POWER SERIES SOLUTIONS

1.1 SOME BASIC DEFINITIONS

1.1.1 Power series

An infinite series of the form



 a x  x   a 0  a1 x  x0   a 2 x  x0   
n 2
n 0
n 0
…… (1)
is called a power series in x  x0 .
For example, the exponential function has
the power series
 n 2 3
x x x
e 
x
 1 x   
n 0 n ! 2 ! 3! …………. (2)

There

are three possibilities for convergence of
 a n x  x 0  :
n

n 0

1. The series may converge only for x  x 0 .

2. The series may converge for all real number
x. 
1 n
For example the series  n!
n 0
x

converge for all x   .

3. There is a positive number r , called the
radius of convergence of the series, such that


 a x  x 
n 0
n 0
n
converges if x  x0  and
r diverges if x  x 0  r.
 
x0  r , x0  ris called the open interval of
convergence.
For series (2) the interval of convergence is
( ,  ) i.e. the real line.
The series may or may not converge at
and .
 r use thex0ratio
xWe0 may  r test to find the radius of
convergence of a power series.
Ratio Test
Suppose bn  0 for n  0, 1, 2,and that
bn 1
lim L
n  b
n

in which L may be infinity.


Then
b
n 0
n

converges absolutely if L < 1 and diverges if L > 1

If L = 1 , then this test allows no conclusion.
Question 1
Find the interval of convergence of

(1) n

n  0 9 ( n  1)
n
( x  2 ) 2n

Solution
Let
(1) n
bn  n ( x  2) 2n

9 (n  1)
 n 1
(1) 2n2
n 1
( x  2)
bn 1 9 (n  2)

bn (1) n
( x  2) 2n

9 (n  1)
n

(1) n
2
 x2
9(n  2)
Then
bn1 n 1 2
lim  lim x2
n  b n 9 ( n  2)
n
 2 n 1 1 2
 x  2 lim  x2
n 9 ( n  2) 9

Let
1 2
L x2
9
L  1  x  2  9 or x  2  3
2
If
Therefore, the power series converges for
 1  x  5.
Does the series converge at endpoints?
At x  1, the series is

(1) n 
(1) n


n 0 9 ( n  1)
n
( 3) 2n
 
n 0 n  1

(1) n


n 0 n  1
is the convergent

alternating series.
At x  5 the series is


(1) n
(1)  n


n 0 9 ( n  1)
n
3  
2n

n 0 n  1

which is the same alternating series.

Therefore, the power series

(1) n

n 0 9 n
(n  1)
( x  2) 2n

converges on  1, 5 and diverges for x  1and

x  5.
1.1.2 Analytic function
A function f (x) defined on an interval
containing the point x  x0 is said to be
analytic at x0 if its Taylor series

f (n)
x 0 
 x  x 0 
n
………… (3)
n 0 n!

exists and converges to f (x) for all x in the

interval of convergence of (3).
Remarks:
x
• All polynomial functions, e , sin x, cos x ,cosh x
and sinh x are analytical everywhere.
• A rational function is analytic everywhere
except at those values of x at which its
denominator is zero.
e.g. x is analytic everywhere except
x  3x  2
2

at x = 1 and x = 2.
• Not every function has a Taylor series
expansion about a point.
• A function must be infinitely differentiable at
x0 to be analytic at x0 (but this is not a
sufficient condition).
 x2
e.g. f ( x)  e is infinitely differentiable at
x  0 but its Taylor series doesn’t converge
to the function.
1.1.3 Ordinary and Singular points
A point x  x0 is called an ordinary point of the
equation
y   p ( x ) y   q ( x ) y  0
……. ( 4)
if both the functions p (x) and q (x) are
analytic at x  x0 .
If the point x  x 0 is not an ordinary point of
the d.e (4), then it is called a singular point of
the d.e (4).
There are two type of singular points :
• Regular singular points.
&
• Irregular singular points.

A singular point x  x0 of the d.e (4) is called a

regular singular point of the d.e (4) if both
( x  x0 ) p ( x ) ( x  x0 ) q ( x )
2

x  x0 and are analytic at

A singular point which is not regular is called an
irregular singular point.

Question 2
Determine whether x  0 is an ordinary or a
regular singular point of the equation
 d 2
y  dy 
2 x  2   7 x( x  1)   3 y  0.
2

 dx   dx 
Solution
2
Dividing by 2x , the given d.e becomes

 d 2 y  7( x  1)  dy  3
 2     2 y 0
 dx  2 x  dx  2 x

comparing (i) with standard equation

y  p( x) y  q( x) y  0
we have

7( x  1) 3
p( x)  and q( x)   2
2x 2x
Since both p (x ) and q (x ) are undefined at
x0 x0
, so both
x are
0 not analytic at .
x0
Therefore, is not an ordinary point and
so is a singular point.

Also,( x  0) p ( x )  7( x  1)
2 and
3
( x  0) q ( x )  
2

2,
Showing that both ( x  0) p ( x ) and
( x  0) q ( x) are analytic at x  0 . Therefore
2

x  0 is a regular singular point.

Question 3
Show that x  0 and x  1 are singular points
of x ( x  1) y   ( x  1) y   2 y  0 where the
2 2 2

first is irregular and the other is regular.

Solution
Dividing by x ( x  1) , the given equation
2 2

becomes
2
d y x  1 dy 2
 2  2 y0
dx 2
x ( x  1) dx x ( x  1) 2

…….. (i).
Comparing (1) with standard equation
y  p ( x) y  q ( x) y  0
x 1 2
We get p ( x)  2 and q( x)  2
x ( x  1) x ( x  1) 2
Since both p (xand
) qare
(x)undefined at
x  0 and x  so1they are not analytic at
and .
x0 x  1
Thus both and are singular
points. x0 x  1
Also x and
1
( x  0) p ( xshowing
)  that
x ( x  1) is
2 p (x )
( x  0) q ( x ) 
2
,
not analytic at ( x and 2
1) so is an irregular
singular point.
x0 x0
But
x 1 2
( x  1) p ( x)  2 and ( x  1) q ( x)  2 ,
2

x x
showing that both ( x  1) p ( x) and
( x  1) q ( x) are analytic at and therefore is a
2

regular singular point.

 1.2 Power Series Solution
Consider a d.e
y   y  0 ………… (5)

The general solution of (5) is

x
y  Ae  Be
x

x
x
e
But we can expand and e as
2 3
x x
e  1 x 
x
 
2 ! 3!
and
2 3
x x x
e  1 x   
2 ! 3!
This means that we can present the solution of
(5) as
 x 2
x 3
  x 2
x 3

y  A 1  x      B 1  x     .
 2! 3!   2! 3! 
The solution of (5) in the form
2 3
x x
y  1 x   
2 ! 3!
and
2 3
x x
y  1 x   
2 ! 3!
are called power series solutions .
A power series to a d.e is the solution which
gives y (x)as an infinite series involving
constants times powers of x.
The power series solution has the form

y ( x )   a n x  x 0 
n

n 0

There are two types of d.e:

(i) D.e whose solution admit power series
solution.

