MATHEMATICS - III

BITS Pilani Prof. Manoj Kumar Pandey

Department of Mathematics
K K Birla Goa Campus

BITS Pilani, K K Birla Goa Campus

 SOME SPECIAL FUNCTIONS
OF MATHEMATICAL PHYSICS

BITS Pilani, K K Birla Goa Campus

 Special Functions
A special function is a function (usually named after an early investigator
of its properties) with particular use in mathematical physics or some
other branch of mathematics.

(a) Hypergeometric Function

(b) Legendre Polynomials
(c) Bessel Functions
(d) Gamma Functions
(e) Beta Function
(f) Zeta Functions

BITS Pilani, K K Birla Goa Campus

Special Functions

• Legendre Polynomials
• Bessel Functions and Gamma Function

BITS Pilani, K K Birla Goa Campus

 Legendre Polynomials
The Legendre Equation is

(1  x ) y   2 xy   n(n  1) y  0,
2

where n is a non-negative integer.

We have found the solution of this equation near the ordinary point x = 0.
é n(n 1) n(n  2)(n 1)(n  3) 4 ù
y(x)  a0 ê1 x x  ú
ë 2! 3! û
é (n 1)(n  2) 3 (n 1)(n  3)(n  2)(n  4) 5 ù
 a1 ê x  x  x  ú
ë 3! 5! û

which is valid for -1< x< 1.

BITS Pilani, K K Birla Goa Campus

However, the solutions most valuable in the applications
are those bounded near x = 1.

Here we will solve Legendre’s equation near x =1, in terms

of hypergeometric function F(a, b, c, x).

BITS Pilani, K K Birla Goa Campus

We change the independent variable from x to t by
1
t  (1  x)
2
In terms of t, the equation becomes

t (1  t ) y   (1  2 t ) y   n ( n  1) y  0 .....(*)
Here t = 0 corresponds to x = 1 of the Legendre equation.
(*) is a hypergeometric equation with
c  1, a  b  1  2, ab   n ( n  1)

 a = -n, b = n+1, c = 1

BITS Pilani, K K Birla Goa Campus

Therefore, the solution of (*) near t = 0 is

y ( t )  F (  n , n  1,1, t )

Since c=1 which implies the exponents m1= 0 and m2 = 1- c = 0

hence only one solution is possible at this moment.

To find a second linearly independent solution, we take

BITS Pilani, K K Birla Goa Campus

y2 = vy1, where
1   P dt 1  ( 2t 1) / t (1t ) dt
v  2 e  2e
y1 y1
1 1é 1 ù
 2  ê 2 ú
y1 t (1  t ) t ë y1 (1  t ) û
Since y12 is a polynomial with constant term 1, the bracketed
expression on the right is an analytic function of the form

1  a1t  a 2 t 2  

BITS Pilani, K K Birla Goa Campus

 1
We can write v    a1  a 2 t  
t
This yields v  log t  a1t    y 2  y1 (log t  a1t  )

Hence, the general solution of (*) near t=0 is

y  c1 y1  c2 y2 .

The solution will be bounded near t = 0 if and only if c2 = 0.

BITS Pilani, K K Birla Goa Campus

If we replace t by (½)(1-x), we obtain the solution of Legendre’s
equation near x=1.

The solution of Legendre’s equation, which is bounded near

x = 1 are precisely constant multiplies of the polynomial

F(-n, n+1, 1, ½(1-x)).

BITS Pilani, K K Birla Goa Campus

 Legendre Polynomials

The nth Legendre polynomial denoted by Pn(x) is defined by

é 1 x ù
Pn (x)  F ên,n 1,1, ú
ë 2 û
2
(n)(n 1) æ 1 x ö (n)(n 1)(n 1)(n  2) æ 1 x ö
 1 2 ç ÷ 2 ç ÷ 
(1!) è 2 ø (2!) è 2 ø
n
(n)(n 1)[n  (n 1)](n 1)(n  2)(2n) æ 1 x ö
 2 ç ÷
(n!) è 2 ø

BITS Pilani, K K Birla Goa Campus

Which gives
n(n 1) n(n 1)(n 1)(n  2)
Pn (x)  1 2
(x 1)  2 2
(x 1) 2

(1!) 2 (2!) 2
(2n)!
 2 n
(x 1) n
.
(n!) 2
with Pn(1)=1.
This is not a suitable form and is a very inconvenient tool.
So we look for a simple form which can be used to compute
Legendre polynomials for different n.

