LIMIT THEOREMS Proof: Let us first prove (a).2 We proceed in the same way to compute the following limit.1. This implies x∈D e whenever 0 < |x − x̄| < δ . combining all x+7 parts of Theorem 3.2.  x2 + 2x − 3  Example 3. then x̄ is a limit point of D. x→−1 x+1 x→−1 .2. For any such x we have x2 + 6x + 5 (x + 1)(x + 5) = = x + 5.1.3 We now consider x2 + 6x + 5 lim . there is a sequence {uk } in D converging to x̄ such that uk 6= x̄ for every k. Let {xn } be a sequence in D that converges to x̄ and xn 6= x̄ for every n. and hence x̄ is a limit point of D. n→∞ n→∞ It follows from Theorem 2.2 again. x ∈ D. x ∈ D. Since m 6= 0. 1 + (2x − 1)2 limx→0 1 + limx→0 (2x − 1)2 1 + 1 2 lim = = = . x→−1 x+1 Since the limit of the denominator is 0 we cannot apply directly part (d) of Theorem 3. it follows from an easy application of Theorem 3. x2 + 6x + 5 lim = lim x + 5 = 4. Instead.2.1. lim f (xn ) = ` and lim g(xn ) = m. e The rest of the proof of (d) can be completed easily following the proof of (a).6 that there exists δ > 0 with g(x) 6= 0 whenever 0 < |x − x̄| < δ . Then uk ∈ D e for all k sufficiently large.2.68 3. In this case. we get limx→−2 (x2 + 2x − 3) limx→−2 x2 + limx→−2 2x − limx→−2 3 lim f (x) = = x→−2 limx→−2 (x + 7) limx→−2 x + limx→−2 7 (limx→−2 x)2 + 2 limx→−2 x − limx→−2 3 (−2)2 + 2(−2) − 3 3 = = =− . x+1 x+1 Therefore.1 that lim ( f (xn ) + g(xn )) = ` + m. Let us now show that if m 6= 0. By Theorem 3.

This implies | f (un ) − f (um )| < ε for all m.3). a) and B+ (a. s ∈ D and 0 < |r − x̄| < δ . Similarly. Given ε > 0.1. δ ) contains an infinite number of elements of A. Given ε > 0. Given a subset A of R. there exists `0 ∈ R such that lim f (xn ) = `0 . 2 Thus. we say that a is a left limit point of A if for any δ > 0. B− (a. n→∞ Fix any ε > 0 and let δ > 0 satisfy (3. It now follows from Theorem 3. a is called a right limit point of A if for any δ > 0. Let us prove the converse. δ ) = (a. there exists δ > 0 such that | f (r) − f (s)| < ε whenever r.2. n→∞ We now prove that f has limit ` at x̄ using Theorem 3. there exists δ > 0 such that ε | f (x) − `| < whenever x ∈ D and 0 < |x − x̄| < δ . Then | f (un ) − f (xn )| < ε for such n. for r.2 (Cauchy’s criterion) Let f : D → R and let x̄ be a limit point of D. By the previous argument. 0 < |s − x̄| < δ . s ∈ D and 0 < |r − x̄| < δ . and hence there exists ` ∈ R such that lim f (un ) = `. Thus. 69 Theorem 3. Fix a sequence {un } in D such with limn→∞ un = x̄ and un 6= x̄ for every n.3) Proof: Suppose limx→x̄ f (x) = `. Thus. Then f has a limit at x̄ if and only if for any ε > 0.1. Definition 3. B+ (a.2 that limx→x̄ f (x) = `. { f (un )} is a Cauchy sequence. s ∈ D with 0 < |r − x̄| < δ and 0 < |s − x̄| < δ . There exists K ∈ N such that |un − x̄| < δ and |xn − x̄| < δ for all n ≥ K. n ≥ N. (3.2. Letting n → ∞. there exists δ > 0 such that | f (r) − f (s)| < ε whenever r. δ ) = (a − δ . δ ) contains an infinite number of elements of A. Define B− (a. a + δ ). Then there exists N ∈ N satisfying 0 < |un − x̄| < δ for all n ≥ N.2.  The rest of this section discussed some special limits and their properties. Let {xn } be a sequence in D such that limn→∞ xn = x̄ and xn 6= x̄ for every n. 0 < |s − x̄| < δ . we have |` − `0 | ≤ ε. ` = `0 since ε is arbitrary.2 Let a ∈ R and δ > 0. we have | f (r) − f (s)| ≤ | f (r) − `| + |` − f (s)| < ε. .

lim f (x) = lim− −1 = −1. if x ≥ −1.5 Consider the function f : R → R given by ( x + 4. We have lim f (x) = lim + x2 − 1 = 0.2. x→−1− x→−1 The following theorem follows directly from the definition of one-sided limits. Note first that 0 is a limit point of the set D = R \ {0} → R.2. Theorem 3. f (x) = 2 (3. The right-hand limit of f at x̄ can be defined in a similar way and is denoted limx→x̄+ f (x). we have f (x) = x/x = 1. x→x̄+ x→0 Similarly. for x < 0 we have f (x) = −x/x = −1.3 Let f : D → R and let x̄ be both a left limit point of D and a right limit point of D. Definition 3.4 Consider the function f : R \ {0} → R given by |x| f (x) = . x→x̄− x→0  Example 3. Then lim f (x) = ` x→x̄ if and only if lim f (x) = ` and lim− f (x) = `. there exists δ > 0 such that | f (x) − `| < ε for all x ∈ B− (x̄. Since. We say that ` is the left-hand limit of f at x̄.2.4) x − 1.3 (One-sided limits) Let f : D → R and let x̄ be a left limit point of D. Therefore. x→x̄+ x→x̄ . δ ). x Let x̄ = 0.70 3.2 LIMIT THEOREMS It follows from the definition that a is a limit point of A if and only if it is a left limit point of A or it is a right limit point of A. x→−1+ x→−1 and lim f (x) = lim − x + 4 = 3.2. We write lim f (x) = ` x→x̄− if for every ε > 0. we have lim f (x) = lim+ 1 = 1.  Example 3. for x > 0. if x < −1.

δ ) ∪ B+ (x̄. δ ) ∩ D. x1 < x2 implies f (x1 ) ≥ f (x2 ). we get ` − ε < f (x1 ) ≤ f (x) ≤ ` < ` + ε for all x ∈ (x1 .4 (monotonicity) Let f : (a. sup f (x) = lim− f (x) ≤ f (x̄) ≤ lim+ f (x) = inf f (x). we write lim f (x) = −∞ x→x̄ if for every M ∈ R. b). (1) We say that f is increasing on (a. We write lim f (x) = ∞ x→x̄ if for every M ∈ R. the set { f (x) : x ∈ (a. x̄). ` = sup f (x) a<x<x̄ is a real number. For any ε > 0. limx→x̄− f (x) = `. (2) We say that f is decreasing on (a. x2 ∈ (a.5 (infinite limits) Let f : D → R and let x̄ be a limit point of D. there exists δ > 0 such that f (x) < M for all x ∈ B0 (x̄. Strict mono- tonicity can be defined similarly using strict inequalities: f (x1 ) < f (x2 ) in (1) and f (x1 ) > f (x2 ) in (2).4 that limx→0 x does not exists. there exists a < x1 < x̄ such that ` − ε < f (x1 ). for all x1 . δ ) = (x̄ − δ . Then limx→x̄− f (x) and limx→x̄+ f (x) exist.2. Thus.2. b). since the one-sided limits do not agree. Moreover. b) and x̄ ∈ (a. b) if.2. b).  Let B0 (x̄. . x̄) = B− (x̄.4 Suppose f : (a. Theorem 3. Using the increasing monotonicity. δ ) ∩ D. b). b) → R. for all x1 . x2 ∈ (a. 71 |x|  Example 3. We will show that limx→x̄− f (x) = `. there exists δ > 0 such that f (x) > M for all x ∈ B0 (x̄. x̄)} is nonempty and bounded above. Similarly. x̄ + δ ) \ {x̄}.6 It follows from Example 3. Let δ = x̄ − x1 > 0. b) → R is increasing on (a. Definition 3.2. a<x<x̄ x→x̄ x→x̄ x̄<x<b Proof: Since f (x) ≤ f (x̄) for all x ∈ (a. by the definition of the least upper bound. If f is increasing or decreasing on (a. Therefore. x1 < x2 implies f (x1 ) ≤ f (x2 ). Definition 3. b) if. The rest of the proof of the theorem is similar. δ ) = B− (x̄. we say that f is monotone on this interval. δ ).2.

choose δ such that 0 < δ < |M|+1 . We want to identify c so that . where D is not bounded below. note that |M| + 1 > M. Let f : D → R.3. Then.7 We show that limx→1 (x−1)2 = ∞ directly from Definition 3. there exists c ∈ R such that | f (x) − `| < ε for all x > c.2.8 We prove from the definition that 3x2 + x 3 lim = . x ∈ D. is equivalent to q 1 |M|+1 > |x − 1|.5. Next note that (x−1)2 > |M| + 1. x ∈ D.2. 1  Example 3. 1 Let M ∈ R. We write lim f (x) = ` x→−∞ if for every ε > 0. Let ε > 0. we work backwards from (x−1)2 > M to an inequality for |x − 1|.2. if 0 < |x − 1| < δ we have 1 1 1 2 > 2> 1 = |M| + 1 > M.2 and Theorem 2. with the difference that x need not be an integer. q 1 Now. x→−∞ 2x2 + 1 2 The approach is similar to that for sequences. We write lim f (x) = ` x→∞ if for every ε > 0.6).  Example 3. 1 As in the case of finite limits.2 LIMIT THEOREMS Infinite limits of functions have similar properties to those of sequences from Chapter 2 (see Definition 2. We want to find δ > 0 that will guarantee (x−1)2 > M whenever 0 < |x − 1| < δ .72 3. there exists c ∈ R such that | f (x) − `| < ε for all x < c. (x − 1) δ |M|+1 as desired. Definition 3. where D is not bounded above.3. We can also define lim f (x) = ±∞ and lim f (x) = ±∞ x→∞ x→−∞ in a similar way. 1 To simplify calculations.6 (limits at infinity) Let f : D → R.2.

2 .

.

3x + x 3 .

.

2x2 + 1 − 2 .

(3. < ε.5) .

.

. for all x < c.

73 .

2 .

.

3x +x 3 .

. |2x−3| Now.

2x2 +1 − 2 .

we 2 +1) have 3/(4ε) < −x = |x|. 3. 3. = 2(2x2 +1) . Therefore. if x < c. Then. (3. We set c < min{0. − 3/(4ε)}.6 will be guaranteed p p if 1/ε < 4x2 /3 or. equivalently 3/(4ε) < |x|. Then. so |2x − 3| > 3. . 1/ε < 2(2x p |2x−3| and. simplifying.5 is equivalent to 1 2(2x2 + 1) < . hence. since 4x3 < 2(2x |2x−3| . Thus.6) ε |2x − 3| 2 +1) 2 We first restrict x to be less than 0.

2 .

.

3x + x 3 .

|2x − 3| .

2x2 + 1 − 2 .

. = 2(2x2 + 1) < ε.

.

y ∈ D \ {x̄}.2. where k ≥ 0 is a constant.2.2. Exercises 3. Suppose that | f (x) − f (y)| ≤ k|x − y| for all x. x −1 xm − 1 (b) limx→1 n .5 I Let f : D → R and let x̄ be a limit point of D.2. 3. then lim[ f (x)]n = [lim f (x)]n . for any n ∈ N . . x−3 x2 + 4x + 3 (b) limx→−3 x2 − 9 3. m. n ∈ N.2.2 Let f : D → R and let x̄ is a limit point of D.2. Prove that if limx→x̄ f (x) exists.3 Find the following limits: √ x−1 (a) limx→1 2 . Prove that limx→x̄ f (x) exists. where m.4 Find the following limits: √ √ (a) limx→∞ ( 3 x3 + 3x2 − x2 + 1). n ∈ N. where [x] denotes the greatest integer that is less than or equal to x. x→x̄ x→x̄ 3. 3.1 Find the following limits: 3x2 − 2x + 5 (a) limx→2 . m x−1 √ √ x− 3 x (d) limx→1 . √ √ (b) limx→−∞ ( 3 x3 + 3x2 − x2 + 1). where m. n ≥ 2.6 Determine the one-sided limits limx→3+ [x] and limx→3− [x]. x −1 √n x−1 (c) limx→1 √ . x−1 3.

2.7 Find each of the following limits if they exist: x+1 (a) limx→1+ . .3 CONTINUITY 3. 74 3.

− x 1 .

(b) limx→0+ .

x3 sin(1/x).

if x ≤ 1.2.. ∞). Figure 3. b). Prove the following. (a) If f is bounded above. Find the value of a such that limx→1 f (x) exists.9 Determine all values of x̄ such that the limit limx→x̄ (1 + x − [x]) exists.3. 3. If f is continuous at every point x ∈ D. let f be the function given by ( x2 . then limx→b− f (x) exists and is a real number. or (−∞.1 Let D be a nonempty subset of R and let f : D → R be a function. (c) limx→1 (x − [x]). ∞).1: Definition of continuity. (b) If f is not bounded above. there exists δ > 0 such that if x ∈ D and |x − x0 | < δ . if x > 1. 3. .3 CONTINUITY Definition 3. (a. f (x) = ax − 1. then | f (x) − f (x0 )| < ε. 3.2.8 For a ∈ R. The function f is said to be continuous at x0 ∈ D if for any real number ε > 0. 3. then limx→b− f (x) = ∞. we say that f is continuous on D (or just continuous if no confusion occurs). b) → R is increasing.10 Let a. b ∈ R and suppose f : (a. State and prove analogous results in case f is bounded below and in case that the domain of f is one of (−∞.2.

Then f is continuous at x0 if and only if for any sequence {xk } in D that converges to x0 . The proofs of the next two theorems are straightforward using Theorem 3. x→x0 x→x0 it follows that f is continuous at x0 . We will use Theorem 3. by Theorem 3. δ ) ∩ D = {x0 }. This shows that f is continuous at x0 . Remark 3.1 Note that the above definition of continuity does not mention limits. This allows to include in the definition. If f is continuous at x0 and g is continuous at f (x0 ).3 again we conclude f + g is continuous at x0 .3. Then (a) f + g and f g are continuous at x0 .1 (a). If x0 is an isolated point of D.3.3. Let {xk } be a sequence in D that converges to x0 . we have the following theorem which follows directly from the definitions of continuity and limit. the sequence { f (xk )} converges to f (x0 ). Then if |x − x0 | < δ .3. | f (x) − f (x0 )| = 0 < ε for any epsilon. Theorem 3. Choose δ = ε/3.2.3. δ ) ∩ D. . Suppose f and g are continuous at x0 .3. every function is continuous at an isolated point of its domain. f (c) If g(x0 ) 6= 0.1. g Proof: We prove (a) and leave the other parts as an exercise.3.3.3 Let f : D → R and let x0 ∈ D. from the results of the previous theorem.2 Let f : D → R and let x0 ∈ D be a limit point of D.  Theorem 3.we get that { f (xk ) + g(xk )} converges to f (x0 ) + g(x0 ).3.3. 75  Example 3.3.5 Let f : D → R and let g : E → R with f (D) ⊂ E.3 we obtain that { f (xk )} converges to f (x0 ) and {g(xk )} converges to g(x0 ). we have lim f (x) = lim (3x2 − 2x + 1) = 3x02 − 2x0 + 1 = f (x0 ).3. Therefore. Theorem 3. then g ◦ f is continuous at x0 . points x0 ∈ D which are not limit points of D.1 Let f : R → R be given by f (x) = 3x + 7.2 and we leave its proof as an exercise. The following theorem follows directly from the definition of continuity. Since. then there is δ > 0 such that B(x0 . (b) c f is continuous at x0 for any constant c. Since f and g are continuous at x0 . Theorem 3. By Theorem 2. then (defined on D e = {x ∈ D : g(x) 6= 0}) is continuous at x0 . Theorem 3. Fix x0 ∈ R. using Theorem 3.2 and Theorem 3.3. Therefore.3. g : D → R and let x0 ∈ D. x→x0  Example 3.2 Let f : R → R be given by f (x) = 3x2 − 2x + 1. Then f is continuous at x0 if and only if lim f (x) = f (x0 ). To study continuity at limit points of D. k→∞ k→∞ Since {xk } was arbitrary. Let x0 ∈ R and let ε > 0. we have | f (x) − f (x0 )| = |3x + 7 − (3x0 + 7)| = |3(x − x0 )| = 3|x − x0 | < 3δ = ε. It follows that for x ∈ B(x0 .4 Let f . lim ( f + g)(xk ) = lim f (xk ) + g(xk ) = f (x0 ) + g(x0 ) = ( f + g)(x0 ).

76 3.3 CONTINUITY Exercises 3. . x 3. √ (c) f (x) = x on [0. b ∈ R.1. The domain of all the functions is R. (b) f (x) = x2 − 3 on R.3. on R.1 Prove.2 Determine the values of x at which each function is continuous. using definition 3. a.3. that each of the following functions is continuous on the given domain: (a) f (x) = ax + b.3. 1 (d) f (x) = on R \ {0}. ∞).

.

.

.

sin x .

.

if x 6= 0. . (a) f (x) = .

x .

1. n→∞ 2(1 + x2n ) 3.  (b) f (x) = |x| 1. if x = 0.  cos πx .   sin x . Find the value of a such that f is continuous. 3.3.3.7 Prove Theorem 3. (Hint: treat separately the cases when x0 is a limit point of D and when it is not.6 Prove parts (b) and (c) of Theorem 3.5 Prove Theorem 3.8 I Explore the continuity of the function f in each case below. Prove that if f is continuous at x0 .3. if x = 0.4.3 Let f : R → R be the function given by ( x2 + a.3. if x 6= 0.3. π (e) f (x) = lim sin . .4 Let f : D → R and let x0 ∈ D. (d) f (x) = 2 |x − 1|. if |x| ≤ 1. f (x) = ax − 1. (c) f (x) = x 0.3. x ∈ R. if x 6= 0. then | f | is continuous at this point.5. if x ≤ 2. 3. if x > 2.) 3.3.3. Is the converse true in general? 3.3.   x sin 1 . if |x| > 1. if x = 0.3. 3.

1]. h : [0. if x ∈ Qc ∩ [0. f (x) = h(x). if x ∈ Qc ∩ [0. if x ∈ Q ∩ [0. f (D) is compact. Prove that f (x) = g(x) for all x ∈ R. 1]. x2 . k→∞ k→∞ Therefore. Find all the points on [0. 1] → R be the function given by ( x.  f (x) = q q 0. Since D is compact. if x = p .4. and discontinuous at every rational point. Then f (D) is a compact subset of R. . 1]. g are continuous functions on R and f (x) = g(x) for all x ∈ Q. p.10 B Consider k distinct points x1 . if x ∈ Q ∩ [0. Theorem 3. .5  . 1] → R be continuous functions and define ( g(x). .11 Suppose that f . q ∈ N. In particular. there exists an ∈ D such that yn = f (an ). 3. k. there exists a subsequence {ank } of {an } and a point a ∈ D such that lim ank = a ∈ D.6. 1] at which the function is continuous. Prove that if g(a) = h(a).3. 77 (a) Let g. 1] : f (x) ≥ ε} is finite.9 B Consider the Thomae function defined on (0. and discontinuous at all other points. . if x is irrational.3 that lim ynk = lim f (ank ) = f (a) ∈ f (D). 3.1 Let D be a nonempty compact subset of R and let f : D → R be a continuous function.  (a) Prove that for every ε > 0.3. Find a function defined on R that is continuous at each xi . i = 1. .4 PROPERTIES OF CONTINUOUS FUNCTIONS Recall from Definition 2.3. f (x) = 1 − x. . f (D) is closed and bounded. k ≥ 1. the set Aε = {x ∈ (0.3. . Proof: Take any sequence {yn } in f (D). then f is continuous at a. k→∞ It now follows from Theorem 3. for some a ∈ [0. .6. xk ∈ R. 1]. (b) Let f : [0. (b) Prove that f is continuous at every irrational point. The final conclusion follows from Theorem 2. Then for each n. 1] by  1 . 3. 3.3 that a subset A of R is compact if and only if every sequence {an } in A has a subsequence {ank } that converges to a point a ∈ A. 1]. where p and q have no common factors.

nk by the squeeze theorem (Theorem 2. we say that f has an absolute maximum at x̄ if f (x) ≤ f (x̄) for every x ∈ D. m ≤ f (xn ) < m + 1/n.4 PROPERTIES OF CONTINUOUS FUNCTIONS Definition 3. For each n.3 The proof of Theorem 3.4 If f : [a.4. On the other hand.2: Absolute maximum and absolute minimum of f on [a. Corollary 3.1 We say that the function f : D → R has an absolute minimum at x̄ ∈ D if f (x) ≥ f (x̄) for every x ∈ D. .5).4. by continuity we have limk→∞ f (xnk ) = f (x̄). Figure 3. m ∈ R. there exists an element x̄ ∈ D and a subsequence {xnk } that converges to x̄ ∈ D as k → ∞.4. there exists xn ∈ D such that an = f (xn ) and.6.2 — Extreme Value Theorem.  Remark 3. Thus. hence. A = f (D) is closed and bounded (see Theorem 2. since an ∈ A = f (D). b] → R is continuous. Let m = inf A = inf f (x). Theorem 3. The proof is similar for the case of absolute maximum.1. b]. x∈D In particular. We conclude that f (x̄) = m ≤ f (x) for every x ∈ D.4.4.78 3. f has an absolute minimum at x̄. Then f has an absolute minimum and an absolute maximum on D. we have provided a direct proof instead.6) we conclude limk→∞ f (xnk ) = m. Because 1 m ≤ f (xnk ) < m + for every k. Proof: Since D is compact. Similarly.2 can be shortened by applying Theorem 2. By the compactness of D. Suppose f : D → R is continuous and D is a compact set. For every n ∈ N. However. then it has an absolute minimum and an absolute maximum on [a.4. there exists an ∈ A such that m ≤ an < m + 1/n.6. b].