(ii) D.e whose solution don’t admit power series

solution .
 1.2.1 Why power series solution.
Consider the d.e
y  e y  x ; y (0)  4.
x 2
…… (6)

On solving wex get


x
e 2 em e x
ye m e dm  4 ee .
0

This solution is difficult to analyze, or even to

evaluate at specific points.
Problem (6) has a power series solution
1 4
y ( x)  4  4 x  x  x  
3

12

which is easily to make good approximations to

function values.
1.2.2 Finding Power Series Solutions about the
ordinary point x  x 0 .
Let the given d.e be

y   p ( x) y   q ( x) y  0 ……….(1)

If x  x0 is an ordinary point of (1), then (1)

has two trivial solutions of the form


 a x  x 
n
n 0 ……… (2)
n 0
and these power series converge in some
interval of convergence .
To get the coefficients in (2), we take x  x0  R
.
an '(3)
………..
s

y   a n x  x 0 
From (3) we obtain n

and n 0
….……… (4)

 
y    na n x  x0  y    n(n  1)a n x  x0 
n 1 n2

n 1 n2
Substituting y, y  and y  in (1) results the
equation
A0  A1 ( x  x 0 )  A2 ( x  x 0 )    An ( x  x 0 )    0
2 n

………. (5)
where the coefficients A0 , A1 , A2 , etc are
now some functions of the coefficients
a 0 , a1 , a 2 ,  etc.
Since (5) is an identity, all the coefficients
A0 , A1 , A2 ,  of (5) must be zero, i.e.,
A0  0, A1  0, A2  0,  , , An  0.
……….. (6)
Solving equation (6) we obtain the coefficients
of (3) in terms of a0 and a1 .
Putting these coefficients in (3), we obtain the
required series solution of (1) in powers of
x  x0 .

Question 4
Find the power series solution of the equation
x  1y  xy  xy  0 in powers of x .
2
Solution
 
Given: x 2  1 y   xy   xy . 0 ………. (1)
Dividing (1) by x 2 gives
1
x x
y   2 y  2 y . 0 ………. (2)
x  1
Comparing (2) with
x  1 ,
y   p ( x) y   q ( x) y  0
we see that
x and x
p( x)  2 q( x)   2
showing that x and 1 are analyticxat 1 .
p (x) q(x) x0
So x  0 is an ordinary point.


Let y   n
a x n

n 0
……….. (3)

From (3)
 

y    na n x  x0  y    n(n  1)a n x  x0 
n 1 n2

n 0 and n 0

……….. (4)
Substituting y, y and y in (1) we get
  
( x  1) n(n  1)an x
2 n2
 x  nan x n 1
 x  an x  0
n

n2 n 1 n 0
    

 n
n2
( n  1) a n x n

n2
 n ( n  1) a n n2
x 
n 1
 n  n 0
na x n
 a x
n 0
n 1

Shifting indices in the second and last

summations, the above equation becomes
   

 n
n2
( n  1) a n x n

n 0
 ( n  2 )( n  1) a n2 x n

n 1
  n  n1  0
na x n
 a
n 1
x n

When we collect terms from n  2 on under one

summation we’ve
2a 2  (6a3  a1  a 0 ) x

  n(n  1)a n  (n  2)(n  1)a n  2  na n  a n 1  x n  0
n2
Equating coefficients of like powers of x on both
sides of the last equation we get
From :0 so that .
x 2a 2  0 a2  0
From : so that
1 6a3  a1  a 0  0 a 0  a1
x a3 
From :
6

x n n (n  1) a n  (n  for n  1) .a n  2
2) (all
for all .
 n a n  a n 1  0 n 2
……… (5).
an1  n an
2
n 2
an 2 
(n  1) (n  2)
Equation (5) is known as recurrence relation.
When
a1  4a2 a1
n  2; a4   ( a2  0)
12 12 .
9a 3 9  a 0  a1 
n  3; a5    
20 20  6 
3  a0  a1 
  (a 0  a1 )  a3  
40  6 
 a 2 , a3 , a 4, a5 , 
Putting the above values of we’ve
1 1 3
y  a 0  a1 x  (a 0  a1 ) x  a1 x  (a 0  a1 ) x 5  
3 4

6 12 40
Or
 1 3 3 5   1 3 1 4 3 5 
y  a0  1  x  x    a1  x  x  x  x  
 6 40   6 12 40 
which is the required solution near x = 0, where
a 0 and a1 are arbitrary constants.
Question 5
Find the power series solution in powers of
of the initial value problem
x 1
xy   y   2 y  0,
Solution y (1)  1, y (1)  2.
Given: …….. (1)
Comparing (1) with here
and , which are analytic at
xy   y   2 y  0.
y   p ( x) y   q ( y )  0,
1 2
p( x)  q( x) 
x x
x  1.
Hence x = 1 is an ordinary point of (1).

Let y   a n ( x  1) n . …….. (2)
n 0

 y   na n ( x  1)
 n 1

n 1
 …..…. (3)
and y    n (n  1) a n ( x  1) n  2
n2
Putting (2) and (3) in (1) results
  
x  n (n  1) a n ( x  1) n2
  n a n ( x  1) n 1
 2 a n ( x  1)  0
n

n2 n 1 n 0
Or
  
( x  1)  1 n (n  1) a n ( x  1) n2   n a n ( x  1) n1  2 a n ( x  1) n  0
n2 n 1 n 0

 

 n (n  1) a
n2
n ( x  1) n 1
  n (n  1) a n ( x  1)
n2
n2

 
  n a n ( x  1) n 1
 2 a n ( x  1)  0 n

n 0 n 0

 

Or  (
n 1
n  1) n a n 1 ( x  1) n
  ( n  2) ( n  1
n 0
) a n2 ( x  1) n

 
  (n  1) a n 1 ( x  1) n  2 a n ( x  1) n  0
n 0 n 0
Combining the terms for n , 1we’ve
2 a 2  a1  2 a 0

  (n  1) n a n 1  (n  2) (n  1) a n  2  (n  1) a n 1  2 a 2 ( x  1) n .
n 1

Equating the various powers of x – 1, we get

so that a1  2 a 0....(4)
2 a 2  a1  2 a 0  0 a2   .
2
(n  1) n a n 1  (n  2) (n  1) a n  2  (n  1) a n 1  2 a n  0 ,
for all .
n 1
Or

