SMA 300: REAL ANALYSIS I

Purpose of the Course

The purpose of this course is to enable learners to develop a deeper and more rigorous
understanding of the real number system including defining terms and proving theorems
about open and closed sets, sequences, series, limits and continuity of functions.

Course Learning Outcomes

By the end of this course the learner will be able to:

i) Construct the real number system and apply the concept of completeness in the
set of real numbers and its topology.
ii) Prove and apply theorems associated with sequences of real numbers.
iii) Classify a given series of real numbers as convergent or divergent.
iv) Discuss properties of countable and uncountable sets.
v) Distinguish between point wise and uniform continuity of real-values functions.

Course Content
The Real number System: field and order structure, principle of induction, Rational &
Irrational numbers, Bounded sets, Supremum and Infimum, Completeness in the set of
real numbers. Topology of the Real Numbers: Open sets, limit points, closed sets and
closure of sets. Sequences: Limit point of a sequence. Limit superior and Limit Inferior,
Convergent sequences, Monotonic sequence, Cauchy sequence, principle of convergent
sequences. Series: Partial sum of series, Convergence of series, Absolute and conditional
convergence of series. Tests of convergence: Comparison tests, Cauchy’s Root test,
D’Alembert’s Ratio test, Integral test, Alternating series test. Countable and uncountable
sets; Functions: Limits, Continuity and Uniform Continuity.
(Pre-requisite SMA 200: Calculus II)

Core Reading Materials

1. Bartle, R.G., and Sherbert, D.R., Introduction to Real Analysis, 4th Ed., Wiley, 2001.

i
2. Rudin, W., Principles of Mathematical Analysis, 3rd Ed., McGraw-Hill, 1976.
3. Trench, W.F., Introduction to Real analysis, Pearson Education, 2003.
4. Lewin, J., An Interactive Introduction to Mathematical Analysis, Revised Ed.,
Cambridge University Press, 2014.

Recommended Reading Materials

1. Lewin, J and Lewin, M., An Introduction to Mathematical Analysis, 2nd Ed. McGraw
–Hill, New York, 1993.
2. Nikolsky, S.M., A course of Mathematical Analysis, Vol.2, MIR, Moscow, 1987.

ii
 THE REAL NUMBER SYSTEM

Field and Order structure of Real Numbers

Notation
We shall denote by:

 : the set of all int egers

 : the set of all positive int egers
Q : the set of all rational numbers
Q  : the set of all irrational numbers
 : the set of all real numbers
 : For all (or for every)
 : There exists (or there is)
!: There exists a unique element
: Implies that
: If and only if

Algebraic Structure of Real Numbers

Let + denote addition and . denote multiplication. Then
1. For all x, y  , we have
i) x  y 
ii) x. y   (Closure)
2. For all x, y  , we have
i) x  y  y  x
ii) x. y  y.x (Commutative law)

3. For all x, y , z   , we have

i)  x  y   z  x   y  z 
ii)  x. y  .z  x. y.z  (Associative law)
4. There exist distinct elements 0 and 1 in  such that
x0  x
x.1  x, x 
Number 0 is called the additive identity
1 is called the multiplicative identity
5. For all x , the exists a real number y such that x  y  0 . The number y  is
called the additive inverse of x.
Eg if x  7 the y  7 .

1
 6. For all x such that x  0 , the exists a real number y  such that x. y  1 .
The number y denote by x 1 is called the multip[licative inverse of x.
Eg if x  2, then 2 1  1 2 .
Definition 1.1: A set with two operations + and . satisfying all the above properties is
called a field. Thus the set of all real numbers is a field.

Ordering property of real numbers

Definition 1.2: Let F be a field, an order relation  on F is a relation such that
i)  x, y  F , x  y and / or y  x
ii) if x  y and y  x then x  y
iii) if x  y and y  z then x  z
and if x  y we write x  y or y  x .
A field with an ordered relation satisfying (i), (ii) and (iii) is called an ordered field.

Example 1.1
The set of all real numbers  is an ordered field.

Rational and Irrational Numbers

Rational Numbers

The continuum
We extend the natural numbers by including the number zero and all the negative whole
numbers to obtain the set of integers ...  2, 1, 0,1, 2,...
As long as we are in equation of the type 2x  5  0 we have no whole number solutions,
so we need more numbers to include fractions. The next set is called the set of rational
numbers (i.e. we extend  to the set of rational numbers)

Definition 1.3: A rational number is a number of the form r  m n , such that  m, n  =1

and, m, n  , n  0 , that is m and n have no common factor other than 1. The set of

rational numbers is denoted by Q , that is Q  m n :  m, n   1, m, n  , n  0 .
The set of rational numbers is closed under both addition and multiplication.

Lemma 1.1
Between any two rational numbers there exists always a third number.
v v
Proof: let v1 and v2 be two rational numbers, then v3  1 2 is a rational number
2
between v1 and v2 .

2
In fact, this process can be repeated infinitely many times, and so between any two
rational numbers, there are infinitely many rationals.

1.2.2.2 Irrational numbers

Definition 1.4: The set of numbers that are not rational cannot be expressed in the form
m and are called irrational numbers and are denoted by Q c .
n

Theorem 1.2: There does not exist a rational number r such that r 2  2 , in other words
r  2 is an irrational number.
Proof
Suppose to the contrary that r  2 is rational, then
m
2 where  m, n  =1, m, n  .
n
2
 2  m 2  m 2  2n 2 , we see that m 2 is even. This implies that m is also
n
 
even (because if m  2k 1 is odd then its square m 2  2 2k 2  2k  1 is also odd). Now
let m=2k for k   , then 4k 2  2n 2 , so that n 2  2k 2 . Therefore, n 2 is even, and so n is
even.
Then both m and n are even, hence have a common factor 2, which is a contradiction.
Hence 2 is irrational.

Definition 1.5: The set of all rational and irrational numbers is called the set of real
numbers or the continuum denoted by . Thus   Q  Q c , the set  is closed under
addition and multiplication.

1.2.2.4 Intervals
Let a and b be any two real numbers such that a  b . Then
a) a, b  x : x  , a  x  bis called an open interval
b) a, b  x  ; a  x  b is called a closed interval
c) a, b  x : x  , a  x  b is called half-open/ half-closed interval.
d) a, b  x : x   : a  x  bis called half-closed/ half-open interval

Bounded Sets
Definition: Let S be a non-empty subset of .
a) The set S is said to be bounded above if there exists a number x   such that
s  x for all s  S . Such a number x is called an upper bound of S.
b) The set S is said to be bounded below if there exists a number y   such that
y  s for all s  S. Such that a number y is called a lower bound of S.
A set S is said to be bounded if it is both bounded above and bounded below. A set is said
to be unbounded if it is not bounded

3
Example
i) The set S   x  : x  4 is bounded above. The set has no lower bound.
Note that all real numbers greater than 4 are also upper bound of S.
ii) The set S   x  : x  2 is bounded below. The set has no upper bound. Note
that all real numbers less than 2 are also lower bounds of S.
iii) Lastly the set S   x  : 2  x  4 isbounded, with lower bound 2 and upper
bound 4.