BITS Pilani, K K Birla Goa Campus

Legendre polynomials

Any polynomial of degree n satisfying Legendre

equation
(1  x 2 ) y  2 xy  n(n  1) y  0

With Pn(1) = 1 is called a Legendre polynomial.

BITS Pilani, K K Birla Goa Campus

Rodrigues Formula

The nth Legendre polynomial is given by the

Rodrigues formula
1 dn 2
Pn ( x )  n n
( x  1) n
, n  0,1, 2,....
2 n ! dx

BITS Pilani, K K Birla Goa Campus

 dn
• To show  x  1
n
which is a polynomial of
y( x) 
dx n
2

degree n, is a solution of the Legendre

equation.

• Leibnitz’s rule for higher-order derivative

( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )

BITS Pilani, K K Birla Goa Campus

 1 dn
Rodrigues formula Pn ( x )  n n
( x 2
 1) n
provides a
2 n ! dx
relatively easy method for computing the successive Legendre
polynomials. (useful for small n)

Use it to calculate the first 5 Legendre polynomials.

BITS Pilani, K K Birla Goa Campus

Legendre Polynomials

Few Legendre Polynomials

P0 ( x)  1, P1 ( x)  x,
1 2 1 3
P2 ( x)  (3x 1), P3 ( x)  (5x  3x),
2 2
1 1
P4 ( x)   35x  30x  3 , P5 ( x)   63x5  70x3 15x  ,....
4 2

8 8

BITS Pilani, K K Birla Goa Campus

Legendre Polynomials

BITS Pilani, K K Birla Goa Campus

Legendre Polynomials

BITS Pilani, K K Birla Goa Campus

Problems:

Use the Rodrigues formula to prove the following relations:

n ( n  1)
( i ) P n (  1 )  (  1 ) , ( ii ) P n ( 1 ) 
n

2
( iii ) ( 2 n  1 ) P n ( x )  P n 1 ( x )  P n 1 ( x )
hences show that
P 0  3 P1  5 P 2    ( 2 n  1 ) P n  P n 1  P n

Hint: (i) and (ii) Expand using Leibnitz’s rule.

(iii) Expand the first term on the right side . Take n= 1,2 .. and
sum all terms to get the second part.

BITS Pilani, K K Birla Goa Campus

Generating Function

The function on the left side of


1
  Pn ( x)t n
1  2 xt  t 2 n 0

is called the generating function of the Legendre polynomials.

For small values of t (t ≠ 1).

BITS Pilani, K K Birla Goa Campus

Recursion Formula

Important Recursion Formula :

( n  1) Pn 1 ( x )  (2 n  1) xPn ( x )  nPn 1 ( x ), n  1, 2, ....

Hint: Differentiate the generating function with respect to t and equate

the coefficient of the general term.

BITS Pilani, K K Birla Goa Campus

Problems:

Use the generating function to verify the following results

(i ) Pn (1)  1, (ii ) Pn (  1)  (  1) n (iii ) P2 n 1 (0)  0,

æ 1  3 (2 n  1) ö
(iv ) P2 n (0)  (  1) ç
n
÷.
è 2  n!
n
ø

BITS Pilani, K K Birla Goa Campus

Additional Problems

Show that

(i ) Pn (  x )  (  1 ) n Pn ( x )
( ii ) nP n ( x )  x Pn ( x )  Pn 1 ( x )
( iii ) ( n  1 ) Pn  Pn 1  x Pn

