Chapter 2 – Sequences

Course Title: Real Analysis 1 Course Code: MTH321

Course instructor: Dr. Atiq ur Rehman Class: MSc-II
Course URL: www.mathcity.org/atiq/fa14-mth321

Sequence
A sequence is a function whose domain of definition is the set of natural
numbers.
Or it can also be defined as an ordered set.
Notation:
An infinite sequence is denoted as

{sn } or sn : n  or s1, s2 , s3 ,... or simply as sn  ,
n 1

e.g. i) n  1,2,3,  .

1   1 1 
ii)    1, , ,  .
n  2 3 
iii) (1)n1  1, 1,1, 1, .
Subsequence
It is a sequence whose terms are contained in given sequence.
 
A subsequence of {sn } is usually written as {snk } .
n 1

Increasing Sequence
A sequence sn  is said to be an increasing sequence if sn1  sn  n  1 .
Decreasing Sequence
A sequence sn  is said to be an decreasing sequence if sn1  sn  n  1 .
Monotonic Sequence
A sequence sn  is said to be monotonic sequence if it is either increasing or
decreasing.
s
sn  is monotonically increasing if sn1  sn  0 or n1  1,  n  1 .
sn
s
sn  is monotonically decreasing if sn  sn1  0 or n  1,  n  1 .
sn1
Strictly Increasing or Decreasing
sn  is called strictly increasing or decreasing according as
sn1  sn or sn1  sn  n  1 .
 Sequences and Series -2-

Bernoulli’s Inequality
Let p  , p  1 and p  0 then for n  2 we have
1  p   1  np .
n

Proof:
We shall use mathematical induction to prove this inequality.
If n = 2
L.H.S  (1  p)2  1  2 p  p 2 ,
R.H.S  1  2 p ,
 L.H .S  R.H .S ,
i.e. condition I of mathematical induction is satisfied.
Suppose 1  p   1  kp .................. (i) where k  2
k

1  p   1  p 1  p 
k 1 k
Now
 1  p 1  kp  using (i)
 1  kp  p  kp 2
 1  (k  1) p  kp 2
 1  (k  1) p ignoring kp 2  0 ,
 1  p   1   k  1 p .
k 1

Since the truth for n  k implies the truth for n  k  1 therefore condition II of
mathematical induction is satisfied. Hence we conclude that 1  p   1  np . 
n

Example:
 1 n 
Prove that 1    is an increasing sequence.
 n  
n
 1
Let sn  1   where n  1 .
 n
1
To prove that this sequence is an increasing sequence, we use p  , n  2 in
n2
Bernoulli’s inequality to have
n
 1  n
1  2   1  2
 n  n
n
  1  1   1
  1  1     1 
  n  n   n
1 n 1 n n 1 n 1
 n 1 
n
 1  1  n   1 
 1    1          1  
 n  n  n   n 1  n 1
 sn  sn1  n  1.
This shows that sn  is increasing sequence. 
 Sequences and Series -3-

Example:
 1  
n 1

Prove that a sequence 1    is a decreasing sequence.

 n  
n 1
 1
Let tn  1   ; n  1.
 n
1
We use p  2 in Bernoulli’s inequality.
n 1
n
 1  n
1  2  1 2 ……….. (i)
 n 1 n 1
where
1 n2  n  n 
1 2  2   
n  1 n  1  n  1  n  1 
 1  n  1   n 
 1  2    …………… (ii)
 n  1  n   n  1 
n n
 1   n 
Now tn1  1    
 n 1  n 1
n
 1  n  1  
  1  2   from (ii)
  n  1  n  
1   n 1
n n

 1  2   
 n 1  n 
n  n  1 
n

 1  2   from (i)
 n  1  n 
 1  n  1 
n
n n 1
 1     
 n  n  n2  1 n2 n
n 1
 n 1
   tn ,
 n 
i.e. tn1  tn .
Hence the given sequence is decreasing sequence. 
 Sequences and Series -4-