t = max{c − δ /2. Because c < b (since f (b) > 0). 79 Corollary 3. b] : f (x) ≤ 0}. we can find s ∈ (c. since t < c = sup A. b]. and the fact that the interval [a. This is a contradiction because s ∈ A and s > c. in particular.4. b) such that f (s) < 0 (indeed s = min{c + δ /2.8 — Intermediate Value Theorem. . Proof: Let ε = f (c) > 0. b) such that f (c) = 0. b]. This set is also bounded since A ⊂ [a. Suppose f (c) > 0. Theorem 3. there exists δ > 0 such that if x ∈ D and |x − c| < δ . b] → R be a continuous function. then | f (x) − f (c)| < ε. This implies. Proof: We prove only the case f (a) < 0 < f (b) (the case f (a) > 0 > f (b) is completely analogous). b] → R be a continuous function. c = sup A exists and a ≤ c ≤ b. (c + b)/2} will do). δ ) ∩ D.4. We are going to prove that f (c) = 0 by showing that f (c) < 0 and f (c) > 0 lead to contradictions. It follows immediately from Theorem 3. Then there exists a number c ∈ (a. Define A = {x ∈ [a. The same conclusion follows if f (a) > γ > f (b). there exists t ∈ (a. b] is compact (see Example 2. Let f : [a.6. The proof is now complete. On the other hand. c) (in fact. Since a < c (because f (a) < 0). Therefore. that f (x) > f (c) − ε = 0 for every x ∈ B(c.4. Then there exists δ > 0 such that f (x) < 0 for all x ∈ B(c. δ ) ∩ D.2. b]. Lemma 3. b) such that f (c) = γ.  Theorem 3. Then there exists c ∈ (a.  Remark 3. c) such that f (x) > 0 for all x ∈ (t.4).4.4. (a + c)/2} will do). We conclude that f (c) = 0. Suppose f (a) · f (b) < 0 (this means either f (a) < 0 < f (b) or f (a) > 0 > f (b)). This is a contradiction. The following result is a basic property of continuous functions that is used in a variety of situations. Then there exists δ > 0 such that f (x) > 0 for all x ∈ B(c. Suppose f (c) > 0. Suppose f (c) < 0. δ ) ∩ [a. But then t < t 0 and f (t 0 ) ≤ 0. Suppose f (a) < γ < f (b).4.5 Let f : D → R be continuous at c ∈ D. there exists t 0 ∈ A with t < t 0 ≤ c. This set is nonempty since a ∈ A. δ ) ∩ [a.6 An analogous result holds if f (c) < 0.4 is sometimes referred to as the Extreme Value Theorem.7 Let f : [a. By the continuity of f at c. Then there exists δ > 0 such that f (x) > 0 for every x ∈ B(c.

instead.3. Second Proof of Theorem 3. Moreover. If f (c1 ) = γ. The proof is now complete. b]}. there exists c ∈ (a. Now consider the function f : [0.4.3. x ∈ [a. given n ∈ N. b] such that f (c) = γ.4. b] → R be a continuous function. Proof: Define ϕ(x) = f (x) − γ.3. Now. Let m = min{ f (x) : x ∈ [a. 1 + a) such that f (x) = a.7. b]. Moreover.2 We show now that. the function f is continuous (as the sum of continuous functions).  Example 3. it follows from the Intermediate Value Theorem that there is x ∈ (0. By Theorem 3.1 We will use the Intermediate Value Theorem to prove that the equation ex = −x has at least one real solution. We can now apply the Intermediate Value Theorem to the function f on the interval [−1.8: We construct a sequence of nested intervals as follows.1 that such an x is unique. First define the function f : R → R by f (x) = ex + x. relies on the Nested Intervals Theorem (Theorem 2.80 3. M]. as desired. We will assume known that the exponential function is continuous on R and that ex < 1 for x < 0. we are done.4. Notice that the given equation has a solution x if and only if f (x) = 0.9 Let f : [a. and let I1 = [a. there exists c ∈ [a. Then ϕ is continuous on [a. b]. 0] such that f (c) = 0.  Corollary 3.) We present below a second proof of Theorem 3. b].3). Since f (0) = 0 and f (1 + a) > a. b) such that ϕ(c) = 0. Let n ∈ N and let a ∈ R with a > 0. First observe that (1 + a)n ≥ 1 + na > a (see Exercise 1. ∞) → R given by f (x) = xn .8 that does not depend on Theorem 3.4.4 PROPERTIES OF CONTINUOUS FUNCTIONS Figure 3. note that f (−1) = e−1 + (−1) < 1 − 1 = 0 and f (0) = 1 > 0. Set a1 = a. but. This is equivalent to f (c) = γ.7.4.4. The point c is the desired solution to the original equation. b1 = b. (We show later in Example 4.3: Illustration of the Intermediate Value Theorem.4. Otherwise. Let c1 = (a + b)/2. ϕ(a)ϕ(b) = [ f (a) − γ][ f (b) − γ] < 0. 0] with γ = 0 to conclude that there is c ∈ [−1. That is. xn = a. every positive real number has a positive n-th root. Then for every γ ∈ [m. b]} and M = max{ f (x) : x ∈ [a.  Example 3.7). . either f (c1 ) > γ or f (c1 ) < γ.

Set c2 = (a2 + b2 )/2. we can define the new function f : D → R by the same function notation. again we are done.5 give f (c) ≤ γ and f (c) ≥ γ. Now set I2 = [a2 . d]. then c ∈ (a. In this case. n→∞ Since f (an ) < γ < f (bn ) for all n.4. b]. f ([a. f (a2 ) < γ < f (b2 ). set a2 = c1 and b2 = b1 . either f (c2 ) > γ or f (c2 ) < γ. or we construct a sequence of closed bounded intervals {In } with In = [an . Note that in either case. we get lim f (an ) = f (c) and n→∞ lim f (bn ) = f (c). Condition (ii) implies that limn→∞ (bn − an ) = 0. b]. b]. (ii) bn − an = (b − a)/2n−1 . b3 ]. By the continuity of f . . the proof is complete. and the inverse function f −1 defined on [c. T as we see from the proof of that theorem. Moreover. set a2 = a1 and b1 = c1 . Note that. b). b] → R be strictly increasing and continuous on [a. Otherwise. condition (iii) above and Theorem 2. hence. Theorem 3. In the first case. and (iii) f (an ) < γ < f (bn ). Now set I3 = [a3 . is a continuous function from [c.10 Let f : [a. Proceeding in this way.  Now we are going to discuss the continuity of the inverse function. Note that in either case. b2 ]. either we find some cn0 such that f (cn0 ) = γ and.3. If f (c2 ) = γ. f (a3 ) < γ < f (b3 ).1. d] by f −1 ( f (x)) = x where x ∈ [a. The function f : D → E is said to be continuous at a point x̄ ∈ D if the corresponding function f : D → R is continuous at x̄. In the second case. For a function f : D → E. where E is a subset of R.3. (i) In ⊃ In+1 . b] such that ∞n=1 In = {c}. In the second case. bn ] such that for all n. we proceed as follows. since f (a) < γ < f (b). there exists c ∈ [a. By the Nested Intervals Theorem (Theorem 2. 81 In the first case. set a3 = a2 and b3 = c2 . d] onto [a. an → c and bn → c as n → ∞. Let c = f (a) and d = f (b). Then f is one-to-one. The proof is now complete. It follows that f (c) = γ. b]) = [c. set a3 = c2 and b3 = b2 . part (b)).

by the continuity of f . Exercises 3.8). In this case we can show using the Inter- mediate Value Theorem that f ((a.7) Since the sequence {xk` } is bounded. Suppose by contradiction that {xk } does not converge to x̄. d] and fix any sequence {yk } in [c.4. then limx→a+ f (x) = c ∈ R. d) to (a. Fix any ȳ ∈ [c.2 Let f . Using the Intermediate Value Theorem we obtain that f ((a. (Assume known that the function cos x is continuous.7). d] that converges to ȳ.4. .82 3. b) with some additional considerations.2. we will again call the new subsequence {xk` }.5 Let f : [a.4. Suppose f (c) > γ. d].4 we can show that both limx→a+ f (x) = c and limx→b− f (x) = d exist in R (see Exercise 3.4 PROPERTIES OF CONTINUOUS FUNCTIONS Proof: The first two assertions follow from the monotonicity of f and the Intermediate Value Theorem (see also Corollary 3. but limx→b− f (x) = ∞ (again see Exercise 3. δ ) ∩ D.8) On the other hand. 3.4.3 Prove that the equation cos x = x has at least on solution in R. 3. This implies x0 = x̄. (3. The other possibilities lead to similar results. (3.2.1 Let f : D → R be continuous at c ∈ D and let γ ∈ R.2. Then f −1 (ȳ) = x̄ and f −1 (yk ) = xk for every k. { f (xk` )} converges to f (x0 ). ∞) and we can proceed as above to prove that f has a continuous inverse from (c.4 Prove that the equation x2 − 2 = cos(x + 1) has at least two real solutions.4. it follows that f (x0 ) = ȳ = f (x̄).  Remark 3. g be continuous functions on [a. b)) = (c. To simplify the notation.4. If f : (a.11 A similar result holds if the domain of f is the open interval (a. d). b] → [a. b) → R is increasing. following the argument in Theorem 3. b] be a continuous function. Since f (xk` ) = yk` → ȳ as ` → ∞. it has a further subsequence that converges to x0 ∈ [a. Prove that there exists δ > 0 such that f (x) > γ for every x ∈ B(c. We will prove that f −1 is continuous on [c. b). If : (a. b). which contradicts (3. (Assume known that the function cos x is continuous. but not bounded above.4.) 3. Suppose f (a) < g(a) and f (b) > g(b). Then there exist ε0 > 0 and a subsequence {xk` } of {xk } such that |xk` − x̄| ≥ ε0 for every `. we get |x0 − x̄| ≥ ε0 > 0. bounded below. continuous. b) such that f (x0 ) = g(x0 ). b) → R is increasing and bounded. b]. Let x̄ ∈ [a. b)) = (c.10). b] and xk ∈ [a.9). We can now proceed as in the previous theorem to show that f has a continuous inverse from (c. A similar theorem can be proved for strictly decreasing functions. ∞) to (a. b] be such that f (x̄) = ȳ and f (xk ) = yk for every k. Taking limits in (3.10). Prove that there exists x0 ∈ (a.) 3. b].

b].4. 3.4.2 Let f : R → R be given by f (x) = 7x − 2. 83 (a) Prove that the equation f (x) = x has a solution on [a.1 If f : D → R is uniformly continuous on D. prove that there exists ` ∈ [0. then | f (u) − f (v)| < ε. . b] and x1 . 3. The following result is straightforward from the definition. Suppose further that f is monotone. 3. b].6 B Let f be a continuous function on [a. b]. . 1]. Prove that the equation f (x) = x has a unique solution on [a. We will show that f is uniformly continuous on R. (a) Prove that f (0) = 0.  Example 3. b].8 I Let f . is uniformly continuous on its domain. v ∈ D regardless of the choice of δ . (b) Suppose further that | f (x) − f (y)| < |x − y| for all x. A function f : D → R is called uniformly continuous on D if for any ε > 0. (b) Given two positive numbers a and b with a < b. xn ∈ [a. b]. 1] such that f (x0 ) = g(x0 ) = x0 . Theorem 3. . v ∈ D and |u − v| < δ . Prove that there exists c ∈ [a.1 Let D be a nonempty subset of R.5 UNIFORM CONTINUITY We discuss here a stronger notion of continuity. there exists δ > 0 such that if u. x 6= y.5. Let ε > 0 and choose δ = ε/7. if u. . b] with f (x1 ) + f (x2 ) + · · · f (xn ) f (c) = . 1) such that | f (x)| ≤ `|x| for all x ∈ [a. v ∈ R and |u − v| < δ .4.1 Any constant function f : D → R. given ε > 0.5. n 3. Indeed. 1] → [0.5. y ∈ [a. . x2 . Definition 3. Prove that there exists x0 ∈ [0. Then.  Example 3. | f (u) − f (v)| = 0 < ε for all u. 1] be continuous functions such that f (g(x)) = g( f (x)) for all x ∈ [0.5. g : [0. then f is continuous at every point x0 ∈ D. we have | f (u) − f (v)| = |7u − 2 − (7v − 2)| = |7(u − v)| = 7|u − v| < 7δ = ε.7 B Suppose f is a continuous function on R such that | f (x)| < |x| for all x 6= 0.

Let ε > 0.3 Let f : [−3. Then.84 3.5. Now set δ = ε/6. we have | f (u) − f (v)| = |u2 − v2 | = |u − v||u + v| ≤ 6|u − v| < 6δ = ε. 2]. 2] we have |u + v| ≤ |u| + |v| ≤ 6. We will show that f is uniformly continuous on R. 2  Example 3. v ∈ [−3.4 Let f : R → R be given by f (x) = x2x+1 . 2] → R be given by f (x) = x2 . Let ε > 0. 2] satisfying |u − v| < δ . v ∈ [−3. for u. This function is uniformly continuous on [−3. First observe that for u.5.5 UNIFORM CONTINUITY  Example 3. We observe first that .

2 v2 .

.

.

.

u2 (v2 + 1) − v2 (u2 + 1) .

.

.

.

.

.

u |u − v||u + v| |u − v|(|u| + |v|) .

u2 + 1 − v2 + 1 .

= .

= 2 ≤ 2 .

2 2 (u + 1)(v + 1) .

(u2 + 1)(v2 + 1) v + 1 u2 + 1 (where we used that |x| ≤ x2 + 1 for all x ∈ R. which can be easily seen by considering separately the cases |x| < 1 and |x| ≥ 1). Now. (u + 1)(v + 1) (u + 1)(v2 + 1) 2 |u − v|((u2 + 1) + (v2 + 1))   1 1 ≤ ≤ |u − v| 2 + ≤ 2|u − v|. set δ = ε/2. v ∈ R satisfying |u − v| < δ we have . In view of the previous calculation. given u.

2 v2 .

.

.

.

u | f (u) − f (v)| = .

2 .

u + 1 v2 + 1 . − ≤ 2|u − v| < 2δ = ε.

Definition 3.5. Then the function f (x) = x is Lipschitz continuous on D and. For any 1/α ε > 0.5. then the function f is called Lipschitz continuous. hence. for any u.2 (Hölder continuity). The number α is called the Hölder exponent of the function. we have √ √ |u − v| 1 | f (u) − f (v)| = | u − v| = √ √ ≤ √ |u − v|. whenever u. ∞). thus. Theorem 3. .5. u+ v 2 a √ which shows f is Lipschitz with ` = 1/(2 a). Indeed. v ∈ D. If ` = 0.  √  Example 3.2 If a function f : D → R is Hölder continuous. there are constants ` ≥ 0 and α > 0 such that | f (u) − f (v)| ≤ `|u − v|α for every u. The proof is now complete. v ∈ D. If α = 1. Proof: Since f is Hölder continuous. v ∈ D. uniformly continuous on this set.5 (1) Let D = [a. Let D be a nonempty subset of R. with |u − v| < δ we have | f (u) − f (v)| ≤ `|u − v|α < `δ α = ε. where a > 0. uniformly continuous. A function f : D → R is said to be Hölder continuous if there are constants ` ≥ 0 and α > 0 such that | f (u) − f (v)| ≤ `|u − v|α for every u. then it is uniformly continuous. let δ = ε` . Then. then f is constant and. v ∈ D. Suppose next that ` > 0.

∞).4: The square root function. but it is Hölder continuous on D and. 85 Figure 3. f is also uniformly continuous on this set. hence. Then there exists a constant ` > 0 such that √ √ | f (u) − f (v)| = | u − v| ≤ `|u − v| for every u. Indeed. Then f is not Lipschitz continuous on D. suppose by contradiction that f is Lipschitz continuous on D. for every n ∈ N. v ∈ D. Thus. we have . (2) Let D = [0.

.

.

.

.

1 .

.

1 .

.

√ − 0.

≤ ` .

− 0.

. .

n .

.

n .

9) The inequality in (3. We are going to prove that | f (u) − f (v)| ≤ |u − v|1/2 for every u.9) holds obviously for u = v = 0. f is not Lipschitz continuous on D. For u > 0 or v > 0. This is a contradiction. (3. This implies √ n ≤ ` or n ≤ `2 for every n ∈ N . v ∈ D. Let us show that f is Hölder continuous on D. Therefore. we have √ √ | f (u) − f (v)| = | u − v| .

.

.

u−v .

= .

√ .

√ .

u+ v.

(3. Note that one can justify the inequality p |u| + |v| √ √ ≤1 u+ v by squaring both sides since they are both positive. Thus. .9) is satisfied. p p |u − v| = |u − v| √ √ u+ v p |u| + |v| p ≤ √ √ |u − v| u+ v p = |u − v|.

1) → R be given by 1 f (x) = .  Example 3. x Figure 3. Then x. The following example illustrates this point. y ∈ (0. but . x = δ0 . and y = 2δ0 . 1).5: Continuous but not uniformly continuous on (0. ∞). 1) and |x − y| = δ0 < δ .5 UNIFORM CONTINUITY While every uniformly continuous function on a set D is also continuous at each point of D. Let ε = 2 and δ > 0.5. Set δ0 = min{δ /2. 1/4}. We will show that f is not uniformly continuous on (0.86 3. the converse is not true in general.6 Let f : (0. We already know that this function is continuous at every x̄ ∈ (0. 1).

.

.

.

.

.

.

.

.

1 1 .

.

y − x .

.

δ0 .

.

1 .

| f (x) − f (y)| = .

− .

= .

.

.

.

.

= .

.

= .

.

x y xy . ≥ 2 = ε.

.

2δ02 .

.

2δ0 .

Then f is uniformly continuous on D if and only if the following condition holds (C) for every two sequences {un }. Let ε > 0.3 Let D be a nonempty subset of R and f : D → R.3. it follows that limn→∞ ( f (un ) − f (vn )) = 0. 1). The following theorem offers a sequential characterization of uniform continuity analogous to that in Theorem 3. {vn } in D such that limn→∞ (un − vn ) = 0.3. Choose δ > 0 such that | f (u) − f (v)| < ε whenever . This shows f is not uniformly continuous on (0. {vn } be sequences in D such that limn→∞ (un − vn ) = 0.5. Theorem 3. Proof: Suppose first that f is uniformly continuous and let {un }.

By the continuity of f . by way of contradiction. Then clearly. v ∈ D with |u − v| < δ and | f (u) − f (v)| ≥ ε0 . Suppose D is compact. Then there exists ε0 > 0 such that for any δ > 0. limn→∞ (un − vn ) = 0. for every n ∈ N. Then there exists ε0 > 0 such that for any δ > 0. Consider the two sequences un = 1/(n + 1) and vn = 1/n for all n ≥ 2. n→∞ n→∞ 1/(n + 1) 1/n n→∞ The following theorem shows one important case in which continuity implies uniform continuity. vn ∈ D with |un − vn | ≤ 1/n and | f (un ) − f (vn )| ≥ ε0 . Theorem 3. there exist u.7 Using this theorem. there exist u. we have | f (un ) − f (vn )| < ε. Proof: Suppose by contradiction that f is not uniformly continuous on D. which contradicts the assumption. assume condition (C) holds and suppose. that f is not uniformly continuous.  . we also have vnk → u0 as k → ∞. but { f (un ) − f (vn )} does not converge to zero.5. there exist un . Let N ∈ N be such that |un − vn | < δ for n ≥ N. Then 1 |unk − vnk | ≤ . Therefore. Thus. The proof is now com- plete. f (unk ) → f (u0 ) and f (vnk ) → f (u0 ). vn ∈ D with |un − vn | ≤ 1/n and | f (un ) − f (vn )| ≥ ε0 .4 Let f : D → R be a continuous function. which is a contradiction. there exist un . Thus. v ∈ D with |u − v| < δ and | f (u) − f (v)| ≥ ε0 . v ∈ D and |u − v| < δ . 87 u. To prove the converse.6 is not uniformly continuous.5.5. we can give an easier proof that the function in Example 3. Since D is compact. for every n ∈ N. nk for all k and. For such n. This shows limn→∞ ( f (un ) − f (vn )) = 0. It follows that for such sequences. there exist u0 ∈ D and a subsequence {unk } of {un } such that unk → u0 as k → ∞. { f (unk ) − f (vnk )} converges to zero. but   1 1 lim ( f (un ) − f (vn )) = lim − = lim (n + 1 − n) = 1 6= 0.   Example 3. hence. limn→∞ (un − vn ) = 0. Then f is uniformly continuous on D.

5 Let a. |u − a| < δ . f˜ : [a. b) → R is uniformly continuous if and only if f can be extended to a continuous function f˜ : [a.1 Prove that each of the following functions is uniformly continuous on the given domain: (a) f (x) = ax + b. b] → R (that is. More precisely.  if x ∈ (a. 3. f˜ is continuous at every x ∈ (a. Therefore. then we also have | f (u) − f (v)| < ε when we restrict u.2 holds.5. Moreover. then f is uniformly continuous on A.2 to conclude limx→a+ f (x) exists. ∞). b). Proof: Suppose first that there exists a continuous function f˜ : [a. if u. where a > 0. b ∈ R.2 I Prove that each of the following functions is not uniformly continuous on the given domain: (a) f (x) = x2 on R. v to be in A. v ∈ (a.2. there is a continuous function f˜ : [a. b] → R by   f (x). We will show first that limx→a+ f (x) exists.b) .5.5. Thus f˜ is the desired continuous extension of f . v ∈ (a. By Theorem 3. if x = b. 3. For the converse. (b) f (x) = 1/x on [a.2. so. v ∈ D with |u − v| < δ .5 UNIFORM CONTINUITY We now prove a result that characterizes uniform continuity on open bounded intervals. .  Exercises 3. b].2.b) ). b) → R is uniformly continuous. We first make the observation that if f : D → R is uniformly continuous on D and A ⊂ D. b).  limx→b− f (x). b) and |u − v| < δ0 .5.b) = f and.5. Theorem 3. Set δ = δ0 /2. x (c) f (x) = ln(x) on (0. and |v − a| < δ we have |u − v| ≤ |u − a| + |a − v| < δ + δ = δ0 and. Now define. Note that the one sided limit corresponds to the limit in Theorem 3. Choose δ0 > 0 so that | f (u) − f (v)| < ε whenever u. We will check that the ε-δ condition of Theorem 3.3 Determine which of the following functions are uniformly continuous on the given domains. Let ε > 0. ∞). In a similar way we can show that limx→b− f (x) exists.3.88 3. b). it follows from our early observation that f is uniformly continuous on (a. the restriction f|A : A → R is uniformly continuous on A (see Section 1. limx→a+ f˜(x) = limx→a+ f (x) = f˜(a) and limx→b− f˜(x) = limx→b− f (x) = f˜(b). hence. b] → R such that f = f˜|(a. A function f : (a. 1 (b) f (x) = sin on (0.2.2.4. b] → R such that f = f˜|(a.2 for the notation). b). This follows by noting that if | f (u) − f (v)| < ε whenever u. 1). so f˜ is also continuous at a and b by Theorem 3. Then. suppose f : (a. We can now invoke Theorem 3. on R. | f (u) − f (v)| < ε. if x = a.  By its definition f˜|(a. b ∈ R and a < b. the function f˜ is uniformly continuous on [a. ˜f (x) = limx→a+ f (x). a.

(a) Prove that if f is uniformly continuous. 3. 3.6.4 Let D ⊂ R and k ∈ R. 89 (a) f (x) = x sin( 1x ) on (0. then { f (xn )} is also a Cauchy sequence. g : D → R such that f g is not uniformly conitnuous on D. Definition 3.δ )∩D . Prove that if f . then it is uniformly continuous. δ ) = (x̄ − δ . b ∈ R and let f : (a. Recall that B0 (x̄. x→∞ (a) Prove that f is bounded on [a. 3.5.5. 1).5.7 B Let a.5.δ )∩D Similarly. x→x̄ δ >0 x∈B0 (x̄. 3. x (b) f (x) = on [0. ∞).8 B Let f be a continuous function on [a. the limit inferior of the function f at x̄ is defined by lim inf f (x) = sup inf f (x). Prove that there exists x0 ∈ [a. then f + g and k f are uniformly continuous on D. g : D → R are uniformly continuous on D. ∞). δ ) = B− (x̄. δ ) ∪ B+ (x̄. Suppose that f is uniformly continuous on D. then f is bounded. 3. bounded. x̄) ∪ (x̄. and monotone. The limit superior of the function f at x̄ is defined by lim sup f (x) = inf sup f (x). (c) Suppose further that c > f (a).5. x→x̄ δ >0 x∈B0 (x̄. (b) Prove that f is uniformly continuous on [a. 1).6 LIMIT SUPERIOR AND LIMIT INFERIOR OF FUNCTIONS We extend to functions the concepts of limit superior and limit inferior.6 Let D be a nonempty subset of R and let f : D → R. x+1 1 (c) f (x) = on (0. (b) Prove that if f is continuous. x̄ + δ ). Prove that if {xn } is a Cauchy sequence with xn ∈ D for every n ∈ N. ∞)}. 1). |x − 2| 3.5 Give an example of a subset D of R and uniformly continuous functions f .1 Let f : E → R and let x̄ be a limit point of D. |x − 1| 1 (d) f (x) = on (0. ∞). ∞) such that f (x0 ) = inf{ f (x) : x ∈ [a. Suppose lim f (x) = c. ∞). b) → R.

