IMM2001 MATHEMATICS - II MODULE -1

INEQUALITIES
1) WEIERSTRASS 'S INEQUALITY
If a 1 , a 2 ,......... ., a n are positive numbers, less than 1, whose sum is denoted by s n , then

1
1  s n  (1  a 1 )( 1  a 2 )......... .......... ....( 1  a n )  (I)
1  sn

and
1
1  s n  (1  a 1 )( 1  a 2 )......... .......... ....( 1  a n )  (II)
1  sn

where, in the last inequality (II), it is supposed that sn  1.

Proof:

(1  a 1 )( 1  a 2 )  1  a 1  a 2  a 1 a 2  1  ( a 1  a 2 )  a 1 a 2  1  ( a 1  a 2 )

So, we get

(1  a 1 )( 1  a 2 )  1  ( a 1  a 2 ) (1)

Similarly, we have

(1  a 1 )( 1  a 2 )( 1  a 3 )  (1  a 1  a 2  a 1 a 2 )( 1  a 3 )  1  ( a 1  a 2 ) (1  a 3 ) (by (1))

 1  a 3  a1  a1a 3  a 2  a 2 a 3

 1  (a1  a 2  a 3 )  a1a 3  a 2 a 3

 1  (a1  a 2  a 3 )

So, we get

(1  a 1 )( 1  a 2 )( 1  a 3 )  1  ( a 1  a 2  a 3 ) (2)

Similarly, we get,

(1  a 1 )( 1  a 2 )( 1  a 3 )( 1  a 4 )  1  ( a 1  a 2  a 3  a 4 ) (3)

.....................................................................................

...................................................................................

(1  a 1 )( 1  a 2 )( 1  a 3 )( 1  a 4 ).......( 1  a n )  1  ( a 1  a 2  a 3  a 4  ......  a n )  1  s n

That is, we get

1
IMM2001 MATHEMATICS - II MODULE -1

(1  a 1 )( 1  a 2 )( 1  a 3 )( 1  a 4 ).......( 1  a n )  1  s n (A)

Similarly, if we take,

(1  a 1 )( 1  a 2 )  1  a 1  a 2  a 1 a 2  1  ( a 1  a 2 ) (4)

then, we get,

(1  a 1 )( 1  a 2 )( 1  a 3 ).  1  ( a 1  a 2  a 3 ) (5)

In the same way, we can get

(1  a 1 )( 1  a 2 )( 1  a 3 )( 1  a 4 ).......( 1  a n )  1  ( a 1  a 2  a 3  a 4  ......  a n )  1  s n

That is, we get

(1  a 1 )( 1  a 2 )( 1  a 3 )( 1  a 4 ).......( 1  a n )  1  s n (B)

Now, if we take
2
(1  a 1 )( 1  a 1 )  1  a 1  1 (as 0  a 1  1 )

This implies that

1
1  a1  (6)
1  a1

and

1
1  a1  (7)
1  a1

So, we have the following results from the equation (6)

1
1  a1 
1  a1

1
1  a2 
1  a2

....................

1
1  an 
1  an

Multiplying these, we get

2
IMM2001 MATHEMATICS - II MODULE -1

1
(1  a 1 )( 1  a 2 )......... (1  a n ) 
(1  a 1 )( 1  a 2 )........( 1  an )

which by equation (B), reduces to

1
(1  a 1 )( 1  a 2 )......... (1  a n )  (C)
(1  s n )

Now, we get the following results from equation (7),

1
1  a1 
1  a1

1
1  a2 
1  a2

....................

1
1  an 
1  an

Multiplying these, we get

1
(1  a 1 )( 1  a 2 )......... (1  a n ) 
(1  a 1 )( 1  a 2 )........( 1  an )

which by equation (A), reduces to

1
(1  a 1 )( 1  a 2 )......... (1  a n )  (D)
(1  s n )

with sn  1.

Combining equations (A) and (C), we get

1
1  s n  (1  a 1 )( 1  a 2 )......... .......... ....( 1  a n )  (I)
1  sn

and

Combining equations (B) and (D), we get

1
1  s n  (1  a 1 )( 1  a 2 )......... .......... ....( 1  a n )  (II)
1  sn

3
IMM2001 MATHEMATICS - II MODULE -1

2) Inequalities depend on the fact that the square of a real number is positive. Here, we
discuss some examples based on it.
1
Example 2.1: If a  0 and b  0 , then show that a  b  ab i.e.
2