(n  1) a n 1  (n  2) (n  1) a n  2  2 a n  0
2
for all .
n 1
Or
(n  1) a n 1  2 a n
2
an2  , for all n  1.
(n  1) (n  2)
……. (5)
Given y  1 and y   2 when x  1.
Therefore, substituting x  1 in (2) and (3), gives
a 0  1 and a1  2 . ……. (6)
Using (6), (4) yields
22 .……(7)
a2    2
2
From (5) when
2 2 a 2  2 a1
n  1; a3  
23
4 (2)  2(2)
  a1  2 and a 2  2 
6
2

3
3 2 a3  2 a 2
n  2; a4  
3 4
2
9    2(2)
 3  2
  a 2  2 and a3  
12  3
1

6
 4 2 a 4  2 a3
n  3; a5  
45
 1 2
16     2 
 6 3  2 1
  a3  and a 4   
20  3 6
1

15

and so on.
Putting the obtained values for a ‘s in (2) we’ve
2 1 1
y  1  2 ( x  1)  2 ( x  1)  ( x  1)  ( x  1)  ( x  1) 5  
2 3 4

3 6 15
 Frobenius Method
Consider the d.e
. .…… (1)
y   p( x) y   q( x) y  0
If is a regular singular point of equation (1)
then we assume a solution of the form
x0
, where …….. (2)
Series (2) is called a Frobenius series , and the



strategy of attempting a solutionnmof this form is
ycalled  method
( x) the  x0 )
a n (ofxFrobenius.
n 0
a0  0
A Frobenius series need not to be a power
Series, since m may be negative or may be non
integer.
Question 6
Solve in series x y   5 xy   ( x  4) y  0
2

about x  0.
Solution
2
Dividing the given d.e by x both sides we’ve
5 x4
y  y  2 y  0
x x
Now
5 x4
p ( x)  ; q( x)  2
x x
So x  0 is a singular point of the given d.e.

( x  x 0 ) p ( x)  5 and ( x  x 0 ) q ( x)  x  4
2

 x0  0 is a regular singular point.

Let 
y ( x)   a n x nm
…… (1)
n 0
 
 y ( x )   ( n  m) a n x
 n  m 1

n 0


y ( x)   (n  m) (n  m  1) a n x
 n m2

n 0

Substituting the derivatives into the given d.e

we get
 

 (n  m) (n  m  1) an x nm
  5 ( n  m) a
n 0
n x nm

n 0
 

a
n 0
n x n  m 1
  n
4 a x nm

n 0
 0.
…….. (2)
On shifting indices on the 3rd summation, (2)
Becomes

 
 (
n 0
n  m ) ( n  m  1) a
. n x nm
  5 ( n  m ) a n x nm

n 0

 
When we combine terms we write
a
n 1
nm
n 1 x   4a
nm

n 0
n x 0

.
m (m  1)a 0 x m  5ma 0 x m  4a 0 x m 


 ( n  m ) ( n  m  1) a n  5 ( n  m ) a n  a n 1  4 a n x nm
0
n 0
m(m  1)  5m  4a0 x m 


 ( n  m ) ( n  m  1) a n  5 ( n  m ) a n  a n 1  4 a n x nm
0
n 0
…….. (3)
m
 a 0  0 we set the coefficient of x to
m(m  1)  5m  4  0 . …….. (4)
or m  4m  42  0
2

or m  2   0
or m  2
(4) is called an indicial equation.
Inserting m  2 in the coefficient of x nm we
get
n  2(n  3) a n  5 (n  2) a n  a n1  4a n 0

or
1
an   an 1
(n  2) (n  3)  5 (n  2)  4

for n  1, 2,
On simplifying we’ve
 1
an   2 an1 for n  1, 2, 3,
n
Some of the coefficients are
a1   a 0 .
1 1 1
a 2   a1  a 0  2 a 0 .
4 4 2
1 1 1
a3   a 2   a0   a0 .
9 49 (2  3) 2

1 1 1
a 4   a3  a0  a0 .
16 4  9  16 ( 2  3  4) 2
and so on.
In general
1
a n  (1) 2
a 0
n
for n  1, 2, 3, 
( n !)
From (1)

y ( x)  a 0 x   a n x n  m
m

n 1


m  2 
1  
 a0 x 2
  (1) n
a x n2
1
n 1 ( n ! ) 2 0 & a n  (1) n a0 
 ( n !) 2

 
1
 a0  (1) n
ax n2

n0 (n!) 2 0
only one solution 

Theorem ( A second solution in the Frobenius

method)
Suppose 0 is a regular singular point of
y  p ( x) y  q ( x) y  0 .
Let m1 andm2 be roots of the indicial equation. If
these are real, suppose m1  m2 . Then
1. If m1  m2 is not integer, there are two linearly
independent Frobenius solutions:
 

y1 ( x)   a n x n  m1
and y 2 ( x)   a n x n  m2
n 0
n 0


with a 0  0 and a  0. These solutions are
0

valid in some interval (0, r ) or ( r , 0) .

2. If m1  m2  0 , there is a Frobenius solution

y1 ( x)   an x nm 1
with a0  0
n0
as well as a second solution 
y 2 ( x)  y1 ( x) ln x   a x
n
n  m1

n 1
Further, y1and yform a fundamental set of
2
solutions on some interval (0,. r )
3. If m  m is a positive integer, then there is a
1 2
Frobenius series solution:

y1 ( x)   a n x
n  m1
In this case there is a second
n 0
solution of the
form

y 2 ( x)  ky1 ( x) ln x   a x 
n
n  m2

n 0
where y1 and y2 form a fundamental set of
solutions on some interval (0, r ) .

Referring to question 6,since from the indicial

equation m  2   0 we attempt a second
2

solution of the form:


y 2 ( x)  y1 ( x) ln x   a x 
n
n2

n 1 …… (*)
1 
 y2 ( x)  y1 ( x) ln x  y1 ( x)   (n  2) an x
 n 3

x n 1
 
1 1
y2 ( x)  y1 ( x) ln x  2 y1 ( x)  y ( x) 2   (n  2) (n  3) an x
    n4

x x n 1

upon substitution y2 , y 2 and y2 into the d.e

yields
 
 n4 
x  y1 ln x  2 x y1  x y1   (n  2) (n  3) a n x 
2 1 2

 n 1 
 
 n 3 
 5 x  y1 ln x  x y1   (n  2) a n x 
1

 n 1 

 
 n2 
 ( x  4)  y1 ln x   a n x   0.
 n 1 
 
x y ln x  2 xy1  y1   (n  2) (n  3) a x
2  n2
n
n 1

 5 xy ln x  5 y1   5 (n  2) an x
  n2

n 1

 ( x  4) y1 ln x   a x

 n 1
n   4an x n  2  0.
n 1 n 1

Rearranging the above equation we get,

x 2

y1  5 y1 ln x  ( x  4) y1 ln x  2 xy   4 y1
 
  (n  2) (n  3) a x 
n
n2
  5 ( n  2) a x 
n
n2

n 1 n 1
  
 a x 
n
n 1
  4a x 
n
n2
 0.
n 1 n 1

The first term is zero because y1 is a solution of

the d.e.  
1 1
Substituting y  a  (1)
(n!)
x
and
0
y    ( n n
2) ( 1) 2
(n!)
n2
x 1
n 1
n
2
n 3