This proves (2).10). there exists δ > 0 such that ` ≤ g(δ ) = sup f (x) < ` + ε.δ )∩D Thus. there exists xδ ∈ B0 (x̄. ∞) → (−∞. Next note that for any ε > 0 and δ > 0. there exists xδ ∈ B0 (x̄. Then ` = inf g(δ ). then lim supx→x̄ f (x) is a real number. we have ` − ε < ` ≤ g(δ ) = sup f (x). δ ) ∩ D. ∞] defined by g(δ ) = sup f (x).δ )∩D Thus. x∈B0 (x̄. For any ε > 0. Proof: Suppose ` = lim supx→x̄ f (x). Then ` = lim supx→x̄ f (x) if and only if the following two conditions hold: (1) For every ε > 0. f (x) < ` + ε for all x ∈ B0 (x̄. δ ) ∩ D with ` − ε < f (xδ ). while lim supx→x̄ f (x) = ∞ in the other case. δ ) ∩ D such that ` − ε < f (xδ ). δ ) ∩ D. there exists δ > 0 such that f (x) < ` + ε for all x ∈ B0 (x̄. . x∈B0 (x̄.6.δ )∩D It is clear that g is increasing and lim sup f (x) = inf g(δ ). x→x̄ δ >0 We say that the function f is locally bounded above around x̄ if there exists δ > 0 and M > 0 such that f (x) ≤ M for all x ∈ B(x̄. δ >0 where g is defined in (3. (2) For every ε > 0 and for every δ > 0. if f is locally bounded above around x̄. δ ) ∩ D.6 LIMIT SUPERIOR AND LIMIT INFERIOR OF FUNCTIONS Consider the extended real-valued function g : (0.1 Let f : D → R and let x̄ be a limit point of D. Theorem 3. Similar discussion applies for the limit inferior.10) x∈B0 (x̄. which proves condition (1).90 3. Clearly. (3.

11). δk0 ) ∩ D such that ` − εk < f (xk ). let δ > 0 be as in (1) of Theorem 3. x→x̄ Again.δ )∩D This implies lim sup f (x) = inf g(δ ) ≤ ` + ε. if {yk } is a sequence in D that converges to x̄.  . ` = lim supx→x̄ f (x).6. 1 Proof: For each k ∈ N. Fix any ε > 0 and let δ > 0 satisfy (1). 91 Let us now prove the converse. Now let {yk } be a sequence in D that converges to x̄. By (1) of Theorem 3. take εk = . For any ε > 0.6. let ε > 0. Therefore. let xδ be as in (2). Then g(δ ) = sup f (x) ≤ ` + ε. yk 6= x̄ for every k. x∈B0 (x̄. It follows that `0 ≤ `.1. Since yk ∈ B0 (x̄. and lim f (xk ) = `. Therefore. and limk→∞ f (yk ) = `0 . k→∞ Moreover. (3. Moreover. {xk } is a sequence that satisfies the conclusion of the corollary. x∈B0 (x̄. we get lim sup f (x) ≤ `.δ )∩D This implies ` − ε ≤ inf g(δ ) = lim sup f (x).2 Suppose ` = lim supx→x̄ f (x). δ >0 x→x̄ It follows that ` ≤ lim supx→x̄ f (x).  Corollary 3.6. This implies `0 ≤ ` + ε. Therefore.11) Let δk0 = min{δk . ` − ε < f (xδ ) ≤ sup f (x) = g(δ ).1. Then there exists a sequence {xk } in D such that {xk } converges to x̄. Then δk0 ≤ δk and limk→∞ δk0 = 0. yk 6= x̄ for every k. there exists δk > 0 such that k f (x) < ` + εk for all x ∈ B0 (x̄. δ ) ∩ D when k is sufficiently large. Given δ > 0. 1k }. and limk→∞ f (yk ) = `0 . Suppose (1) and (2) are satisfied. x→x̄ δ >0 Since ε is arbitrary. we have f (yk ) < ` + ε for such k. From (2) of Theorem 3.6. xk 6= x̄ for every k. then `0 ≤ `. there exists xk ∈ B0 (x̄.1. f (xk ) < ` + εk by (3. δk ) ∩ D.

92 3. Since limk→∞ f (xk ) = ∞. xk → x̄. Then the previous corollary shows that A 6= 0/ and lim supx→x̄ f (x) = max A.5 Let f : D → R and let x̄ be a limit point of D.6.δk )∩D there exists xk ∈ B0 (x̄. for δk = . xk 6= x̄ for every k. Theorem 3. Thus. Then inf g(δ ) = ∞. g(δ ) = sup f (x) ≥ M. x∈B0 (x̄. Proof: Suppose lim supx→x̄ f (x) = ∞. since k g(δk ) = sup f (x) = ∞. it follows that limk→∞ f (xk ) = −∞. limk→∞ f (xk ) = ∞.δ )∩D This implies g(δ ) = ∞. g(δ ) = ∞ for every δ > 0.10). and limk→∞ f (xk ) = ∞. and hence lim supx→x̄ f (x) = ∞.6. for every M ∈ R. Therefore. Suppose lim supx→x̄ f (x) is a real number. xk 6= x̄ for every k.6 LIMIT SUPERIOR AND LIMIT INFERIOR OF FUNCTIONS Remark 3. δ >0 where g is the extended real-valued function defined in (3. Let us prove the converse. For any δ > 0. f (xk ) → `}. The latter is equivalent to limx→x̄ f (x) = −∞. Following the same arguments. x∈B0 (x̄.3 Let f : D → R and let x̄ be a limit point of D. Define A = {` ∈ R : ∃{xk } ⊂ D. xk 6= x̄ for every k. δ ) ∩ D whenever k is sufficiently large. . we have xk ∈ B0 (x̄.6. 1 Given k ∈ N. Thus.  Theorem 3. Then lim sup f (x) = −∞ x→x̄ if and only if for any sequence {xk } in D such that {xk } converges to x̄. we can prove similar results for inferior limits of functions. Then lim sup f (x) = ∞ x→x̄ if and only if there exists a sequence {xk } in D such that {xk } converges to x̄.4 Let f : D → R and let x̄ be a limit point of D. there exists K ∈ N such that f (xk ) ≥ M for every k ≥ K. δk ) ∩ D such that f (xk ) > k.

6. Theorem 3.6. (2) For every ε > 0 and for every δ > 0. Define B = {` ∈ R : ∃{xk } ⊂ D.10 Let f : D → R and let x̄ be a limit point of D. Then lim inf f (x) = ∞ x→x̄ if and only if for any sequence {xk } in D such that {xk } converges to x̄.6. δ ) ∩ D. it follows that limk→∞ f (xk ) = ∞. xk 6= x̄ for every k. xk 6= x̄ for every k.11 Let f : D → R and let x̄ be a limit point of D. Suppose lim infx→x̄ f (x) is a real number. xk 6= x̄ for every k.6. there exists δ > 0 such that ` − ε < f (x) for all x ∈ B0 (x̄. Remark 3. Theorem 3. f (xk ) → `}. The latter is equivalent to limx→x̄ f (x) = ∞. Then there exists a sequence {xk } in D such that xk converges to x̄. Corollary 3. xk 6= x̄ for every k. δ ) ∩ D such that f (x) < ` + ε. yk 6= x̄ for every k. Then B 6= 0/ and lim infx→x̄ f (x) = min B. and lim f (xk ) = `. if {yk } is a sequence in D that converges to x̄.8 Let f : D → R and let x̄ be a limit point of D.6. x→x̄ x→x̄ .6. Then ` = lim infx→x̄ f (x) if and only if the following two conditions hold: (1) For every ε > 0. there exists x ∈ B0 (x̄. Then lim f (x) = ` x→x̄ if and only if lim sup f (x) = lim inf f (x) = `. 93 Theorem 3. and limk→∞ f (yk ) = `0 . then `0 ≥ `.7 Suppose ` = lim infx→x̄ f (x).9 Let f : D → R and let x̄ be a limit point of D. k→∞ Moreover. xk → x̄.6 Let f : D → R and let x̄ be a limit point of D. Then lim inf f (x) = −∞ x→x̄ if and only if there exists a sequence {xk } in D such that {xk } converges to x̄. and limk→∞ f (xk ) = −∞. Theorem 3.

there exists δ > 0 such that ` − ε < f (x) < ` + ε for all x ∈ B0 (x̄. f : D → R. The proof for the limit inferior is similar. δ ) ∩ D. x cos x (d) lim infx→0 . and x̄ be a limit point of D. x 3.1 Let D ⊂ R.7 LOWER SEMICONTINUITY AND UPPER SEMICONTINUITY The concept of semicontinuity is convenient for the study of maxima and minima of some discontinuous functions.6.1 and Theorem 3. there exists δ > 0 such that f (x) < f (x̄) + ε for all x ∈ B(x̄. 3.s. Prove that lim infx→x̄ f (x) ≤ lim supx→x̄ f (x).s. The converse follows directly from (1) of Theorem 3.c.6. δ ) ∩ D. x→x̄ Then for every ε > 0. (3. δ ) ∩ D.  Exercises 3.7 LOWER SEMICONTINUITY AND UPPER SEMICONTINUITY Proof: Suppose lim f (x) = `. It is clear that f is continuous at x̄ if and only if f is lower semicontinuous and upper semicontin- uous at this point.6. x cos x (c) lim supx→0 . 94 3. Since this also holds for every 0 < δ 0 < δ .7. We say that f is lower semicontinuous (l. .12) Similarly.1 Let f : D → R and let x̄ ∈ D. we get ` − ε < g(δ 0 ) ≤ ` + ε. we say that f is upper semicontinuous (u. Definition 3. x   1 (b) lim infx→0 sin .c.) at x̄ if for every ε > 0. lim supx→x̄ f (x) = ` since ε is arbitrary.) at x̄ if for every ε > 0. δ >0 Therefore. It follows that ` − ε ≤ inf 0 g(δ 0 ) ≤ ` + ε. there exists δ > 0 such that f (x̄) − ε < f (x) for all x ∈ B(x̄.6.2 Find each of the following limits:   1 (a) lim supx→0 sin .6.

95 Figure 3. f is upper semicontinuous at x̄ if and only if lim sup f (x) ≤ f (x̄).δ )∩D Thus.7.7: Upper semicontinuity. Then there exists δ0 > 0 such that f (x̄) − ε < f (x) for all x ∈ B(x̄. This implies f (x̄) − ε ≤ h(δ0 ). Theorem 3.1 Let f : D → R and let x̄ ∈ D be a limit point of D. x→x̄ Proof: Suppose f is lower semicontinuous at x̄. Then f is lower semicontinuous at x̄ if and only if lim inf f (x) ≥ f (x̄). x∈B0 (x̄. x→x̄ δ >0 . Let ε > 0. where h(δ ) = inf f (x). δ0 ) ∩ D. Figure 3. x→x̄ Similarly. lim inf f (x) = sup h(δ ) ≥ h(δ0 ) ≥ f (x̄) − ε.6: Lower semicontinuity.

s. at x̄ if and only if for every sequence {xk } in D that converges to x̄. k→∞ Similarly. the function f is lower semicontinuous at x̄. we obtain lim infx→x̄ f (x) ≥ f (x̄). δ ) when k is sufficiently large.  . f is u.s. Thus. we obtain a sequence {xk } in D that converges to x̄ with k f (x̄) − ε̄ ≥ f (xk ) for every k. k→∞ This is a contradiction. there exists δ > 0 such that (3.c. δ ) ∩ D with f (x̄) − ε̄ ≥ f (xδ ). Suppose lim inf f (x) = sup h(δ ) ≥ f (x̄) x→x̄ δ >0 and let ε > 0. we have xk ∈ B(x̄. by way of contra- diction. Since ε is arbitrary. The proof for the upper semicontinuous case is similar.c. This implies f (x) > f (x̄) − ε for all x ∈ B0 (x̄. at x̄ if and only if for every sequence {xk } in D that converges to x̄.12) holds. there exists xδ ∈ B(x̄. We now prove the converse. It follows that f (x̄) − ε ≤ lim infk→∞ f (xk ). it follows that f (x̄) ≤ lim infk→∞ f (xk ). Since {xk } converges to x̄. k→∞ Proof: Suppose f is l. Then f is l.  Theorem 3. at x̄. Then there exists ε̄ > 0 such that for every δ > 0. δ >0 there exists δ > 0 such that h(δ ) > f (x̄) − ε.7 LOWER SEMICONTINUITY AND UPPER SEMICONTINUITY Since ε is arbitrary. Since this is also true for x = x̄. lim sup f (xk ) ≤ f (x̄). f (x̄) − ε < f (xk ) for such k. that f is not l.s.7.2 Let f : D → R and let x̄ ∈ D. Suppose lim infk→∞ f (xk ) ≥ f (x̄) and assume. lim inf f (xk ) ≥ f (x̄).c.96 3. We now prove the converse. Then for any ε > 0. This implies f (x̄) − ε̄ ≥ lim inf f (xk ). at x̄. 1 Applying this for δk = .c. δ ) ∩ D.s. Since sup h(δ ) > f (x̄) − ε.

2 lim inf f (xk` ) ≥ f (x0 ). . For every a ∈ R. This shows f is bounded below. `→∞ This is a contraction because lim inf`→∞ f (xk` ) = −∞. define La ( f ) = {x ∈ D : f (x) ≤ a} and Ua ( f ) = {x ∈ D : f (x) ≥ a}. hence. Then f has an absolute minimum on D. Then f is lower semicontinuous if and only if La ( f ) is closed in D for every a ∈ R. the sequence {uk } has a convergent subsequence {uk` } that converges to some x̄ ∈ D. Let {uk } be a sequence in D such that { f (uk )} converges to γ. Proof: We first prove that f is bounded below. Define γ = inf{ f (x) : x ∈ D}. `→∞ `→∞ This implies γ = f (x̄) and.  The following theorem is proved similarly.4 Suppose D is a compact subset of R and f : D → R is upper semicontinuous.s. by Theorem 3. That means there exists x̄ ∈ D such that f (x) ≥ f (x̄) for all x ∈ D. That is. The proof is now complete. Theorem 3. We say that f is lower semicontinuous on D (or lower semicontinu- ous if no confusion occurs) if it is lower semicontinuous at every point of D.7.7.7.7. Since f is l.c. Since D is compact. f is upper semicontinuous if and only if Ua ( f ) is closed in D for every a ∈ R.7.3 Suppose D is a compact set of R and f : D → R is lower semicontinuous. there exists xk ∈ D such that f (xk ) < −k. Since the set { f (x) : x ∈ D} is nonempty and bounded below. Then f has an absolute maximum on D. there exists a subsequence {xk` } of {xk } that converges to x0 ∈ D. By the compactness of D. Similarly.5 Let f : D → R. Then γ = lim f (uk` ) = lim inf f (uk` ) ≥ f (x̄) ≥ γ.. f (x) ≥ f (x̄) for all x ∈ D. γ ∈ R. Suppose by contradiction that for every k ∈ N. Theorem 3.2 Let f : D → R. Theorem 3. 97 Definition 3. there exists x̄ ∈ D such that f (x) ≤ f (x̄) for all x ∈ D.

Let us prove the converse.98 3.b is open since it is the intersection of two open sets La and Ub . f is lower semicontinuous. By Theorem 3. For every a ∈ R. there exists δ > 0 such that B(x̄. f is lower semicontinuous by Corollary 3.7. Since f is lower semicontinuous at x̄. Then the set G = {x ∈ D : f (x) > f (x̄) − ε} = D \ L f (x̄)−ε( f ) is open in D and x̄ ∈ G.b = Lb ∩Ua . Corollary 3. The proof for the upper semicontinuous case is similar. Thus. since xk ∈ La ( f ). b)) is an open set in D. Ua ( f ) is open in D as it is a union of open sets in D.7. k→∞ Thus.7. Therefore. Then f is lower semicontinuous and upper semicontinuos.10. f (x̄) ≤ lim inf f (xk ) ≤ a. It follows that f (x̄) − ε < f (x) for all x ∈ B(x̄. Then Oa. Then f is lower semicontinuous if and only if Ua ( f ) is open in D for every a ∈ R. Proof: Suppose f is continuous.  . f is upper semicontinuous if and only if La ( f ) is open in D for every a ∈ R.6. we have Ua ( f ) = {x ∈ D : f (x) > a} = ∪n∈N f −1 ((a. x̄ ∈ La ( f ). we will prove that for every sequence {xk } in La ( f ) that converges to a point x̄ ∈ D.6. Then f is continuous if and only if for every a.7 LOWER SEMICONTINUITY AND UPPER SEMICONTINUITY Proof: Suppose f is lower semicontinuous.b = {x ∈ D : a < f (x) < b} = f −1 ((a. For every k. It follows that La ( f ) is closed. δ ) ∩ D. Theorem 3. f (xk ) ≤ a. Similarly. b ∈ R with a < b. Fix a. a + n)) Thus. we also define La ( f ) = {x ∈ D : f (x) < a} and Ua ( f ) = {x ∈ D : f (x) > a}. Fix any x̄ ∈ D and ε > 0.7 Let f : D → R. the set Oa.  For every a ∈ R.6. b ∈ R with a < b. Using Corollary 2. δ ) ∩ D ⊂ G.6 Let f : D → R. the set Oa. We now prove the converse. Therefore. we get x̄ ∈ La ( f ).7. We will only show that f is lower semicontinuous since the proof of upper semicontinuity is similar.

g : D → R be lower semicontinuous functions and let k > 0 be a constant. 3.1 Let f be the function given by ( x2 .4 I Let f : R → R be a lower semicontinuous function such that lim f (x) = lim f (x) = ∞. if x 6= 0. f (x) = 1.7. 3. Prove that f is lower semicontinuous. . if x = 0.3 Let f . 3. Prove that f is upper semicontinuous. if x = 0. Prove that f + g and k f are lower semicontinous functions on D.2 Let f be the function given by ( x2 .7. f (x) = −1. if x 6= 0.7. 99 Exercises 3.7. x→∞ x→−∞ Prove that f has an absolute minimum at some x0 ∈ R.

.

1 DEFINITION AND BASIC PROPERTIES OF THE DERIVATIVE Let G be an open subset of R and consider a function f : G → R. and f is said to be differentiable at a.1. a is a limit point of G \ {a} (see Example 2. then f (x) − f (a) f 0 (a) = lim . DEFINITION AND BASIC PROPERTIES OF THE DERIVA- TIVE THE MEAN VALUE THEOREM SOME APPLICATIONS OF THE MEAN VALUE THEOREM L’HOSPITAL’S RULE TAYLOR’S THEOREM CONVEX FUNCTIONS AND DERIVATIVES NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBD- IFFERENTIALS 4. Thus. In this case. For every a ∈ G. We say that the function f defined on G is differentiable at a if the limit f (x) − f (a) lim x→a x−a exists (as a real number). x→a x→a x−a Definition 4. it is possible to discuss the limit f (x) − f (a) lim φa (x) = lim . the function f (x) − f (a) φa (x) = x−a is defined on G \ {a}.1 Let G be an open subset of R and let a ∈ G. the function f 0 : G → R is called the derivative of f on G. Thus. In this case. we discuss basic properties of the derivative of a function and several major theorems. if f is differentiable at a. Since G is an open set.6. including the Mean Value Theorem and l’Hôpital’s Rule. DIFFERENTIATION In this chapter. 4. the limit is called the derivative of f at a denoted by f 0 (a). x→a x−a We say that f is differentiable on G if f is differentiable at every point a ∈ G.6). .

1 DEFINITION AND BASIC PROPERTIES OF THE DERIVATIVE  Example 4.3 Let G be an open subset of R and let f . g : G → R. (a) The function f + g is differentiable at a and ( f + g)0 (a) = f 0 (a) + g0 (a).1 (a) Let f : R → R be given by f (x) = x and let a ∈ R. x→a x→a x−a Therefore. Then f (x) − f (a) x 2 − a2 (x − a)(x + a) lim = lim = lim = lim (x + a) = 2a.1 is not true. hence. but it is not differentiable at this point (as shown in the example above). Then f (x) − f (a) x−a lim = lim = lim 1 = 1.1. x→ 0+ x−0 x→ 0 x x→ 0 x and f (x) − f (0) |x| −x lim = lim− = lim− = −1.2.3. Theorem 4. f is differentiable at every a ∈ R and f 0 (a) = 2a. Suppose both f and g are differentiable at a ∈ G.1 Let G be an open subset of R and let f be defined on G.1. x→ 0− x−0 x→ 0 x x→ 0 x f (x)− f (0) Therefore. x→ a x−a x→ a x − a x→ a x−a x→ a Thus. x→ a x−a x→ a x − a x→ a It follows that f is differentiable at a and f 0 (a) = 1. (b) Let f : R → R be given by f (x) = x2 and let a ∈ R. (b) For a constant c.1. f is continuous at a by Theorem 3. Theorem 4.2 The converse of Theorem 4. f is not differentiable at 0.  Remark 4. . the absolute value function f (x) = |x| is continuous at 0. Proof: We have the following identity for x ∈ G \ {a}: f (x) = f (x) − f (a) + f (a) f (x) − f (a) = (x − a) + f (a). x−a Thus. (c) Let f : R → R be given by f (x) = |x| and let a = 0.1. limx→ 0 x−0 does not exist and. If f is differentiable at a ∈ G.   f (x) − f (a) lim f (x) = lim (x − a) + f (a) = f 0 (a) · 0 + f (a) = f (a). Then f (x) − f (0) |x| x lim = lim+ = lim+ = 1. Then the following hold. For instance.102 4. then f is continuous at this point. the function c f is differentiable at a and (c f )0 (a) = c f 0 (a).1.

we obtain (d). Let us prove (c). lim g(x) = g(a). g (g(a))2 Proof: The proofs of (a) and (b) are straightforward and we leave them as exercises. Moreover. Furthermore. The proof is now complete. we can write ( f g)(x) − ( f g)(a) f (x)g(x) − f (a)g(x) + f (a)g(x) − f (a)g(a) = x−a x−a ( f (x) − f (a))g(x) f (a)(g(x) − g(a)) = + . let g : R → R be given by g(x) = x.1(a) and Theorem 4. (4.1.1) x→a Thus.2 Let f : R → R be given by f (x) = x2 and let a ∈ R. f (d) Suppose additionally that g(a) 6= 0. ( f g)(x) − ( f g)(a) lim = f 0 (a)g(a) + f (a)g0 (a). Proceeding by induction.1. Then the function is differentiable at a and g  0 f f 0 (a)g(a) − f (a)g0 (a) (a) = . Then f = g · g so f 0 (a) = (gg)0 (a) = g0 (a)g(a) + g(a)g0 (a) = 2g0 (a)g(a) = 2a. x−a x−a By Theorem 4. x→a x−a This implies (c).1). . g f (x) f (a) f (a) f (a) h(x) − h(a) g(x) − g(x) + g(x) − g(a) = x−a x−a 1 f (a) g(x) ( f (x) − f (a)) + g(x)g(a) (g(a) − g(x)) =  x−a  1 f (x) − f (a) g(x) − g(a) = g(a) − f (a) .1. 103 (c) The function f g is differentiable at a and ( f g)0 (a) = f 0 (a)g(a) + f (a)g0 (a). Indeed.3(c) we can provide an alternative derivation of a formula for f 0 (a). Then h is defined on I.1. g(x)g(a) x−a x−a Taking the limit as x → a.1.3(a)(b) we obtain the familiar formula for the derivative of a polynomial p(x) = an xn + · · · + a1 x + a0 as p0 (x) = nan xn−1 + · · · + 2a2 x + a1 . we can obtain the derivative of g : R → R given by g(x) = xn for n ∈ N as g0 (a) = nxn−1 .   Example 4. the function g is continuous at a and. Next we show (d). hence. there exists an open interval I containing a such that f g(x) 6= 0 for all x ∈ I. Since g(a) 6= 0. using this and Theorem 4. Using Example 4. by (4.1. Let h = . For every x ∈ G \ {a}.