Arithmetic mean > geometric mean

 
2
Proof: a  b  0

This implies that a  b  2 ab  0

1
This implies that a  b  ab
2

that is, Arithmetic mean > geometric mean of two positive numbers

Example 2.2: If a, b, c are positive numbers and not all are equal, then prove that

a  b  c  ab  bc  ca  9 abc

Proof:

a  b  c  ab  bc  ca   9 abc

 a b  abc  a c  ab  b c  abc  abc  bc  ac  9 abc

2 2 2 2 2 2

 a (b  c )  b(a  c )  c (b  a )  6 abc
2 2 2 2 2 2

 a (b  c  2 bc )  b ( a  c  2 ac )  c ( b  a  2 ab )
2 2 2 2 2 2

 a (b  c )  b (c  a )  c(a  b)
2 2 2

 0

Thus, we have the required inequality.

3) Sometimes, a special arrangement of terms or factors are required to get useful

inequalities
n
Example 3.1: Prove that n!  n 2
, where n is any positive integer
Proof:
n !  1 . 2 . 3 ........( n  1 ) n (1)
Similarly, we can also write
n !  n ( n  1 )........ 3 . 2 . 1 (2)
Equations (1) and (2) on multiplying give,
 n !  1 . n ( 2 .( n  1 )).  3 .( n  2 ) .......  r .( n  r  1 ) ....  n . 1 
2
(3)
Here, 1 r  n (4)
which implies that

4
IMM2001 MATHEMATICS - II MODULE -1

r 1 0 and r  n  0

This implies that

( r  1 )( r  n )  0

 r  r  nr  n  0
2

 nr  r  r  n
2

 r ( n  r  1)  n

Here, from,
r ( n  r  1)  n (5)
Putting r  1, 2 , 3 ,......, n in equation (3), we can prove that each pair in equation (3) is
greater than n , thus we get

 n !
2
 n . n .......... n  n
n

or
n
n!  n 2
.

Example 3.2: If a, b, c are positive numbers, where the sum of any two is greater than
the third, then prove that
1 1 1 1 1 1
    
b  c  a c  a b a  b  c a b c

Proof: It is given that the sum of any two of three positive numbers greater than the third,
so
a  b  c , b  c  a , and a  c  b , with a , b , c  0 .

With this,
a ( b  c  a )( a  b  c )  0 (1)
So, if we take,
 a  (b  c )
2 2 2
a

 ( a  b  c )( a  c  b )

 ( a  b  c )( a  c  b )
2
or, a

This implies that

 2 ( a  b  c )( a  c  b )
2
2a

or, a ( a  b  c  a  c  b )  2 ( a  b  c )( a  c  b )

(a  b  c  a  c  b) 2
or, 
( a  b  c )( a  c  b ) a

or, we get

5
IMM2001 MATHEMATICS - II MODULE -1

1 1 2
   (A)
a  b  c c  a b a

Similarly, we get
1 1 2
   (B)
a  c b b  c  a c

and

1 1 2
  (C)
b  c  a a  b  c b

Adding equations (A), (B) and (C), we get,

 1 1 1  1 1 1
2     2   
b  c  a c  a  b a  b  c a b c

or
1 1 1 1 1 1
    
b  c  a c  a b a  b  c a b c

4) If a, b, x are positive numbers, then

a  x a
 if a  b
b  x b

and
a  x a
 if a  b
b  x b

Proof: Let us consider,

a  b

This implies, ax  bx (all are positive numbers)\

This implies, ax  ab  bx  ab

This implies, a(x  b)  b(x  a)

or,
a  x a

b  x b

a  x a
Similarly, we get  with a  b.
b  x b

Example 4.1: Prove that

1 1 3 5 2n  1 1
 . . ......... 
2 n 1 2 4 6 2n 2n  1

6
IMM2001 MATHEMATICS - II MODULE -1

Solution: Let
1 . 3 . 5 ......( 2 n  1 )
un  (1)
2 . 4 . 6 ......( 2 n )

a  x a
By inequality (4), if a  b , then  ,
b  x b

11 1 2 1
for, 1  2,  , that is, 
2 1 2 3 2

Similarly, we can get

4 3 2n 2n  1
 ,......., 
5 4 2n  1 2n

With this, we can write,

2 4 2n 1 3 2n  1
. .....  . ......  un
3 5 2n  1 2 4 2n

Thus, we get,
2 . 4 . 6 ......( 2 n )
un  (2)
3 . 5 ......( 2 n  1 )

