Series Solutions of Linear

Differential Equations
 Solutions about Ordinary Point
 Review of Power Series
Recall from that a power series in x – a has the form


n
c (
n 0
x  a ) n
 c0  c1 ( x  a )  c 2 ( x  a ) 2


Such a series is said to be a power series centered at

a.
 Analytic at a Point
A function f is analytic at a point a, if it can be
represented by a power series in x – a with a positive
radius of convergence. For example:
x x2 x3 x5
e  1     , sin x  x    
x

1! 2! 3! 5!
x2 x4 x6
cos x  1      (2)
2! 4! 6!
 Arithmetic of Power Series
Power series can be combined through the operations
of addition, multiplication and division.
e x sin x
 x2 x3 x 4  x3 x5 x7 
 1  x       x     
 2 6 24  6 120 5040 
 (1) x  (1) x 2      x 3      x 4      x 5  
1 1 1 1 1 1 1
 6 2  6 6  120 12 24 
x3 x5
 x  x   
2

3 30

Copyright © Jones and Bartlett ；滄海書局 Ch5_5

 Example 1 Adding Two Power Series
Write  n2 n(n  1)cn x   n0 cn x n1 as one power series.
 n2 

Solution:
Since series starts with series starts with
x for n = 3 x for n = 0
↓ ↓
   

 n(n  1)cn x
n2
n2
  cn x
n 0
n 1
 2c2 x   n(n  1)cn x
0

n 3
n2
  cn x n 1
n 0

we let k = n – 2 for the first series and k = n + 1 for the

second series,
 Example 1 (2)
then we can get the right-hand side as
same
 
2c2   (k  2)(k  1)ck 2 x   ck 1 x k
k
(3)
k 1 k 1
same

We now obtain
 

 n(n  1)cn x
n 2
n2
  cn x n1
n 0
 (4)
 2c2   [(k  2)(k  1)ck 2  ck 1 ]x k

k 1
 A Solution
 Suppose the linear DE
a2 ( x) y  a1 ( x) y  a0 ( x) y  0 (5)

is put into
y  P( x) y  Q( x) y  0 (6)

Definition :Ordinary and Singular Points

A point x0 is said to be an ordinary point of (5) if both
P(x) and Q(x) in (6) are analytic at x0. A point that is
not an ordinary point is said to be a singular point.
 Polynomial Coefficients
 Since P(x) and Q(x) in (6) is a rational function,

P(x) = a1(x)/a2(x), Q(x) = a0(x)/a2(x)

It follows that x = x0 is an ordinary point of (5) if

a2(x0)  0.
Theorem Existence of Power Series Solutions
If x = x0 is an ordinary point of (5), we can always find
two linearly independent solutions in the form of power
series centered at x0, that is,
y   n0 cn ( x  x0 ) n


A series solution converges at least of some interval

defined by |x – x0| < R, where R is the distance from x0
to the closest singular point.
 Example 2 Power Series Solutions

Solve y"xy  0
Solution:
We know there are no finite singular points.
n0 n and  n2
  n2
Now, y  c x n
y"  n ( n  1) c n x

then the DE gives

 
y  xy   cn n(n  1) x n2
 x  cn x n
n2 n 0
  (7)
  cn n(n  1) x n2
  cn x n1

n2 n 0

Ch5_15
 Example 2 (2)
From the result given in (4),

y  xy  2c2   [(k  1)(k  2)ck 2  ck 1 ]x k  0 (8)
k 1

Since (8) is identically zero, it is necessary all the

coefficients are zero, 2c2 = 0, and
( k  1)( k  2)ck 2  ck 1  0 , k 1, 2 , 3 ,
(9)
Now (9) is a recurrence relation, since
(k + 1)(k + 2)  0, then from (9)
ck 1
ck  2  , k 1, 2 , 3 , (10)
(k  1)(k  2)
 Example 2 (3)

Thus we obtain
c0
k  1, c3  
2．3

k  2, c1
c4  
3．4
c2
k  3, c5    0 ← c2 is zero
4．5
c3 1
k  4, c6    c0
5．6 2．3．5．6
c4 1
k  5, c7    c1
6．7 3．4．6．7
 Example 2 (4)
c5
k  6, c8    0 ← c5 is zero
7．8
c6 1
k  7, c9    c0
8．9 2．3．5．6．8．9

c7 1
k  8, c10    c1
9．10 3．4．6．7．9．10
c8
k  9, c11   0 ← c8 is zero
10．11

and so on.
 Example 2 (5)