Then there exists a function u : G → R satisfying f (x) − f (a) = [ f 0 (a) + u(x)](x − a) for all x ∈ G and limx→a u(x) = 0. This implies g( f (x)) − g( f (a)) = [g0 ( f (a)) + v( f (x))][ f 0 (a) + u(x)] for all x ∈ G1 \ {a}.2) and limt→ f (a) v(t) = 0.  . Proof: Since f is differentiable at a.  Theorem 4. Proof: Define  f (x) − f (a) − f 0 (a). Applying (4. the function u satisfies the conditions of the lemma.1.1. x→a x→a x−a Therefore.1. x = a. we have f (x) − f (a) lim u(x) = lim − f 0 (a) = f 0 (a) − f 0 (a) = 0.5 — Chain rule. (4. Lemma 4. Then the function g ◦ f is differentiable at a and (g ◦ f )0 (a) = g0 ( f (a)) f 0 (a).2) for t = f (x).1 DEFINITION AND BASIC PROPERTIES OF THE DERIVATIVE The following lemma is very convenient for studying the differentiability of the composition of functions. and limx→a u(x) = 0. hence. we have g( f (x)) − g( f (a)) = [g0 ( f (a)) + v( f (x))][ f (x) − f (a)]. x−a By the continuity of f at a and the property of v. Suppose f is differentiable at a. Thus.4. where G1 and G2 are two open subsets of R with f (G1 ) ⊂ G2 .104 4. x ∈ G \ {a}  u(x) = x−a 0. there exists a function u defined on G1 with f (x) − f (a) = [ f 0 (a) + u(x)](x − a) for all x ∈ G1 . we have limx→a v( f (x)) = 0 and. Since f is differentiable at a. g( f (x)) − g( f (a)) lim = g0 ( f (a)) f 0 (a). there exists a function v defined on G2 with g(t) − g( f (a)) = [g0 ( f (a)) + v(t)][t − f (a)] for all t ∈ G2 . since g is differentiable at f (a). Suppose f is differentiable at a and g is differentiable at f (a). Similarly. x→a x−a The proof is now complete. g( f (x)) − g( f (a)) = [g0 ( f (a)) + v( f (x))][ f 0 (a) + u(x)](x − a) for all x ∈ G1 .4 Let G be an open subset of R and let f : G → R. by Lemma 4. Let f : G1 → R and let g : G2 → R.

(a) Prove that f is differentiable at 0.1.2 Let G be an open set and let f : G → R be a differentiable function. if x > 0. but rather compute the appropriate limit directly (see Example 4. f (0) = f . if x ≤ 0.1. g : G2 → R.1.3.1. g(G2 ) ⊂ G3 . f 00 = ( f 0 )0 . Exercises 4. Definition 4. . As a convention. and h : G3 → R such that f (G1 ) ⊂ G2 . That is. x √ (c) f (x) = x (here assume x > 0) 4. We denote by Cn (I) the set of all n times continuously differentiable functions on I.3. Definition 4.1. given functions f : G1 → R.1.1. it follows from Theorem 4. (b) Is f 0 continuous? Is f 0 differentiable? 4.2 Compute the following derivatives directly from the definition. If the function f 0 : G → R is also differentiable. if x > 0. if x ≤ 0.1. we say that f is three times differentiable if f (2) is differentiable. f (x) = 0.3 Let f : R → R be given by ( x2 . For example. we say that f is twice differentiable (on G). 105  Example 4.1. 4.1.5 that (g ◦ f )0 (a) = g0 ( f (a)) f 0 (a) = 15(3a4 + a + 7)14 (12a3 + 1). We can define in this way n times differentiability and the nth derivative of f for any positive integer n.1). The second derivative of f is denoted by f 00 or f (2) . If f is n times differentiable on I and the nth derivative is continuous. and ( f (2) )0 is called the third derivative of f and is denoted by f 000 or f (3) . Then h = g ◦ f .3 Consider the function h : R → R given by h(x) = (3x4 + x + 7)15 . We denote by C1 (I) the set of all continuously differentiable functions on I. Similarly. do not use Theo- rem 4. Since h(x) is a polynomial we could in principle compute h0 (x) by expanding the power and using Example 4. Thus.1.1. The function f is said to be continuously differentiable if f is differentiable on I and f 0 is continuous on I. Find f 0 (x) for all x ∈ R. However. we can extended the result to the composition of more than two functions in a straightforward way. 1 (b) f (x) = (here assume x 6= 0).5 provides a shorter way.2.4 Let ( xα .1 Prove parts (a) and (b) of Theorem 4. g : R → R by f (x) = 3x4 + x + 7 and g(x) = x15 . b ∈ R. Define f . Given a ∈ R. and h is differentiable at g( f (a)).4 By iterating the Chain Rule.  Example 4. we obtain that h ◦ g ◦ f is differentiable at a and (h ◦ g ◦ f )0 (a) = h0 (g( f (a)))g0 ( f (a)) f 0 (a). then f is called n times continuously differentiable. (a) f (x) = mx + b where m.3 Let I be an open interval in R and let f : I → R.1. f (x) = 0. Theorem 4.1. f is differentiable at a. g is differentiable at f (a).

x−a 4.1. if x = 0. n→∞ f (a) 4.3 and 4. if x 6= 0. if x ∈ / Q.8 Let f . 4. find f 0 .10 B Let f be differentiable at a and f (a) > 0. ( x2 .1. (a) f (x) = x 0. p (c) f (x) = cos( |x|).1.5 to compute the derivatives of the following functions at the indicated points (see also Example 4.1.1.11 I Consider the function  x2 sin 1 + cx. then g(x) = ( f (x) − f (a))2 is differentiable at a. (b) f (x) = x 0. if x 6= 0. 4. Justify your answer. f (x) = x 0.1.7 Determine if each of the following functions is differentiable at 0. (d) f (x) = x|x|. if x 6= 0.  x2 sin 1 .1. if x ∈ Q.106 4.1. if x = 0.9 Let G be an open subset of R and a ∈ G. .) 3x4 + 7x (a) f (x) = at a = −1. if x = 0. (b) Determine the values of α for which f is differentiable on R.5 Use Theorems 4. Find the following limits: x f (a) − a f (x) (a) limx→a .  x sin 1 . In this case.1. 4. g be differentiable at a. Find the following limit: !n f (a + n1 ) lim . (Assume known that the function sin x is differentiable at all points and that its derivative is cos x. (a) f (x) = 3 x . 2x2 + 3 (b) f (x) = sin5 (3x2 + π2 x) at a = π 8 4. (b) f (x) = [x] sin2 (πx).6 Determine the values of x at which each function is differentiable. 4. x−a f (x)g(a) − f (a)g(x) (b) limx→a .1.4).1 DEFINITION AND BASIC PROPERTIES OF THE DERIVATIVE (a) Determine the values of α for which f is continuous on R. Prove that if f : G → R is Lipschitz continuous.

2. h→0 h 4. We start with some basic definitions of minima and maxima. Find the limit f (x0 + ch) − f (x0 − ch) lim . b) and let c be a constant.2. .2 THE MEAN VALUE THEOREM In this section. δ ). given by f (x) = |x|3 . and B− (a. If f has a local minimum or maximum at a ∈ I and f is differentiable at a.” His method presented in the theorem below is now known as Fermat’s Rule. The Mean Value Theorem we study in this section was stated by the French mathematician Augustin Louis Cauchy (1789–1857). a). Theorem 4. (b) Prove that for every α > 0. Definition 4. 107 where 0 < c < 1. the function f 0 changes its sign on (−α. Similarly. Fermat’s method can also be used to prove Rolle’s Theorem. In January 1638.1. we say that f has a local (or relative) maximum at a ∈ D if there exists δ > 0 such that f (x) ≤ f (a) for all x ∈ B(a. a + δ ). 4. α). b) and let c be a constant. δ ).1 Let D be a nonempty subset of R and let f : D → R.14 Prove that f : R → R. is in C2 (R) but not in C3 (R) (refer to Defini- tion 4. a + δ ) and (a − δ . h 4.1. then f 0 (a) = 0. δ ) ∩ D. δ ) denote the intervals (a − δ . (a) Prove that the function is differentiable on R. B+ (a. A simple method for identifying local extrema of a function was found by the French mathematician Pierre de Fermat (1601–1665).12 Let f be differentiable at x0 ∈ (a. We say that f has a local (or relative) minimum at a ∈ D if there exists δ > 0 such that f (x) ≥ f (a) for all x ∈ B(a.13 Let f be differentiable at x0 ∈ (a.) 4. respectively. (a. An important application of differentiation is solving optimization problems. Recall that for a ∈ R and δ > 0. one of the most important tools of calculus and one of the most beautiful results of mathematical analysis. Let I be an open interval and f : I → R. we focus on the Mean Value Theorem. which follows from a simpler version called Rolle’s Theorem. Pierre de Fermat described his method for finding maxima and minima in a letter written to Marin Mersenne (1588–1648) who was considered as “the center of the world of science and mathematics during the first half of the 1600s.1. δ ) ∩ D. Prove that (a) limn→∞ n f (x0 + n1 ) − f (x0 ) = f 0 (x0 ).   f (x0 + ch) − f (x0 ) (b) limh→0 = c f 0 (x0 ).3).1 — Fermat’s Rule. the sets B(a.1. (Hint: the key issue is differentiability at 0.

x→a x−a x→a x−a Similarly. f (x) − f (a) ≤ 0 for all x ∈ B− (a.1: Illustration of Fermat’s Rule. Then there exists c ∈ (a. Then f (x̄1 ) ≤ f (x) ≤ f (x̄2 ) for all x ∈ [a. b ∈ R with a < b and f : [a. Suppose f is continuous on [a. we have f (x) − f (a) ≥ 0 for all x ∈ B+ (a. Since B+ (a. b]. b]} and f (x̄2 ) = max{ f (x) : x ∈ [a. δ ).2. δ ). b] and differentiable on (a.  Theorem 4. x−a It follows that f (x) − f (a) f (x) − f (a) f 0 (a) = lim = lim− ≤ 0.4. f 0 (a) = 0. x→a x−a x→a x−a Therefore. b] → R. b]. (4. b) such that f 0 (c) = 0.108 4. b) with f (a) = f (b). The proof is similar for the case where f has a local maximum at a.3) Proof: Since f is continuous on the compact set [a. δ ). x−a Taking into account the differentiability of f at a yields f (x) − f (a) f (x) − f (a) f 0 (a) = lim = lim+ ≥ 0. Let a. δ ) is a subset of B(a. b] and x̄2 ∈ [a. Then there exists δ > 0 sufficiently small such that f (x) ≥ f (a) for all x ∈ B(a. (4. by the extreme value theorem (Theorem 3.2 — Rolle’s Theorem.4) . b] such that f (x̄1 ) = min{ f (x) : x ∈ [a.2) there exist x̄1 ∈ [a.2 THE MEAN VALUE THEOREM Figure 4. b]}. δ ). Proof: Suppose f has a local minimum at a.

b]. By (4. By Theorem 4.2: Illustration of Rolle’s Theorem. f 0 (x̄1 ) = 0 or f 0 (x̄2 ) = 0.3: Illustration of the Mean Value Theorem. b) such that f (b) − f (a) f 0 (c) = . (4. If x̄1 ∈ (a. b] → R.4). then f (x̄1 ) = f (x̄2 ) because f (a) = f (b).3 — Mean Value Theorem. 109 Figure 4. so f 0 (c) = 0 for any c ∈ (a.3) holds with c = x̄1 or c = x̄2 .2. b). b ∈ R with a < b and f : [a. b). b] and differentiable on (a.5) b−a Proof: The linear function whose graph goes through (a.2. If both x̄1 and x̄2 are the endpoints of [a. b). Theorem 4. f (a)) and (b. Figure 4. b) or x̄2 ∈ (a. b−a . f (b)) is f (b) − f (a) g(x) = (x − a) + f (a). then f has a local minimum at x̄1 or f has a local maximum at x̄2 . Let a.1. f is a constant function. and (4.  We are now ready to use Rolle’s Theorem to prove the Mean Value Theorem presented below. Suppose f is continuous on [a. Then there exists c ∈ (a.

b) such that h0 (c) = 0. Let f (x) = 1 + 4x for all x ≥ 2. Then f 0 (x) = cos x. there exists c ∈ (a. x].2. √  Example 4. there exists c ∈ (0. and h satisfies the assumptions of Theorem 4.2. Then 4 2 f 0 (x) = √ =√ .5) holds. .   Example 4. Since | cos c| ≤ 1 we conclude | sin x| ≤ |x| for all x > 0. (4.1 We show that | sin x| ≤ |x| for all x ∈ R. We apply the Mean Value Theorem to f on the interval [2. x−0 | sin x| Therefore. Another application of the Mean Value Theorem shows there exists c ∈ (x. 3 Rearranging terms provides the desired inequality. Since equality holds for |x| x = 0.110 4.2. x].2 THE MEAN VALUE THEOREM Define   f (b) − f (a) h(x) = f (x) − g(x) = f (x) − (x − a) + f (a) for x ∈ [a. Since f 0 (2) = 2/3 and f 0 (c) < f 0 (2) for c > 2 we conclude that √ 2 1 + 4x − 3 < (x − 2). = | cos c| ≤ 1. there exists c ∈ (2. 0) such that sin 0 − sin x = cos c. 0−x | sin x| Then. x > 0.2 We show that √ 1 + 4x < (5 + 2x)/3 for all x > 2. It follows that | sin x| ≤ |x| for x < 0. b−a it follows that f (b) − f (a) f 0 (c) − = 0. By the Mean Value Theorem applied to f on the interval [0. Since f (b) − f (a) h0 (x) = f 0 (x) − . Now. b−a Then h(a) = h(b). Then. Thus. b]. Next suppose |x| x < 0. Let f (x) = sin x for all x ∈ R. again. x) such that √ 1 + 4x − 3 = f 0 (c)(x − 2). since f (2) = 3.2. = | cos c|. b−a Thus. we conclude that | sin x| ≤ |x| for all x ∈ R. x) such that sin x − sin 0 = cos c. 2 1 + 4x 1 + 4x Now. fix x ∈ R. fix x ∈ R such that x > 2.

this implies (4.  The following theorem shows that the derivative of a differentiable function on [a. we introduce a couple of definitions. Then h(a) = f (b)g(a) − f (a)g(b) = h(b). in addition.6) Proof: Define h(x) = [ f (b) − f (a)]g(x) − [g(b) − g(a)] f (x) for x ∈ [a.2. there exists c ∈ (a. Since h0 (x) = [ f (b) − f (a)]g0 (x) − [g(b) − g(a)] f 0 (x). b] satisfies the intermediate value property although the derivative function is not assumed to be continuous. and h satisfies the assumptions of Theorem 4. b). b]. Thus.2. Theorem 4. and f : [a. a < b. b] → R. Definition 4. Theorem 4.6). both f+0 (a) and f−0 (b) exist. If the limit f (x) − f (a) lim x→a+ x−a exists. b] if f 0 (x) exists for each x ∈ (a. b] and f+0 (a) < λ < f−0 (b). 111 A more general result which follows directly from the Mean Value Theorem is known as Cauchy’s Theorem. b) such that h0 (c) = 0. Then there exists c ∈ (a. (4. To give the theorem in its greatest generality.2. b ∈ R with a < b.2. we say that f has a right derivative at a and write f (x) − f (a) f+0 (a) = lim+ .2.2 Let a. b ∈ R with a < b. Sup- pose f is differentiable on [a. x→a x−a If the limit f (x) − f (b) lim− x→b x−b exists. x→b x−b We will say that f is differentiable on [a. . Then there exists c ∈ (a. Let a. Suppose f and g are continuous on [a. b) such that [ f (b) − f (a)]g0 (c) = [g(b) − g(a)] f 0 (c). Let a. we say that f has a left derivative at b and write f (x) − f (b) f−0 (b) = lim− . b ∈ R. b) such that f 0 (c) = λ .4 — Cauchy’s Theorem. b] and differentiable on (a. b) and.5 — Intermediate Value Theorem for Derivatives.

Similarly. b]. b].2. b] and g0+ (a) < 0 < g0− (b). x→a x−a It follows that there exists δ1 > 0 such that g(x) < g(a) for all x ∈ (a. . b). it follows that c ∈ (a. it attains its minimum at a point c ∈ [a.2 THE MEAN VALUE THEOREM Figure 4. b] → R by g(x) = f (x) − λ x. b]. there exists δ2 > 0 such that g(x) < g(b) for all x ∈ (b − δ2 .112 4. Since g is continuous on [a.6 The same conclusion follows if f+0 (a) > λ > f−0 (b). Proof: Define the function g : [a.4: Right derivative.  Remark 4. g(x) − g(a) lim+ < 0. b) ∩ [a. a + δ1 ) ∩ [a. From the observations above. Thus. b]. equivalently. This implies g0 (c) = 0 or. that f 0 (c) = λ . Then g is differentiable on [a.

2.) x 4. 113 Exercises 4. for x > 1.2. Prove that the equation n x + ∑ (ak sin kx + bk cos kx) = 0 k=1 has a solution on (−π.1 B Let f and g be differentiable at x0 .3 I Prove that | sin(x) − sin(y)| ≤ |x − y| for all x.2. 4. n. √ (a) 1 + x < 1 + 12 x for x > 0. (b) ex > 1 + x. b]. Suppose f (x0 ) = g(x0 ) and f (x) ≤ g(x) for all x ∈ R . .2.5 B Let f and g be differentiable functions on [a. . 4. .4 B Let n be a positive integer and let ak . (Assume known that the derivative of ex is itself. Prove that there exists c ∈ (a. . (Assume known that the derivative of ln x is 1/x. Prove that f 0 (x0 ) = g0 (x0 ). y ∈ R.) x−1 (c) < ln x < x − 1. π). b]. 4. Suppose g(x) 6= 0 and g0 (x) 6= 0 for all x ∈ [a. b) such that .2 Prove the following inequalities using the Mean Value Theorem. bk ∈ R for k = 1.2. for x > 0.

.

.

.

1 .

f (a) f (b) .

.

= 1 .

f (c) g(c) .

.

.

g (c) ..

f 0 (c) g0 (c) .

g(b) − g(a) g(a) g(b) .

.

0 .

. . an satisfy n ak ∑ 2k + 1 = 0.6 BLet n be a fixed positive integer. 2 n Prove that the equation a1 + a2 x + a3 x2 + · · · + an xn−1 = 0 has a solution in (0. . (b) Suppose a0 . a2 .2. k=0 . 1). where the bars denote determinants of the two-by-two matrices. . a1 . (a) Suppose a1 . . . . 4. an satisfy a2 an a1 + + · · · + = 0. .

4. If f 0 (x) = 0 for all x ∈ (a.2. there exists c ∈ (a1 . b1 − a1 which is a contradiction. b). x2 − x1 This implies f (x1 ) < f (x2 ). b] and differentiable on (a. b]. By Theorem 4. Therefore. f (x) (a) Show that if limx→∞ f 0 (x) = a.2. b ∈ R and a < b. then limx→∞ = ∞.2 Let f be differentiable on (a. Prove that if both limx→∞ f (x) and limx→∞ f 0 (x) exist. Proof: Suppose by contradiction that f is not constant on [a. π2 ).2.  . we assume that a. In the proposition below. x (c) Are the converses in part (a) and part (b) true? 4. Fix any x1 . we show that it is possible to use the derivative to determine whether a function is constant.1 Let f be continuous on [a. (i) If f 0 (x) > 0 for all x ∈ (a. b]. The proof is based on the Mean Value Theorem. x2 ) such that f (x2 ) − f (x1 ) = f 0 (c) > 0. b). By Theorem 4. x f (x) (b) Show that if limx→∞ f 0 (x) = ∞. f is strictly increasing on (a. ∞) → R be a differentiable function. b). The proof of (ii) is similar.8 B Let f : [0. x2 ∈ (a. Proof: Let us prove (i). then f is constant on [a.3 SOME APPLICATIONS OF THE MEAN VALUE THEOREM Prove that the equation n ∑ ak cos(2k + 1)x = 0 k=0 has a solution on (0. b).3.2. b1 ) such that f (b1 ) − f (a1 ) f 0 (c) = 6= 0.3. Proposition 4. (ii) If f 0 (x) < 0 for all x ∈ (a. b). Then there exist a1 and b1 such that a ≤ a1 < b1 ≤ b and f (a1 ) 6= f (b1 ). there exists c ∈ (x1 . 114 4. b). b) with x1 < x2 . Proposition 4. b).3.  The next application of the Mean Value Theorem concerns developing simple criteria for monotonicity of real-valued functions based on the derivative. then limx→∞ = a.7 Let f : [0.3 SOME APPLICATIONS OF THE MEAN VALUE THEOREM In this section. then f is strictly increasing on (a. then limx→∞ f 0 (x) = 0 4. b). ∞) → R be a differentiable function.3. then f is strictly decreasing on (a.

k→∞ f (xk ) − f (x̄) f (x̄) xk − x̄ . Suppose f 0 (x) > 0 for all x ∈ (a.5 that f 0 (x) > 0 for all x ∈ (a. It remains to prove the differentiability of the inverse function f −1 and the representation of its derivative (4. Let g = f −1 .2). b).1 Let n ∈ N and f : [0. this shows that every positive real number has exactly one n-th root (refer to Example 3. ∞) → R be given by f (x) = xn . For each yk . Suppose f is differentiable on I = (a. Then f is one-to-one. f 0 (x) > 0 for all x > 0 and.3. f (I) is an open interval. g(yk ) = xk for all k. or f 0 (x) < 0 for all x ∈ (a. 1 ( f −1 )0 (y) = .  Example 4.4.11 that f (I) is an open interval and f −1 is continuous on f (I). y→ȳ y − ȳ f (x̄) Fix any sequence {yk } in f (I) that converges to ȳ and yk 6= ȳ for every k.5: Strictly Increasing Function.2. (4. there exists xk ∈ I such that f (xk ) = yk . It follows from Theorem 3. it is one-to-one. There- fore.4. In particular. That is. so. We will show that g(y) − g(ȳ) 1 lim = 0 . Proof: It follows from Theorem 4. Moreover. 115 Figure 4. Then f is strictly increasing on this interval and. Then g(yk ) − g(ȳ) xk − x̄ lim = lim k→∞ yk − ȳ k→∞ f (xk ) − f (x̄) 1 1 = lim = 0 . hence. Then f 0 (x) = nxn−1 .3.10 and Remark 3. b). and the inverse function f −1 : f (I) → I is differentiable. Theorem 4.4.3 — Inverse Function Theorem. f is strictly increasing. Fix any ȳ ∈ f (I) with ȳ = f (x̄). b) and f 0 (x) 6= 0 for all x ∈ (a.7). b). b). It follows from the continuity of g that {xk } converges to x̄.7) f 0 (x) where f (x) = y.