Definition: Of all the upper bounds of a set S, the least is called a least upper bound or
supremum of S denoted by sup S.
That is u is a supremum of S if
i) u is an upper bound of S, and
ii) if v is any upper bound of S, then u  v
iii)  0,  s  S : u   s  u .

Definition (Bounded set): If S is bounded below, then of all the lower bounds, the
greatest is called a greater lower bound, or an infimum of S denoted by inf S. That is u is
an infimum of S if:
i) u is a lower of S and
ii) if v is a lower bound of S, then u  v
iii)  0,  s  S : u  s  u   .
If a supremum of a set belongs to the set, then it is called the maximal element of the set.
Similarly if an infimum of a set belongs to the set, then it is called the minimum element
of the set.

Example
a) For the set S   2,3 the
Supremum = maximal element = 3 and the
Infimum = minimum element = -2
b) For the set S  1,3 , S has neither maximal or minimum element. The supremum
of S is 3 while the infimum is 1.

Completeness Axiom of the set of real numbers 

Every non-empty subset S of  which is bounded above has a supremum in  and
every non-empty subset S of  which is bounded below has an infimum in  .
The property of  as an ordered field taken together with the completeness axiom of 
makes  a complete ordered field.

Example 1.4
The set of rational numbers does not satisfy the completeness axiom. For consider a non-
empty subject S of Q, S  x  Q : Q  x 2  2. Then S is bounded and sup S  2 ,
which does not belong to Q. i.e 2  Q .

4
Exercise
Find the supremum, infimum, maximal and minimum elements of the following sets:

a) S  r  Q : 0  r  2 
1 1 
b) S    : n, m   
 m n 
m 
c) S   : m, n   , m  n 
n 
 1 
d) S  1  : n   
 n 
 n 1 
e) S   1 1   : n   
  n 

Density of Rational Numbers in 

The set of rational numbers is “dense” in  , in the sense that given any two real
numbers, there is a rational number between them (Infact, there are infinitely many
rational numbers).

Proposition 1.3: If S   1 n : n   then inf S  0

Proof-clear
Lemma 1.4: If t  0,  n   such that 0  1 n  t .
Proof: Since inf  1 n : n    0, and t  0 , then t is not a lower bound for the set
 1 n : n  . Thus there exists n  N such that 0  1 n  t .

Theorem 1.5 (The Density): If x and y are any real numbers with x  y then there exists
a rational number r  Q such that x  r  y.
Proof
Assume that x  0. Since y  x  0, it follows from Lemma 1.4 that  n   such that
1  y  x . Therefore, we have nx  1  ny. Since nx  0,  m   such that
n

m 1  nx  m (show this). Therefore m  nx  1  ny and so nx  m  ny. Thus, the

rational number r  m n satisfies x  r  y.

Corollary : If x and y are real numbers with x  y, then there exists an irrational number
such that x  z  y .
Proof
x y
By Theorem 1.5 between the real numbers and , there is a rational number r  0
2 2

5
 x y
(why?) such that r .
2 2
Then z  r 2 is an irrational number (why?) such that x  z  y.

Assessment Questions

Let A and B be any two bounded nonempty subsets of real numbers and let
A  B  a  b : a  A, b  B. Show that sup  A  B   sup A  sup B.

6
 TOPOLOGY OF REAL NUMBERS
Open sets

Definition 2.1: A subset N of  is said to be a neighbourhood of a point x   if there

exists an open interval a, b  x , x  such that x   a, b   . We call
a, b  ( x , x ) an   neighbourhood of x for some  0 .
Example 2.1
1. The set  is nbd (neighbourhood) of each of its points
2. The set of rational numbers is not a nbd of each of its points (why?)
3. An open interval  a, b  is a nbd of its points.
4. The empty set  is a nbd of each of its points.

Lemma 2.1
If N1 and N 2 are nbds of a point x   then N1  N 2 is also a nbd of x.
Proof (Exercise).

Definition 2.2: Let A be a subset of . A point x  A is called an interior point of A if 

an open interval  a, b  such that x   a, b   A. The set of all interior points of a set A is
called the interior set denoted by A  or int A.

Example 2.2
1. Let A  2,4, a closed interval; then 7 2 is an interior point of A, but neither 2
nor 4.
2. Note that 2,4  2,4  but 2,4  2,4


3.    ,    , Q   ,   

Definition 2.3: A subset B of  is open in  if for each x  B there exists a nbd V of x

such that x V  B. That is B is open if it contains an open interval about each of its
points. We also see that a set B is open if and only if B   B .

Example 2.3
i) Every open interval is an open set.
ii) The closed interval is not open
iii) The empty set is open, since  has no element, so there is no element in  which
is not an interior point.
iv) The sets , ,Q are not open sets.
v) The set  is open since    
The following basic result describes the manner in which open sets relate to the
operations of the union and intersection of sets in  .

7
Theorem 2.2 (Open sets)
a) The union of an arbitrary collection of open subsets in  is open.
b) The intersection of any finite collection of open sets in  is open.
Proof
a) Let U  A where A is open for each  . If U   , then U is open. If
 1
U   , let x  U , then x  A for some  . Since A is open, there exist
an open interval  a, b  such that x   a, b   A  x   a, b   U (since
A  U   ) . Hence U is a nbd of x implying that U is open, since x
was arbitrary.
n
b) Let  Ai 1 be a finite family of open sets. We show that B 
n
Ai is open.
i 1

If B   , then B is open. Let x  B, then x  Ai for all i  1,..., n.

Since Ai is open, then x  x  i , x  i   Ai for some i  0 for all i.
Letting  min i  , i  1, 2,..., n. Then x   x , x   B . So
x  ( x   , x   )  Ai  B. Hence B is open.
Note that, the intersection of an arbitrary collection of open sets need not be open.
For example:
 1 1 
a) Let Gn   ,  , for n  .
 n n
 1 1   1 1 
i. e G1   1, 1 , G2   ,  , G3   ,  ,…
 2 2  3 3
Then clearly, Gn is open for each n  . However, the intersections.

G Gn  0 , which is not open for 0 is not an interior point of 0 since
n 1

 0 , 0   0 for every  0 .

b) Let An   2  1 n ,3  1 n  for n   .