Then the power series solutions are

y = c0y1 + c1y2

c0 3 c1 4 c0
y  c0  c1 x  0  x  x 0 x6
2． 3 3． 4 2． 3． 5． 6
c1
 x 7  0  ....
3． 4． 6． 7

Copyright © Jones and Bartlett ；滄海書局 Ch5_19

 Example 2 (6)

1 3 1 1
y1 ( x)  1  x  x 
6
x9  
2．3 2．3．5．6 2．3．5．6．8．9

(1) k
 1  x 3k
k 1 2．3 (3n  1)(3n)

1 4 1 1
y2 ( x)  1  x  x 
7
x10  
3．4 3．4．6．7 3．4．6．7．9．10

(1) k
 x x 3k 1
k 1 3．4 (3n)(3n  1)

Copyright © Jones and Bartlett ；滄海書局 Ch5_20

 Example 3 Power Series Solution
Solve ( x 2  1) y" xy' y  0
Solution:
Since x2 + 1 = 0, then x = i, −i are singular points. A
power series solution centered at 0 will converge at least
for |x| < 1. Using the power series form of y, y’ and y”,
then
  
( x 2  1) n(n  1)cn x n2  x  ncn x n1   cn x n
n2 n1 n 0
   
  n(n  1)cn x   n(n  1)cn x
n n2
  ncn x   cn x n
n

n2 n2 n 1 n 0
 Example 3 (2)

 2c2 x 0  c0 x 0  6c3 x  c1 x  c1 x   n(n  1)cn x n
n2     
k n
  
  n(n  1)cn x n2
  ncn x   cn x n
n

n4      n2   n2 

k n2 k n k n


 2c2  c0  6c3 x   [k (k  1)ck  (k  2)(k  1)ck 2  kck  ck ]x k
k 2

 2c2  c0  6c3 x   [(k  1)( k  1)ck  (k  2)(k  1)ck  2 ]x k  0
k 2
 Example 3 (3)

From the above, we get 2c2 － c0 = 0, 6c3 = 0 , and

(k  1)(k  1)ck  (k  2)(k  1)ck 2  0

Thus c2 = c0/2, c3 = 0, ck+2 = (1 – k)ck/(k + 2)

Then 1 1 1
c 4   c2   c0   2 c0
4 2．4 2 2!
2
c5   c3  0 ← c3 is zero
5
3 3 1．3
c6   c4   c0  3 c0
6 2．4．6 2 3!
4
c7   c5  0 ← c5 is zero
7
Copyright © Jones and Bartlett ；滄海書局 Ch5_23
 Example 3 (4)
5 3．5 1．3．5
c8   c6   c0   4 c0
8 2．4．6．8 2 4!
6
c9   c7  0 ← c7 is zero
9
7 3．5．7 1．3．5．7
c10   c8  c0  5
c0
10 2．4．6．8．10 2 ．5!

and so on.

Ch5_24
 Example 3 (5)

Therefore,
y  c0  c1 x  c2 x 2  c3 x 3  c4 x 4  c5 x 5
 c6 x 6  c7 x 7  c8 x 8  c9 x 9  c10 x10  

 c0 1  x 2  2 x 4  3 x 6  4 x 8  c x
1 1 1．3 1．3．5 1．3．5．7 10
5
x   1
 2 2 2! 2 3! 2 4! 2 5! 
 c0 y1 ( x)  c1 y2 ( x)

1 2  n 1 1．3．5 ( 2 n  3) 2 n
y1 ( x)  1  x   (1) n
x , | x | 1
2 n2 2 n!

y2 ( x)  x

Ch5_25
Yours
 Example 4 Three-Term Recurrence Relation

If we seek a power series solution y(x) for

y  (1  x) y  0
we obtain c2 = c0/2 and the recurrence relation is
ck  ck 1
ck  2  , k 1, 2 , 3 ,
(k  1)(k  2)
Examination of the formula shows c3, c4, c5, … are
expresses in terms of both c1 and c2. However it is more
complicated. To simplify it, we can first choose c0  0,
c1 = 0.