It is also clear that f ((0. (c) Find a nondifferentiable function that satisfies the condition above for α = 1. Prove that if f 0 (x) is bounded. then f is Lipschitz continuous and. then f is a constant function.8) (a) Prove that f is uniformly continuous on R.116 4. ∞)) = (0.2 Let n ∈ N and consider the function f : (0.   Example 4.3. ∞). f (x) g(x) Prove that f (1) g(1) >1> .4 Let f . ∞). Prove that f (x) ≤ g(x) for all x ≥ x0 . (4.3.2 I Let f : R → R.1 (a) Let f : R → R be differentiable. ∞) 1 1 ( f −1 )0 (y) = = . 4. uniformly continuous.3 SOME APPLICATIONS OF THE MEAN VALUE THEOREM The proof is now complete. n( y)n−1 n Exercises 4. ∞) is differentiable and given y ∈ (0. the value f −1 (y) is the unique positive real number whose n-th power is y.3. Suppose there exist ` ≥ 0 and α > 0 such that | f (u) − f (v)| ≤ `|u − v|α for all u. Then f is differentiable and f 0 (x) = nxn−1 6= 0 for all x ∈ (0. f 0 ( f −1 (y)) n( f −1 (y))n−1 Given y > 0. ∞) → R given by f (x) = xn . 4.3 B Let f and g be differentiable functions on R such that f (x0 ) = g(x0 ) and f 0 (x) ≤ g0 (x) for all x ≥ x0 . We call √ f −1 (y) the (positive) n-th root of y and denote it by n y. (b) Prove that if α > 1. g : R → R be differentiable functions satisfying (a) f (0) = g(0) = 1 f 0 (x) g0 (x) (b) f (x) > 0. v ∈ R . in particular.3. ∞) → R which is differentiable and uniformly continuous but such that f 0 (x) is not bounded. g(1) f (1) . We also obtain the formula 1 ( f −1 )0 (y) = √ . g(x) > 0 and > for all x. (b) Give an example of a function f : (0.3. 4. It follows from the Inverse Function Theorem that f −1 : (0. ∞) → (0.

{ck } converges to x̄. (4. By Theorem 4. δ ) ∩ [a. c) such that f 00 (d) < 0.1 Suppose f and g are continuous on [a.1 to show that 2x + sin x lim = 1.9) x→x̄ g(x) Proof: Let {xk } be a sequence in [a.4 L’HOSPITAL’S RULE We now prove a result that allows us to compute various limits by calculating a related limit involving derivatives. b]. b ∈ R with a < b.4. All four theorems in this section are known as l’Hospital’s Rule.1. with ck between xk and x̄. b]. Theorem 4. Suppose further that there exists δ > 0 such that g0 (x) 6= 0 for all x ∈ B(x̄. there exists a sequence {ck }. b] that converges to x̄ and such that xk 6= x̄ for every k. we have f (xk ) f 0 (ck ) = 0 . such that [ f (xk ) − f (x̄)]g0 (ck ) = [g(xk ) − g(x̄)] f 0 (ck ). x 6= x̄.2.1 We will use Theorem 4. Thus.2.9) follows from Theorem 3.4. we assume a.   Example 4. (4. b. g(xk ) g (ck ) Under the assumptions that g0 (x) 6= 0 for x near x̄ and g(x̄) = 0.4. b). c ∈ I with a < b < c such that f (a) < f (b) and f (b) > f (c). x→0 x2 + 3x . Prove that there exists d ∈ (a. k→∞ g(xk ) k→∞ g (ck ) x→x̄ g (x) Therefore. 4. and g0 (ck ) 6= 0 for sufficiently large k.3. b] and differentiable on (a. For this section. If f 0 (x) lim = `.5 B Let f be twice differentiable on an open interval I.3. Suppose that there exist a. where x̄ ∈ [a.1. x→x̄ g0 (x) then f (x) lim = `.6 B Prove that the function f defined in Exercise 4.11 is not monotone on any open interval containing 0.6).4. 4. f (xk ) f 0 (ck ) f 0 (x) lim = lim 0 = lim 0 = `. By the squeeze theorem (Theorem 2. Suppose f (x̄) = g(x̄) = 0. we also have g(xk ) 6= 0 for sufficiently large k. for each k. Since f (x̄) = g(x̄) = 0.1. 117 4.

the desired limit is 14 as well.118 4. x→1 4x4 − 2x − 2 Here f (x) = 3x3 − 2x2 + 4x − 5 and g(x) = 4x4 − 2x − 2.4.1 that limx→0 2x+sin x2 +3x x = 1 as we wanted to show. Now. b] f (x) and differentiable on (a.  Example 4. 1].1 hold. Moreover. δ ) ∩ [a.1 to get f 0 (x) 2 + cos x limx→0 2 + limx→0 cos x 2 + 1 lim = lim = = = 1. x→0 g0 (x) x→0 g00 (x) x→0 cos x Applying L’Hospital’s rule one more time we get x2 f (x) f 0 (x) lim = lim = lim 0 = 2. b] = [−1. b] \ {x̄}. g(x) is well defined on [a. and x̄ = 0. 9x2 − 4x + 4 9 lim = . Now g0 (x) = sin x and.4. f 00 (x) 2 lim 00 = lim = 2. Finally we calculate the limit of the quotient of derivatives using Theorem 3. g0 (x) 6= 0 for x near zero. we get g (x) = 2x + 3 6= 0 for x ∈ B(x̄. g(x) = x2 + 3x. f 0 (0) = 0 = g0 (0) and g00 (x) = cos x 6= 0 for x near 0.1 we may apply L’Hospital’s rule to determine first the limit of f 0 (x)/g0 (x) and then apply the rule again to determine the original limit. taking 0 δ = 7/3.4. We may take [a. so that f and g are continuous on [a. Here f (x) = 2x + sin x.  Example 4. Thus f (1) = g(1) = 0. x→0 1 − cos x Here f (x) = x2 and g(x) = 1 − cos x so both functions and all its derivatives are continuous. x→0 g (x) x→0 cos x By L’Hospital’s rule we get f 0 (x) f 00 (x) 2 lim = lim = lim = 2. for example. x 6= 0.2. b) and. Since g0 (1) = 14 6= 0 and g0 is continuous we have g0 (x) 6= 0 for x near 1. b]. Moreover. x→x̄ g0 (x) x→0 2x + 3 limx→0 2x + 3 3 It now follows from Theorem 4. f 0 (x) = 9x2 − 4x + 4 and g0 (x) = 16x3 − 2. so.4. x→0 1 − cos x x→0 g(x) x→0 g (x) . x→1 16x3 − 2 14 9 Thus.2 We will apply L’Hospital’s rule to determine the limit 3x3 − 2x2 + 4x − 5 lim . Also. Moreover.4 L’HOSPITAL’S RULE First we observe that the conditions of Theorem 4. furthermore. Consider the limit x2 lim .4.3 If the derivatives of the functions f and g themselves satisfy the assumptions of Theorem 4.

if x = 0. we can assume that f (x) 6= 0 and g(x) 6= 0 for all x ∈ B(x̄. (4. x→0 x + 3x2 x→0 1 + 3x limx→0 (1 + 3x) Theorem 4. f 0 is not continuous at 0 (since limx→0 f 0 (x) does not exist) and. b). b). if x 6= 0.11) x→x̄ g(x) Proof: Since limx→x̄ f (x) = limx→x̄ g(x) = ∞.4. However. if x = 0. x 6= x̄. f (x) = 0. x→0 g(x) x→0 x + 3x2 Using the derivative rules at x 6= 0 and the definition of derivative at x = 0 we can see that f is differentiable and ( 2x sin 1x − cos 1x . f (x) f (x) We will show that limx→x̄+ = `. L’Hospital’s rule cannot be applied in this case. Suppose f . and x̄ ∈ (a. b ∈ R. b) \ {x̄} → R are differentiable on (a. If ` ∈ R and f 0 (x) lim = `.4 Let g(x) = x + 3x2 and let f : R → R be given by ( x2 sin 1x . if x 6= 0. g : (a. (4. The proof that limx→x̄− = ` is completely analogous. one can choose K > 0 and a positive δ1 < δ such that . a < b. g(x) g(x) Fix any ε > 0. We need to find δ0 > 0 such that | f (x)/g(x) − `| < ε whenever x ∈ B+ (x̄. Suppose further that there exists δ > 0 such that g0 (x) 6= 0 for all x ∈ B(x̄. choosing a smaller positive δ if necessary. hence. δ ) ∩ (a.10). Now consider the limit f (x) x2 sin 1x lim = lim . δ0 ) ∩ (a. b) \ {x̄} and assume limx→x̄ f (x) = limx→x̄ g(x) = ∞.10) x→x̄ g0 (x) then f (x) lim = `.4. f 0 (x) = 0.2 Let a. 119  Example 4. b). b). x2 sin 1x On the other hand limx→0 x+3x2 does exist as we can see from x2 sin 1x x sin 1x limx→0 x sin 1x lim = lim = = 0. δ ) ∩ (a. From (4.

0 .

.

0 .

.

f (x) .

.

≤ K and .

f (x) − `.

< ε .

.

.

.

g0 (x) .

.

g0 (x) .

Fix α ∈ B+ (x̄. 2 (4. x 6= x̄. δ1 ) ∩ (a. b) = B+ (x̄. δ1 ) ∩ (a. for . δ2 ) ∩ (a. b) (in particular. Moreover. we can find δ2 > 0 such that δ2 < min{δ1 .12) whenever x ∈ B(x̄. α > x̄). δ2 ). α − x̄} and f (x) 6= f (α) for x ∈ B+ (x̄. Since limx→x̄ f (x) = ∞. b).

applying (4. 2K For any x ∈ B+ (x̄. we can write [ f (x)− f (α)]g0 (c) = [g(x) − g(α)] f 0 (c) for some c ∈ (x.12) we get that. g(x) g (c) Since c ∈ B(x̄ : δ1 ) ∩ (a.2. since g0 (z) 6= 0 if x < z < α. Rolle’s theorem (Theorem 4. γ). define g(α) 1− g(x) Hα (x) = for x ∈ B+ (x̄. for all x ∈ B+ (x̄. δ2 ). α) (note that. applying Theorem 4. Thus. Therefore. γ) ∩ (a. f (α) 1− f (x) Since limx→x̄ f (x) = limx→x̄ g(x) = ∞. . in particular.2. b)). c ∈ B(x̄ : δ1 ) ∩ (a.4 L’HOSPITAL’S RULE such x. δ2 ) we can write. for x ∈ B+ (x̄. we have that limx→x̄+ Hα (x) = 1. α]. γ) = B+ (x̄. b).2) guarantees that g(x) 6= g(α). g(x) g(x) − g(α) f (α) 1− f (x) Now. γ). For such c we get f (x) f 0 (c) = 0 Hα (x). g(α) 1− f (x) f (x) − f (α) g(x) = . there exists a positive γ < δ2 such that ε |Hα (x) − 1| < whenever x ∈ B+ (x̄. b).120 4.4 on the interval [x.

.

.

0 .

.

f (x) .

.

f (c) .

.

.

g(x) − `.

= .

.

.

g0 (c) H α (x) − `.

.

.

0 0 (c) .

.

f (c) f .

= .

.

0 (Hα (x) − 1) + 0 − `.

.

g (c) g (c) .

0 .

.

0 .

.

f (c) .

.

f (c) .

≤ .

0 .

|Hα (x) − 1| + .

0 .

.

.

− `.

.

  Example 4.5 Consider the limit ln x2 lim 1 . We have √ 2/x 3 5 x √3 lim 2 1 = lim −3 = lim −3 x2 = 0. We analyze the quotient of the derivatives. g(x) = 1 + √31 2 . x→0 1 + √ 3 2 x Here f (x) = ln x2 . limx→x̄ f (x) = limx→x̄ g(x) = ∞. x→0 − √ 3 3 5 x→0 x x→0 x . and we may take as (a.4. g (c) g (c) ε ε <K + = ε. b) any open inteval containing x 0. Moreover. Clearly f and g satisfy the differentiability assumptions and g0 (x) 6= 0 for all x 6= 0. x̄ = 0. 2K 2 Setting δ0 = γ completes the proof.

∞) and lim f (x) = lim g(x) = 0. the results can also be modified to include the case when x̄ is an endpoint of the domain of the functions f and g.4.4. Suppose g0 (x) 6= 0 for all x ∈ (a. x→∞ x→∞ If ` ∈ R and f 0 (x) lim = `.2 that ln x2 lim 1 = 0.4. Theorem 4.4.4.5 Let f and g be differentiable on (a.4 Let f and g be differentiable on (a.1. x→∞ g0 (x) then f (x) lim = `. Moreover.4. x→∞ g(x) . The following theorem can be proved following the method in the proof of Theorem 4.1 and Theorem 4.4.4. x→∞ x( π − arctan x) 2 1/x Writing the quotient in the form π we can apply Theorem 4.3 The proofs of Theorem 4. x→∞ − 21 x→∞ x2 x +1 In view of Theorem 4. x→∞ g(x)  Example 4.2 show that the results in these theorems can be applied for left-hand and right-hand limits. x→0 1 + √ 3 2 x Remark 4.6 Consider the limit 1 lim .4. ∞) and lim f (x) = lim g(x) = ∞.2. ∞). The following theorem can be proved following the method in the proof of Theorem 4. Suppose g0 (x) 6= 0 for all x ∈ (a.4 the desired limit is also 1.4. 121 It now follows from Theorem 4. x→∞ x→∞ If ` ∈ R and f 0 (x) lim = `.4. ∞). Theorem 4.4. x→∞ g0 (x) then f (x) lim = `. We now compute the limit 2 −arctan x of the quotient of the derivatives −1/x2 x2 + 1 lim = lim = 1.

3 Use the relevant version of L’Hospital’s rule to compute each of the following limits. x→∞ x Clearly the functions f (x) = ln x and g(x) = x satisfy the conditions of Theorem 4.4. 3x2 + 2x + 7 (a) lim . x→0 x − tan x 4.7 Consider the limit ln x lim . x→0 sin x cos x x−1 (c) lim √ √ .5 that limx→∞ lnxx = 0.2 For the problems below use L’Hospital’s rule as many times as appropriate to determine the limits. x→0 ln(1 + x) ln x (e) lim .1 Use L’Hospital’s Rule to find the following limits (you may assume known all the relevant derivatives from calculus): x3 − 4x (a) lim . We have f 0 (x) 1/x lim = lim =0 x→∞ g0 (x) x→∞ 1 It follows from Theorem 4. x→0 x3 x − sin x (d) lim .4. x→1 sin(πx) 4. x→0 cot x . x→1 x + 1 − 2 ex − e−x (d) lim .4.4.4 L’HOSPITAL’S RULE  Example 4. x→0 1 − sin x x − arctan x (c) lim . x→0 x sin x (x − π2 )2 (b) lim . 1 − cos 2x (a) lim . x→−2 3x2 + 5x − 2 ex − e−x (b) lim .4.4. Exercises 4. x→∞ 4x2 − 6x + 1 − ln x (b) lim+ .122 4.5.

we prove a result that lets us approximate differentiable functions by polynomials. (Hint: first rewrite as a quotient. k=0 k! Proof: Let x̄ be as in the statement and let us fix x 6= x̄.  x2 e−x2 . (b) f (x) = π x−1  sign x + . .) x→∞ 4. if x 6= 0.4 Prove that the following functions are differentiable at 1 and -1.4. Prove that lim P(x)e−x = 0. b]. b] → R is a function such that f (n) is continuous on [a. b] with x 6= x̄. 4. if |x| > 1. if x = 0. b].5 B Let P(x) be a polynomial. Let x̄ ∈ [a. Suppose f : [a. 4 2 4. (a) f (x) = 1  . f (x) = 0. Theorem 4. Prove that f ∈ Cn (R) for every n ∈ N. ln(1 + 1x ) x→∞ √ (d) lim xe−x .6 I Consider the function ( 1 e− x2 . Let n be a positive integer. Since x − x̄ 6= 0.1 — Taylor’s Theorem.5 TAYLOR’S THEOREM In this section. b). 123 π 2 − arctan x (c) lim . if |x| ≤ 1. and f (n+1) (x) exists for all x ∈ (a.4. if |x| > 1. e  arctan x. there exists a number c in between x̄ and x such that f (n+1) (c) f (x) = Pn (x) + (x − x̄)n+1 . Then for any x ∈ [a.5. (n + 1)! We will now show that λ = f (n+1) (c). (n + 1)! where n f (k) (x̄) Pn (x) = ∑ (x − x̄)k . x→∞ 4. there exists a number λ ∈ R such that λ f (x) = Pn (x) + (x − x̄)n+1 . if |x| ≤ 1.4.

Suppose f (n) exists and continuous on (a. Since f 0 (x) = cos x. there exists c in between x̄ and x such that g0 (c) = 0. b).3 Let n be an even positive integer. c = x = x̄. π/2].1 We will use Taylor’s theorem to estimate the error in approximating the function f (x) = sin x with it 3rd Taylor polynomial at x̄ = 0 on the interval [−π/2.5.066.124 4. for x ∈ [−π/2. for any c ∈ R we have | f (4) (c)| = | sin c| ≤ 1.5. The following hold: . k=0 k! (n + 1)! Then n f (k) (x̄) λ λ g(x̄) = f (x) − ∑ (x − x̄)k − (x − x̄)n+1 = f (x) − Pn (x) − (x − x̄)n+1 = 0. Consider the function n f (k) (t) λ g(t) = f (x) − ∑ (x − t)k − (x − t)n+1 .  Example 4.5. 3! More over. . 4! 4! Theorem 4. .5 TAYLOR’S THEOREM for some c in between x̄ and x. Remark 4. π/2] we get (for some c between x and 0). Let x̄ ∈ (a. The proof is now complete. = f (n−1) (x̄) = 0 and f (n) (x̄) 6= 0. we have ! n 0 0 f (k+1) (c) k f (k) (c) k−1 λ g (c) = − f (c) + ∑ − (x − c) + (x − c) + (x − c)n k=1 k! (k − 1)! n! λ 1 = (x − c)n − f (n+1) (c)(x − c)n n! n! = 0. b) satisfy f 0 (x̄) = . a direct calculation shows that x3 P3 (x) = x − . Therefore.2 The conclusion of Taylor’s theorem still holds true if x = x̄.  The polynomial Pn (x) given in the theorem is called the n-th Taylor polynomial of f at x̄. Taking the derivative of g (keeping in mind that x is fixed and the independent variable is t) and using the product rule for derivatives. This implies λ = f (n+1) (c). | f (4) (c)| π/2 | sin x − P3 (x)| = |x| ≤ ≤ 0. k=0 k! (n + 1)! (n + 1)! and n f (k) (x) λ g(x) = f (x) − ∑ (x − x)k − (x − x)n+1 = f (x) − f (x) = 0. In this case. f 00 (x) = − sin x and f 000 (x) = − cos x. k=0 k! (n + 1)! By Rolle’s theorem.

5.5. δ ) for sufficiently large k. Then there exists δ > 0 such that f (x) ≥ f (x̄) for all x ∈ B(x̄.  Example 4. b) that converges to x̄ with xk 6= x̄ for every k. b). Proof: We will prove (a). Suppose f (n) (x̄) > 0. Since {ck } converges to x̄. Now. n! Since n is even and c ∈ B(x̄. for the converse. n! Since xk ∈ B(x̄. f has a local minimum at x̄. (b) f (n) (x̄) < 0 if and only if f has a local maximum at x̄. by the way. It follows from the previous theorem that f has a local minimum at 0. we have f (x) ≥ f (x̄). A direct calculations shows f 0 (1) = f 00 (1) = f 000 (1) = f (4) (1) = 0 and f (5) (1) < 0. suppose that f has a local minimum at x̄. Fix a sequence {xk } in (a. It follows from the previous theorem that f has a local maximum at 1. Then f (0) = f 0 (0) = 0 and f 00 (0) = 2 > 0. with ck between xk and x̄ for each k. Then f 0 (x) = 2x cos x −x2 sin x and f 00 (x) = 2 cos x − 4x sin x − x2 cos x. Thus. δ ). such that f (n) (ck ) f (xk ) = f (x̄) + (xk − x̄)n . δ ). there exists a sequence {ck }. It follows that f (n) (ck ) f (xk ) − f (x̄) = (xk − x̄)n ≥ 0. By Taylor’s theorem. we have f (xk ) ≥ f (x̄) for such k. n! This implies f (n) (ck ) ≥ 0 for such k. 0 is not a global minimum.3 Consider the function f (x) = −x6 + 2x5 + x4 − 4x3 + x2 + 2x − 3 defined on R. there exists c in between x̄ and x such that f (n) (c) f (x) = f (x̄) + (x − x̄)n . Fix any x ∈ B(x̄. . that since f (0) = 0 and f (π) < 0. f (n) (x̄) = limk→∞ f (n) (ck ) ≥ 0. δ ) ⊂ (a. δ ) ⊂ (a. 125 (a) f (n) (x̄) > 0 if and only if f has a local minimum at x̄.2 Consider the function f (x) = x2 cos x defined on R. Since f (n) (x̄) > 0 and f (n) is continuous at x̄. b).   Example 4. By Taylor’s theorem and the given assumption. Notice. there exists δ > 0 such that f (n) (t) > 0 for all t ∈ B(x̄. The proof of (b) is similar.