Then An is open for each n  . However,  An  2,3 which is not open.
n 1

Exercise
Find the interior of
 1 
a) A  1  n : n   
 10 
m 
b) B=  : m, n  , m  n 
n 
 1 
c) C= 1  : n   
 n 
d) D   n, n  1 : n  

8
 e) E   , 2    2,3  4  7,  

Closed sets
Definition (Limit point): A real number p is a limit point (or an accumulation point or a
cluster point) of S if every neighbourhood of p contains at least one point of S different
from p. That is if N is the neighbourhood of p, then p is a limit point of S if
 N   p S  .
The set of all limit points of S is denoted by S ' and is called the derived set of S.

Example 2.4
Let 0,1  0,1. Then every member of 0,1 is a limit point of 0,1. Thus though 0,1
do not belong to 0,1 they are its limit points for the nbd 1 , 1  of 1 contains
infinitely many elements of 0,1 .
Note that from this we see that a limit point may or may not belong to the set.

Example
a) The sets ,  and  have no limit points.
b) Every real number is a limit point of the sets Q and .

Lemma 2.3
Finite sets have no limit points.

Proof: Consider the set A  x1 , x2 ,..., xn  . Let p be an arbitrary real number and define
d1  x1  p ,..., d n  xn  p .
Let r  min d1 ,..., d n . Then the nbd N p  2r  of p contains no point of A. Hence p is not
a limit point of A. Since p was arbitrary, A has no limit points.
Next we have a characterization of a limit point of a set.

Theorem 2.4
A real number p is a limit point of a subset S of  iff every nbd of p contains infinitely
many points of S.
Proof
( ) Let p be a limit point of S and Let N p be a nbd of p, we show that N p has infinitely
many points. Suppose to the contrary that N p has a finite number of points of S distinct
from p. Then by the above lemma p is not a limit point of S, which is a contradiction?
Hence N p contains infinitely many points of S. Thus every nbd of p has infinitely many
of S, since N p was arbitrary.
 Conversely, given that every nbd of p has infinitely many points of S we have that
every nbd of p contains a point of S different from p. Hence p is a limit point of S.

9
Definition 2.5 (Closed set): Let X be a set and S  X . Then S is said to be closed in X if
it contains all its limit points. That is, if S is closed and s is a limit point of S then s  S.

Example 2.6
1. The  is closed for there is no element which is not a limit point of  .
2. The set a, b is closed since a, b are limits points of A and a, b  A.
3. Every finite set is closed for there is no limit point that does not belong to the set
since finite sets have no limit points.

Theorem 2.5
A set A is closed iff its complement Ac is open.
Proof
Suppose A is closed. We show that Ac is open. If Ac   , then Ac is open. Suppose that
Ac   and let x  Ac , then x  A . Since A is closed x is not a limit point of A so there
exists a neighbouhood N x of x such that N x A   , which implies that N  Ac . Thus
x  N  Ac implying that Ac is open.
Conversely, suppose that Ac is open. We want to show that A is closed. Let x be a limit
point of A. Then every neighbourhood of x in such that N  A   , thus x  Ac since Ac
is open. There x  A as required. Hence A is closed.

Theorem 2.6
a) The intersection of an arbitrary collection of closed sets in  is closed.
b) The union of any finite collection of closed sets in  is closed.
Proof
a) Let S  A , where each A is closed. We will shoe that S c is open. Now, by
 I

De Morgan’s Laws,
c
 
S   A   Ac .
c

  I   I
Since A is closed  , A c is open for all  and so Ac  S c is open Theorem
 I

2.2. Hence S is closed.

n
b) Similarly, let S  Ai , where each Ai is closed. Now, by De Morgan’s Laws,
i 1
c
 n  n
S   Ai  
c
Aic .
 i 1  i 1

10
 n
Since Ai is closed i , Ai c is open for all i and since Aic is open Theorem 2.2,
i 1
c
S is open. Hence S is closed.

Remark
The finiteness condition in part b) cannot be removed. For consider the example

An  1 n ,1 which is closed but  An  0,1 which is not closed.
n 1

Definition (Closure): Let S  , the set of all limit points of S is denoted by S ' and is
called the derived set. The closure of S denoted by S is defined to be S  S  S ' .
Note that if S is closed then S  S.

Definition (Dense set): A subset A of  is said to be dense in  if A   . That is

every point of X is a limit point of A.

Example
The set of rational numbers is dense in  .
Proof

For let x , if x  Q we are done, otherwise if x  Q then every neighbourhood

N x of x contains at least one other rational number. Therefore x is a limit point of Q
and so Q is dense in , i.e Q   .

Exercise
1. Find the derived sets of the following sets and state whether they are closed or not:
a) S  all int egers
b) S   n, n  1 : n  
c)
 1 
d) S   x : x  , n   
 n 
  1 n

e) S  1  : n  
 n 
 1 1 1 1 1 1 
f) S  1, 1,1 , 1 ,1 , 1 ,1 , 1 ,...
 2 2 3 3 4 4 

2. If A is open and B is closed, prove that A\B is open.

11
 SEQEUNCES

Convergence of Sequences

Loosely speaking a sequence is a collection of real numbers obeying a particular pattern

or definite rule or order.

Definition: A sequence of real numbers is a function whose domain is the set of natural
numbers   1, 2,... and range is a subset of  . That is f :    where for any
n   there exists a unique number xn such that f  n   xn .
We denote a sequence by  xn  or by listing the elements  x1 , x2 ,..., xn ,... and call xn the
nth term of the sequence.

Example 3.1
1 1 1 1 
a) x   , , , ... is the sequence of the even numbers.
2 4 6 8 
 
b) x  1, 1 2 , 1 4 , 18 , 116 ,...

 2  , give the formula of the sequence.

 1 n

Definition (Bounded sequence): A sequence  xn  is said to be bounded if there is a real

number M such that
xn  M n .

Example
The sequence s  1  2  is bounded for x
n n  1 or 0  xn  1 .

But the sequence s  2, 4, 6,8,...

 2n is not bounded.

Definition 3.2 (Limit point and convergence of sequences): A sequence  xn  of real

numbers is said to converge to x in  if for every  0 there exist a number N  N   
such that
xn  x  , n  N .
The number x is called the limit point of the sequence and we write
lim xn  x or xn  x as n   .
n

12
If the limit of a sequence exists then the sequence is said to be convergent otherwise the
sequence in said to be divergent.

Example
1
Show by first principles that lim    0 .
n
n 
Proof
We want to show that given  0 , there is a number N  N  such that
xn  0 , n  N .
Now,
1
xn  0   0 
n
1

n
 n 1

Certainly 1 n  ,  n  1 .
Therefore, choose N such that N  1 .