Bounded Sequence
A sequence sn  is said to be bounded if there exists a positive real number 
such that sn    n  .
If S and s are the supremum and infimum of elements forming the bounded
sequence sn  we write S  sup sn and s  inf sn .
All the elements of the sequence sn such that sn    n  lie with in the
strip  y :    y   . But the elements of the unbounded sequence can not be
contained in any strip of a finite width.
Examples
 (1) n 
(i) un     is a bounded sequence
 n 
(ii) vn   sin nx is also bounded sequence. Its supremum is 1 and infimum is 1 .
(iii) The geometric sequence ar n1 , r  1 is an unbounded above sequence. It is
bounded below by a.
 n 
(iv)  tan  is an unbounded sequence.
 2 
Convergence of the Sequence
A sequence sn  of real numbers is said to convergent to limit ‘s’ as n   , if
for every positive real number   0 , there exists a positive integer n0 , depending
upon  , such that sn  s    n  n0 .

Theorem
A convergent sequence of real number has one and only one limit (i.e. Limit of
the sequence is unique.)
Proof:
Suppose sn  converges to two limits s and t, where s  t .
s t
Put   then there exits two positive integers n1 and n2 such that
2
sn  s    n  n1
and sn  t    n  n2 .
 sn  s   and sn  t   hold simultaneously  n  max(n1, n2 ) .
Thus for all n  max(n1, n2 ) we have
s  t  s  sn  sn  t

 sn  s  sn  t
     2
 Sequences and Series -5-

 s t 
 s  t  2 
 2 
 s t  s t
Which is impossible, therefore the limit of the sequence is unique. 
Note: If sn  converges to s then all of its infinite subsequence converge to s.
Cauchy Sequence
A sequence  xn  of real number is said to be a Cauchy sequence if for given
positive real number  ,  a positive integer n0 ( ) such that
xn  xm    m, n  n0
Theorem
A Cauchy sequence of real numbers is bounded.
Proof:
Let sn  be a Cauchy sequence.
Take   1 , then there exits a positive integers n0 such that
sn  sm  1  m, n  n0 .
Fix m  n0  1 then
sn  sn  sn0 1  sn0 1
 sn  sn0 1  sn0 1
 1  sn0 1  n  n0
  n  1 , and   1  sn0 1 ( n0 changes as  changes)
Hence we conclude that Sn  is a Cauchy sequence, which is bounded one. 
Note:
(i) Convergent sequence is bounded.
(ii) The converse of the above theorem does not hold.
i.e. every bounded sequence is not Cauchy.
Consider the sequence sn  where sn  (1)n , n  1 . It is bounded sequence because
(1)n  1  2  n  1.
But it is not a Cauchy sequence if it is then for   1 we should be able to find a
positive integer n0 such that sn  sm  1 for all m, n  n0 .
But with m  2k  1 , n  2k  2 when 2k  1  n0 , we arrive at
sn  sm  (1)2 n2  (1)2 k 1
 11  2 1 is absurd.
Hence sn  is not a Cauchy sequence. Also this sequence is not a convergent
sequence. (it is an oscillatory sequence).
……………………………
 Sequences and Series -6-

Divergent Sequence
A sn  is said to be divergent if it is not convergent or it is unbounded.
e.g.  n  is divergent, it is unbounded.
2

(ii) (1)  tends to 1 or -1 according as n is even or odd. It oscillates finitely.

n

(iii) (1) n is a divergent sequence. It oscillates infinitely.

n

Note: If two subsequence of a sequence converges to two different limits then the
sequence itself is a divergent.

Theorem
If sn  un  tn  n  n0 and if both the sn  and tn  converge to same limits as
s, then the sequence un  also converges to s.
Proof:
Since the sequence sn  and tn  converge to the same limit s, therefore, for
given   0 there exists two positive integers n1, n2  n0 such that
sn  s    n  n1
tn  s    n  n2
i.e. s    sn  s    n  n1
s    tn  s    n  n2
Since we have given
sn  un  tn  n  n0
 s    sn  un  tn  s    n  max(n0 , n1, n2 )
 s    un  s    n  max(n0 , n1, n2 )
i.e. un  s    n  max(n0 , n1, n2 )
i.e. lim un  s . 
n

Example
1
Show that lim n  1n
n

Solution
Using Bernoulli’s Inequality
n
 1  n
1   1  n 1  n.
 n n
Also
2

1   1  
2 n

 n
n 2 1

 1      1    
n
 n  1,
n

 n   n  
 Sequences and Series -7-

2
 1 
1
 1  n  1 
n
 ,
 n 
2
 1 
1
 lim1  lim n  lim 1 
n
 ,
n n n
 n
1
 1  lim n  1.n
n
1
i.e. lim n  1 .n

n
……………………..