4.5.5. at x̄ = 0 for cos x.5. f (x + h) + f (x − h) − 2 f (x) lim = f 00 (x).6 CONVEX FUNCTIONS AND DERIVATIVES Exercises 4. (a) f (x) = x3 sin x at x̄ = 0. This is called the Taylor expansion with Peano’s remainder. we have f (x̄ + h) = Pn (h) + g(h).4 Suppose f is twice differentiable on (a.) 4.3 to determine if the following functions have a local minimum or a local maximum at the indicated points.5. 126 4. 4. wher g is a function that satisfies limh→0 g(h) hn = 0. π/2]. b).6 Suppose f is n times differentiable on (a. b). P5 (x). h→0 h2 4. Determine an upper bound for the error |P5 (x) − cos x| for x ∈ [−π/2. b).5 I (a) Suppose f is three times differentiable on (a. 4. h→0 hn (Thus.2 Find the 5th Taylor polynomial.5. b).1 B Use Taylor’s theorem to prove that m xk ex > ∑ k! k=0 for all x > 0 and m ∈ N. Prove that 2 f (x̄ + h) − f (x̄) − f 0 (x̄) 1!h − f 00 (x̄) h2! f 000 (x̄) lim = . b). . h→0 h3 3! (b) State and prove a more general result for the case where f is n times differentiable on (a.5.6 CONVEX FUNCTIONS AND DERIVATIVES We discuss in this section a class of functions that plays an important role in optimization problems.5. b) and x̄ ∈ (a. 4. (b) f (x) = (1 − x) ln x at x̄ = 1. Define n hn Pn (h) = ∑ f (n) (x̄) n! for h ∈ R k=0 Prove that f (x̄ + h) − Pn (h) lim = 0. Show that for every x ∈ (a. b) and x̄ ∈ (a.3 Use Theorem 4.

(c) f : R → R. we only need to prove the implication by induction. This follows from the triangle inequality and other basic properties of absolute value. The conclusion holds for n = 2 by the definition of convexity. .1 Let I be an interval of R. . i=1 . Definition 4. . . n. Theorem 4. . we get f (λ x + (1 − λ )y) = (λ x + (1 − λ )y)2 = λ 2 x2 + 2λ (1 − λ )xy + (1 − λ )2 y2 ≤ λ 2 x2 + λ (1 − λ )(x2 + y2 ) + (1 − λ )2 y2 = λ (λ x2 + (1 − λ )x2 ) + (1 − λ )(λ y2 + (1 − λ )y2 ) = λ x2 + (1 − λ )y2 = λ f (x) + (1 − λ ) f (y).6.6: A Convex Function. . i = 1. Then k ∑ λi = 1 − λk+1 . if 0 < λ < 1 and x.1 The following functions are convex. f (x) = |x|. k + 1.6. Fix λi ≥ 0. i = 1. Let k be such that the conclusion holds for any n with 2 ≤ n ≤ k. . . i = 1. . y. We say that f is convex on I if f (λ u + (1 − λ )v) ≤ λ f (u) + (1 − λ ) f (v) for all u. f (x) = x2 . A function f : I → R is convex if and only if for every λi ≥ 0. with ∑ni=1 λi = 1 (n ≥ 2) and for every xi ∈ I. .1 Let I be an interval of R and let f : I → R. (4. k + 1. Here note first that 2xy ≤ x2 + y2 for all real numbers x. This is straightforward.13) i=1 i=1 Proof: Since the converse holds trivially. 127 Figure 4.6. n. with k+1 ∑i=1 λi = 1 and fix every xi ∈ I. 1). . (a) f : R → R. f (x) = x. i = 1. . y ∈ R. . ! n n f ∑ λ i xi ≤ ∑ λi f (xi ). . . Then. We will show that it also holds for n = k + 1. .  Example 4. v ∈ I and for all λ ∈ (0. (b) f : R → R.

hence. Thus. for such n.6 CONVEX FUNCTIONS AND DERIVATIVES If λk+1 = 1. For any x ∈ I.  Theorem 4. . . k.  Theorem 4.6. . Suppose f is differentiable at x̄.128 4. this shows f has an absolute minimum at x̄.2 Let I be an interval and let f : I → R be a convex function. Proof: Clearly if f has a global minimum at x̄. then λi = 0 for all i = 1. λi /(1 − λk+1 ) ≥ 0 and k λi ∑ 1 − λk+1 = 1. δ ) ∩ I. Then f 0 (x̄)(x − x̄) ≤ f (x) − f (x̄) for all x ∈ I. for each i = 1. . 1 1 f (x̄) ≤ f (xn ) ≤ (1 − ) f (x̄) + f (x). Since x was arbitrary. The proof is now complete. xn ∈ B(x̄. Then f has a local minimum at x̄ if and only if f has an absolute minimum at x̄. then it also has a local minimum at x̄. Then there exists δ > 0 such that f (u) ≥ f (x̄) for all u ∈ B(x̄. δ ) ∩ I when n is sufficiently large. . we have xn = (1 − n1 )x̄ + 1n x → x̄.6. f (x̄) ≤ f (x). n n This implies that for a sufficient large n. Then. Suppose 0 ≤ λk+1 < 1. . suppose that f has a local minimum at x̄. Conversely. .3 Let I be an open interval and let f : I → R be a convex function. i=1 where the first inequality follows from the definition of convexity (or is trivial if λk+1 = 0) and the last inequality follows from the inductive assumption. . k. (4. Thus.13) holds.14) . we have 1 1 f (x̄) ≤ f (x) n n and. i=1 It follows that ! k+1 ∑ki=1 λi xi   f ∑ λi xi = f (1 − λk+1 ) + λk+1 xk+1 i=1 1 − λk+1  k  ∑i=1 λi xi ≤ (1 − λk+1 ) f + λk+1 f (xk+1 ) 1 − λk+1 ! k λi = (1 − λk+1 ) f ∑ xi + λk+1 f (xk+1 ) i=1 1 − λk+1 k λi ≤ (1 − λk+1 ) ∑ f (xi ) + λk+1 f (xk+1 ) i=1 1 − λk+1 k+1 = ∑ λi f (xi ). and (4.

Then f has an absolute minimum at x̄ if and only if f 0 (x̄) = 0. b−a . f has an absolute minimum at x̄. By Theorem 4. x−a b−a b−x Proof: Let x−a t= . we obtain f (x) ≤ t f (b) + (1 − t) f (a).1. t→0 t which completes the proof.  Corollary 4. b−a By convexity of f . f (x̄ + t(x − x̄)) − f (x̄) f 0 (x̄)(x − x̄) = lim+ ≤ f (x) − f (x̄). Suppose f 0 (x̄) = 0.6. Proof: Suppose f has an absolute minimum at x̄.6. Fix a. b−a Then t ∈ (0.6. 1). 1) and   x−a f (x) = f (a + (x − a)) = f a + (b − a) = f (a + t(b − a)) = f (tb + (1 − t)a). Let us prove the converse. This implies f (x) ≥ f (x̄) for all x ∈ I.3 that 0 = f 0 (x̄)(x − x̄) ≤ f (x) − f (x̄) for all x ∈ I. Then f (x) − f (a) f (b) − f (a) f (b) − f (x) ≤ ≤ . b.  Lemma 4. Suppose f is differentiable at x̄. 129 Proof: For any x ∈ I and t ∈ (0. f 0 (x̄) = 0. It follows from Theorem 4. Thus.5 Let I be an open interval and suppose f : I → R is a convex function. x−a f (x) − f (a) ≤ t f (b) + (1 − t) f (a) − f (a) = t[ f (b) − f (a)] = ( f (b) − f (a)).2. x ∈ I with a < x < b. we have f (x̄ + t(x − x̄)) − f (x̄) f (tx + (1 − t)x̄) − f (x̄) = t t t f (x) + (1 − t) f (x̄) − f (x̄) ≤ t = f (x) − f (x̄). Since f is differentiable at x̄. Thus.4 Let I be an open interval and let f : I → R be a convex function.

Let us prove the converse. f (b) − f (a) ≤ f 0 (b). By Lemma 4.6. This implies t f (xt ) − t f (x1 ) = f 0 (c1 )t(1 − t)(x2 − x1 ).  Theorem 4. 1). for any x ∈ (a. x−a b−a Similarly. b−a Similarly. . f 0 (a) ≤ f 0 (b). Fix a < b with a.5. Fix x1 < x2 and t ∈ (0.6. where xt = tx1 + (1 − t)x2 . x−b f (x) − f (b) ≤ t f (b) + (1 − t) f (a) − f (b) = (1 − t)[ f (a) − f (b)] = [ f (b) − f (a)].6 CONVEX FUNCTIONS AND DERIVATIVES Equivalently. b−a It follows that f (b) − f (a) f (b) − f (x) ≤ .6 Let I be an open interval and let f : I → R be a differentiable function. b−a Therefore. By the Mean Value Theorem (Theorem 4. Suppose f 0 is increasing. (1 − t) f (xt ) − (1 − t) f (x2 ) = f 0 (c2 )t(1 − t)(x1 − x2 ). x−a b−a This implies.130 4. b−a b−x The proof is now complete. b ∈ I. there exist c1 and c2 such that x1 < c1 < xt < c2 < x2 with f (xt ) − f (x1 ) = f 0 (c1 )(xt − x1 ) = f 0 (c1 )(1 − t)(x2 − x1 ). we have f (x) − f (a) f (b) − f (a) ≤ . b). f (x) − f (a) f (b) − f (a) ≤ . that f (b) − f (a) f 0 (a) ≤ . taking limits. Then x1 < xt < x2 .2. Then f is convex if and only if f 0 is increasing on I. and f 0 is an increasing function. Proof: Suppose f is convex. f (xt ) − f (x2 ) = f 0 (c2 )(xt − x2 ) = f 0 (c2 )t(x1 − x2 ).3).

The conclusion then follows directly from Theorem 4. δ ). v ∈ B(x̄.15) In particular. b−a u−a u−v d −v d −c Using a similar approach for the case u < v. b. b−a u−v d −c Then (4. Suppose f is twice differentiable on I. Therefore. d satisfying a < b < x̄ < c < d with a. Choose δ > 0 such that B(x̄.2 that f 00 (x) ≥ 0 for all x ∈ I if and only if the derivative function f 0 is increasing on I. b−a u−v d −c Choose ` ≥ 0 sufficiently large so that f (b) − f (a) f (u) − f (v) f (d) − f (c) −` ≤ ≤ ≤ ≤ ` for all u. c.8 Let I be an open interval and let f : I → R be a convex function. δ ) with v < u. v ∈ B(x̄.7 Let I be an open interval and let f : I → R be a function. v ∈ B(x̄. Since f 00 (x) ≥ 0 for all x.  . we see that f (b) − f (a) f (u) − f (a) f (u) − f (v) f (d) − f (v) f (d) − f (c) ≤ ≤ ≤ ≤ .  √  Example 4. it follows from the corollary that f is convex. Then it is locally Lipschitz continuous in the sense that for any x̄ ∈ I. δ ).6. The proof is now complete. c). f is convex. (4. f is continuous.2 Consider the function f : R → R given by f (x) = √ x2 + 1. we have t f (xt ) −t f (x1 ) = f 0 (c1 )t(1 −t)(x2 − x1 ) ≤ f 0 (c2 )t(1 −t)(x2 − x1 ) = (1 −t) f (x2 ) − (1 −t) f (xt ). δ ) ⊂ (b. Then by Lemma 4. 131 Since f 0 (c1 ) ≤ f 0 (c2 ). we get f (b) − f (a) f (u) − f (v) f (d) − f (c) ≤ ≤ for all u. there exist ` ≥ 0 and δ > 0 such that | f (u) − f (v)| ≤ `|u − v| for all u. Proof: Fix any x̄ ∈ I. Rearranging terms.6. Let u.15) holds.5.  Corollary 4. Now. f 0 (x) = x/ x2 + 1 and f 00 (x) = 1/(x2 + 1)3/2 .6. Proof: It follows from Proposition 4. we get f (xt ) ≤ t f (x1 ) + (1 − t) f (x2 ). Then f is convex if and only if f 00 (x) ≥ 0 for all x ∈ I. Theorem 4.6. v ∈ B(x̄. Choose four numbers a.6. δ ). The proof is now complete.6.3. d ∈ I.

A number u ∈ R is called a subderivative of the function f at x̄ if u · (x − x̄) ≤ f (x) − f (x̄) for all x ∈ R . 4. where c ≥ 0 is a constant.3 I Let I be an interval and let f : I → R be a convex function.6. are nonnegative real numbers. 4.16) The set of all subderivatives of f at x̄ is called the subdifferential of f at x̄ and is denoted by ∂ f (x̄).2 Let f : R → R be a convex function. . where n ≥ 2. x ∈ [0. π4 ).7. Prove that φ ◦ f is convex on I. where b is a constant. 2 (b) If a1 . (4. g} are convex functions on I. .  x  e (d) f (x) = − ln . 4. then a1 + a2 + · · · + an ≥ (a1 · a2 · · · an )1/n . Suppose that φ is a convex. (b) Find two convex functions f and g on an interval I such that f · g is not convex. 132 4. a2 .7 NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBDIFFERENTIALS Exercises 4. an . we introduce a new concept that is helpful in the study of optimization problems in which the objective function may fail to be differentiable. x ∈ R.1 Let f : R → R be a convex function. 4. . b ∈ R.6. increasing function on an interval J that contains f (I). x ∈ (−∞. prove that the function defined by g(x) = f (ax + b). .5 B Prove the following: (a) If a. 1).7 NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBDIFFERENTIALS In this section. for x ∈ R is also a convex function on R. g : I → R be convex functions. . x ∈ R. Definition 4. 1 + ex (e) f (x) = x sin x. x ∈ (− π4 . (b) f (x) = xk . then a+b √ ≥ ab. f + g.1 (a) Let I be an interval and let f .6.6. Prove that c f . ∞) and k ≥ 1 is a constant. and max{ f .6. n 4. b are nonnegative real numbers. (c) f (x) = − ln(1 − x). Given a.4 B Prove that each of the following functions is convex on the given domain: (a) f (x) = ebx .

u · 1 ≤ |1| = 1 and u · (−1) = −u ≤ | − 1| = 1. Fix a ∈ R.6. Then u · x ≤ |u · x| = |u||x| ≤ |x| for all x ∈ R . Thus.7.1 Let f : R → R be a convex function. ∞) with x1 < x2 . ∂ f (0) = [−1. In particular. we have u · x = u(x − 0) ≤ f (x) − f (0) = |x| for all x ∈ R . x<x̄ x>x̄ where φx̄ is defined in (4. Proof: This lemma follows directly from Lemma 4.1 Let f (x) = |x|. x2 ∈ (−∞.7. u ∈ [−1. This implies u ∈ ∂ f (0). a) ∪ (a. Lemma 4.17). . Then. for any u ∈ ∂ f (0). 1]. 1]. ∞).7. we have |u| ≤ 1.  Example 4. Therefore.7: A nondifferential convex function. for x1 . sup φx̄ (x) = f−0 (x̄) ≤ f+0 (x̄) = inf φx̄ (x). Moreover. we have φa (x1 ) ≤ φa (x2 ). It follows that ∂ f (0) ⊂ [−1.5. a) ∪ (a. Define the slope function φa by f (x) − f (a) φa (x) = (4. Then f has left derivative and right derivative at x̄. For any u ∈ [−1. Then ∂ f (0) = [−1.  Theorem 4.17) x−a for x ∈ (−∞. 1]. Indeed.2 Let f : R → R be a convex function and let x̄ ∈ R. 133 Figure 4. 1]. 1].

x→x̄+ x>x̄ Thus.134 4.18) Proof: Suppose u ∈ ∂ f (x̄).2. This implies f−0 (x̄) ≤ f+0 (x̄). Then ∂ f (x̄) = [ f−0 (x̄). x>x̄ Similarly.1.1 again. we see that φx̄ (x) ≤ φx̄ (y) whenever x < x̄ < y. x − x̄ .7 NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBDIFFERENTIALS Figure 4.4. f+0 (x̄)].7. ∞) and bounded below by φx̄ (x̄ − 1).  Theorem 4. the slope function φx̄ defined by (4. Proof: By Lemma 4. we have u · (x − x̄) ≤ f (x) − f (x̄) for all x > x̄. Moreover.8: Definition of subderivative. x<x̄ Applying Lemma 4. The proof is complete.3 Let f : R → R be a convex function and let x̄ ∈ R. (4.17) is increasing on the interval (x̄.7. This implies f (x) − f (x̄) u≤ for all x > x̄. f+0 (x̄) exists and f+0 (x̄) = inf φx̄ (x). lim φx̄ (x) = inf φx̄ (x).7. By the definition (4. f−0 (x̄) exists and f−0 (x̄) = sup φx̄ (x). By Theorem 3.16). the limit f (x) − f (x̄) lim φx̄ (x) = lim+ x→x̄+ x→x̄ x − x̄ exists and is finite.

By Theorem 4. we must have f−0 (x̄) = f+0 (x̄).16) holds and.  .  Corollary 4. f (x) − f (x̄) u ≤ lim+ = f+0 (x̄). Then f−0 (x̄) = f+0 (x̄) = f 0 (x̄). ∂ f (x̄) = [ f−0 (x̄).7. x − x̄ It follows that u · (x − x̄) ≤ f (x) − f (x̄) for all x ≥ x̄.3. 135 Thus.7. hence. we have u · (x − x̄) ≤ f (x) − f (x̄) for all x < x̄. one has f (x) − f (x̄) u ≤ φx̄ (x) = for all x > x̄. Then f is differentiable at x̄ if and only if ∂ f (x̄) is a singleton. f is differentiable at x̄. f (x) − f (x̄) u≥ for all x < x̄.4 Let f : R → R be a convex function and x̄ ∈ R. ∂ f (x̄) is a singleton. x→x̄ x − x̄ Similarly. u ∈ ∂ f (x̄). So ∂ f (x̄) ⊂ [ f−0 (x̄). By Theorem 4. Thus. (4. f+0 (x̄)] = { f 0 (x̄)}. Conversely. Proof: Suppose f is differentiable at x̄. take u ∈ [ f−0 (x̄). one also has u · (x − x̄) ≤ f (x) − f (x̄) for all x < x̄. f+0 (x̄)]. To prove the opposite inclusion. Thus. Similarly.18) holds. f+0 (x̄)].7. ∂ f (x̄) = { f 0 (x̄)}. Thus.2 sup φx̄ (x) = f−0 (x̄) ≤ u ≤ f+0 (x̄) = inf φx̄ (x). In this case. Therefore. if ∂ f (x̄) is a singleton. Thus. (4. x − x̄ This implies u ≥ f−0 (x̄). x<x̄ x>x̄ Using the upper estimate by f+0 (x̄) for u.

if x > b.7 NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBDIFFERENTIALS  Example 4. g : R → R be convex functions and let α > 0. ∂ f (b) = [−a. if x = b. we have  {−a}.3.7. Then f is a convex function and f−0 (b) = −a. Since f is differentiable on (−∞. f+0 (b) = a. The proof for the second formula is similar.7. g0+ (x̄)] = ∂ f (x̄) + ∂ g(x̄). ∂ ( f + g)(x̄) = [( f + g)0− (x̄). Thus.5 Let f . Proof: It is not hard to see that f + g is a convex function and ( f + g)0+ (x̄) = f+0 (x̄) + g0+ (x̄) ( f + g)0− (x̄) = f−0 (x̄) + g0− (x̄).  ∂ f (x) = [−a. ( f + g)0+ (x̄)] = [ f−0 (x̄) + g0− (x̄). where a > 0. a]. b ∈ B} and αA = {αa : a ∈ A}.7.136 4.2 Let f (x) = a|x − b| + c. Then f + g and α f are convex functions and ∂ ( f + g)(x̄) = ∂ f (x̄) + ∂ g(x̄) ∂ (α f )(x̄) = α∂ f (x̄). if x < b.  {a}. Figure 4. f+0 (x̄) + g0+ (x̄)] = [ f−0 (x̄). By Theorem 4. f+0 (x̄)] + [g0− (x̄). ∞).7. b) and (b. a]. Theorem 4.2 Let A and B be two nonempty subsets of R and let α ∈ R.  Definition 4. Define A + B = {a + b : a ∈ A.  .9: Set addition.

.7. . ∂ f (x̄) = [m. . M]. The following result may be considered as a version of the first derivative test for extrema in the case of non differentiable functions. . Similarly we verify that f+0 (x̄) = M and f−0 (x̄) = m. . . Define f (x) = max{ fi (x) : i = 1. but h(x) = x3 is not a convex function.7. n. .7.5. i=1 Then f is a convex function. f+0 (x̄)]. Then f has an absolute minimum at x̄ if and only if 0 ∈ ∂ f (x̄) = [ f−0 (x̄).7. M]. f is a convex function. Theorem 4. where 0 0 m = min fi− (x̄) and M = max fi+ (x̄). n. we get ( ∑ai <x̄ µi − ∑ai >x̄ µi . Define n f (x) = ∑ µi |x − ai |. This implies f (λ u + (1 − λ )v) = max fi (λ u + (1 − λ )v) ≤ λ f (u) + (1 − λ ) f (v).  Remark 4. n : fi (u) = f (u)}.3 Let a1 < a2 < · · · < an and let µi > 0 for i = 1. if x̄ ∈ / {a1 . i∈I(x̄) i∈I(x̄) Proof: Fix u. we have fi (λ u + (1 − λ )v) ≤ λ fi (u) + (1 − λ ) fi (v) ≤ λ f (u) + (1 − λ ) f (v). . .7 The product of two convex functions is not a convex function in general.3. .7. For instance. n. . Then f is a convex function. . By Theorem 4. f (x) = x and g(x) = x2 are convex functions. Moreover. By Theorem 4. . i = 1. 1). The proof is now complete. . . Then f (x̄) ≤ f (x) for all x ∈ R . an } ∂ f (x̄) = ∑ai <x̄ µi − ∑ai >x̄ µi + [−µi0 . µi0 ]. v ∈ R and λ ∈ (0. . Proof: Suppose f has an absolute minimum at x̄.6 Let fi : R → R. n} and I(u) = {i = 1.7. . . . . . . For any i = 1. if x̄ = ai0 . .8 Let f : R → R be a convex function. Theorem 4. . 137  Example 4. be convex functions. 1≤i≤n Thus. . ∂ f (x̄) = [m.

7.11: Subdifferential of g(x) = ∑2k i=1 |x − ai |. ak+1 ]. Conversely. i=1 Then 0 ∈ ∂ g(x̄) if and only if x̄ ∈ [ak . 0 · (x − x̄) = 0 ≤ f (x) − f (x̄) for all x ∈ R . It follows from (4.10: Subdifferential of f (x) = ∑2k−1 i=1 |x − ai |. if a1 < a2 < · · · < a2k and 2k g(x) = ∑ |x − ai |. f has a unique absolute minimum at ak . It follows from the subdifferential formula in Example 4. Define 2k−1 f (x) = ∑ |x − ai |.16) that 0 ∈ ∂ f (x̄). Thus. ak+1 ]. g has an absolute minimum at any point of [ak . by (4.138 4. f has an absolute minimum at x̄.3 that 0 ∈ ∂ f (x̄) if and only if x̄ = ak .7 NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBDIFFERENTIALS This implies 0 · (x − x̄) = 0 ≤ f (x) − f (x̄) for all x ∈ R . Thus.7. Figure 4.16). Figure 4. if 0 ∈ ∂ f (x̄). i=1 for x ∈ R. Thus. . again.   Example 4.4 Let k be a positive integer and a1 < a2 < · · · < a2k−1 . Similarly.