Example 3.4
Show that lim xn  1
n 

1
where xn  1   1
n

n2
Proof
Let  0 be given. Then
xn  1  
1
1   1  1 
n

n2
 1
n


n2
1
 n2 

1
n

Choose N such that, N  1 then

xn  1   ,  n  N .

13
Example 3.6
 3n  2 
Show that lim  3.
n 
 n 1 
Proof
Given  0 we have,

3n  2
 3 
n 1
3n  2  3n  3 1 1
   
n 1 n 1 n 1
1
 
n 1
 n 1  1

 n  1 1

Choose N such that N  1  1 , then
xn  3 ,  n  N .

Properties of sequences

Theorem 3.1
A sequence  xn  in  has a unique limit.
Proof
Suppose  xn  has two limits x1 and x2 . That is lim xn  x1 and lim xn  x2 .Then given
n  n 

any   0,  numbers N1 and N 2 such that,


xn  x1  , n  N1
2
and

xn  x2  , n  N 2 .
2
By the triangle inequality for all n  N  max  N1 , N 2  ,
 
x1  x2  x1  xn  xn  x2  x1  xn  xn  x2    .
2 2
Since   0 is arbitrary, this implies that
x1  x2  0  x1  x2  0  x1  x2 .

14
Theorem 3.2
Every convergent sequence of real numbers is bounded.

Proof
Suppose that  xn  converges to x i.e. lim xn  x and let  1 be given. Then  a
n

natural number N  N  such that

xn  x  1,  n  N .
 xn  xn  x  x  xn  x  x  1  x , n  N
Let M  max  x1 , x2 ,..., xN ,1  x .
Then xn  M , n . Hence  xn  is bounded.

Remark:
Conversely a bounded sequence need not be convergent for instance the sequence
  
 xn    1 1  1 n is bounded since xn  2, n ; but not convergent since
n

 
 xn   2, 3 2 ,  4 3 , 5 4 ,  6 5 , 7 6 ,... diverges.

Theorem 3.3
Let  xn  and  yn  be sequences of real numbers such that
lim xn  x and lim yn  y . Then
n n

a) lim  xn  yn   x  y.
n 

b) lim kxn  kx, k  constant.

n

c) lim  xn yn   xy.
n 

d) lim  n   x , provided that yn  0 for all n and y  0.

x
n   yn  y
Proof
a) To show that lim  xn  yn   x  y , we need to make the difference
n 

( xn  yn )  ( x  y) small for all n sufficiently large.

Using the triangle inequality, we have
( xn  yn )  ( x  y)  xn  x  yn  y

 xn  x  yn  y .
Now, given any   0 , since xn  x , there exists N1 such that n  N1 implies
xn  x   . Similarly, since yn  y , there exists N 2 such that n  N 2 implies
2
yn  y   . Thus, if we let N  max  N1 , N 2  , then n  N implies
2

15
 ( xn  yn )  ( x  y )  xn  x  yn  y       .
2 2
Therefore
lim  xn  yn   x  y.
n 

b) Exercise

c) We use the inequality,

xn yn  xy   xn yn  xn y    xn y  xy 
 xn yn  xn y  xn y  xy
 xn yn  y  y xn  x
Since the sequence  xn  converges, it must be bounded. Thus there exists M 1  0 such
that
xn  M1 , n.
Letting M  max M1 , y  , we obtain the inequality,
xn yn  xy  M yn  y  M xn  x .
Now, given any   0 , there exists N1 and N 2 such that


xn  x  , when n  N 1 and
2M

yn  y  , when n  N 2 .
2M
Now let N  max  N1 , N 2  , then
       
xn yn  xy  M yn  y  M xn  x  M  M      ..
 2M   2M  2 2
 xn yn  xy, since  0 was arbitrary.

 xn  1  , it suffices from part c) to show that

xn
d) Since
yn  yn 
lim 1  1 .
n  yn y
That is give  0 we must make
1 1 y  yn
   , for all n sufficiently large.
yn y yn y
Since y  0,  N1 , such that n  N1 
y
yn  y  . Thus for n  N1 , we have
2

16
 y y
yn  y   y  y n   y  y  y n  y  
2 2
(From x  y  x  y )
1
There also exists N 2 such that n  N 2 implies that yn  y   y
2

2
Let N  max  N1 , N2 . Then n  N implies that
1 1 y  yn 2 1 2
   2 . y  
yn y yn y y 2
Hence lim 1  1 .
yn y
Thus
lim  n   x .
x
n   yn  y

3.2.3 Subsequences and Cauchy Sequences

Subsequences

Definition 3.3 (Subsequence): Let X be a set and  xn  be a sequence in X. Let  nk  be a

sequence of positive integers such that n1  n2  ... . Then the sequence x 
nk is a
subsequence of  xn  .
Example 3.7
     
If  xn   1, 1 2 , 1 3 , 1 4 , 1 5 ,... then xnk  1 2 , 1 4 , 1 6 ,... is a subsequence of  xn  .

Theorem 3.4
A sequence  xn  of real numbers converges to x if and only if every subsequence of
 xn  converges to x.

Proof
 Suppose xn  x . We show that every subsequence  xnk  of  xn  converges to x .
Thus for every   0 there exists an integer N such that
xn  x  , n  N .
Take nk  N , then
xnk  x  , nk  N .
Since x 
nk is any subsequence of  xn  , it follows that every subsequence of  xn 
converges to x whenever xn  x

17
 Conversely, let every subsequence of  xn  converge to x . Then xn  x for  xn  is a
subsequence of itself.

Example 3.8
Consider the sequence defined by
1, if n is even
xn  
 0, if n is odd
That is  xn   0,1,0,1,0,... . Then the subsequence
x2n   1,1,1,... and converges to 1.
But the subsequence
x2n1  0, 0, 0,... and converges to 0.
Since 0  1, then  xn  diverges.

3.2.3.2 Cauchy sequences

Definition 3.4: A sequence  xn  of real numbers is said to be a Cauchy sequence if for
every   0 there exists an integer N  N    such that
xm  xn   ,  m, n  N .

Example 3.9
1
Show that the sequence  n  is Cauchy.
2 
Proof
We need to show that given   0 , there exists an integer N  N    such that
xm  xn  ,  m, n  N .
Let m  n  p, where p   , then

1 1 1  1 
 xm  xn  n p
  n 1  p    .
2 2 n
2  2 
 1  1
Since  1  p   1 , then n  
 2  2
1
ln  

 2n   n    .
1
 ln 2
1
ln  

Choose N    , hence an integer N exists such that
ln 2
xm  xn  ,  m, n  N .

18
 1
Hence  n  is Cauchy.
2 

Theorem 3.5
Every convergent sequence is Cauchy.