Ch5_28
 Example 4 (2)

Then we have
1
c2  c0
2
c1  c0 c0 1
c3    c0
2．3 2．3 6
c2  c1 c0 1
c4    c0
3．4 2．3．4 24
c3  c2 c0  1 1  1
c5       c0
4．5 4．5  6 2  30
and so on. Next, we choose c0 = 0, c1  0, then
1
c2  c0  0
2
Ch5_29
 Example 4 (3)
c1  c0 c1 1
c3    c1
2．3 2．3 6
c2  c1 c1 1
c4    c1
3．4 3．4 12
c3  c2 c1 1
c5    c1
4．5 4．5．6 120

and so on. Thus we have y = c0y1 + c1y2, where

1 2 1 3 1 4 1 5
y1 ( x)  1  x  x  x  x  
2 6 24 30
1 3 1 4 1 5
y2 ( x )  x  x  x  x 
6 12 120
Copyright © Jones and Bartlett ；滄海書局 Ch5_30
 Example 5 ODE with Nonpolynomial
Coefficients
Solve y"(cos x) y  0
Solution:
We see x = 0 is an ordinary point of the equation. Using
 n 0 n ,

the Maclaurin series for cos x, and using y  c x n

we find
y  (cos x) y

 x 2
x 4
x 6
 
  n(n  1)cn x n2  1        cn x n
n2  2! 4! 6!  n 0
 1  2  1  3
 2c2  c0  (6c3  c1 ) x  12c4  c2  c0  x   20c5  c3  c1  x  
 2   2 
0
Copyright © Jones and Bartlett ；滄海書局 Ch5_31
 Example 5 (2)

It follows that
1 1
2c2  c0  0 , 6c3  c1  0 , 12c4  c2  c0  0 , 20c5  c3  c1  0
2 2
and so on. This gives c2 = –1/2c0, c3 = –1/6c1, c4 =
1/12c0, c5 = 1/30c1,…. By grouping terms we get the
general solution y = c0y1 + c1y2, where the convergence
is |x| < , and
1 2 1 4
y1 ( x)  1  x  x  
2 12
1 3 1 5
y2 ( x)  1  x  x  
6 30

Ch5_32
 5.2 Solutions about Singular Points
 A Definition
A singular point x0 of a linear DE
a2 ( x) y  a1 ( x) y  a0 ( x) y  0 (1)

is further classified as either regular or irregular. This

classification depends on

y  P( x) y  Q( x) y  0 (2)

Copyright © Jones and Bartlett ；滄海書局 Ch5_33

Definition 5.2.1 Regular/Irregular Singular Points
A singular point x0 is said to be a regular singular
point of (1), if p(x) = (x – x0)P(x), q(x) = (x – x0)2Q(x)
are both analytic at x0. A singular point that is not
regular is said to be irregular singular point.

Ch5_34
 Polynomial Coefficients

 If x – x0 appears at most to the first power in the

denominator of P(x) and at most to the second power
in the denominator of Q(x), then x – x0 is a regular
singular point.
 If (2) is multiplied by (x – x0)2,

( x  x0 ) 2 y  ( x  x0 ) p( x) y  q( x) y  0 (3)

where p, q are analytic at x = x0

Ch5_35
 Example 1 Classification of Singular Points
 It should be clear x = 2, x = – 2 are singular points of

(x2 – 4)2y” + 3(x – 2)y’ + 5y = 0

According to (2), we have

3
P( x) 
( x  2)( x  2) 2
5
Q( x) 
( x  2) 2 ( x  2) 2

Copyright © Jones and Bartlett ；滄海書局 Ch5_36

 Example 1 (2)
For x = 2, the power of (x – 2) in the denominator of P
is 1, and the power of (x – 2) in the denominator of Q is
2. Thus x = 2 is a regular singular point.

For x = −2, the power of (x + 2) in the denominator of P

and Q are both 2. Thus x = − 2 is a irregular singular
point.

Ch5_37
Theorem 5.2.1 Frobenius’ Theorem
If x = x0 is a regular singular point of (1), then there
exists one solution of the form
 
y  ( x  x0 ) r
n
c (
n 0
x  x0 )n
 n
c (
n 0
x  x0 ) nr
(4)

where the number r is a constant to be determined.

The series will converge at least on some interval
0 < x – x0 < R.