9 Let f : R → R be a convex function and let a < b. Theorem 4.2. b) such that f (b) − f (a) ∈ ∂ f (c). Figure 4. g has a local minimum at some c ∈ (a. 139 The following theorem is a version of the Mean Value Theorem (Theorem 4.   0 f (b) − f (a) ∂ h(c) = {h (c)} = − . Observe that the function   f (b) − f (a) h(x) = − (x − a) + f (a) b−a is differentiable at c and.7. g also has an absolute minimum at c. b) and. (4. b−a .3) for nondifferen- tiable functions.19) b−a Proof: Define   f (b) − f (a) g(x) = f (x) − (x − a) + f (a) .12: Subdifferential mean value theorem. hence. hence. b−a Then g is a convex function and g(a) = g(b). Thus. Then there exists c ∈ (a.

Proof: Suppose f is Lipschitz continuous on R. v−u for some c in between u and v. x. The proof is now complete.1 B Find subdifferentials of the following functions: (a) f (x) = a|x|. h→0 h h→0 h Similarly.7. Conversely. This implies | f (u) − f (v)| ≤ `|u − v|. . b−a This implies (4.7 NONDIFFERENTIABLE CONVEX FUNCTIONS AND SUBDIFFERENTIALS By Theorem 4. a > 0. (b) f (x) = |x − 1| + |x + 1|.2 Find the subdifferential of the function f (x) = max{−2x + 1. This inequality obviously holds for u = v. v ∈ R with u 6= v. ∂ f (x) = [ f−0 (x). Therefore.7. f is Lipschitz continuous.7. Applying Theorem 4.9.   f (b) − f (a) 0 ∈ ∂ g(c) = ∂ f (c) − . 4. f+0 (x)] ⊂ [−`. Then for any x ∈ R. we get f (v) − f (u) ∈ ∂ f (c) ⊂ [−`. 4.19). Then f is Lipschitz continuous if and only if there exists ` ≥ 0 such that ∂ f (x) ⊂ [−`. Thus. `].7. fix any u. Find all absolute minimizers of the function. Then there exists ` ≥ 0 such that | f (u) − f (v)| ≤ `|u − v| for all u. v ∈ R .10 Let f : R → R be a convex function.7.  Exercises 4.3 I Let f (x) = ∑nk=1 |x − k|. f−0 (x) ≥ −`. 2x − 1}. f (x + h) − f (x) `|h| f+0 (x) = lim+ ≤ lim+ = `.8 and the subdifferential sum rule.7. `].140 4. `] for all x ∈ R .  Corollary 4.

7.7.4 Let f : R → R be a convex function. 141 4. ∞) for all x ∈ R. b ∈ R and define the function g by g(x) = f (ax + b). 4. Fix a. . Prove that f is monotone increasing on R. Suppose that ∂ f (x) ⊂ [0.5 B Let f : R → R be a convex function. for x ∈ R Prove that ∂ g(x̄) = a∂ f (ax̄ + b).

.

which strictly contains A. Therefore. 5. We have shown that B ⊂ f ( f −1 (B)) and the equality holds. Now fix b ∈ B. Since f is one-to-one. This implies A ⊂ f −1 ( f (A)).1 Exercise 1. This inclusion does not require the surjectivity of f . (X \Y ) ∩ Z = Z \ (Y ∩ Z). Consider f (x) = x2 . x ∈ f −1 (B) and. 2]. hence. Then f (A) = [0. f −1 ( f (A)) ⊂ A and the equality holds. It is also not hard to find an example of a function f and a set B for which the equality in part (b) does not hold true. b = f (x) ∈ B and.1. there exists x ∈ X such that f (x) = b ∈ B. Applying basic rules of operations on sets yields (X \Y ) ∩ Z = Y c ∩ Z = Z \Y. Thus. Since f is onto. b ∈ f ( f −1 (B)). Without the injectivity of f . Therefore. Thus. f ( f −1 (B)) ⊂ B. . 2].1. x ∈ R. SECTION 1. we have f (a) ∈ f (A) and. a = a0 ∈ A. the equality in part (a) is no longer valid. so there exists a0 ∈ A such that f (a) = f (a0 ).2. Note that this inclusion does not require the injectivity of f . f −1 ( f (A)) = [−2. and A = [−1. (a) For any a ∈ A. Solutions and Hints for Selected Exercises SECTION 1. Now fix any a ∈ f −1 ( f (A)). Then f (a) ∈ f (A).2. so. 4] and. (b) Fix any b ∈ f ( f −1 (B)).2 Exercise 1. a ∈ f −1 ( f (A)). Then b = f (x) for some x ∈ f −1 (B). hence. and Z \ (Y ∩ Z) = (Z \Y ) ∪ (Z \ Z) = (Z \Y ) ∪ 0/ = Z \Y. hence.

observe that the two numbers and are the roots of the quadratic 2 2 equation x2 = x + 1. an+2 = αan+1 + β an for n ∈ N . Consider the sequence {an } defined by a1 = a. an+1 = an + an−1 √ √ √ √ 1 h 1 + 5 n  1 − 5 n i 1 h 1 + 5 n−1  1 − 5 n−1 i =√ − +√ − 5 2 2 5 2 2 √ √ √ √ 1 h 1 + 5 n−1 1 + 5     1 − 5 n−1 1 − 5   i =√ +1 − −1 . Let c1 and c2 be two constants such that c1 x1 + c2 x2 = a. Suppose that √ √ 1 h 1 + 5 k  1 − 5 k i ak = √ − 5 2 2 for all k ≤ n. c1 (x1 )2 + c2 (x2 )2 = b.1) follows easily. Then we can prove by induction that xn = c1 (x1 )n + c2 (x2 )n for all n ∈ N . √ √ √ 1 h1+ 5 1− 5i 1 2 5 √ − =√ = 1. (5. A more general result can be formulated as follows. √ √ 1+ 5 1− 5 In this exercise. It is also easy to verify that the conclusion holds for n = 2. For n = 1. . 5 2 2 5 2 Thus. Suppose that the equation x2 = αx + β has two solutions x1 and x2 .144 Solutions and Hints for Selected Exercises SECTION 1. (5. a2 = b.3 Exercise 1. 2 2 2 2 2 2 Therefore. where n ≥ 2. the conclusion holds for n = 1. 5 2 2 2 2 Observe that √ √ √ √ √ √ 1+ 5 3 + 5  1 + 5 2 1− 5 3 − 5  1 − 5 2 +1 = = and +1 = = .6.1) 5 2 2 By the definition of the sequence and the induction hypothesis. Let us show that √ √ 1 h 1 + 5 n+1  1 − 5 n+1 i an+1 = √ − .3.

8. Therefore. 3 3 Exercise 1. xn = c1 2n + c2 (−1)n . y ∈ R. . ||x| − |y|| ≤ |x − y|. . Thus. This implies −(|x| − |y|) ≤ |x − y|. for k = 1. where c1 and c2 are constants such as c1 (2) + c2 (−1) = 1. consider the sequence a1 = 1. This implies |x| − |y| ≤ |x − y|. c1 (2)2 + c2 (−1)2 = 1. . . Solving the equation x2 = x + 2 yields two solutions x1 = 2 and x2 = (−1). to prove that |a| ≤ m. 2.4. It is not hard to see that c1 = 1/3 and c2 = −1/3. In general. we only need to show that a ≤ m and −a ≤ m. . For example. where m ≥ 0.4 Exercise 1. n. |x| = |x − y + y| ≤ |x − y| + |y|. 1 1 an = 2n − (−1)n for all n ∈ N . Similarly. For any x. 145 This is a very useful method to find a general formula for a sequence defined recursively as above. k k−1 k SECTION 1. we have       n n n+1 + = . |y| = |y − x + x| ≤ |x − y| + |x|.5. Therefore. a2 = 1. an+2 = an+1 + 2an for n ∈ N . Hint: Prove first that.3.

This proves condition (2’).1.6. Thus. Let x = . r Since 1/r > 1.6.1 applied to the sets A and B. In particular.12.1 applied to the set A + B that sup A + sup B = sup(A + B) as desired.2. so. Then by Theorem 2. hence. sup A + sup B satisfies condition (1’). sup A and sup B exist and are real numbers. x = a+b ≤ sup A+sup B.9. (5.1 Exercise 2. Let us first show that A + B is bounded above. lim a2n = ` and lim a2n+1 = `. We will now show that sup A + sup B is the supremum of the set A + B by showing that sup A + sup B satisfies conditions (1’) and (2’) of Proposition 1. we get m > 1 and. by the completeness axiom. Using ε2 in part (2’) of Proposition 1.5. SECTION 1.2) is satisfied. (a) Suppose that limn→∞ an = `. Choose N1 ∈ N such that |a2n − `| < ε whenever n ≥ N1 . there exist a ∈ A and b ∈ B such that x = a+b. which shows that A + B is bounded above.2) n→∞ n→∞ Now suppose that (5. and choose N2 ∈ N such that |a2n+1 − `| < ε whenever n ≥ N2 . Set n = m − 1 and then we get 1 1 <r≤ . there exits a ∈ A and b ∈ B such that ε ε sup A − < a and sup B − < b. By Theorem 1.146 Solutions and Hints for Selected Exercises SECTION 1. We have just shown that sup A + sup B is an upper bound of A + B and. It follows that m − 1 ∈ N. Since A and B are nonempty and bounded above. a ≤ sup A for all a ∈ A and b ≤ sup B for all b ∈ B.5.1. Now let ε > 0.2(d). n+1 n SECTION 2. It follows from Proposition 1. there exists m ∈ Z such that r 1 m−1 ≤ < m. m ≥ 2.5 Exercise 1. .5. Fix any ε > 0. For any x ∈ A+B.6 1 Exercise 1.4. 2 2 It follows that sup A + sup B − ε < a + b.5.1.

limn→∞ an = `. This problem is sometimes very helpful to show that a limit exists. Therefore. so we can conclude that {xn } converges to 2 − 1. consider the sequence defined by x1 = 1/2. we can find n2 ∈ N such that `ε |an − `| < for all n ≥ n2 . n2 }. Exercise 2. For example. 2 + xn √ √ that {x2n+1 } and {x2n } both converge to We will see later 2 − 1. 4 Choose n0 = max{n1 . 147 Let N = max{2N1 .8. one has `ε + `ε4 . 1 xn+1 = for n ∈ N .1. Given any ε > 0. Consider the case where ` > 0. By the definition of limit. Then |an − `| < ε whenever n ≥ N. For any n ≥ n0 . we can find n1 ∈ N such that |an | > `/2 for all n ≥ n1 . 2N2 + 1}. (b) Use a similar method to the solution of part (a).

.

.

an+1 .

|an − an+1 | |an − `| + |an+1 − `| 4 .

an − 1.

= ≤ < = ε. .

.

1). (a) The limit is calculated as follows: √  √  p  n2 + n − n n2 + n + n lim n2 + n − n = lim √ n→∞ n→∞ n2 + n + n n = lim √ n→∞ n2 + n + n n = lim p n→∞ n2 (1 + 1/n) + n 1 = lim p = 1/2. The conclusion is no longer true if ` = 0.2.2 Exercise 2. limn→∞ aan+1 n = 1. n→∞ 1 + 1/n + 1 . SECTION 2.3. A counterexample is an = λ n where λ ∈ (0. If ` < 0. consider the sequence {−an }. |an | |an | ` 2 Therefore.

Thus. we can find the following limit: p p  3 lim an3 + bn2 + cn + d − αn2 + β n + γ . an < an+1 for all n ∈ N. By induction. Suppose that ak < ak+1 for k ∈ N. Then p √ ak+1 = 2 + ak < 2 + 2 = 2. which implies p p ak + 2 < ak+1 + 2. √ p √ (b) Clearly. By induction. a1 = 2 < 2 + 2 = a2 . ak+1 < ak+2 . Let ` = limn→∞ an .3 Exercise 2. √ By the monotone convergence theorem. {an } is an increasing sequence. n→∞ n→∞ Using a similar technique. n→∞ where a > 0 and α > 0. we have √ ` = 2 + ` or `2 = 2 + `. Suppose that ak < 2 for k ∈ N. a1 < 2. (a) Clearly.148 Solutions and Hints for Selected Exercises (b) The limit is calculated as follows: √  p √  3 3 2 −n 3 3 + 3n2 )2 + n 3 n3 + 3n2 + n2 p 3  n + 3n (n lim n3 + 3n2 − n = lim p √ n→∞ n→∞ 3 (n3 + 3n2 )2 + n 3 n3 + 3n2 + n2 ) 3n2 = lim p √ n→∞ 3 (n3 + 3n2 )2 + n 3 n3 + 3n2 + n2 3n2 = lim p p n→∞ 3 n6 (1 + 3/n)2 + n 3 n3 (1 + 3/n) + n2 3n2 = lim p  n→∞ 2 3 p n (1 + 3/n)2 + 3 (1 + 3/n) + 1 3 = lim p  = 1. Then ak + 2 < ak+1 + 2. Since an+1 = (c) 2 + an and limn→∞ an+1 = `.3. . SECTION 2. an < 2 for all n ∈ N. n→∞ 3 p 2 3 (1 + 3/n) + (1 + 3/n) + 1 (c) We use the result in par (a) and part (b) to obtain p p p p  3 3 lim ( n3 + 3n2 − n2 + 1) = lim n3 + 3n2 − n + n − n2 + 1 n→∞ n→∞ p   p  3 = lim n3 + 3n2 − n + lim n − n2 + 1 = 1 − 1/2 = 1/2. Therefore.1. limn→∞ an exists.

Thus. (a) The limit is 3. we see that an > 0 for all n ∈ N. . Moreover. {an } is monotone deceasing √ (for n ≥ 2) and bounded below.5. 2 Thus. 3 By induction. The limit is 2 − 1. 2 where ` > 0. Exercise 2. c ≥ 0. s 1 1 1 1 1 1 an+1 = (2an + 2 ) = (an + an + 2 ) ≥ 3 an · an · = 1. 3 an 3 an 3 a2n We also have. The limit is b. (b) Let {an } be the given sequence. Then 1 an+1 = . Observe that an+1 = 2an . b ≥ 0 to show that an+1 ≥ b for all n ∈ N. −a3n + 1 −(an − 1)(a2n + an + 1)   1 1 an+1 − an = 2an + 2 − an = = < 0. 149 Solving this quadratic equation yields ` = −1 or ` = 2. Observe that an + bn p bn+1 = ≥ an bn = an+1 for all n ∈ N .12.1. 3 an 3a2n 3a2n Thus. We√can show that limn→∞ an = 1. {an } √ 2 1 + 1 + 4c √ converges to this limit.3. (c) We use the well-known inequality a+b+c √ 3 ≥ abc for a.2. {an } converges by Exercise 2. {a2n } √is monotone increasing and bounded above. Therefore.3. for n ≥ 2. Exercise 2. b. √ Exercise 2.3. limn→∞ an = 2.3. In fact. The number is obtained by solving the equation ` = c + `. The limit is 2. (b) The limit is 3. a+b (d) Use the inequality 2 ≥ ab for a. √ an+1 = c + an for n ∈ N . Then follow √ part 3 to show that {an } is monotone decreasing. √ 1 + 1 + 4c We can prove that {an } is monotone increasing and bounded above by . 2 + an Show that {a2n+1 } is monotone decreasing and bounded below. (a) Let {an } be the given sequence. p √ an+1 = an bn ≥ an an = an for all n ∈ N. Define a more general sequence as follows: a1 = c > 0. Then show that {an } is monotone increasing and bounded above.

SECTION 2. This sequence converges to 1. lim sup(an + bn ) = lim αn . (d) A Cauchy sequence.1. lim sup(an + bn ) ≤ lim sup an + lim sup bn .5). when lim supn→∞ an = ∞ and lim supn→∞ bn = −∞. (b) Define αn = inf (an + bn ). and {bn } is decreasing and bounded below by a1 . SECTION 2.5 Exercise 2.3. (a) Not a Cauchy sequence. 2 2 It follows that {an } is monotone increasing and bounded above by b1 . 2 Therefore. (c) Consider an = (−1)n and bn = (−1)n+1 . γn = sup bk . (c) A Cauchy sequence.150 Solutions and Hints for Selected Exercises an + bn bn + bn bn+1 = ≤ = bn for all n ∈ N .5.7. (a) Define αn = sup(an + bn ). .5. lim sup bn = lim γn .1. but use part (b) of Exercise 2.4.3. lim sup an = lim βn . Note that the right-hand side is not well-defined. x = y. This sequence converges to 0. n→∞ n→∞ n→∞ Therefore. Here we use the fact that in R a sequence is a Cauchy sequence if and only if it is convergent. This sequence converges to 0 (see Exercise 2.4 Exercise 2. Then √ x+y x= xy and y = . γn = inf bk . n→∞ n→∞ n→∞ This conclusion remains valid for unbounded sequences provided that the right-hand side is well- defined. See Example 2. (b) A Cauchy sequence. k≥n k≥n k≥n Proceed as in part (a).1. αn ≤ βn + γn for all n ∈ N . This implies lim αn ≤ lim βn + lim γn for all n ∈ N . Let x = limn→∞ an and y = limn→∞ bn . βn = inf ak .4.5. k≥n k≥n k≥n By the definition. n→∞ n→∞ n→∞ n→∞ n→∞ n→∞ By Exercise 2. βn = sup ak . for example.

We have ( |x − 1/2|. assume that lim f (x) = `.6 Exercise 2. the given limit does not exists. limx→1/2 f (x) = 1/2. Therefore.6. f (x) is near 1/2 no matter whether x is rational or irrational. | f (x) − 1/2| = |1 − x − 1/2|. if x ∈ Q. f (x) = Q(x). Thus. If a is a solution of the equation P(x) = Q(x). define the function ( P(x). By contradiction. Thus.. SECTION 3. (c) By a similar method to part (b).2(c) we get that A ∪ B is closed.6.1 Exercise 3. Then | f (x) − 1/2| < ε whenever |x − 1/2| < δ . From Theorem 2.1. n→∞ n→∞ This is a contradiction. P(a) = Q(a). Solving this problem suggests a more general problem as follows. respectively. However. Suppose A and B are compact subsets of R. we can show that limx→1 f (x) does not exists. mB be upper and lower bounds for A and B.1. mA . We justify this using the sequential criterion for limits (Theorem 3. We have shown that A ∪ B is both closed and bounded. choose δ = ε.5. A and B are closed and bounded. Then.6. mB } are upper and lower bounds for A ∪ B. hence. MB } and m = min{mA . and choose also a sequence {sn } of irrational numbers that converges to 0. if x 6∈ Q.3. Then f (rn ) = rn and f (sn ) = 1 − sn and. Given any ε > 0. by Theorem 2. ` = lim f (rn ) = 0 n→∞ and ` = lim f (sn ) = lim (1 − sn ) = 1.6. (a) Observe that when x is near 1/2. x→0 where ` is a real number. Choose a sequence {rn } of rational numbers that converges to 0. if x 6∈ Q. In particular. let MA .e. Given two polynomials P and Q. f (x) is near 1. when f is near 0 and x is irrational.5 that A ∪ B is compact. A ∪ B is bounded.2). f (x) is near 0. 151 SECTION 2. Then M = max{MA . then the limit limx→a f (x) exists and the limit is this common value. Moreover. if x ∈ Q. It now follows from Theorem 2. MB .6. . i. (b) Observe that when x is near 0 and x is rational. | f (x) − 1/2| = |x − 1/2| for all x ∈ R. For all other points the limit does not exist.

(a) Observe that f (a) = g(a) = h(a) and. 2(k + 1) Therefore. (b) Apply part (a). f (x1 ) and f (x2 ) are close to each other. Given any ε > 0.     p 1 p 1 Aε = {x ∈ (0. In the case when x is near a and x is rational. This suggests the use of the Cauchy ε criterion for limit to solve the problem. 1]. choose δ = . Prove that f does not have a limit at any a ∈ R. If x1 . 1] except for a finite number of x ∈ Q. we can choose a sufficiently small neighborhood of a to void such x.5. limx→x̄ f (x) exists. Determine all a ∈ R at which limx→a f (x) exists. q ∈ N. where ( x2 . x2 ∈ D \ {x̄} 2(k + 1) with |x1 − x̄| < δ and |x2 − x̄| < δ . So f (x) is close to f (a) for all x ∈ (0. 1] : f (x) ≥ ε} = x = ∈ Q : f (x) = ≥ ε = x = ∈ Q : q ≤ . f (x) = 1/q where p. if x ∈ x 6∈ Q. 1] such that f (x) ≥ ε.3 Exercise 3.2 Exercise 3. It follows that | f (x) − f (a)| ≤ |g(x) − g(a)| + |h(x) − h(a)| for all x ∈ [0. (a) For any ε > 0. f (x) = x + 2. we have f (a) = 0. hence. 1]. 1]. so. At any irrational number a ∈ (0.9. ( |g(x) − g(a)|.3. SECTION 3. Exercise 3. if x ∈ 6 Q. if x ∈ Qc ∩ [0. 2.8.3. Therefore. We will see in part (a) that for any ε > 0. then ε | f (x1 ) − f (x2 )| ≤ k|x1 − x2 | ≤ k(|x1 − x̄| + |x2 − x̄|) < k(δ + δ ) = 2k < ε. | f (x) − f (a)| = |h(x) − h(a)|. there is only a finite number of x ∈ (0. if x ∈ Q. SECTION 3.2. limx→a f (x) = f (a) and. If x is near a and x is irrational. The given condition implies that if both x1 and x2 are close to x̄. hence. if x ∈ Q. 1]. Since a is irrational. then they are close to each other and. it is obvious that f (x) = 0 is near f (a). q q q ε . Consider the function ( x2 + 1. f is continuous at a. f (x) = −x.152 Solutions and Hints for Selected Exercises Similar problems: 1. if x ∈ Q ∩ [0.

1]. and discontinuous at every rational point. . f (x) = 0.10. f is not continuous at b. . . xn }. In particular. (a) Observe that | f (1/n)| ≤ 1/n for all n ∈ N . the set Aε is finite. so we can write Aε = {x ∈ (0. Since a is irrational. p. n Then the conclusion follows from the Intermediate Value Theorem.7. n}). .   f (x) = 1. 1]. where xi ∈ Q for all i = 1. . n (more precisely. Choose a sequence of irrational numbers {sn } that converges to b. for some n ∈ N. Therefore. we consider the domain of f to be the interval (0. Exercise 3. f is continuous at a. the number of q ∈ N such that q ≤ ε is finite. x2 . but the conclusion remain valid for other intervals. SECTION 3. Since 0 < q ≤ 1. we can choose δ = min{|a − xi | : i = 1. . a + δ ) for all i = 1. Then f (b) = 1q . Let α = min { f (x) : x ∈ [a. In this problem. if x ∈ Q. . 1] : f (x) ≥ ε} = {x1 . the sequence { f (sn )} does not converge to f (b).4.3. Now fix any rational number b = qp ∈ (0. we can show that the function defined on R by 1 p  q . Therefore. . f (x1 ) + f (x2 ) + · · · + f (xn ) α≤ . . if x = 0. where p and q have no common factors. q ∈ N. . Since f (sn ) = 0 for all n ∈ N. . Then f (x1 ) + f (x2 ) + · · · + f (xn ) nβ ≤ = β. by part (a).6.4 Exercise 3. . Exercise 3. . b]}. is continuous at every irrational point. if x = q . Aε is finite. 153 1 p Clearly. Consider ( (x − a1 )(x − a2 ) · · · (x − ak ). . 1]. (b) Fix any irrational number a ∈ (0. we have p ≤ q. we can choose δ > 0 such that xi ∈ / (a − δ .    0. if x ∈ Qc . if x is irrational. n. . . n n Similarly.4. b]} and β = max { f (x) : x ∈ [a. Then f (a) = 0. Therefore. Then | f (x) − f (a)| = f (x) < ε whenever |x − a| < δ . . Given any ε > 0.