Example
 1 1 1 
Show that lim    ............  0
n ( n  1) 2 (n  2) 2 (2n) 2 

Solution
Consider
 1 1 1 
sn      
 (n  1) (n  2)
2 2
(2n) 2 
and
n n
 s n 
(2n)2 n2
1 1
  sn 
4n n
1 1
 lim  lim sn  lim
n 4n n n n

 0  lim sn  0
n
 lim Sn  0 
n

Theorem
If the sequence sn  converges to s then  a positive integer n
1
such that sn  s .
2
Proof:
1
We fix   s  0
2
  a positive integer n1 such that
sn  s   for n  n1
1
 sn  s  s
2
 Sequences and Series -8-

Now
1 1
s  s  s
2 2
 s  sn  s  s   sn  s 
1
 s  sn . 
2
Theorem
Let a and b be fixed real numbers if sn  and tn  converge to s and t
respectively, then
(i) asn  btn  converges to as + bt.
(ii) sn tn  converges to st.
s  s
(iii)  n  converges to , provided tn  0  n and t  0 .
 tn  t
Proof:
Since sn  and tn  converge to s and t respectively,
 sn  s    n  n1 
tn  t    n  n2 
Also    0 such that sn    n  1 ( sn  is bounded )
(i) We have
 asn  btn    as  bt   a  sn  s   b tn  t 
 a  sn  s   b  tn  t 
 a  b  n  max (n1, n2 )
 1 ,
where 1  a   b  a certain number.
This implies asn  btn  converges to as + bt.
(ii) sn tn  st  sn tn  sn t  sn t  st
 sn  tn  t   t  sn  s   sn   tn  t   t   sn  s 
   t   n  max (n1, n2 )
 2 , where  2     t  a certain number.
This implies sn tn  converges to st.
1 1 t  tn
(iii)  
tn t tn t
tn  t  1
  1
 n  max (n1, n2 ) tn  t
tn t 2 t t 2
 Sequences and Series -9-

 
 2
 3 , where  3  2
a certain number.
1 t 1 t
2 2
1 1
This implies   converges to .
 tn  t
s   1  1 s
Hence  n    sn   converges to s   .  from (ii)  
 tn   tn  t t
Theorem
For each irrational number x, there exists a sequence rn  of distinct rational
numbers such that lim rn  x .
n

Proof:
Since x and x + 1 are two different real numbers
 a rational number r1 such that
x  r1  x  1
Similarly  a rational number r2  r1 such that
 1
x  r2  min  r1 , x    x  1
 2
Continuing in this manner we have
 1
x  r3  min  r2 , x    x  1
 3
 1
x  r4  min  r3 , x    x  1
 4
……………………………..........
……………………………..........
……………………………..........
 1
x  rn  min  rn1 , x    x  1
 n
This implies that  a sequence rn  of the distinct rational number such that
1 1
x   x  rn  x  .
n n
Since
 1  1
lim  x    lim  x    x .
n
 n  n  n
Therefore
lim rn  x . 
n
 Sequences and Series - 10 -

Theorem
Let a sequence sn  be a bounded sequence.
(i) If sn  is monotonically increasing then it converges to its supremum.
(ii) If sn  is monotonically decreasing then it converges to its infimum.
Proof
Let S  sup sn and s  inf sn
Take   0
(i) Since S  sup sn
  sn0 such that S    sn0
Since sn  is  (  stands for monotonically increasing )
 S    sn0  sn  S  S   for n  n0
 S    sn  S   for n  n0
 sn  S   for n  n0
 lim sn  S
n
(ii) Since s  inf Sn
  sn1 such that sn1  s  
Since sn  is  . (  stands for monotonically decreasing )
 s    s  sn  sn1  s   for n  n1
 s    sn  s   for n  n1
 sn  s   for n  n1
Thus lim sn  s 
n