Ch5_38
 Example 2 Two Series Solutions
 Because x = 0 is a regular singular point of
3xy  y  y  0 (5)

we try to find a solution y  n0 cn x . Now,

n r


y   (n  r )cn x nr 1
n 0


y   (n  r )(n  r  1)cn x nr 2
n 0
 Example 2 (2)

3xy  y  y
  
 3 (n  r )(n  r  1)cn x n r 1
  ( n  r ) cn x n  r 1
  cn x n  r
n 0 n 0 n 0
 
  (n  r )(3n  3r  2)cn x n  r 1
  cn x n  r
n 0 n 0

 
r n
 
 x r (3r  2)c0 x   (n  r )(3n  3r  2)cn x   cn x 
1 n 1

 n1          n1  
 k n1 k n 

r k
 x r (3r  2)c0 x   [(k  r  1)(3k  3r  1)ck 1  ck ]x   0
1

 k 0 

Ch5_40
 Example 2 (3)

which implies r(3r – 2)c0 = 0

(k + r + 1)(3k + 3r + 1)ck+1 – ck = 0, k = 0, 1, 2, …
Since nothing is gained by taking c0 = 0, then
r(3r – 2) = 0 (6)
and
ck
ck 1  , k  0, 1, 2, 
(k  r  1)(3k  3r  1) (7)

From (6), r = 0, 2/3, when substituted into (7),

Ch5_41
 Example 2 (4)
ck
r1 = 2/3, ck 1  , k = 0, 1, 2, …
(3k  5)(k  1)
(8)
ck
ck 1  ,
(k  1)(3k  1)
r2 = 0, k = 0, 1, 2, …
(9)

Ch5_42
 Example 2 (5)

From (8) From(9)

c0 c0
c1  c1 
5．1 1．1
c1 c0 c1 c0
c2   c2  
8．2 2! 5．8 2．4 2!1．4
c2 c0 c2 c0
c3   c3  
11．3 3! 5．8．11 3．7 3!1．4．7
c3 c0 c3 c0
c4   c4  
14．4 4! 5．8．11．14 4．10 4!1．4．7．10
 
c0 (1) n c0
cn  cn 
n! 5．8．11  (3n  2) n!1．4．7  (3n  2)

Ch5_43
 Example 2 (6)

These two series both contain the same multiple c0.

Omitting this term, we have
 
1  (10)
y1 ( x)  x 2 / 3 1   xn 
 n1 n! 5．8．11  (3n  2) 
0 n

1
y2 ( x)  x 1   x 
 n1 n!1．4．7  (3n  2)  (11)
By the ratio test, both (10) and (11) converges for all
finite value of x, that is, |x| < . Also, from the forms of
(10) and (11), they are linearly independent. Thus the
solution is
y(x) = C1y1(x) + C2y2(x), 0 < x < 

Ch5_44
 Indicial Equation
 Equation (6) is called the indicial equation, where
the values of r are called the indicial roots, or
exponents.

 If x = 0 is a regular singular point of (1), then p =

xP(x) and q = x2Q(x) are analytic at x = 0.

Ch5_45
Thus the power series expansions
p(x) = xP(x) = a0 + a1x + a2x2 + …
q(x) = x2Q(x) = b0 + b1x + b2x2 + … (12)
are valid on intervals that have a positive radius of
convergence.
By multiplying (2) by x2, we have
x 2 y  x[ xP( x)] y  [ x 2Q( x)] y  0 (13)

After some substitutions, we find the indicial equation,

r(r – 1) + a0r + b0 = 0 (14)

Ch5_46
 Example 3 Two Series Solutions
Solve 2 xy" (1  x) y ' y  0
Solution:
Let y   n0 cn x nr , then


2 xy  (1  x) y  y
 
 2 (n  r )(n  r  1)cn x n r 1
  (n  r )cn x n r 1
n0 n 0
 
  ( n  r ) cn x n r
  cn x n  r
n 0 n 0
 
  (n  r )(2n  2r  1)cn x n  r 1
  (n  r  1)cn x n r
n 0 n 0

Copyright © Jones and Bartlett ；滄海書局 Ch5_47

 Example 3 (2)

 
r n
 
 x r (2r  1)c0 x   (n  r )(2n  2r  1)cn x   (n  r  1)cn x 
1 n 1

 n1          n0      
 k  n 1 k n 

r k
 x r (2r  1)c0 x   [(k  r  1)(2k  2r  1)ck 1  (k  r  1)ck ]x 
1

 k 0 

which implies r(2r – 1) = 0 (15)

(k  r  1)(2k  2r  1)ck 1  (k  r  1)ck  0, k  0, 1, 2,  (16)

Ch5_48
 Example 3 (3)

From (15), we have r1 = ½ , r2 = 0.