Proof
Suppose xn  x . Then for every   0 there exists an integer N such that

xn  x  , n  N .
2
Take m  n , then we have

xm  x  ,  m  n
2
By the triangle inequality,
 
xm  xn  xm  x  x  xn  xm  x  x  xn     , m, n  N .
2 2
Thus  xn  is Cauchy.

Remark 3.1
Not every Cauchy sequence is convergent.

Example 3.10
 xn   
1
Let X   0,1   and consider the sequence  in X. Then lim xn  0  X .
n n 

1 
However,  xn     is Cauchy. For let   0 be given, then
n
1 1 1
 0      n  .
n n n
1
Hence taking N  , we have

xn   ,  n  N .
Let m  n , then
xm  xn   ,  m, n  N . Hence  xn  is Cauchy.

Monotonic Sequences, Lim sup and Lim inf

Definition 3.5: Let  xn  be a sequence of real numbers, then we say that  xn 

a) Is monotonic increasing, xn  if xn  xn1 , n  N .
b) Is monotonic decreasing, xn  if xn  xn1 , n  N .

19
A sequence  xn  is monotonic if it is increasing or decreasing.

Example 3.10
 
i) The sequence 1, 2, 4, 8, ....  2n is monotonic increasing.

   
ii) The sequence 1, 1 , 1 , 1 , ....  1 n is monotonic decreasing.
3 9 27 3

Theorem 3.5
Let  xn  be a monotonic sequence. Then  xn  is convergent if and only if  xn  is
bounded.

Proof
Exercise

Definition 3.6 (Limit Superior, Limit Inferior): Let  xn  be a bounded sequence. A

subsequential limit of  xn  is any real number that is the limit of some subsequence of
 xn  . If S is the set of all subsequential limits of  xn  , then we define the limit superior
of  xn  to be
Lim sup xn  Lim xn  sup S .
Similarly, we define the limit inferior of  n to be
Lim inf xn  Lim xn  inf S .

Example 3.11
1
Let xn   1 
n
. We see that
n
1
xn   1   2 for all n, so the sequence  xn  is bounded. The first few terms are
n

0, 3 , 2 , 5 , 4 , 7 , 6 , .
2 3 4 5 6 7
The subsequence

 x2 n    3 2 , 5 4 , 7 6 ,  converges to 1 and the subsequence

 x2 n1  0, 2 3 , 4 5 , 6 7 ,  convergences to -1. Since these are the

subsequential limits, we have

Lim sup xn  Lim xn  sup 1, 1  1

20
 and
Lim inf xn  Lim xn  inf 1, 1  1 .

Theorem 3.6
Let  xn  be a sequence of real numbers, then  xn  converges if and only if
Lim xn = Lim xn and the number is real.

Proof
We know that xn  x iff every subsequence of  xn  converges to x . Thus S   x ,
where S is the set of subsequential limits of  xn  . Therefore sup S  inf S  x , that is

Lim xn = Lim xn . Conversely let Lim xn = Lim xn , then S   x . Therefore, every

subsequence of  xn  converges to x . But  xn  is a subsequence of itself. Hence  xn 
converges to x .

3.3 Exercise

1. Show that

 4n 2  3  4
lim  2
n  5n  2n
 .
  5

n 1
2. Prove that xn  does not converge to 0.
n2

21
 SERIES

Introduction to Infinite Series

Definition 4.1 (Infinite Series): Let  an  be an infinite sequence of real numbers. The
expression


a
n 1
n  a1  a2  a3 

is called an infinite series. The number an is called the nth term of the series. If we define

s1  a1
s2  a1  a2
s32  a1  a2  a3
.
.
.
n
sn  a1  a2   an   ai .
i 1

th
Then the sum sn is called the n partial sum of the series. The series a
n 1
n is said to be

convergent if the sequence sn  of partial sums converges and define

a
n 1
n = lim sn .
n

Otherwise the series is said to be divergent.

Example 4.1

1
The infinite series  n  n  1 converges to 1.
r 1

Solution
We have the partial series given by
1 1 1 1
sn     ... 
(1)(2) (2)(3) (3)(4) n  n  1
1 1   1 1   1 1  1 1 
             ...    
1 2   2 3   3 4   n n 1 
1
 1
n 1

22
  1 
Now lim sn  lim 1  1.
n  n 
 n  1 

Some Special Series

Geometric Series
One of the most useful series is the geometric series


 ar  a  ar  ar 2  ar 3   ar n 1 
n
,
n 0
where r is called the common ratio.

Now, the nth partial sum

n a 1  r n 
sn   ar  a  ar  ar  ...  ar
k 2 n 1
 , r  1.
k 0 1 r
a
If r  1 , then r n  0 as n   . Hence the geometric series converges to
1 r
provided r  1 . If r  1 , the series diverges.

Example 4.2
Find the sum of the series
 

 2
n
a) 1 n 1 b)
n 1
3 n0
3

Solution

a)  13
n 1
n 1  1
32
1
33
1
34
1
35


1
3  13  13  13   2 3

3
1 1 n
n 1
3
Now,


 13
n 1
n is a geometric series with r  1 3 and since r  1 3  1 this series converges and

 1 1

n 1
1  3
3n 1  1
 31
2 2
3 3


2
n 2 3
b)  1 2 2 2  is a geometric series with r  2  1 , hence the
n 0
3 3 3 3 3
series diverges.

23
Theorem 4.1 (nth Term Divergence Test)
If  an is a convergent series, then lim an  0 .
n

Proof
If  a n converges, then the sequence  sn  of partial sums has a finite limit, say a.
But an  sn  sn 1 ,
so
lim an  lim sn  lim sn1  a  a  0 .
n n n

The converse is not true in general, as we shall see later that  1n diverges (harmonic
series) but lim 1  0 .
n  n

Remark:
But if for the series a n , lim 1  0 , then the series
n  n
a n diverges.

Example 4.3
Determine whether the following series converges or diverges;
 
n 1
a)  b) 
n  0 2n  5 n 1 n ( n  1)

Solution:
a) We use the nth term divergence test

n n 1
an  and lim an  lim  lim  1 0
2n  5 n  n  2n  5 n 
2
5 2
n
1 1
b) an  and lim an 0
n(n  1) n  n(n  1)

Therefore, the series may or may not converge.

P - series
The series of the form

1 1 1 1 1
n 1 n
p
 p p p  p ,
1 2 3 n
where p is a constant is called a p  series.
The series converges if p  1 and diverges if p  1 .
If p  1 we have the series

24
 
1 1 1 1
 n  1 2  3 
n 1 n
 , 

called the harmonic series.

Convergence Tests

The Integral Test:

Theorem 4.2 (Integral test)

Let f be a continuous positive decreasing function of x defined for all x  1 by

f  n   an . Then the series  an and the integral  f  x  dx both converges and diverges
1
simultaneously.