n 1

(we’ve a0  1take ) in the above equation results

 
1 1

n 1
2 ( n  2) ( 1) n

( n !) 2
x n2
 
n 1
4 ( 1) n

( n !) 2
x n2


 
  (n  2) (n  3) a x 
n
n2
  5 (n  2) a x
n
n2

n 1 n 1
  
 a x 
n
n 1
  4a x
n
n2
0
n 1 n 1
 

Let us shift index to write  a x   a x

 n 1  n2
n n 1
n 1 n2
and separate terms for n  1 from other
Summations we’ve
 
1 1
2 x   2 (n  2) (1)
1 n
2
x n2
 4 x 1
  4 ( 1) n
2
x n2

n2 ( n ! ) n2 ( n ! )


 2a x   (n  2) (n  3) a x

n
1 
n
n2
 5a x 
n
1

n2
   
  5 (n  2) a n x n  2   a n1 x n  2  4a n x 1   4a n x n  2  0
n2 n2 n2

or  2 ( n  2) 4 (  1) n

 2
(1) 
n
2 
 ( n !) ( n !) 

 2 x 1  a1 x 1    (n  2) (n  3) a n  x n2  0
n2  
 5 (n  2) a n  a n1  4a n 
 
 
we set the coefficient of each power of x equal
to zero.
From the coefficient of x 1 :
 
 2  a  0 or
1 a 2
1
n2
From the coefficient of x we get:
(1) n
2 (n  2)  4 2  (n  2) (n  3)  5a n  4a n  a n1  0
( n !)
or
2n(1) n

2
 (n  2) (n  2)  4a n  a n 1  0
 

( n !)
or 2n (1) n 2  
2
 n a n  a n 1  0
( n !)
Or
 1  2 (1) n
a   2 a n 1 
n 2
 0 for n  2, 3, 4, 
n ( n !)

When
n  2; 1 
a   a1 
2
4
2
2(2 !) 2
1
  (2) 
4
1
4
 a1  2 
3

4
 1  2 1 3 2    3
n  3; a   a2        a 
4 
3 2 2
9 3(3!) 9  4  108
11

108
1  2 1  11  2  25
n  4; a   a3 
4 2
   
16 4( 4 ! ) 16  108  1152 3456
  1  2 1   25  2 137
n  5; a   a4 
5 2
   
25 5(5!) 25  3456  72000 432 000

and so on.
From (*)
 1    
y 2 ( x)  y1 ( x) ln x  a x  a  a x  a x  a x  
1 2 3 4
2
5
3


Substituting values for the coefficients a we’ve n

2 3 11 25 2 137
y 2 ( x)  y1 ( x) ln x    x x  x3  
x 4 108 3456 432 000

(1) n n  2 2 3 11 25 2 137
 ln x  2
x    x  x  x 3

n  0 ( n !) x 4 108 3456 432 000
The general solution y ( x)  Ay1 ( x)  By2 ( x) , is
given by

(1) n n  2  2 3 11 25 2 137 
y ( x)  A  B ln x  2
x  B   x x  x  
3

n  0 ( n !)  x 4 108 3 456 432 000 

where A and B are arbitrary constants.

Question 7
Solve x y  x y  2 y  0 near x = 0.
2 2

Solution
2
Dividing by x both sides of
x y  x y  2 y  0
2 2
……. (1)
gives 2
y  y  2 y  0
x
Let 2
p ( x)  1 and q ( x)   2
x
q (x ) is not analytic at x = 0, hence x = 0 is a
singular point of (1).

Now,
2
( x  0 ) q ( x )  x   2  2
2 2

x
So ( x  0) q( x) is analytic at 0, and therefore 0
2

is a regular singular point of (1).

Let

y ( x)   a n x n  m
n 0 …….. (2)

 y ( x )   ( n  m) a n x
 n  m 1

n 0


y ( x)   (n  m) (n  m  1) a n x nm2

n 0
Substituting y, y  andin y(1) results
  

 (n  m) (n  m  1)a n x
n 0
nm
  ( n  m) a n x
n 0
n  m 1
 2 a n x n  m  0
n 0
When we shift index in the 2nd term so that
its power of x becomes n +m, and separating
terms of n = 0 from 1st and 3rd terms we’ve


m (m  1) a0 x   (n  m) (n  m  1)an x
m n m

n 1 .
 
  (n  m  1) an1 x n m
 2 a0 x  2  a n x
m nm
0
n 1 n 1
Rearranging
m (m  1)  2a0 x m

  (n  m) (n  m  1) an  (n  m  1) an1  2an  x n m
0
n 1
…….. (3)
The indicial equation is
m (m  1)  2  0 or m  m  2  0
2

The roots are m1  2 and m2  1 .

Now,
m1  m2  3
Thus the two independent solutions are
 
y1 ( x)   a n x n  m1
and y 2 ( x)  ky1 ( x) ln x   a x

n
n  m2

n 0 n 0

To get 1y ( x ) we set the coefficient of x n m

equal
to 0. i.e
(n  2) (n  1) an  (n  1) an1  2 an  0

Or n 1
an   an1 for n  1, 2,
n (n  3)
Let’s find different values of an ' s .
2 1
n  1; a1     a0 .
4 2
3 3  1 3
n2 a2  a1    a 0   a 0 .
2(5) 10  2  20
4 2 3  1
n3 a3  a2    a0    a0 .
3(6) 9  20  30
5 5  1  1
n  4; a4   a3     a 0   a0 .
4 (7 ) 28  30  168
6 3  1  1
n  5; a5   a4   a0   a0 .
5(8) 20  168  1120
7 7  1  1
n  6; a6   a5    a0    a0 .
6 (9) 54  1120  8 640
From (2)
m 1 m 2 m 3 m 4
y ( x)  a0 x  a1 x  a2 x  a3 x  a4 x  
m

Putting m  2 and the respective values for

coefficients of a n we obtain
 1  3  3  4  1  5
y1 ( x)  a 0 x    a 0  x    a 0  x    a 0  x
2

 2   20   30 
 1  6  1  7  1  8
 a0  x   a0  x    a0  x  
 168   1120   8640 
 1 3 2 1 3 1 4
 1 x  x  x  x 
2 20 30 168
y1 ( x)  a 0 x 2  
 1 x 5  1 x 6   
 1120 8640 
Substituting m  1 in the recurrence relation (3)
We get
  
( n  1) ( n  2) a  (n  2) a
n n 1  2a  0,
n for n  1, 2, 

   1 
n  1;  a  2a  0
0 1 or a   a0 .
1
2

n  2; a  0.
2
   
n  3; 2 a  a  2a  0
3 2 3 or a  0.
3

a n  0 for n2
.
Putting min 1 and using
y 2 ( xthe
) obtained
Coefficients for we get 
an
y 2 ( x)  a 0 x 1  a1 x 0  a 2 x  a3 x 2  
1 1   1 1
The general solution is 
 a0   a0  a0   
x 2  x 2

Where are arbitrary constants.