Foe r1 = ½ , we divide by k + 3/2 in (16) to obtain
 ck
ck 1  , k  0, 1, 2,  (17)
2(k  1)

Foe r2 = 0 , (16) becomes

 ck
ck 1  , k  0, 1, 2,  (18)
2k  1

Ch5_49
Example 3 (4)

From (17) From (18)

 c0  c0
c1  c1 
2． 1 1
 c1 c0  c1 c0
c2   2 c2  
2．2 2 ．2! 3 1．3
 c2  c0  c2  c0
c3   3 c3  
2．3 2 ．3! 5 1．3．5
c c c c0
c4  3  4 0 c4  3 
2．4 2 ．4! 7 1．3．5．7
 
(1) n c0 (1) n c0
cn  n cn 
2 n! 1．3．5．7  (2n  1)

Ch5_50
 Example 3 (5)

Thus for r1 = ½
 
(  1) n
 
( 1) n
y1 ( x)  x1/ 2 1   n x n    n x n1/ 2
 n1 2 n!  n0 2 n!

for r2 = 0

(1) n
y2 ( x)  1   xn , | x |  
n 1 1．3．5．7  ( 2n  1)

and on (0, ), the solution is y(x) = C1y1(x) + C2y2(x).

Copyright © Jones and Bartlett ；滄海書局 Ch5_51

 Example 4 Only One Series Solutions
Solve xy" y  0
Solution:
From xP(x) = 0, x2Q(x) = x, and the fact 0 and x are
their own power series centered at 0, we conclude a0 =
0, b0 = 0. Then form (14) we have r(r – 1) = 0, r1 = 1, r2
= 0. In other words, there is only a single series solution

(1) n n1 x 2 x3 1 4
y1 ( x)   x  x   x  ...
n 0 n!( n  1)! 2 12 144

Copyright © Jones and Bartlett ；滄海書局 Ch5_52

 Three Cases

(1) If r1, r2 are distinct and do not differ by an integer,

there exists two linearly independent solutions of
the form:
 
y1 ( x)   cn x n  r1
and y2 ( x)   bn x nr2
n 0 n 0

Copyright © Jones and Bartlett ；滄海書局 Ch5_53

(2) If r1 – r2 = N, where N is a positive integer, there
exists two linearly independent solutions of the
form: 
y1 ( x)   cn x nr1 , c0  0
n 0

(19)
y2 ( x)  Cy1 ( x) ln x   bn x nr2 , b0  0
n 0 (20)

Copyright © Jones and Bartlett ；滄海書局 Ch5_54

(3) If r1 = r2, there exists two linearly independent
solutions of the form:

y1 ( x)   cn x nr1 , c0  0 (21)
n 0

y2 ( x)  y1 ( x) ln x   bn x nr2 (22)
n 0

Ch5_55
 Finding a Second Solution

 If we already have a known solution y1, then the

second solution can be obtained by
e
 Pdx

y2 ( x)  y1 ( x)  2 dx (23)
y1 ( x)

Copyright © Jones and Bartlett ；滄海書局 Ch5_56

 Example 5 Example 4 Revised—Using a
CAS
Find the general solution of xy" y  0
Solution:
From the known solution in Example 4,
1 2 1 3 1 4
y1 ( x)  x  x  x  x 
2 12 144

we can use (23) to find y2(x). Here please use a CAS for
the complicated operations.

Ch5_57
 Example 5 (2)

e
 0 dx
dx
y2 ( x)  y1 ( x)  2
dx  y1 ( x)  2
[ y1 ( x)]  x  x 2  x3 
1 1 1 
 2 x 4
 
12 144
dx
 y1 ( x) 
 x 2  x 3  5 x 4  7 x 5  
 12 12 
1 1 7 19
 y1 ( x)   2    x   dx
 x x 12 72 
 y1 ( x)    ln x  x  x  
1 7 19 2
 x 12 144 
 1 7 19 2 
y2 ( x)  y1 ( x) ln x  y1 ( x)   x  x  
 x 12 144 
Ch5_58
 5.3 Special Functions
 Bessel’s Equation of order v
x 2 y  xy   ( x 2  v 2 ) y  0 (1)
where v  0, and x = 0 is a regular singular point of
(1). The solutions of (1) are called Bessel functions.

 Lengender’s Equation of order n

(1  x 2 ) y   2 xy  n(n  1) y  0 (2)
where n is a nonnegative integer, and x = 0 is an
ordinary point of (2). The solutions of (2) are called
Legender functions.