Example 4.4
Prove that  1 converges if p  1 and diverges if p  1 .
np
Proof:
Let f x   1 and p  1 , then f satisfies the hypothesis of the integral test.
xp
Now,
  n
1
n
 x  p 1 
For  f  x  dx   p dx  lim  x  p dx  lim  
n  n   p  1
1 1
x 1  1
 n1 p 1 
 lim   
n  1  p 1 p 

 n1 p 1  1
 lim     
n  1  p p 1  p 1

(Because p  1 , 1  p  0 ).
1
Since
p 1
is a finite number, then  1n p converges if p  1 by integral test.

Next, if p  1 ,
 
1  n1 p 1 
1 f  x  dx  
dx   lim     .
p n  1  p 
1
x  1 p 
(Since p  1 , then 1  p  0 ). Hence the corresponding series diverges by integral test if
p  1.
When p  1 we get  1 n , called the harmonic series which is divergent for
 n
1
1   n 1 x dx  lim
 ln x 1
n
f x dx lim
n 

lim  ln n  ln1   .
n 

25
 
Thus  1n
n 1
diverges.

Example 4.5
1
1. Apply the integral test on the series n 2
1
.

Solution:
1
Let f  x   , then clearly f satisfies the hypothesis of the integral test.
x 12

Now,

 n
1
 f  x  dx  lim 
n
dx  lim  tan 1 x 
1
n 
1
x 1
2 n  1

 lim  tan 1 n  tan 1 1

n 

 lim  tan 1 n   
n   4

   
2 4 4


1
Since  f  x  dx converges, then  n
1
2
1
converges by integral test

1
2. n
n 1
2
converges since it is a p-series with p  2  1 .

Comparison test

Theorem 4.3
Let  an and  bn be infinite series of non-negative terms. That is, a n  0 and bn  0
for all n. Then
a) If  an converges and 0  bn  a n for all n, then  bn converges.
b) If a n diverges and 0  an  bn for all n, then b
n diverges.

Example 4.6
Consider the series
 1 n  12 . Now for all n  
we have
1 1
0 
n  1 nn  1
2

26
Since the series  1
n  n  1
is convergent, so this implies that the series

1
 n  1 2
also converges.

Example 4.6
Determine whether each of the following series converges or diverges;

n
a)  2
n 1 n  1

1
b) 
n 1 2 n  1

1
c) 
n 0 n !

Solution
  
n n 1
i)  2  2  32 ,
n 1 n  1 n 1 n n 1 n

1
i.e  3 2 converges being a p-series with p  3 2  1 .
n 1 n

n
Hence by comparison test n
n 1
2
1
converges.

1 1
ii)  for all n  1.
2n  1 2n
Therefore,

 
1 1 1  1

n 1 2 n  1
 n 1 2 n
  
2 n 1 n

1
i.e.  diverges being the harmonic series.
n 1 n

1
Hence  diverges by comparison test.
n 1 2 n  1

1 1 1 1
iii) 
n 0 n !
 11    
2 6 n!
1 1 1 1
 11    n
2 4 8 2

1  1 
 1  n  1    3   , geometric series with
n 0 2  1  1 
 2
a  1 and r  1 .
2

27
 
1
Hence  n ! converges by comparison test.
n 0

Absolute Convergence and Conditional Convergence

Absolute convergence
Definition 4.2: Let  an  be a sequence in  . If the series a n converges then the
series a n is to be absolutely convergent in .

Theorem 4.4
If a series is absolutely convergent then it is convergent.

Proof
Since an  an , for all n we have
 a  a
n n 
Hence  a n converges by Comparison Test.

Corollary 4.7
If  an diverges, then a n also diverges.

Example 4.7
 1
n

  1  1
n
The series 1 2 converges absolutely for .
n n2 n2

 1
n

Thus  n2
 1
n2
which is converges being a p-series with p  2. .

 1n
Hence the series  n2
converges absolutely.

The Alternating Series

Definition (Alternating Series) 4.3: .Alternating series is a series in which terms
alternate between positive and negative values;


  1
n 1
an  a1  a2  a3  a4  .
n 1

Theorem 4.8 (Alternating Series Test/Leibniz’s Theorem)

28
If a n  is a decreasing sequence of positive numbers and lim an  0 , then the series
n

  1
n 1
an converges.

Example 4.8
The series   1 1 n is convergent.
n 1

Solution:
   
Since the sequence 1 n is decreasing and lim 1 n  0 , we have that the alternating
n 

series   1 1 n converges.

n 1

Example 4.9
 1
n

 1 
The series 
n2 n n
converges since the sequence  n  is decreasing and has
n 
lim 1
n 
 n   0 (by L’ hospital’s Rule).
n

Conditional Convergence
Definition 4.4 (Conditional Convergence): A convergent sequence that is not
absolutely convergent is said to be conditionally convergent. That is if  an converges
but a n diverges, then a n is said to be conditionally convergent.

Example 4.10
 1
n 1

Determine the conditional convergence of 
n 1 n
.

Solution:

 1
n 1

1 1 1

n 1 n
 1
2

3

4
 .

Since the sequence  1  is decreasing and lim  1   0 , we have that the

 n n   n
 1
n 1

alternating series 
n 1 n
converges.

Now,
 1
n 1
  
1 1

n 1 n

n 1
 1 
n n 1 n 2

29
 
1
That is 
n 1
1
2
diverges being a p-series with p  1 2  1 .
n
 1
n 1

Hence 
n 1 n
is conditionally convergent.

The Ratio and Root Test

Theorem 4.4 (Ratio Test)

Let  an be a series of nonzero terms
an 1
a) If lim  1 , then the series converges absolutely.
n  an
an 1
b) If lim  1 , then the series diverges.
n  an
an 1
c) If lim  1 , then the test provides no information.
n  an
Example 4.11
n2
Determines whether or not the series  2n converges or diverges.
Solution
 n  1
2

an  n 2 , an1
2 n

2n1
1  n 1 
 
2
a 1 2
lim n 1  lim    lim 1  1  1 1.
n  a n  2
 n  n  2 n 2
n

Hence the series converges by Ratio Test.

Example 4.12
Determine the convergence or divergence of the series  1n 2 .
Solution
an  1 2 , an 1  1
 n  1
2
n
 2 
2

an 1  n    1   1.
lim  lim    lim
n  a n 
 n 1  n    1 1  
n
 n  
The test provides no information.

Example 4.13

30
 n!
Determine the convergence or divergence of the series n n
.

an 1 (n  1)! n n
lim  lim .
n  a n  ( n  1) n 1 n !
n

(n  1)n n
 lim
n  ( n  1) n 1

nn
 lim
n  ( n  1) n

n
 n 
 lim  
n  n  1
 
n
 1 
 lim 1    e   (L’ Hospital’s Rule)
1
n 
 n  1 
Hence the series converges by Ratio Test.
Note: The test is usually effective with series involving factorial expressions.