154 Solutions and Hints for Selected Exercises .

.

.

f (x) .

(b) Apply the Extreme Value Theorem for the function g(x) = .

x .

4. 1]. f has a unique fixed point. Since f is bounded. if x1 ≤ x2 . (a) Let f : D → R. if x and y are close to each other. The behavior of the graph of the squaring function suggests the argument below to show that f (x) = x2 is not uniformly continuous on R. f (x0 ) = g(x0 ) = x0 . 1 Define two sequences {xn } and {yn } as follows: xn = n and yn = n + for n ∈ N. It is similar for the case where x1 > x0 .5. taking limits we have f (x0 ) = x0 . Indeed. f has a fixed point in [0. Exercise 3. We can prove by induction that g(xn ) = xn for all n ∈ N.8. b].5. if x1 ≥ x2 . In fact. Define the sequence {xn } as follows: x1 = c. 1] such that f (x1 ) = x1 . 1] such that g(c) = c. n 1 2   1 | f (xn ) − f (yn )| = n + − n2 = 2 + 2 ≥ 2 for all n ∈ N .2. 1].3 we see that if there exist two sequences {xn } and {yn } in D such that |xn − yn | → 0 as n → ∞. so the previous argument does not work. By Exercise 3. Since f is monotone increasing.5. Thus. then {xn } is monotone decreasing.5. From Theorem 3. Then g(x0 ) = lim g(xn ) = lim xn = x0 . If x1 < x0 . then f is not uniformly continuous on D. 1]. xn+1 = f (xn ) for all n ≥ 1. SECTION 3. In this case. we have f (g(x0 )) = g( f (x0 )) = g(x0 ). there exists c ∈ [0. 1] such that lim xn = x0 . then {xn } is monotone increasing. Consider the case where f is monotone increasing. which means that there exists x0 ∈ [0. Therefore. However. 1]. in order for f to be uniformly continuous on D.4. n→∞ Since f is continuous and xn+1 = f (xn ) for all n ∈ N. However. which yields a contradiction.1). f could have several fixed points on [0. Roughly speaking.3. by the monotone convergence theorem (Theorem 2. 1] such that f (x0 ) = x0 . 1] such that f (x0 ) = x0 . then f (x) and f (y) must be close to each other. g(x0 ) is also a fixed point of f and. First consider the case where f is monotone decreasing on [0. The proof is complete in this case. Since f is monotone decreasing. n→∞ Therefore. {xn } is a monotone sequence. then x1 = f (x1 ) ≥ f (x0 ) = x0 . Since f (g(x)) = g( f (x)) for all x ∈ [0. suppose that there exists x1 ∈ [0. Then n 1 |xn − yn | = → 0 as n → ∞.4.5 Exercise 3. there exists x0 ∈ [0. x0 is the unique point in [0. by Exercise 3. but {| f (xn ) − f (yn )|} does not converge to 0. hence. on the interval [a. g(x0 ) = x0 = f (x0 ). n n .

n 1 1 (b) Use xn = and yn = . The conclusion follows from the Extreme Value Theorem for the function f on [a. 2 Then we can show that | f (u) − f (v)| < ε whenever |u − v| < δ . u. u. it is uniformly continuous on this interval. By a similar method. there exists b > a such that f (x) > f (a) whenever x > b. c + 1]. Thus. Since f is continuous on [a. v ∈ [a.5. ∞)} = inf{ f (x) : x ∈ [a. A more challenging question is to determine whether a polynomial of degree greater than or equal to two is uniformly continuous on R. b + 1]. 155 Therefore.5. there exists 0 < δ < 1 such that ε | f (u) − f (v)| < whenever |u − v| < δ . (b) Fix any ε > 0. by the definition of limit. Thus. n ∈ N. v ∈ [a. we can use xn = n + and yn = n for n ∈ N instead. π/2 + 2nπ 2nπ (c) Use xn = 1/n and yn = 1/(2n). Exercise 3. (a) Applying the definition of limit. b]. n ∈ N. ∞). there exists a > 0 such that f (x) ≥ f (0) whenever |x| > a. we can show that the function f (x) = xn . SECTION 3. f is not uniformly continuous on R.6 Exercise 3. Hint: For part (a) use Theorem 3. Therefore.7.5. {| f (xn ) − f (yn )|} doesr not converge to 0 and. For part (b) prove that the function can be extended to a continuous function on [a. 2 Since f is continuous on [a. f is bounded on [a. b]}. we can use xn = 3 n + and yn = 3 n for n ∈ N to prove that f is not uniformly n continuous on R.5. b].5.7. b] and then use Theorem 3. Since limx→∞ f (x) = limx→−∞ f (x) = ∞. n ≥ 2.5. we find b > a such that ε | f (x) − c| < whenever x > b. inf{ f (x) : x ∈ [a. it is bounded on this interval. .8. Following the It is natural to ask whether the function 1 √ solution for part (a). hence.4. 3 r f (x) = x is uniformly continuous on R. 1 √ In this solution. ∞). (c) Since limx→∞ f (x) = c > f (a). is not uniformly continuous on R. we find b > a such that c − 1 < f (x) < c + 1 whenever x > b. Exercise 3.

we conclude the function is differentiable at any a 6= 0. By the definition of the derivative.10.1. a] at some point x̄ ∈ [−a. (a) Using the differentiability of sin x and Theorem 4.3 and the fact that cos x is the derivative of sin x. we can use this theorem to prove that any polynomial with even degree has an absolute minimum on R. 0 f (x) = x c. Since R is a not a compact set. x→a x−a x→0 Using Theorem 4. we cannot use the extreme value theorem directly. Since limx→0 (−|x|) = limx→0 |x| = 0. SECTION 4.3. a].1. which implies −|x| ≤ x sin(1/x) ≤ |x|.7. we can use any number γ in the range of f instead of f (0). Use the identity !n f (a + 1n ) 1 lim = lim exp(n[ln( f (a + )) − ln( f (a)]).3.1 Exercise 4. n→∞ f (a) n→∞ n Exercise 4. it has an absolute minimum on [−a. we only need to show the differentiability of the function at a = 0. Observe that in this solution. f has an absolute minimum at x̄. So.156 Solutions and Hints for Selected Exercises Since f is lower semicontinuous. if x 6= 0. consider the limit f (x) − f (a) x2 sin(1/x) + cx lim = lim = lim [x sin(1/x) + c].1. For example.11. then f (x) ≥ f (0) ≥ f (x̄). . Obviously. Therefore. x→0 It now follows that f (x) − f (a) f 0 (0) = lim = lim [x sin(1/x) + c] = c. If |x| > a.1. we have |x sin(1/x)| = |x| | sin(1/x)| ≤ |x|. Since any continuous function is also lower semicontinuous. a]. by Theorem 3. the derivative of f can be written explicitly as  2x sin 1 − cos(1/x) + c. the result from this problem is applicable for continuous functions. x→a x−a x→0 x x→0 For any x 6= 0. In particular. f (0) ≥ f (x̄). applying the squeeze theorem yields lim x sin(1/x) = 0. if x = 0. f (x) ≥ f (x̄) for all x ∈ [−a.

then f is not differentiable at 0. Define the function ( xn ϕ(1/x). Show that if n ≥ 2. Let ϕ be a bounded function on R. Show that the functions below are differentiable on R: ( x3/2 cos(1/x). f (x) = 0. Note that the function h(x) = cx does not play any role in the differentiability of f . if x 6= 0. if x < 0 and ( 2 x2 e−1/x . Show that if n = 1 and limx→∞ ϕ(x) does not exists. if x 6= 0. it is important to see that the conclusion remains valid if we replace the function f by  xn sin 1 .e. where n ≥ 2.2 Exercise 4. if x = 0. Define the function ( xn ϕ(1/x). f (x) = 0. n ∈ N. 2. f (x) = 0. g(x) = x 0. if x 6= 0. Similar problems: 1. where n ≥ 2. (b) Hint: Observe that     0 1 0 1 f = −1 + c < 0 and f = 1 + c > 0. if x ≥ 0. f (x) = 0. if x 6= 0.2. .1.. Define the function h(x) = f (x) − g(x). if x = 0. 157 From the solution. n ∈ N. 2nπ (2n + 1)π SECTION 4. We can generalize this problem as follows. Then f is differentiable at a = 0. if x = 0. Suppose that ϕ is bounded and differentiable on R. the function is differentiable on R and find its derivative. there is M > 0 such that |ϕ(x)| ≤ M for all x ∈ R . i. if x = 0.

the equality can be rewritten as . Exercise 4.2. In the case where a 6= b. The inequality holds obviously if a = b. which implies f 0 (x0 ) = g0 (x0 ).3. Thus. h0 (x0 ) = f 0 (x0 ) − g0 (x0 ) = 0.158 Solutions and Hints for Selected Exercises Then h has an absolute maximum at x0 .

.

.

sin(b) − sin(a) .

.

.

. ≤ 1.

b−a .

.

.

.

sin(b) − sin(a) .

The quotient .

.

.

f (a)) and (b. f (b)). We need b−a . is the slope of the line connecting (a.

b−a which implies . Figure 5. there exists c ∈ (a. the function satisfies all assumptions of the Mean Value Theorem on this interval with f 0 (x) = cos(x) for all x ∈ (a. The quotient also reminds us of applying the Mean Value Theorem for the function f (x) = sin(x). By the Mean Value Theorem. b] → R by f (x) = sin(x). b) such that f (b) − f (a) = f 0 (c) = cos(c). Consider the case where a < b and define the function f : [a. Clearly. which can also be seen from the figure.1: The function f (x) = sin(x). to show that the absolute value of the slope is always bounded by 1. b).

.

.

f (b) − f (a) .

.

b − a .

= | cos(c)| ≤ 1. .

.

It is essential to realize that the most important property required in solving this problem is the boundedness of the derivative of the function. it is possible to solve the following problems . The solution is similar for the case where a > b. Thus. It follows that | f (a) − f (b)| ≤ |a − b|.

Prove that | ln(1 + e2a ) − ln(1 + e2b )| ≤ 2|a − b| for all a. Exercise 4. π) such that f (c) = 0. b ∈ R. Prove that | cos(a) − cos(b)| ≤ |a − b| for all a. 159 with a similar strategy. 1. Let us define f : [−π.4. Now. consider the function n  x2  cos(kx) sin(kx) g(x) = + ∑ −ak + bk . π] → R by n f (x) = x + ∑ (ak sin kx + bk cos kx). b ∈ R. k=1 We want to find c ∈ (−π. 2 k=1 k k Observe that g(−π) = g(π) and g0 = f . 2.2. Use the identity f (a) f (b) − g(b) . Exercise 4. The conclusion follows from Rolle’s Theorem.2.5.

.

1 .

f (a) f (b) .

f (a)g(b) − f (b)g(a) g(a) .

.

= = . g(b) − g(a) .

g(a) g(b) g(b) − g(a) 1 1 .

Then use the identity f (x)− f (x0 ) f (x0 ) f (x) f (x) − f (x0 ) + f (x0 ) x−x0 + x−x 0 = = x0 . f (x) a−ε < < a + ε. x x − x0 + x0 1 + x−x 0 and the mean value theorem to show that.2. Exercise 4. Apply Rolle’s theorem to the function x2 xn f (x) = a1 x + a2 + · · · + an 2 n on the interval [0. (a) Given ε > 0. 1]. Apply Rolle’s theorem to the function n sin(2k + 1)x f (x) = ∑ k=0 2k + 1 on the interval [0. 2. (c) Consider f (x) = sin(x). first find x0 large enough so that a − ε/2 < f 0 (x) < a + ε/2 for x > x0 .6. 1. Exercise 4. x (b) Use the method in part (a). g(a) − g(b) f (x) 1 Then apply the Cauchy mean value theorem for two functions φ (x) = g(x) and ψ(x) = g(x) on the interval [a. b].8. . π/2].2. for x large.

  α1 ε choose δ = and get `+1 ε | f (u) − f (v)| ≤ `|u − v|α < `δ α = ` <ε `+1 whenever |u − v| < δ . (b) We will prove that f is a constant function by showing that it is differentiable on R and f 0 (a) = 0 for all a ∈ R. Note that we use ` + 1 here instead of ` to avoid the case where ` = 0. Fix any a ∈ R. Given any ε > 0. (a) We can prove that f is uniformly continuous on R by definition. .160 Solutions and Hints for Selected Exercises SECTION 4. for x 6= a.3 Exercise 4.2.3. Then.

.

.

f (x) − f (a) .

`|x − a|α α−1 .

x − a .

. ≤ |x − a| = `|x − a| .

.

Since α > 1. Therefore. Exercise 4. Use proof by contradiction. It follows that h(x) ≥ h(x0 ) = g(x0 ) − f (x0 ) = 0 for all x ≥ x0 . Thus. Define the function h(x) = g(x) − f (x). Then h0 (x) = g0 (x) − f 0 (x) ≥ 0 for all x ∈ [x0 . . ∞).8) when α ≤ 1.3. h is monotone increasing on this interval. Exercise 4.3.6. we see that it is only interesting to consider the class of functions that satisfy (4. by the squeeze theorem. From this problem. g(x) ≥ f (x) for all x ≥ x0 .4 Exercise 4. f (x) − f (a) lim = 0. (c) We can verify that the function f (x) = |x| satisfies the requirement. x→a x−a This implies that f is differentiable at a and f 0 (a) = 0. Suppose that P(x) = a0 + a1 x + · · · + an xn .4.3. Exercise 4.5. SECTION 4. Then apply L’Hospital’s rule repeatedly. It is an exercise to show that the function f (x) = |x|1/2 satisfies this condition with ` = 1 and α = 1/2. Apply the mean value theorem twice.3.5.

4.  Based on these calculations. Given that the conclusion is true for some n ∈ N.  where P is a polynomial. for x 6= 0 we have       (n+1) −2 0 1 2 1 − 12 1 − 12 f (x) = −x P + 3P e x =Q e x . if x 6= 0. x→0 x−0 x→0 x Letting t = 1/x and applying L’Hospital rule yields 1 e− x 2 t 1 lim+ = lim t 2 = lim = 0. 161 Exercise 4. if x = 0. f 00 (x) = x4 x6 0. In a similar way. x Consider the limit 1 f (x) − f (0) e− x 2 lim = lim . Moreover. if x 6= 0. The conclusion is true for n = 1 as shown above. Now we proceed to prove this conclusion by induction. successive applications of l’Hôpital’s rule give f (n) (x) − f (n) (0)   1 1 − 12 tP(t) lim+ = lim+ P e x = lim t 2 = 0. if x = 0. It is an easy exercise to write the explicit formula of Q based on P. We first consider the case where n = 1 to get ideas for solving this problem in the general case. 0 f (x) = x3 0. x→0 x It follows that f is differentiable on R with   2 e− x12 . x→0 x−0 x→0 x x t→∞ e In a similar way. x→0 x−0 . we predict that f is n times differentiable for every n ∈ N with    P 1 e− x12 .6. 1 e− x2 lim− = 0. if x = 0. we can show that f is twice differentiable on R with    6 + 2 e− x12 . x x x x where Q is also a polynomial. x→0 x t→∞ e t→∞ 2tet 2 Similarly. we can show that f (n) (x) − f (n) (0) lim− = 0. (n) x f (x) = 0. if x 6= 0. From the standard derivative theorems we get that the function is differentiable at any x 6= 0 with 1 2 1 f 0 (x) = 2x−3 e− x2 = 3 e− x2 .

f (n+1) (0) = 0. if f is twice differentiable on (a. (b) With the analysis from part (a).5. we see that if f is n times differentiable on (a. we can also show that the function ( 1 e− x . for any x > 0. b). then f (k) (x̄)hk f (x̄ + h) − ∑n−1 k=0 f (n) (x̄) lim k! = . f (x) = 0. We have proved that for every n ∈ N. h→0 hn+1 n! This conclusion can be proved by induction. there exists c ∈ (0. b). x) such that m f (k) (0) k f (m+1) (c) m+1 f (x) = ex = ∑ x + c k=0 k! (m + 1)! m k m x ec xk = ∑ k! + (m + 1)! cm+1 > ∑ k! . b) and x̄ ∈ (a.1. By Taylor’s theorem.5. .5. f is n times differentiable and.5 Exercise 4. Let f (x) = ex . then f (x̄ + h) − f (x̄) − f 0 (x̄) 1!h f 00 (x̄) lim = . if x ≤ 0 is n times differentiable for every n ∈ N. b). f ∈ Cn (R). SECTION 4. In a similar way. Similarly. h→0 h 1! This conclusion follows directly from the definition of derivative.5. Here we do not need to prove the continuity of f (n) because the differentiability of f (n) implies its continuity. then f (x̄ + h) − f (x̄) f 0 (x̄) lim = . b) and x̄ ∈ (a. h→0 h2 2! We can prove this by applying the L’Hospital rule to get f (x̄ + h) − f (x̄) − f 0 (x̄) 1!h f 0 (x̄ + h) − f 0 (x̄) f 00 (x̄) lim = lim = . (a) Observe that a simpler version of this problem can be stated as follows: If f is differentiable on (a. k=0 k=0 Exercise 4.6. so. h→0 h2 h→0 2h 2! It is now clear that we can solve part (a) by using the L’Hospital rule as follows: 2 f (x̄ + h) − f (x̄) − f 0 (x̄) 1!h − f 00 (x̄) h2! f 0 (x̄ + h) − f 0 (x̄) − f 00 (x̄) 1!h f 000 (x̄) lim = lim = .162 Solutions and Hints for Selected Exercises Therefore. This general result can be applied to obtain the Taylor expansion with Peano’s remainder in Exercise 4. b) and x̄ ∈ (a. h→0 h3 h→0 3h2 3! Note that the last equality follows from the previous proof applied to the function f 0 . if x > 0.

but (φ ◦ f )(x) = |x2 − 1| is nonconvex.5.4. 163 SECTION 4. (a) By Theorem 4. Observe that f (x) = x2 and φ (x) = |x − 1| are convex. consider f (x) = |x| and φ (x) = x p . By the nondecreasing property and the convexity of φ . a]. (b) Use Theorem 4.6.  ∂ f (x) = [−a. x ∈ (0.1. The answer is negative.7 Exercise 4.6 or Corollary 4. b ∈ (0. Given any u. Similarly. ∞). h(x) = ex is also a convex function on R. Observe that in this problem.5. if x > 0. if x = 0. ∞). 2 2 This implies a+b − ln(a) − ln(b) √ − ln( )≤ = − ln( ab). f (v) ∈ J and J is an interval. consider the function f (x) = − ln(x). Since f (u). SECTION 4. A natural question is whether the composition of two convex functions is convex. is convex on R. For example.3 for the function f (x) = − ln(x) on (0. φ ( f (λ u + (1 − λ )v)) ≤ φ (λ f (u) + (1 − λ ) f (v)) ≤ λ φ ( f (u)) + (1 − λ )φ ( f (v)).7.3. if x < 0.7. We can show that f is convex on (0. one has   a+b f (a) + f (b) f ≤ .7. v ∈ I and λ ∈ (0. p > 1.  {−a}. p > 1. we require the nondecreasing property of φ . The result from this problem allows us to generate convex functions. Alternatively. ∞). 2 This inequality holds obviously when a = 0 or b = 0. (a) Use the obvious inequality √ √ ( a − b)2 ≥ 0. We apply the definition to solve this problem. the composition 2 g(x) = |x| p .  {a}. φ ◦ f is convex on I. Therefore. ∞). Therefore.  . a+b √ ≥ ab. We have seen that f is convex on R.6. 2 2 Therefore.6. Exercise 4. we have f (λ u + (1 − λ )v) ≤ λ f (u) + (1 − λ ) f (v) by the convexity of f . For a.6. 1). Use Theorem 4. Exercise 4.6. λ f (u) + (1 − λ ) f (v) ∈ J. The function φ is convex and increasing on [0.6. ∞) which contains the range of the function f .6 Exercise 4.

we can see that f has an absolute minimum at x = 2. Exercise 4.      {2}. then f has an absolute minimum at any point x ∈ [m.7.7. if x̄ < m.164 Solutions and Hints for Selected Exercises (b) By Theorem 4.7. We then conjecture that if n is odd with n = 2m − 1. Consider x̄ = m. ∂ fi (x̄) = {1} if i < m. If n = 1. By Theorem 4. b ∈ R with a < b. if x < −1. hence. The graphing of the function suggests that f has an absolute minimum at any x ∈ [1. 2]. f is monotone increasing on R. Thus. 0). To better understand the problem. In this case. 2].    ∂ f (x) = {0}. ∞). if x > 1.  [0.  Exercise 4. which does not contain 0. Therefore. if x = −1. In the case where n = 3. Obviously. then ∂ f (x̄) ⊂ (−∞. there exists c ∈ (a. we consider some special cases. If n is even with n = 2m.5. if x ∈ (−1. then f (x) = |x − 1| + |x − 2|.5. The subdifferential sum rule yields ∂ f (x̄) = [−1. 1]. b) such that f (b) − f (a) ∈ ∂ f (c) ⊂ [0. we can see that ∂ f (x̄) ⊂ (0. ∞). The case where n is even can be treated similarly.3. Fix a. Therefore. if x = 1. Similarly.   {−2}. Then ∂ fm (x̄) = [−1. Let us prove the first conclusion. f has an absolute minimum at the only point x̄ = m. 0]. f has an absolute minimum at x = 1. then f has an absolute minimum at x = m.  [−2. m + 1].9. f has an absolute minimum at x̄. b−a This implies f (b) − f (a) ≥ 0 and. 1] which contains 0. 1). i=1 i=1 where fi (x) = |x − i|. If n = 2.7. 2m−1 2m−1 f (x) = ∑ |x − i| = ∑ fi (x). ∂ fi (x̄) = {−1} if i > m. f (b) ≥ f (a). If x̄ > m. then f (x) = |x − 1|. .

[Fit09] P. McGraw-Hill. Paris. [Ros68] M. Hiriart-Urruty and C. Academic Press. Wade. Lemaréchal. Inc. Ellipses. Dover Publications. American Mathematical Society. Prentice Hall. Lay. R. Principles of Mathematical Analysis. Fundamentals of Convex Analysis.-C. M. Second Edition. Pearson. [Sto00] M. 1996. [Lay13] S. Fitzpatrick. Stoll. Third Edition. Introduction to Real Analysis. Second Edition. [HUL01] J. 2000.. 1968. [FJ96] J. Pearson. Analysis with an Introduction to Proof. [Mun00] J. Springer. Foundation of Modern Analysis. Dieudonné. R. Fourth Edition. 2009. Topology. Advanced Calculus. Prentice Hall. Second Edition. 2013. 2000. 1976. Munkres. Franchini and J. Analyse I. Bibliography [Die69] J. Rudin. Introduction to Analysis. 2001. 2009. [Rud76] W.-B. Fifth Edition. Rosenlicht. [Wad09] W. 1969. . An Introduction to Analysis. Jacquens.