BITS Pilani, K K Birla Goa Campus

Applications of Legendre
Polynomials

BITS Pilani, K K Birla Goa Campus

Important Result

Every polynomial of degree k can be expressed as a linear

combination of first k+1 Legendre polynomials:

P0 ( x ) , P1 ( x ) ,  , Pk ( x )

That is k
p ( x)   a n Pn ( x).
n 0

BITS Pilani, K K Birla Goa Campus

For example let
p(x)=1+2x+3x2+5x3

1  P0  x  , x  P1 ( x),
1 1 2 1 2
P2 ( x)  (3x  1)  x   P2 ( x)  P0 ( x)  P2 ( x) ,
2 2

2 3 3 3 3
1 3 2 3 2
P3 ( x)  (5 x  3x )  x  x  P3 ( x)  P1 ( x)  P3 ( x)
3 3

2 5 5 5 5

BITS Pilani, K K Birla Goa Campus

Therefore
é1 2 ù é3 2 ù
p ( x )  P0 ( x )  2 P1 ( x )  3 ê P0 ( x )  P2 ( x ) ú  5 ê P1 ( x )  P3 ( x ) ú
ë3 3 û ë5 5 û
 2 P0 ( x )  5 P1 ( x )  2 P2 ( x )  2 P3 ( x )
3
  a n Pn ( x ).
n 0

Therefore, any polynomial p(x) of degree k can be written as

k
p ( x)   a n Pn ( x).
n 0

BITS Pilani, K K Birla Goa Campus

Example

Express

p(x)=5+2x2+x3

In terms of Legendre polynomials.

Solution:
2 4 3 17
p(x)  P3 ( x )  P2 ( x )  P1 ( x )  P0 ( x )
5 3 5 3

BITS Pilani, K K Birla Goa Campus

Orthogonal functions

What are orthogonal functions?

Two functions f and g are said to be orthogonal in the

interval [a,b] if
b

 f ( x ) g ( x ) dx  0
a

How is this related to the more conceivable concept of

orthogonal vectors?

BITS Pilani, K K Birla Goa Campus

 VECTORS FUNCTIONS
Orthogonal Orthogonal Example:
dot product 1 ,  2 We say they are 1  t ,  2  t 2 , [1,1]
Orthogonal if
u  (2,3), w  (3,2) inner product
Unlike vector analysis,
 1 ,  2  0 where the word
uw  0 orthogonal is a
synonym for
Orthogonal set perpendicular, in this
Orthogonal set present context the
u1 , u2 ,, un  1 ,  2 ,,  n  term orthogonal have
We say it is orthogonal set if no geometric
We say it is orthogonal set if
significance.
u m  u n  0 m  n   m ,  n   0 m  n
The only continuous
Example: Example: function orthogonal to
é1 ù é1ù é1ù 1  cos t ,  2  cos 2t ,  3  cos 3t each member of the set
u1  ê 2 ú u2  ê 1 ú u3  ê 0 ú [ ,  ] is the zero function.
ë1 û ë1û ë1û
Show that they are orthogonal
Show that they are orthogonal
BITS Pilani, K K Birla Goa Campus
Properties of Legendre Polynomials
Orthogonality

The most important property of the Legendre polynomials

P0 ( x ), P1 ( x ),  , Pn ( x ), 

is  0 if m  n,
1 
1 m n
P ( x ) P ( x ) dx   2 if m  n.
 2n  1

that is, Legendre polynomials form an orthogonal sequence of

functions in the interval [-1, 1].

BITS Pilani, K K Birla Goa Campus

Proof: Let f(x) be any function with at least n continuous
derivatives on the interval  1 x 1 , and consider the
integral 1
I   f ( x) Pn ( x)dx
1
1 1 dn 2
By Rodrigues formula I  n  f ( x) n ( x  1) n dx
2 n ! 1 dx

after integration by parts, we get

1
1 é d n1
ù 1 1 d n1
I  n ê f (x) n1 (x2 1)n ú  n  f (x) n1 (x2 1)ndx
2 n! ë dx û1 2 n! 1 dx
0
BITS Pilani, K K Birla Goa Campus
Therefore, we get
(  1) 1 d n 1
I  n  f ( x ) n 1 ( x 2  1) n dx
2 n ! 1 dx
continuing to integrate by parts, we obtain

(  1) n 1
I  n  ( x )( x 2  1) n dx
(n)
f
2 n! 1

If f(x) = Pm(x) with m < n, then f(n)(x) = 0 and consequently I = 0.

which proves the first part i.e. the case when m ≠ n (????)