Ch5_59
 The Solution of Bessel’s Equation
 Because x = 0 is a regular singular point, we know
there exists at least one solution of the form
. Then
 n r
y  from c
n 0 n
(1),
x
x 2 y  xy  ( x 2  v 2 ) y
   
  cn (n  r )(n  r  1) x n r   cn (n  r ) x nr   cn x n r 2  v 2  cn x nr
n0 n 0 n 0 n 0

 c0 (r 2  r  r  v 2 ) x r
 
x r
 cn [(n  r )(n  r  1)  (n  r )  v ]x  x
n 1
2 n r
n
c x
n 0
n 2

 
 c0 (r 2  v 2 ) x r  x r  cn [(n  r ) 2  v 2 ]x n  x r  cn x n2 (3)
n 1 n 0

Copyright © Jones and Bartlett ；滄海書局 Ch5_60

From (3) we have the indicial equation r2 – v2 = 0, r1 =
v, r2 = −v. When r1 = v, we have
(1 + 2v)c1 = 0
(k + 2)(k + 2+ 2v)ck+2 + ck = 0
 ck
or ck  2  , k  0, 1, 2,  (4)
(k  2)(k  2  2v)

The choice of c1 = 0 implies c3 = c5 = c7 = … = 0,

so for k = 0, 2, 4, …, letting k + 2 = 2n, n = 1, 2, 3, …,
we have c2 n  2
c2 n   2
2 n( n  v ) (5)

Copyright © Jones and Bartlett ；滄海書局 Ch5_61

Thus
c0
c2   2
2 ．1．(1  v)
c2 c0
c4   2  4
2 ．2(2  v) 2 ．1．2(1  v)(2  v)
c c0
c6   2 4  6
2 ．3(3  v) 2 ．1．2．3(1  v)(2  v)(3  v)

(1) n c0
c2 n  2 n , n  1, 2, 3,  (6)
2 n!(1  v)(2  v)  (n  v)

Copyright © Jones and Bartlett ；滄海書局 Ch5_62

We choose c0 to be a specific value
1
c0 
2v (1  v)
where (1 + v) is the gamma function. There is an
important relation:
(1 + ) = ()
so we can reduce the denominator of (6):
(1  v  1)  (1  v)(1  v)
(1  v  2)  (2  v)(2  v)  (2  v)(1  v)(1  v)

Copyright © Jones and Bartlett ；滄海書局 Ch5_63

 Hence we can write (6) as

(1) n
c2 n  2 n  v , n  0, 1, 2, ...
2 n!(1  v  n)

Copyright © Jones and Bartlett ；滄海書局 Ch5_64

 Bessel’s Functions of the First Kind

 We define Jv(x) by
2 nv

( 1) n  x
J v ( x)     (7)
n0 n! (1  v  n )  2 

and
2 n v

( 1) n  x
J v ( x )     (8)
n0 n! (1  v  n )  2 

In other words, the general solution of (1) on (0, ) is

y = c1Jv(x) + c2J-v(x), v  integer (9)

See Fig 5.3.1.
Ch5_65
Ch5_66
 Example 1 General Solution: v Not an
Integer
 Consider the DE
x 2 y" xy '( x 2  1/4) y  0

We find v = ½, and the general solution on (0, ) is

y  c1 J1/2 ( x)  c2 J 1/2 ( x)

Ch5_67
 Bessel’s Functions of the Second Kind
 If v  integer, then

cos v J v ( x)  J v ( x) (10)
Yv ( x) 
sin v
and the function Jv(x) are linearly independent.
Another solution of (1) is y = c1Jv(x) + c2Yv(x).
 As v  m, m an integer, (10) has the form 0/0. From
L’Hopital’s rule, the function
Ym ( x)  lim Yv ( x)
v m
and Jv(x) are linearly independent solutions of
x 2 y" xy ' ( x 2  m 2 ) y  0

Ch5_68
 Hence for any value of v, the general solution of (1)
is
y  c1 J v ( x )  c2Yv ( x ) (11)

Yv(x) is called the Bessel function of the second

kind of order v. Fig 5.3.2 shows y0(x) and y1(x).
Ch5_70
 Example 2 General Solution: v an Integer
 Consider the DE
x 2 y" xy ' ( x 2  9) y  0

We find v = 3, and from (11) the general solution on

(0, ) is
y  c1 J 3 ( x)  c2Y3 ( x)