 1 3 2 1 3 1 4
 1 x  x  x  x 
2 20 30 168  1 1 
y1 ( x)  a 0 x    a0    .
2

 1 x 5  1 x 6     x 2
 1120 8640 

a0 and a 0
Question 8
Solve xy   y  0 near x = 0.
Solution
Dividing xy   y  0 by x through out gives
1
y   y  0
x
1
q ( x)  is not analytic at x = 0, so x = 0 is a
x
singular point of the given d.e.
We consider a Frobenius
solution
y ( x)   a n x n  m
n 0 ……. (1)
 
 y ( x )   ( n  m) a n x
 n  m 1

n 0

y ( x)   (n  m) ( n  m  1) a n x

 nm 2

n0

Substituting y and y  into the given d.e yields

 

 (n  m) (n  m  1) a
n 0
n x n  m 1
  an x
n 0
nm
0

Shifting indices in the second summation we’ve

 

 (n  m) (n  m  1) a n x   a n1 x  0 .
n 0
n  m 1 n  m 1

n 1
Putting under one summation the above
equation, we’ve

m (m  1) a 0 x m 1   (n  m) (n  m  1) a n  a n 1  x n  m  0
n 1

m (m  1)  0.
…….. (2)
The indicial equation is
m1  m2  1  0  1

Here  m  0 or m  1 (an integer)

The recurrence relation is
(n  m) (n  m  1) a n  a n 1  0 for n  1, 2, 
……… (3)
1
an  a n 1
(n  m) (n  m  1)

When m =1

1
an  an 1 for n  1, 2, 3,
n (n  1)
Some of the coefficient are :

1
a1  a 0
2
1 1 1
a2  a1   a0
2 (3) 2 (3) 2
1 1 1 1 1
a3  a2    a0  a0
3 ( 4) 3 (4) 2 (3) 2 3 (2)  4 (3) (2)

1
an  a0 for n  1, 2, 3,
n!(n  1)!
Putting m = 1 and 1 results
an  a0
n !(n  1) !

1
y1 ( x)  a0  x n 1

or n  0 n !( n  1) !

 1 2 1 3 1 4 
y ( x )  a
Putting m = 0 into
1 0  x 
(3),
x
the
 x 
recurrence
x
relation
 
 2 12 144 
becomes

……. (4)
n (n  1) a n  a n 1  0 for n  1, 2, 
If we put n =1, into (4) we get a 0  0 , contrary
a0  0
to the assumption . Therefore the second
solution has the form 
y 2 ( x)  ky1 ( x) ln x   a n x n  m2
n 0

………. (5)

Differentiating (5) twice1successively
 n 1 we get,
y  ( x)  ky  ln x  ky  n a x
2 1 1
x

n 0
n

1 1
y 2 ( x)  ky1 ln x  2ky1  2 y1   n (n  1)a n x n  2
x x n 0
Substituting y 2 , y 2 and y 2 into the given d.e we
get
 1   
k ( x ln x) y1  2 y1  y1    n (n  1) a n x  ky1 ln x   a n x n  0
 n 1

 x  n2 n 0
……… (6)
Rearranging (6) results
 
k
k ln xxy1  y1   2ky   y1   n (n  1) a n x n 1   a n x n  0
x n 0 n 0
Or  
k
2ky   y1   n (n  1) a n x n 1   a n x n  0
x n 0 n 0

 y1 is a solution therefore, x y   y1  0
……… (7)
inserting

1
y1 ( x )  a0  x n 1
n  0 n ! (n  1) !
and

1
y1  a0  ( n  1)
 x n

n 0 n!(n  1)!

in (7) gives
 
1 1
2ka 0  (n  1) x  ka 0 
n
xn
n 0 n !(n  1) ! n  0 n !( n  1) !
 
  n (n  1) a n x n 1   a n x n  0.
n 0 n 0
 n 1
Shifting indices in the 3rd summation from x
n
to x and writing terms from n = 1 in one
summation notation we’ve
  2ka 0 ka 0  n
2ka 0  a 0 k  a   

  
 a n 1 n (n  1)  a n  x  0.
n 1  n !
0 2
n !(n  1) ! 

Then 
 a
a 0 k  a  0 or k 
0
0

a0
and
2ka0 ka0  
  n ( n  1) a  a n  0.
n ! n !(n  1) !
2 n 1

for n  1, 2, 3
The recurrence relation is
 1   (2n  1)a  
a n 1  a n 
0

n (n  1)  n !(n  1) ! 
The coefficients are
1  3a 
 1 3a 
n  1; a 2  a1  0   a1  0
2 2  2 4
 1   5 a 
 1  5 
n  2; a3  a 2 
0
  a 2  a0 .
2 (3)  2 (3!)  6 72
1 1  3   5 
  a1  a 0   a 0
6 2 4  72
1  7 
 a1  a 0
12 36
  1   7 a 
 1  7 
n  3; a4  a3 
0
  a3  a0 .
3 ( 4)  3!(4 !)  12 1728
1  1  7  7 
  a1  a 0   a0
12 12 36  1728
1  7  7 
 a1  a0  a0
144 432 1728
1  35 
 a1  a0
144 1728

and so on.
From (5)
   
y 2 ( x)  ky1 ( x) ln x  a  a x  a x  a x  
0 1 2
2
3
3
Substituting the values for coefficients and y1 ( x)
in (5) we’ve
a 0  1 2 1 3 1 4 
y 2 ( x)  x  x  x  x   ln x  a 0  a1 x
a0  2 12 144 
1  3  2  1  7  3
  a1  a 0  x   a1  a 0  x
2 4   12 36 
 1  35   4
 a1  a0  x  
 144 1728 

The two solutions y1 ( x) and y 2 ( x) give the

general solution of the given differential
equation.
 Bessel’s Equation
The d.e
x y  xy  x  v y  0