Example 4.13
Determine the convergence or divergence of the series 3 en n
.
Solution
an  3n e n , an1  3n1 e  
 n 1

3n 1 en 3 3
lim n 1 . n  lim   1
n  e 3 n  e e
Hence the series diverges by Ratio Test.

Theorem 4.5 (Root Test)

Given a series  an with all terms positive, and
let   lim n an
n 

a) If   1 then the series converges absolutely.

b) If   1, then the series diverges.
c) If   1 , then the test gives no information.

Example 4.14
The series  2 n e  n converges.

Solution:
1
 2n  n 2 2
For lim  n   lim   1 .
n  e n  e
  e

31
Hence the series converges by Root Test.

Example 4.15
The series  3n e n diverges.
For
1
 3n  n 3 3
lim  n   lim   1 .
n  e n  e
  e

Exercise
1. Determine whether the following series converge or diverge;

1
a) 
n 1 n3  1
 3n
b)  n!
 5 n

c)  2n
 1
n

2. Determine whether the series  converges absolutely, conditionally or
n2 n2  1
not at all.

 1
n

a) 
n2 n2  1
(1) n
b)  n3


1
3. Apply the integral test to the series n
n 1
2
1
.

4. Determine the convergence or divergence of each of the following series. Give

reasons in each case:

2n
n 
3
a) b)
2n n 1 n
n

 1
n 1
sin 2 n n
c)  2 d) 
n n 1
2

32
 n!
e) n n

33
 COUNTABLE AND UNCOUNTABLE SETS

Countable sets

Definition 5.1 (Infinite set): A non-empty subset S of  is said to be infinite if for any
given natural number n, however large there is a subset of S containing more than n
elements.

Example 5.1
The set S  x   : x  2 is infinite.

Definition 5.2 (Countable set): A set S is said to be countable if it is finite or if it is

infinite and its elements can be put into 1-1 correspondence with the set of natural
numbers. That is S can be listed as infinite sequence x1 , x 2 , x3 ,...

Example 5.2
The set of all even (odd) numbers is countable for its elements can be put into 1-1
correspondence with the set of natural numbers by definition. That is, there exists a
function f : N   defined by
f  n   2n for even numbers and f  n   2n  1 for odd numbers that sets up a 1-1
correspondence.

Example 5.3
The set of all integers is countable.

Proof
Consider the following arrangement of the sets and ;
 : 0 1  1 2  2 3  3...
 :1 2 3 4 3 4 5...
Then we can define a function f :    which sets up a 1-1 correspondence as follows:
n , if n is even
 2
f  n   n 1 .
 , if n is odd
 2

Properties of countable sets

Theorem 5.1
A subset of a countable set is countable.
Proof:
Let A be a subset of a countable set B. By definition of a countable set, there exists a 1  1
function f : B  N . Now, restrict to the domain of f to A to give a new function

34
g : B  N . Function g is 1  1 too since f is. So g is a 1  1 function that maps A into N.
Thus A is a countable set.

Theorem 5.2

Let An n1 be a sequence of countable sets. We show that

An is countable.
n 1
Proof
For each An there exists 1-1 correspondence. f n :   An , so we can write

An   An1 , An 2 ,... where f n  j   Anj . We can now arrange the elements of An in a
n 1
rectangular array:
A1 : A11  A12 A13  A14 ....
A2 : A21 A22 A23 A24 ....
A3 : A 31 A32 A33 A34 ...
A4 : A41 A42 A43 A44 ...

By moving along each diagonal of the array in the manner indicated, we obtain a listing

of all elements in An ;
n 1

 A11 , A12 , A21 , A31, A22 , A13 , A14 , A23 , A32 , A41,....
This procedure results in an infinite sequence of elements  z1 , z2 , z3 ,... of A which can

be put into 1-1 correspondence with the set of natural numbers. Hence n is
n 1
countable.

Uncountable set
Definition 5.3 Uncountable sets): A set S is said to be uncountable if is not countable.
That is if there is no one-one correspondence (function) between the elements of S and
the set of natural numbers.

Example 5.4
The set  of real numbers is uncountable.

Proof
Suppose that  countable. Then  can be written as a sequence of numbers
   x1 , x2 , x3 ,.... . Choose a1 , b1  such that x1   a1 , b1  . Having done this choose
a2 , b2 such that a1  a2  b2  b1 and x2  a2 ,b2  . Continuing this way, choose a n 1 , bn 1
such that a n  a n 1  bn 1  bn and xn 1   an 1 , bn 1  .
Note that for each n, xn   an , bn  and the numbers x1 x2 ,.., xn cannot be in the interval
a n , bn  .

35
 
Define A  an n  Z   and   sup A . Clearly  is a real number and so   x N for
some natural number N. Since  is an upper bound of A, aN   and since bN is an
upper bound of A and  is the least upper bound of A, we see that   bN . We therefore
conclude that xN   aN , bN  , contradicting the choice of a N and b N which shows that the
set  is uncountable.
Exercise

1. If the set F is countable and E is non-empty subset of F. Show that E is countable.

2. If the set E is countable, show that the set   a is countable.
3. Show that the set Q  of irrational numbers uncountable.

36
 FUNCTIONS

Limits of functions
Consider the function
x2  4
f  x 
x2
The function is defined for all x except at x  2 . If x  2, then f  x   x  2 .

Investigate what happens to the function f  x  as x  2 .

It can be shown that as x  2, f  x  4 .

In this case 4 is called the limit value of f x  as x  2 written:
 x2  4 
lim f  x   lim    4.
x 2 x 2
 x2 

Definition 6.1 (Limit): Let S   and f : S   be a function. A real number L is said

to be a limit point of f at point a  S if given any   0,  a   0 such that if x  S
f  x   L   whenever x  a   .

Note that f may or may not be defined as x  a .