Multiplying equations (1) and (2), we get,

 1 . 3 . 5 ......( 2 n  1 )   2 . 4 . 6 ......( 2 n ) 
u n .u n   



 2 . 4 . 6 ......( 2 n )   3 . 5 ......( 2 n  1 ) 

or
1
un  (A)
2n  1

Also, from equation (1),

1 . 3 . 5 ......( 2 n  1 )
( 2 n  1) u n  ( 2 n  1)
2 . 4 . 6 ......( 2 n )

or
3 5 7 ( 2 n  1)
( 2 n  1) u n  . . .... (B)
2 4 6 2n

a  x a
Now, from inequality 4), if a  b , then 
b  x b

31 3 4 3
for, 3  2,  , that is,  .
2 1 2 3 2

With this, from equation (B), we can write,

3 5 7 ( 2 n  1) 4 6 (2n  2)
( 2 n  1) u n  . . ....  . ........
2 4 6 2n 3 5 ( 2 n  1)

7
IMM2001 MATHEMATICS - II MODULE -1

or
4 6 (2n  2)
( 2 n  1) u n  . ........ (C)
3 5 ( 2 n  1)

Multiplying equations (B) and (C), we get

2 3 5 7 ( 2 n  1)   4 6 (2n  2) 
( 2 n  1) u n   . . ....   . ........   n 1
2

2 4 6 2n  3 5 ( 2 n  1) 

or
2
( 2 n  1) u n  n 1
2

Thus, we get

n 1 n 1 1
un   
2n  1 2n  2 2 n 1

So, we have,
1
un  (D)
2 n 1

From equations (A) and (D), we get the required result,

1 1 3 5 2n  1 1
 . . ......... 
2 n 1 2 4 6 2n 2n  1

5) CAUCHY SCHWARTZ INEQUALITY

If a 1 , a 2 ,......... ., a n and b 1 , b 2 ,......... ., b n be two sets of positive real numbers, then

 a 1 b1  a 2 b 2  ....  a n b n   a 1  a 2  ....  a n
2
 2 2 2
b 1
2 2
 b 2  ....  b n
2

In short, it can be written as,
2
   2  2 
  a ibi     a i   bi 
 i   i  i 

However, the sign of equality hold only if

a1 a2 an
  .... 
b1 b2 bn

Proof: Let
2

2 2 2
ai  a 1  a 2  ....  a n  A
i

 a i b i  a 1 b 1  a 2 b 2  ....  a n b n  B
i

2
b
2 2 2
i
 b 1  b 2  ....  b n  C
i

8
IMM2001 MATHEMATICS - II MODULE -1

Then, for all values of , we have

 a   a 1   b1   a 2   b 2   a n   b n 
2 2 2 2
i
  bi  .........
i

 a 1
2
 a2
2
 ....  a n
2
   b 2
1
2
 b2
2
 ....  b n
2
  2  a 1
b 1  a 2 b 2  ....  a n b n 
 A  2B   C  0
2

Taking,

 a 
2
  bi  A  2B   C  0
2
i
(1)
i

Now, considering the following cases,

i) For the sign of equality to occur, we must have,
a i   bi  0, i

that is,
ai a1 a2 an
      .... 
bi b1 b2 bn

If we put, a 1    b 1 , a 2    b 2 ,......... ., lhs of equation (1) becomes zero

Hence,
A  2B   C  0
2

or
 C  2B  A  0
2
(2)
we should have, discriminant of (2) as
 4 AC
2
4B

 AC
2
or, B

that is,
2
   2  2 
  a ibi     a i   bi 
 i   i  i 

ii) Otherwise,
A  2B   C  0
2 

This implies,

 2 
2 2
B A B B
C  2       0
 2 2 
 C C C C 

This gives,

9
IMM2001 MATHEMATICS - II MODULE -1

  
2
AC  B
2
B 
C      2   0 (3)
  C  C 

AC  B  0
2
For this equation (2) to hold true,
 AC
2
or B

or
2
   2  2 
  a ibi     a i   bi 
 i   i  i 

6) HOLDER'S INEQUALITY1
If a 1 , a 2 ,......... ., a n and b 1 , b 2 ,......... ., b n be two sets of positive real numbers, where

1 1
  1, with p , q  1, then
p q

1 1

 p  p  q q
 a ibi    ai    bi 
i  i   i 

where the equality holds when

p 1
bi  c a i where c is any constant.

Note: If p  q  2, then Holder's inequality reduces to Cauchy's Schwartz inequality.