2
  2 2
……… (1)
is called Bessel’s equation of order v, for v  0 .
Solutions of Bessel’s equation are called Bessel
functions.
Zero is a regular point of Bessel’s equation, so
we attempt a solution:

y ( x)   a n x nr

n 0
Substituting y, y  and y  into (1) results
m (m  1)  m  v a x  m(m  1)  m  1  v a x
2
0
m 2
1
m 1

    (n  m) (n  m  1)  n  m  v a  a  x

nm
2
n n2 0
n2

……. (2)
Setting the coefficient of each power of x to
zero (a0  0) we obtain the indicial equation
m v  0
2 2

 m  v
Let m  v in the coefficient of x m 1 in equation
(2) we get
(2v  1) a1  0.
a1  0  2v  1  0 as v  0.
nm
From the coefficient of x in the equation (2)

we(nget
 m) (n  m  1)  (n  m)  v a n  a n  2  0
2

for n  2, 3, 4, 
Setting m  v in this equation and solving for
an
we obtain a   1
n a n 2
n ( n  2v )
for n  2, 3, 4,  a1  0
a3  a5 ,since
  aodd equation
 0. yields

For the even-indexed coefficients, we write

1 1
a2 n   a2 n  2   2 a2 n  2
2 n ( 2 n  2v ) 2 n (n  v)
 1
a2 n   2 a2 n  4
2 n (n  v) 2 (n  1) 2 (n  1)  2v 

(1) n
  2 a0 .
2 n (n  1)  2 (1) (n  v) (n  1  v)  (1  v)

(1) n
 2n a0 .
2 n !(1  v) (2  v)  (n  v)
One Frobenius solution of Bessel’s equation of
order v is therefore

( 1)
n
y1 ( x)  a0  2n x 2 n v

n  0 2 n !(1  v) (2  v)  ( n  v)

…….. (3)
These functions are called Bessel functions of
the first kind of order v.
Depending on v, we may or may not obtain two
linearly independent solutions upon using v and
– v in the series solution (3).
The Bessel’s equation of zero order (v =0) is
written
x y   xy   x y  0.
2 2

We have seen the indicial equation of Bessel’s

equation of order v has root r   v .
Thus when v = 0, the indicial equation has only
one root, r = 0.
We know that one Frobenius solution of
Bessel’s equation of order v is given by
 
(1) n
y1 ( x)  a 0  2 n x 2nv

n 0 2 n !(1  v) (2  v)  (n  v)

with c  1 and v  0the above expression

0
becomes

(1) n
y1 ( x)   2 n 2
x 2n
…………. (1)
k  0 2 ( n ! )
We attempt a second, linearly independent
solution of the form
 
y 2 ( x)  y1 ( x) ln x   a x 
n
n

n 1

1 
 y 2 ( x)  y1 ln x  y1   n a n x
   n 1

x n 1

2 y1
y 2  y1 ln x  y1  2 y1   n ( n  1) a n x n  2
x x n2

Substituting y 2 , y 2 and y 2 expressions into

the
 


d.e results1
x 2
ln x y   2 xy   y1   n ( n  1) a  n2
nx  x ln x y1  y1
n2

 
 
  n a n  x 2 ln x y1   a n x n  2  0
n 1 n 1
Rearranging
x 

2
y   xy1  x 2 y1 ln x  2 xy1   n (n  1) a n x n 
n2
 
 na x  a x

n
n 
n
n2
0
n 1 n 1
Or
  
2 xy1   n (n  1) a n x n   n a n x n 1   a n x n  2  0
n2 n 1 n 1

Or
   
2 y1   n a x 2 
n
n 1
 na x 
n
n 1
 na x 
n
n 1
 a x 
n
n 1
0
n2 n2 n 1 n 1
  
2 y1   n 2 a n x n 1  a1   a n x n 1  0
n2 n 1
……… (2)
From (1)

(1) n
y1   2 n
 2
x 2 n 1

n  0 2 ( n !)


n 1
  1 (1)  2 n
n
x 2 n 1

n 1 2 2 n !n (n  1) !
 
(1) n
  2n2 x 2 n 1

n 1 2 n !(n  1) !
Substituting y1 into (2) results

(1) n  
2 2 n 1 x 2 n 1   n 2 a n x n 1  a1   a n x n 1  0
n 1 2 n!(n  1) ! n2 n 1

Let’s shift indices in the last series to write

equation as

(1) n  


n 1 2
2n2
n!(n  1) !
x 2 n 1
 
n2
n 2  n 1
a n x  a1

 
n 3
a 
n2 x n 1
0
or
(1) n
 
 


n 1 2
2n2
n !(n  1) !
x 2 n 1
 a1

 4 a 
2 x  
n 3
n 2 
a n  a 
n2 x n 1
0
……… (3)
Comparing coefficients we’ve
(constant term): a1  0.


The only even powers of x appearing in

equation (3) are from the series

 n x

  n 1
2
a a
n n2
n 3
 
 n a a
2
n n2 0 for n  3, 5, 7, 
  1 
a   2 a n2
n
n
When
 1  
n  3; a   2 a1  0 ( a1  0)
3
3
 1  1
n  5; a5   2 a3    0  0
5 25
All odd-indexed coefficient are zero (all are

a
multiples of 1 which is zero).

a 2 n 1 0 for n  0,1, 2, 
Let’s determine the even-indexed coefficients.
Replacing n by 2j in the second summation of
equation (3) and n with j in the first summation
we’ve
(1) j
 
 


j 1 2
2 j 2
j!( j  1) !
x 2 j 1
 4 a 
2  
j 2
4 j 2 
a 2j  a 
2 j 2 x 2 j 1
0

2 j 1
Combining terms in x we’ve

 (1) j
 2 j 1
(4a 2  1) x    2 j  2
 2  
 4 j a2 j  a2 j 2  x 0
j 2  2 j !( j  1) ! 
Equating the coefficient of each power of x to
zero we get
  1
4a 2  1  0 or a 2 
4
and
(1) j 2  
2 j 2
 4 j a 2 j  a 2 j  2  0.
2 j !( j  1) !

Or
(1) j 1 1 
a 2 j  2j  2 a 2 j 2 for j  2, 3, 4, 
2  j ! j 4 j
2
For 1  1 
j  2 ; a  2 2 1   4
2 4  2
 1  1 1
j  3 ; a 6  2 2 2 1   
2 4 6  2 3
and, in general
 (1) j 1  1 1
a  1     
2  4    2 j 
2j 2
 2 j
2 2

(1) j 1
  j
2 2j
 j ! 2

where
1 1
 j  1     for j  1, 2, 
2 j
Therefore the second solution of Bessel’s
equation of order zero is

(1) n 1
y 2 ( x)  y1 ( x) ln x   2 n  ( n) x 2n
for x  0
n 1 2 ( n !) 2
.
where
1 1
 n   1     for n  1, 2, 
2 n
 Writing Bessel’s Solution in Gamma Function
For x  0, the gamma function  is defined by

( x)   t x 1  t
e dt .
0

The integral converges for all x  0.