For the second part, let’s take f(x) = Pn(x).

BITS Pilani, K K Birla Goa Campus

 (n) (2n)!
Since Pn ( x)  n , we get
2 n!
( 2 n )! 1
I 
2 2 n ( n!) 2  1
(1  x 2 ) n dx

2 ( 2 n )! 1
 2n
2 ( n!) 2  0
(1  x 2 ) n dx .

If we change the variable by x  sin  , and recall the formula

 /2 2 2 n (n!) 2
0 cos  d  (2n  1)! .
2 n 1

We conclude that in this case I  2 /(2n  1) , hence the proof.

BITS Pilani, K K Birla Goa Campus

Alternate way to prove the orthogonality

See problems 2 and 3 of the text book (pp. 347)

BITS Pilani, K K Birla Goa Campus

Problem: 1

Show that 1
2n
1 xPn ( x) Pn1 ( x)dx  4n 2  1

Hint: Use the main recursion formula slide 23 and orthogonality.

BITS Pilani, K K Birla Goa Campus

Problem: 2

If p(x) is a polynomial of degree n  1 such that

1
1
x k p ( x)dx  0 for k  0,1, , n  1.

Show that p(x) = c Pn(x) for some constant c. Is the

converse true?

BITS Pilani, K K Birla Goa Campus

 Legendre Series

An “arbitrary function” f(x) can be expressed as Legendre


series

f ( x)  a P ( x)
n 0
n n

Need to calculate coefficients an for the above expression.

Multiply Pm(x) on both sides and integrate from –1 to 1, we get

f ( x) Pm dx   an  Pm ( x) Pn ( x )dx
1 1
1
n 0
1

Using orthogonality of Legendre polynomials

1 2am
1 f ( x) Pm dx  2m  1
BITS Pilani, K K Birla Goa Campus
Therefore, we have
æ 1ö 1
a n  ç n  ÷  f ( x ) Pn dx .
è 2 ø 1

Legendre expansion theorem: If f(x) and f  ( x ) have at most a

finite number of jump discontinuities on the interval [-1, 1], and
if f(x-) and f(x+) denotes the limits of f(x) from the left and from
the right at a point x, then a n exist and the Legendre series
converges to 1
[ f ( x  )  f ( x  )] for  1  x  1 .
2

BITS Pilani, K K Birla Goa Campus

 Example 1:

Find the first three terms of the Legendre Series of

0 if  1  x  0
(i ) f ( x)  
 x if 0  x 1.

( ii ) f ( x )  e x

BITS Pilani, K K Birla Goa Campus

Interesting Facts about Legendre
Polynomials
• All the roots of the Legendre polynomials are real
and distinct.

• All the roots lie in the interval (-1, 1).

• There is a second set of solutions to Legendre’s

equation, called “Legendre functions of the second
kind”; these diverge at x = 1, -1 and are therefore
often not used.

BITS Pilani, K K Birla Goa Campus

Interesting Facts about Legendre
Polynomials
Let f(x) be a function defined in [-1, 1], now consider the problem
of approximating f(x) as closely as possible in the sense of least
square by polynomials p(x) of degree  n.
1
I   1
[ f ( x )  p ( x )] 2 dx .

The Problem is to minimize I:

It turns out that the minimizing polynomial is:
p ( x)  a0 P0 ( x)  a1 P1 ( x)    an Pn ( x)

æ 1 ö 1
With an  ç n  ÷  f ( x ) P n dx .
è 2 ø 1

BITS Pilani, K K Birla Goa Campus