Note
A monotonic sequence can not oscillate infinitely.
Example:
 1 n 
Show that 1    is bounded sequence.
 n  
 1 n 
Consider sn   1   
 n  
As shown earlier it is an increasing sequence
2n n
 1   1 
Take s2 n  1   , then s2 n  1   ,
 2n   2n 
n n
1  2n  1  1 
     1  
s2 n  2n  1  s2 n  2n  1 
Using Bernoulli’s Inequality we have
 Sequences and Series - 11 -

n

1
 1
n
 1
n

1 1  1   1  n
 
s2 n 2n  1 2n 2  2n  1  2n  1

 s2 n  2  n  1,2,3,
 s2 n  4  n  1,2,3,
 sn  s2n  4  n  1,2,3,
Which show that the sequence sn  is bounded one.
Hence sn  is a convergent sequence the number to which it converges is its
supremum, which is denoted by ‘e’ and 2  e  3 . 
Recurrence Relation
A sequence is said to be defined recursively or by recurrence relation if the
general term is given as a relation of its preceding and succeeding terms in the
sequence together with some initial condition.
Example:
1
Let t1  0 and let tn  be defined by tn1  2  for n  1 .
tn
(i) Show that tn  is decreasing sequence.
(ii) It is bounded below.
(iii) Find the limit of the sequence.
1
Let t1  0 and let tn  be defined by tn1  2  ; n 1
tn
 tn  0  n 1
1
Also tn  tn1  tn  2 
tn
tn2  2tn  1  tn  1
2

   0.
tn tn
 tn  tn1  n  1.
This implies that tn is monotonically decreasing.
Since tn  1  n  1,
 tn is bounded below  tn is convergent.
Let us suppose lim tn  t .
n

 1
Then lim tn1  lim tn  lim  2    lim tn
n n n tn  n

1 2t  1
 2 t  t  2t  1  t 2  t 2  2t  1  0
t t
  t  1  0  t  1.
2

 Sequences and Series - 12 -

Theorem
Every Cauchy sequence of real numbers has a convergent subsequence.
Proof:
Suppose sn  is a Cauchy sequence.
Let   0 then  a positive integer n0  1 such that

snk  snk 1   nk , nk 1, k  1,2,3,.........
2k
Put 
bk  sn1  sn0     s  s 
 sn2  sn1  nk nk 1

 bk  s n1  sn0   s  s    s  s 
n2 n1 nk nk 1

 s n1  sn0   s  s    s  s 
n2 n1 nk nk 1

  
  2
 
2 2 2k
1 1 1 
  2   k 
2 2 2 


2 2 
 1 1  1k 
 


1
 1 k 
 1  12   2 
 
 bk    k 1
 bk  is convergent
bk  snk  sn0  snk  bk  sn0 ,
 
where sn0 is a certain fix number therefore snk which is a subsequence of sn  is
convergent. 

Theorem (Cauchy’s General Principle for Convergence)

A sequence of real number is convergent if and only if it is a Cauchy sequence.
Proof:
Necessary Condition
Let sn  be a convergent sequence, which converges to s .
Then for given   0  a positive integer n0 , such that

sn  s   n  n0
2
Now for n  m  n0
sn  sm  sn  s  sm  s
 sn  s  sm  s
 
 .
 
2 2
Which shows that sn  is a Cauchy sequence.
 Sequences and Series - 13 -

Sufficient Condition
Let us suppose that sn  is a Cauchy sequence then for   0 ,  a positive
integer m1 such that

sn  sm   n, m  m1 ……….. (i)
2
Since sn  is a Cauchy sequence
 
therefore it has a subsequence snk converging to s (say).
  a positive integer m2 such that

snk  s   n  m2 ……….. (ii)
2
Now
sn  s  sn  snk  snk  s
 sn  snk  snk  s
 
 
  n  max(m1, m2 ) ,
2 2
this shows that sn  is a convergent sequence. 

Example
 1 1 1
Prove that 1    ...............   is divergent sequence.
 2 3 n
Let tn  be defined by
1 1 1
tn  1    ...............  .
2 3 n
For m, n , n  m we have
1 1 1
tn  tm    ............. 
m 1 m  2 n
1 m
 ( n  m) = 1  .
n n
In particular if n  2m then
1
tn  tm  .
2
This implies that tn  is not a Cauchy sequence therefore it is divergent. 