It can be shown that
( x  1)  x ( x).
Now
(n  v  1)  (n  v) (n  v)
 (n  v) (n  v  1) (n  v  1)
   (n  v) (n  v  1)  n  v  (n  1) n  v  (n  1) 
 (1  v) (2  v)  (n  1  v) (n  v) (v  1).
Therefore,
(n  v  1)
(1  v) (2  v)  (n  1  v) (n  v)  .
(v  1)
Now replacing (1  v) (2  v)  (n  1  v) (n  v)
(n  v  1)
by in Bessel’s equation solution
(v  1)
 
(1) n
y ( x)  a 0  2 n x 2nv

n  0 2 n !(1  v ) ( 2  v )  ( n  v )

gives 
(1) (v  1)
n
y ( x)  a 0  2 n x 2nv

n 1 2 n !(n  v  1)

It is customary to choose
1
a0  v
2 (v  1)
Then we write the solution y (x) as

(1)n
J v ( x)   2 n  v x 2nv

n 0 2 n !(n  v  1)
J v is called a Bessel function of the first kind of
order v.
The series defining J v (x) converges for all x.
 General Solution of Bessel’s Equation
The indicial equation of Bessel’s equation is
m  v  0 with roots  m.
2 2

The difference of the roots = 2v.

1.If 2v is not an integer, then Jand
v
J are
v

linearly independent, and the general solution

of Bessel’s equation of order v is
y ( x)  a J v  b J v
2. If 2v is an odd positive integer, say 2v  2n  1
1
then v  n  for some positive integer n.
2
The general solution of Bessel’s equation of
order v is
y ( x)  a J n  1 ( x)  b J  n  1 ( x)
2 2
where a and b are arbitrary constants.
1
3. If 2v is an integer but is not of the form n  2
J
for positive integer n then v and J v

are linearly dependent since

J v ( x)  (1) v J v ( x).
To construct the second solution of Bessel’s
equation, linearly independent from J v (x) ,
leads
us to Bessel functions of the second kind.
Question 9
Prove that
J v ( x)  (1) J v ( x)
v

where v is a positive integer.

Solution
We know that
2nv

( 1)n
x
J v ( x)    
n 0 n !( n  v  1)  2 
Replacing v by – v we get
2 n v

(1) n
 x
J v ( x)    
n 0 n !(n  v  1)  2 
2 n v 2 n v
v 1
(1) n
 x 
(1) n
 x
     
n 0 n !(n  v  1)  2  n  v n ! ( n  v  1)  2 

2 n v

(1) n
 x
 0    ( ve integer)   
n  v n ! ( n  v  1)  2 

On putting nvk
vk v2k

(1)  x
J v ( x)    
k  0 (v  k ) ! ( k  1)  2 
 v2k

(1)  x 
k
 (1) n
 
k  0 (v  k ) ! k!  2 

 (1) J v ( x)
v
 Equations Reducible to Bessel’s Functions
Let us reduce the equation

 
2
d y dy
x 2
2
 x  k x v y  0
2 2 2

dx dx
to Bessel’s equation.
Put
dt dy dy dt dy
t  k x,  k,   k
dx dx dt dx dt
2 2
d y d  dy  d  k dy  dt  k 2 d y
2
 k     2
dx dx  dt  dt  dt  dx dt
Thus (1) becomes
 t2   2 d 2 y   1   dy 
 2   k 
2 
 
  k   t  v y  0
2 2

k  dt   k   dt 
Or
 
2
d y dy
t2
2
t  t v y  0
2 2

dx dx
If 2v is non-integer, then the solution of this
equation is y  c1 J v (t )  c 2 J v (t )
Therefore, the solution of the given equation is
y  c1 J v (k x)  c 2 J v (k x)
Question 10
Use the change of variable y  x u solve
2

9 x y   27 xy   (9 x  35) y  0
2 2

Solution
Given: 9 x 2
y   27 x y   ( 9 x 2
 35) y  0
Let yx u 2

 y  2 xu  x u 
 2
y  2 u  4 x u  x u 
  2
,
…….. (1)
Substituting (1) into the given d.e gives
18 x u  36 x u  9 x u  54 x u  27 x u 
2 3
 4
 2 3

 9 x u  35 x u  0
4 2

Adding like terms results

9 x 4 u   9 x 3 u   ( 9 x 4  x 2 ) u  0
Dividing by 9 x 2
we get
 2 1
x u   xu    x   u  0
2

 9
1
This is Bessel’s equation of order v 
3
Since 2v is not an integer, the general solution
for u is
u ( x)  a J 1 ( x)  b J  1 ( x)
3 3

Therefore, the original d.e has general solution

y ( x)  a x J 13 ( x)  b x J  13 ( x) for x  0 .
2 2
Question 11
Prove that
 2 
(a ) J 12 ( x)    sin x
 x 

 2 
(b) J  12 ( x)    cos x
 x 
Solution
We know that
2nv

(1) n
 x
J v ( x)    
n 0 n !(n  v  1)  2  ……..(1)

(n  v  1)  (n  v) (n  v  1)  (1  v) (1  v)
……… (2)
Using (2) in (1) results

(1) n
J v ( x)   2 n  v x 2nv
n 0 2 n !(1  v) (2  v)  (n  v) (v  1)

xv 
(1) n
 v 
2 (1  v) n 0 2 2 n n !(1  v) (2  v)  (n  v)
x 2n

xv  x2 x4 
 v 1  2  4  
2 (1  v)  2 (1  v) 2  2 !(1  v) (2  v) 

………… (3)
 1
(a) Substituting v  in (3), we get
2
 
1
x2  x2 x4 
J 12 ( x)  1    
 1  2  1  1  1 
2 1    2 1   2  2 !1    2  
1 4
2

 2   2  2  2 

x  x2 x4 
J 12 ( x)  1   
 3   3! 5! 
2  
2
x 1 x3 x5 
  x    
1 1 x  3! 5! 
2   
2 2
 1  1 
 sin x  since     
1  2 
2x  
2
 2 
    sin x
 x  proved.