Copyright © Jones and Bartlett ；滄海書局 Ch5_71

 DEs Solvable in Terms of Bessel Functions
 Let t = x,  > 0, in

x 2 y   xy   ( 2 x 2  v 2 ) y  0 (12)
then by the Chain Rule,
dy dy dt dy
 
dx dt dx dt

d 2 y d  dy  dt  2
2 d y
2
     
dx dt  dx  dx  dt 2

Copyright © Jones and Bartlett ；滄海書局 Ch5_72

 Thus, (12) becomes
2 2
 t   2 d y   t  dy   t 2  v 2  y  0
   
     dt
2
dt
2
t 2 d y dy
2
 t   t 2
 v 2
y  0
dt dt

The solution of the above DE is y = c1Jv(t) + c2Yv(t)

Let t = x, we have
y = c1Jv(x) + c2Yv(x) (13)

Ch5_73
 Another equation is called the modified Bessel
equation order v,
x 2 y  xy   ( x 2  v 2 ) y  0 (14)
This time we let t = ix, then (14) becomes
2
d y dy
t 2 2  t  (t 2  2 ) y  0
dt dt
The solution will be Jv(ix) and Yv(ix). A real-valued
solution, called the modified Bessel function of the
first kind of order v is defined by
I ( x)  i  J (ix) (15)

Copyright © Jones and Bartlett ；滄海書局 Ch5_74

 Analogous to (10), the modified Bessel function of
the second kind of order v  integer is defined by
 I  ( x)  I ( x) (16)
K ( x)  
2 sin
and for any integer v = n,
K n ( x)  lim K ( x)
 n

Because Iv and Kv are linearly independent on (0, ),

the general solution of (14) is
y  c1 I ( x )  c2 K ( x ) (17)

Copyright © Jones and Bartlett ；滄海書局 Ch5_75

 We consider another important DE:
1  2a  2 2 2 c2 a 2  p 2 c 2 
y  y   b c x  2
 y  0, p  0 (18)
x  x 

The general solution of (18) is

y  x a [c1 J p (bx c )  c2Y p (bx c )] (19)

We shall not supply the details here.

Copyright © Jones and Bartlett ；滄海書局 Ch5_76

 Example 3 Using (18)

Find the general solution of xy  3 y  9 y  0 on (0, ).

Solution:
Writing the DE as
3 9
y  y  y  0
x x
according to (18)
1 – 2a = 3, b2c2 = 9, 2c – 2 = −1, a2 – p2c2 = 0
then a = −1, c = ½ . In addition we take b= 6, p = 2.
From (19) the solution is
y  x 1[c1 J 2 (6 x1/ 2 )  c2Y2 (6 x1/ 2 )]

Copyright © Jones and Bartlett ；滄海書局 Ch5_77

 Example 4 The Aging Spring Revised
 Recall the model in Sec. 3.8
mx  ke t x  0,   0
You should verify that by letting
2 k t / 2
s e
 m
we have
2
d x dx
s2 2  s  s2 x  0
ds ds

Copright © Jones and Bartlett ；滄海書局 Ch5_78

 Example 4 (2)

The solution of the new equation is

x = c1J0(s) + c2Y0(s),
If we resubstitute
2 k t / 2
s e
 m
we get the solution.
 2 k t / 2   2 k t / 2 
x(t )  c1 J 0  e   c2Y0  e 
 m   m 

Copyright © Jones and Bartlett ；滄海書局 Ch5_79

 Properties

(i) J m ( x)  (1) m J m ( x)

(ii) J m ( x)  (1) m J m ( x)

0 , m  0
(iii) J m ( 0)  
1 , m  0

(iv) lim x0 Ym ( x)  

Copyright © Jones and Bartlett ；滄海書局 Ch5_80

 Example 5 Derivation Using Series
Definition
Derive the formula xJ v ( x)  vJ v ( x)  xJ v1 ( x)
Solution:
It follows from (7)
2 n v

(1) n (2n  v)  x 
xJ v ( x)    
n 0 n! (1  v  n )  2
2 n v 2 nv

(1) n
 x

(1) n  x 
n
 v    2  
n 0 n! (1  v  n)  2  n 0 n! (1  v  n)  2 
2 n v 1

(1) n  x
 vJ v ( x)  x   
(n  1)!(1  v  n) 2
 n1            
k  n 1
2 k v 1

(1) k
 x
 vJ v ( x)  x     vJ v ( x)  xJ v1 ( x)
k 0 k! ( 2  v  k )  2 
Copyright © Jones and Bartlett ；滄海書局 Ch5_81
 The result in example 5 can be written as
v
J v ( x)  J v ( x)   J v1 ( x)
x
which is a linear DE in Jv(x). Multiplying both sides
the integrating factor x-v, then
d v
[ x J ( x)]   x v J ( x) (20)
v v 1
dx
It can be shown
d v
[ x J v ( x)]  x v J v1 ( x) (21)
dx
When y = 0, it follows from (14) that
J 0 ( x)   J1 ( x), Y0( x)  Y1 ( x) (22)