 Sequences and Series - 14 -

Theorem (nested intervals)

Suppose that I n  is a sequence of the closed interval such that I n   an , bn  ,
I n1  I n  n  1 , and  bn  an   0 as n   then I n contains one and only one
point.
Proof:
Since I n1  I n
 a1  a2  a3   an1  an  bn  bn1   b3  b2  b1
an  is increasing sequence, bounded above by b1 and bounded below by a1 .
And bn  is decreasing sequence bounded below by a1 and bounded above by b1 .
 an  and bn  both are convergent.
Suppose an  converges to a and bn  converges to b.
But a  b  a  an  an  bn  bn  b
 an  a  an  bn  bn  b  0 as n   .
 a b
and an  a  bn  n  1 . 
Theorem (Bolzano-Weierstrass theorem)
Every bounded sequence has a convergent subsequence.
Proof:
Let sn  be a bounded sequence.
Take a1  inf sn and b1  sup sn
Then a1  sn  b1  n  1 .
Now bisect interval  a1 , b1  such that at least one of the two sub-intervals contains
infinite numbers of terms of the sequence.
Denote this sub-interval by  a2 , b2  .
If both the sub-intervals contain infinite number of terms of the sequence then
choose the one on the right hand.
Then clearly a1  a2  b2  b1 .
Suppose there exist a subinterval  ak , bk  such that
a1  a2   ak  bk   b2  b1
1
  bk  ak   k  b1  a1 
2
Bisect the interval  ak , bk  in the same manner and choose  ak 1 , bk 1  to have
a1  a2   ak  ak 1  bk 1  bk   b2  b1
1
and bk 1  ak 1  k 1  b1  a1 
2
This implies that we obtain a sequence of interval  an , bn  such that
 Sequences and Series - 15 -

1
bn  an  (b1  a1 )  0 as n   .
2n
 we have a unique point s such that
s   an , bn 
there are infinitely many terms of the sequence whose length is   0 that contain s.
For   1 there are infinitely many values of n such that
sn  s  1
Let n1 be one of such value then
sn1  s  1
Again choose n2  n1 such that
1
sn2  s 
2
 
Continuing in this manner we find a sequence snk for each positive integer k such
that nk  nk 1 and
1
snk  s   k  1,2,3,............
k
Hence there is a subsequence snk   which converges to s. 

Limit Inferior of the sequence

Suppose sn  is bounded below then we define limit inferior of sn  as follow
lim  inf sn   lim uk , where uk  inf sn : n  k
n n
If sn is not bounded below then
lim  inf sn    .
n

Limit Superior of the sequence

Suppose sn  is bounded above then we define limit superior of sn  as follow
lim  sup sn   lim vk , where vk  inf sn : n  k
n n
If sn is not bounded above then we have
lim  sup sn    .
n

Note:
(i) A bounded sequence has unique limit inferior and superior
(ii) Let sn  contains all the rational numbers, then every real number is a
subsequencial limit then limit superior of sn is   and limit inferior of sn is 
 1
(iii) Let sn   (1)n 1  
 n
then limit superior of sn is 1 and limit inferior of sn is 1 .
 Sequences and Series - 16 -

 1
(iv) Let sn  1   cos n .
 n
Then uk  inf sn : n  k
 1   1   1  
 inf 1   cos k , 1   cos( k  1) , 1   cos( k  2) ,................ 
 k   k 1  k 2 
 1 
 1  k  cos k if k is odd
  
1  1  cos( k  1) if k is even
 k  1 
 lim  inf sn   lim uk  1
n n

Also vk  supsn : n  k
 1 
  1   cos(k  1) if k is odd
 k 1

1  1  cos k if k is even
 k 
 lim  inf sn   lim vk  1 
n n



Theorem
If sn  is a convergent sequence then
lim sn  lim  inf sn   lim  sup sn 
n n n

Proof:
Let lim sn  s then for a real number   0 ,  a positive integer n0 such that
n

sn  s    n  n0 ……….. (i)
i.e. s    sn  s    n  n0
If vk  supsn : n  k
Then s    vn  s    k  n0
 s    lim vn  s    k  n0 …………. (ii)
k 
from (i) and (ii) we have
s  lim supsn 
k 

We can have the same result for limit inferior of sn  by taking
uk  inf sn : n  k . 