1
(b) Again substituting v in (3), we’ve
2
 
 12
x  x2 x4 
J  12 ( x)  1    
 12  1   1   1  1  
2    1  2 2    1 2 4  2 !   1    2  
 2   2   2  2  
 2  x2 x4   2   1 
 1        cos x since     
 1   2! 4!
x     x   2 
2 Proved.
Recurrence Formulae
(a ) x J v  v J v  x J v 1
(b) x J v  v J v  x J v 1
(c) 2 J v  J v 1  J v 1

(d )
d v
dx
 v

x J v  x J v 1
Proof
(a)We want to prove that x J v  v J v  x J v 1
We know that
2nv

(1) n
x
J v ( x)    
n 0 n !(n  v  1)  2 

Differentiating with respect to x, we get

2 n  v 1

(1) (v  2n)  x 
n
1
J n     
n 1 n !(n  v  1)  2  2
 2nv v  2 n 1

(1) n
 x 
(1)  2 n  x 
n
xJ n  v     x  
n 1 n !(n  v  1)  2  n 1 2  n ! ( n  v  1)  2 

v  2 s 1

(1) n
 x
 v Jv  x   
n 1 ( n  1) ! ( n  v  1)  2 

Putting n 1  s the last equation becomes

s 1 ( v 1)  2 s

(1)  x
x J n  n J n  x   
s  0 s ! ( n  1)  s  1   2 

 n J n  x J n 1
proved.
(b) We want to prove that x J v  v J v  x J v 1
We know that
n 2 n v

( 1) x
J v ( x)    
n 0 n !( n  v  1)  2 

Differentiating with respect to x, we get

2 n v 1

( 1) (v  2n)  x 
n
1
J n     
n 1 n !( n  v  1)  2  2
2nv

(1) (v  2n)  x 
n
x J n    
n 0 n !(n  v  1)  2 
 (1) ( 2v  2n)  n  x 
 n 2nv

  
n 0 n !( n  v  1)  2 
2nv 2nv

(1) (2v  2n)  x 
n 
(1) n
 x
    n  
n  0 n ! ( n  v  1)  2 
n 0 n !(n  v  1)  2 
2nv

(1)  2  x 
n
    n Jn
n 0 n ! ( n  v )  2 
( n 1 2 n )

(1) n
 x
 x    n Jn
n 0 n !(n  1)  n  1  2 

 x J n 1  n J n
proved.
(C) We want to prove that 2 J n   J n 1  J n 1
We know that
x J v  v J v  x J v 1 (from (a))
x J v  v J v  x J v 1 (from (b))
Adding the two expression, we get
2 x J n   x J n 1  x J n 1
Or 2 J n  J n 1  J n 1 proved.
(d) We want to prove that 2v J v  x ( J v 1  x J v 1 ).
We know that
x J v  v J v 1  x J v 1 (from (a))
 x J v  v J v  x J v 1(from (b))
Subtracting the 2nd expression from the 1st,
we’ve

0  2 n J n  x J n 1  x J n 1
Or proved.
(e) We want
2 n to  xthat
J show
n (J n 1  J n 1 )

i.e.,
d v
dx
  v
x J v   x J v 1
Or
We know that
x J v  v J v  x J v 1
Multiplying by , we’ve
n 1
x
x v
J v  v x  v 1
Jv  x v
J v 1
i.e.,
or
x v
J v  v x  v 1
J v  x v
J v 1

d v
dx
 v
x J v   x J v 1 
Or
A generating Function For J n (x)
We want to demonstrate that is the
Coefficient of in the expansion of . J n (x )
i.e. x 1
n t 
t e 2 t 

Proof x 1 

e
 z 
2 z  J
n  
n ( x) z n

x 1 xz x
 z  
e 2 z
 e e 2 2z
We know that
2 3
z z
e  1 z 
z
 
2 ! 3! ……. (1)
xz 2 3
 xz 1 xz 1 xz
e 2
1        
 2  2 ! 2  3!  2 

2 3
1  x  …….. (2)
x
  x  1 x 
e 2z
 1           
 2 z  2!  2 z  3!  2 z 
On multiplying (1) and (2), we get
x 1
 z    x z  1  x z 2 1  x z 3 
e 2 z
 1            
  2  2 !  2  3!  2  

  x  1  x 2 1  x 3 
1           
  2 z  2 !  2 z  3!  2 z  
…….. (3)
Let us consider the coefficient of z.4
We obtain z when 4z on the left is
4 4
4 x
2 4! x5 z5
multiplied by 1 on the right, and when 2 5 5!
 x
is multiplied by  2 z
on the right, and when
2
x z is multiplied by x
6 6
, and so on.
2 2
6
2 6! 2 2! z

4
In this way we find that the coefficient of z in
this product is
1 1 5 1 1
4
x  6 x  8
4
x  10
6
x 
7

2 4! 2 5! 2 2 !6! 2 3!7 !
2n4

(1) n
 x
  2n4  
n 0 2 n ! ( n  4) !  2  .
Now comparing this series with
2n4

(1) n
 x
J 4 ( x)   2 n  4  
n 0 2 n !(n  4  1) !  2 

2n4

(1) n
 x
  2n4  
n 0 2 n !(n  4) !  2 

The two series are the same!

Similar reasoning establishes that J n (x) is the
n x 1
z
coefficient of in expansion of e 2  z  for any
 z 

nonnegative integer n.
For negative integer, we use the fact that

J v ( x)  (1) J v ( x)
v
 Trigonometric Expansion Involving
Bessel Functions
x 1
 z 
e 2 z
 J 0  z J1  z J 2  z J 3  
2 3

1 2 3
 z J 1  z J 2  z J 3  
…….. (1)
i
Let z  e
x  i 1 
 e  i 
i 2 i 3 i
e 2 e 
 J 0  e J1  e J2  e J3 
 e i J 1  e  2i J  2  e 3i J 3  
But
e i  e  i   v
 sin  and  v
J ( x ) ( 1) J v ( x)
2i

So we’ve
i x sin  i 2 i 3 i
e  J 0  e J1  e J2  e J3 
 i  2 i  3 i
e J1  e J2  e J3 
Or
cos ( x sin  )  i sin( x sin  ) 

J 0  (2 i sin  ) J 1  (2 cos 2 ) J 2  (2 i sin  ) J 3  

When we equate real and imaginary parts,
we’ve
cos ( x sin  )  J 0  2 cos 2 J 2  2 cos 4 ) J 4  

sin ( x sin  )  2 sin  ) J 1  2 sin 3 ) J 3  2 i sin 5 ) J 5  


Let us take   2  
So,
cos ( x cos  )  J 0  2 cos 2 J 2  2 cos 4 ) J 4  
sin ( x cos  )  2 cos  ) J 1  2 cos 3 J 3  2 cos 5 J 5  
Question 12
Prove that
cos x  J 0  2 J 2  2 J 4  
sin x  2 J 1  2 J 3  2 J 5  

Solution
cos ( x sin  )  J 0  2 cos 2 J 2  2 cos 4 ) J 4  
sin ( x sin  )  2 sin  ) J 1  2 sin 3 ) J 3  2 i sin 5 ) J 5  


Putting   2 in (1) and (2), we get
cos x  J 0  2 J 2  2 J 4  
and sin x  2 J 1  2 J 3  2 J 5  