We say that f  x   L as x  a and write
lim f  x   L .
x a

Example 6.1
Let f  x   x 2  1, x  . Prove that lim f  x   5 .
x 2

Solution
We need to show that given   0,  a   0 such that
x  2    f  x  5  .
Now for x   ,

f  x   5  x2  1  5  x2  4  x  2 x  2   x  2

Choose   . Then whenever x  2   we have
x2

f  x  5  x  2 x  2 . x  2  .
x2
Hence lim f  x   5 .
x 2

37
Properties of limits
(Proofs are similar to those as sequences)
a) If lim f  x  exists then it is unique.
x a

b) If lim f  x   L1 and lim g  x   L2 , then

x a x a

i) lim  f  x   g  x    L1  L2
xa

ii) lim  f  x  g  x    L1 L2
xa

L
lim  g  x    1 provided g  x   0 x and L2  0.
f x
iii)
xa   L
2

One-sided limits

Definition 6.2: Let S   and f : S   be a function. If for every x  S

f  x   L as x  a and x  a always, then we say that x  a from the right and
write x  a  and say that
lim f  x   L .
x a
Similarly,
If f  x   L as x  a and x  a always, then we say that x  a from the left and
write x  a  and say that
lim f  x   L .
x a

The limits lim f  x  and lim f  x  are called one-sided limits of f at a.

xa xa

Theorem 6.1
Let S  . Then
lim f  x   L iff lim f  x   lim f  x   L.
x a x a  x a

Example 6.2
x
a) lim does not exist
x 0 x

For
x
lim  1
x 0 x

and
x
lim  1
x  0 x

1 1
b) lim 2 and lim have infinite limit.
x 0 x x 0 x

38
 xx
c) lim has infinite limit.
x  xx

Exercise

 Show that from definition of a limit:

x2  x  1 1
lim  2
a) x1 x  1
x3  8
lim  125
b) x 2 x  x6
2

 Find
4x 4x
a) lim b) lim
x 0 x x 0 x
x2 1
c) lim
x 1 x  1

Continuity of a function

Definition 6.3 (Continuity): A function f : S   is said to be continuous at x  a

where a  S if
i) lim f  x  exists finitely
x a

ii) lim f  x   f  a  .
x a

Example 6.3
Let
 x2 1
 ,x 1
f  x    x 1
2 , x 1

Then
x 2  1  x  1 x  1
lim f  x   lim   2.
x 1 x 1 x 1 x 1
Also
x2 1
lim f  x   lim  x 1  2 .
x 1 x 1 x  1

Thus, lim f  x   2  f 1 . Hence f is continuous x  1.

x 1

Definition 6.4 (Pointwise continuity): A function f : S   is said to be continuous at

a  S if given any   0,    0 (depending both on  and the point) such that

39
 f  x   f  a    whenever x  a   .

Example 6.4
a) The function f  x   x 2 is continuous at every x  a .
For given any   0, a   0
can be found such that
x  a    f  x  f a   .
Now,
f  x   f  a   x 2  a 2   x  a  x  a 
 xa xa  xa

Choose   . Then whenever x  a   we have
xa

x2  a2  x  a x  a  x  a  .
xa
Hence f  x   x 2 is continuous at every x  a .
b) Prove that f x   x 2  2 x  6 or continuous at x  3.

Solution
Given   0 we show that , a   0 such that x  3   implies
f  x   f  3   .
Now,
f  x   f  3  x 2  2 x  6  21  x 2  2 x  15   x  5   x  3
 x 5 x 3   x 5

Choose   , then whenever x  3   we have
x5

f  x   f  3  x  5 x  3  x  5  .
x5
Hence f x   x 2  2 x  6 or continuous at x  3 .

Definition 6.5 (Discontinuity of a function): Points where a function fails to be

continuous are called points of discontinuity of the function. A function is said to be
discontinuous at a point x if it is not continuous at that point.

40
Example 6.5
Let
 x2  a2
 ,x  a
f  x   x  a
2 ,x  a

Then f x  is discontinuous at x  a since
lim f  x   2a  lim f  x   2a  2  f  a  .
x a  x a

Uniform continuity

Recall (Pointwise continuity)

We said that f : S   is said to be continuous at x  a if for every   0,    0
such that f  x   f  a    whenever x  a   , where  may depend on both
x and  .
That is for different points in S corresponding to pre-arranged   0 we get different
values of  . If we can find a uniform  that serves for all point of S corresponding to
given   0 , then we say that f is uniformly continuous on S.

Definition 6.6 (Uniform continuity): A function f defined on S is said to be uniformly

continuous on S if for every   0,  a   0 (depending on  only) such that.
f  x   f  y    whenever x  y   for all x, y  S .

Example 6.6
Prove that f  x   2 x is uniformly continuous on  .
Proof
Given   0 we want to make f  x   f  y    by making x sufficiently close to y. That
is we must show that for each   0,  a   0 such that
x  y    f  x  f  y  
Now,

f  x   f  y   2 x  2 y  2 x  y  2 . Choose   depending only on  . Then
2
whenever x  y   we have

f  x   f  y   2 x  y  2.  
2
Hence f is uniformly continuous on  .

41
Example 6.7
Prove that f  x   x 2 is uniformly continuous on  5,5.

Solution
Now
f  x   f  y   x 2  y 2   x  y  x  y 
 x y x y
 10 x  y  10

Choose   (depending on  alone). Then whenever x  y   we have
10

f  x   f  y   10 x  y  10.  .
10
Thus f  x   x 2 is uniformly continuous on  5,5.

Example 6.8
is uniformly continuous on  2,  .
1
The function f  x  
x
Proof
Given   0 and x, y   2,   we have,
1 1 1
f  x  f  y    x y
x y xy
1 1
x y  

4 4
Choose   4 , then whenever x  y   we have,
1 1
f  x  f  y  x  y  . 4   .
4 4

Negation of the definition of uniform continuity on S

A function f : S   is not uniformly continuous on S if there exists   0 such that

   0,  x, y  S such that x  y   but f  x   f  y    .

Example 6.9
The function f  x   1 x is not uniformly continuous in 0,1 .

Solution (BWOC)
Suppose that f is uniformly continuous in  0,1 so that    0,  a   0 such that
f  x   f  y    whenever x  y   , , x, y   0,1

42
 
Let x   and y  , where 0    1 .
1 
   
Then x  y        .
1   1  
But
1 1 1 1 
f  x  f  y    
x y  

 

Thus f  x   f  y    which contradicts the fact that if is uniformly continuous in
 0,1 .
Example 6.10
Show that f  x   x 2 is not uniformly continuous on  .

Solution
1 1  
Let   1 and   0 be given. Let x  and y   . Then x  y    . But
  2 2
1 1  2 
f  x  f  y  x  y x  y    . 1.
  2  2
Thus f is not uniformly continuous on . (why?) .

Suppose   2  0 , would it as well work. We must show that given any

  0,  x, y   such that
x y   f  x  f  y  2

For x, if we let y  x 
2

Then x  y   .
2
To make
f  x  f  y   x  y x  y  
2
We need to have x  y  . This prompts us to choose   1 8 (How?).
8

6.3 Exercise

x2  4 x  5
1. Let f  x  for x  5. How should f x  defined so that it is
x 5
continuous at?

43
  
2. Let f  x   x sin 1 x for x  0 and f  0   0. Show that f  x  is continuous at
x  0.

3. Prove that f  x   3x  5 is uniformly continuous on  1, 1 .

44