Copyright © Jones and Bartlett ；滄海書局 Ch5_82

 Spherical Bessel Functions
 When the order v is half an odd number, that is,
1/2, 3/2, 5/2, …..
The Bessel function of the first kind Jv(x) can be
expressed as spherical Bessel function:
2 n 1 / 2

(1) n
 x
J1/ 2 ( x )    
n0 n! (1  1 / 2  n)  2 

Since (1 + ) = () and (1/2) = ½, then

(2n  1)!
1   n   2 n1
1

 2  2 n!

Copyright © Jones and Bartlett ；滄海書局 Ch5_83

 Hence
2 n 1/ 2

(1) n
 x 2  (1) n 2 n1
J 1/ 2 ( x )       x
n 0 (2n  1)!   2 x n0 (2n  1)!
n!
2 2 n1 n!

and 2
J1/ 2 ( x)  sin x (23)
x
2
J 1/ 2 ( x )  cos x (24)
x

Copyright © Jones and Bartlett ；滄海書局 Ch5_84

 The Solution of Legendre Equation
 Since x = 0 is an ordinary point of (2), we use
y   n  0 cn x n


After substitutions and simplifications, we obtain

n(n  1)c0  2c2  0
(n  1)(n  2)c1  6c3  0
( j  2)( j  1)c j 2  (n  j )(n  j  1)c j  0

or in the following forms:

Copyright © Jones and Bartlett ；滄海書局 Ch5_85

 n(n  1)
c2   c0
2!
(n  1)(n  2)
c3   c1
3!
(n  j )(n  j  1)
c j2   c j , j  2, 3, 4,  (25)
( j  2)( j  1)

Using (25), at least |x| < 1, we obtain

 n(n  1) 2 (n  2)n(n  1)(n  3) 4

y1 ( x)  c0 1  x  x
 2! 4!
(n  4)(n  2)n(n  1)(n  3)(n  5) 6
 x  
6! 
Copyright © Jones and Bartlett ；滄海書局 Ch5_86
  (n  1)(n  2) 3 (n  3)(n  1)(n  2)(n  4) 5
y2 ( x)  c1  x  x  x
 3! 5!
(n  5)( n  3)(n  1)(n  2)(n  4)(n  6) 7
 x   (26)
7! 

Notices: If n is an even integer, the first series

terminates, whereas y2 is an infinite series.
If n is an odd integer, the series y2 terminates with xn.

Copyright © Jones and Bartlett ；滄海書局 Ch5_87

 Legendre Polynomials
 The following are nth order Legendre polynomials:
P0 ( x)  1, P1 ( x)  x
1 1
P2 ( x)  (3 x 2  1), P3 ( x)  (5 x 3  3 x) (27)
2 2
1  1
P4 ( x)  (35 x  30 x  3), P5 ( x)  (63 x 5  70 x 3  15 x)
2

8 8

Copyright © Jones and Bartlett ；滄海書局 Ch5_88

 They are in turn the solutions of the DEs. See Fig
5.3.5
n  0: (1  x 2 ) y  2 xy  0
n  1: (1  x 2 ) y  2 xy  2 y  0
n  2: (1  x 2 ) y  2 xy  6 y  0
(28)
n  3: (1  x 2 ) y  2 xy  12 y  0
 

Copyright © Jones and Bartlett ；滄海書局 Ch5_89

 Properties

(i) Pn ( x)  (1) n Pn ( x)

(ii) Pn (1)  1

P
(iii) n ( 1)  ( 1) n

(iv) Pn (0)  0, n odd

(v) P 'n (0)  0, n even

 Recurrence Relation
 Without proof, we have
( k  1) Pk 1 ( x )  ( 2k  1) xPk ( x )  kPk 1 ( x )  0 (29)

which is valid for k = 1, 2, 3, …

Another formula by differentiation to generate
Legendre polynomials is called the Rodrigues’
formula:
1 dn 2
Pn ( x)  n n
( x  1) n
, n  0, 1, 2, ... (30)
2 n! dx