7) INEQUALITY BETWEEN ARITHMETIC AND GEOMETRIC MEAN
The arithmetic mean of n positive numbers, is greater than or equal to their geometric
mean, that is,
Arithmetic mean (A)  Geometric mean(G)
Proof: Here, we prove it under two cases:
Case - I: When n is in power of 2, that is, two, four, eight etc.
Because,
(a  b )  0
2

 (a  b )  4 ab  0
2

 (a  b )  4 ab
2

2
 a  b 
    ab
 2 

a  b 
This implies that    ab for two positive numbers, that is, A  G.
 2 

1
Without proof, statement is to be remembered

10
IMM2001 MATHEMATICS - II MODULE -1

Now, from
2
a  b
   ab , (1)
 2 

we can also write,

2
c  d 
   cd , (2)
 2 

and also, again squaring (1) we get

4
a  b
 a b
2 2
  (3)
 2 

Multiplying equations (1) and (2), we get

2 2
a  b c  d 
 B1 B 2 
2
abcd      (4)
 2   2 

a  b  c  d 
where B1   , B 2   ,
 2   2 

Using (3) in (4), we get

4
 a  b c  d 
  
a 
4 4 4
 B1  B 2   b  c  d a  b  c  d 
 B1 B 2 
2
 2 2 
abcd      4 4
  
 2   2  2 .2  4 
 
 

which gives
 a  b  c  d  1

    abcd 4 (5)
 4 

Thus, the result holds for four positive numbers a, b, c, d .

Proceeding in this way, we can show that when n is in power of 2, then

 a  b  c  d  ....  k  1

    abcd n
 n 

so that A  G.
Case - II: When n is not in power of 2, then, consider the set,
a , b , c ,.... k ,  ,  ,......, where  occur r times and nr is in power of 2, that is, include
number of  's (arithmetic means), so that the total numbers becomes in power of 2.
In the same way, we get

11
IMM2001 MATHEMATICS - II MODULE -1

 a  b  c  ....  k  r  
a , b , c ,.... k .   
r

 n  r 

or
G   Α   A 
n r nr

which implies
A  G.
Note: In this inequality, equality holds when all positive numbers are equal.
 1 . 3 . 5 ....( 2 n  1 )
n
Example 7.1: Show that n

Solution: from the inequality A > G, we have

1  3  5  .....( 2 n  1 )  1

 1 . 3 . 5 .....( 2 n  1 )  n (1)
n

Now, from the sum of series, we know that

n
1  3  5  .....( 2 n  1  2  1  ( n  1 ). 2   n
2

Putting this in (1), we get

2 1
n
 1 . 3 . 5 .....( 2 n  1 )  n
n

or

 1 . 3 . 5 ....( 2 n  1 )
n
n

Example 7.2: If a 1 , a 2 ,......... ., a n are positive and ( n  1 ) s  a 1  a 2  ........., an

then show that

.. a n  ( n  1 ) ( s  a 1 )( s  a 2 ).......( s  an )
n
a 1 . a 2 ,.........

Solution:
ns  s  a 1  a 2  ........., an

or
s  s  .... n times  s  a 1  a 2  ........., an

or
( s  a 2 )  ( s  a 3 )  .......  ( s  a n ) a1
 (1)
n 1 n 1

Now, arithmetic mean  geometric mean

so, we have
( s  a 2 )  ( s  a 3 )  .......  ( s  a n )
 n
( s  a 2 )( s  a 3 )....( s  a n ) (2)
n 1

12
IMM2001 MATHEMATICS - II MODULE -1

From (1) and (2), we get

a1
 n
( s  a 2 )( s  a 3 )....( s  a n ) (3)
n 1

Now, there will be n such inequalities, like

a2
 n
( s  a 1 )( s  a 3 )....( s  a n )
n 1

...........................................
Multiplying all such inequalities, we get the result.
SIMPLE CONTINUED FRACTIONS
The expression of the form
1
a1 
1
a3 
a 3  .......

is called a simple continued fraction, where a 1 , a 2 , a 3 ...... are positive integers, where a1

may be zero.
The simple continued fraction in more compact from can be written as:
1 1 1
a1  ......... .....
a 2  a 3  ...... an 

251
Example: Reduce to a simple continued fraction.
802

1
m 251 802
Solution:  
n 802 251

Here, a1  0

Divide 802 by 251, we get quotient a2  3 and remainder r 2  49 .

Divide 251 by 49, we get quotient a3  5 and remainder r3  6 .

Divide 49 by 6, we get quotient a4  8 and remainder r4  1 .

Divide 6 by 1, we get quotient a5  6 and remainder r5  0 .

Stop the division, as you get the remainder as zero.

So,
251 1 1 1 1 1 1 1
 a1  .  0 
802 a2  a3  a4  3 5 8 6 

13