Arab Academy for Science, technology& maritime Transport

College of Engineering &Technology

Department of Basic & Applied Science

Mathematics (V) - BA323

Common for (Electronics, Computer, Electrical and Mechanics)

Notes & Sheets

 TABLE OF CONTENTS

Chapter I: Series solution of Differential Equation .. 5

1 Taylor’s series method.................................................... 6

Sheet 1 ................................................................................ 8

2 Power series solution about an ordinary point .......... 9

Sheet 2 .............................................................................. 15

3 Frobenius Method ........................................................... 16

Sheet 3 .............................................................................. 24

Chapter II: Special Functions ................................. 25

1 Gamma Function ............................................................ 25

2 Beta Function .................................................................. 30

Sheet 4 .............................................................................. 36

3 Bessel Functions ............................................................. 37

Sheet 5 .............................................................................. 48

4 Legendre Polynomials .................................................. 50

Sheet 6 .............................................................................. 60

2
Chapter III: Boundary Value Problems .................. 62

1 Heat Equation ................................................................. 63

2 Wave Equation ............................................................... 73

Sheet 7 .............................................................................. 80

Chapter IV: Conformal Mapping ............................ 82

1 Bilinear Transformation ............................................... 85

2 Schwarz - Christoffel Transformation ...................... 90

Sheet 8 .............................................................................. 94

Examinations ........................................................... 96

3
 Syllabus for Math V (323) & course Assignment
Common for (Electronics, Computer, Electrical and Mechanics)
Text book: Advanced Engineering mathematics, Erwin Kreyszig
Reference book: Advanced Engineering mathematics, Zill
Course Grading: 30 Marks 7th week (10 Quiz+20 Test) +20 marks 12th
(2 quizzes)+10 marks H.W.& attendance+40 marks final Exam

Week Chapter Topics Tutorial

1 1.1 Taylor’s and power series method for solving Sheet 1
ordinary differential equations
2 1.2 Differential equations with variable coefficients, Sheet 2
ordinary and singular points, solution about
ordinary points
3 1.3 Solutions about singular points: regular singular Sheet 3
points, the method of Fronenius – case I
4 1.3 The method of Fronenius – case II and Case III Sheet 3
5 2.1,2.2 Gamma and Beta functions Sheet 4
6 2.3 Bessel differential equation, Bessel function of Sheet 5
the first kind
7 2.4 Legendre differential equations, Legendre Sheet 6
polynomails
8 Revision and mid. term exam
9 3 Boundary value problems. Partial differential Sheet 7 [A]
equations, separation of variables
10 3.1 Heat Equation, Heat transfer in a bar Sheet 7
11 3.2 Wave equation, vibration of a string [B]
12 3.3 Laplac, equation, Potential fields Sheet 8 [C]
13 4 Conformal mappings Complex functions as Sheet 8
mappings
14 4.1 Bilinear transformation (Linear fraction Sheet 8 [A]
transformations
15 4.2 Schwarz – christoffel transformation Sheet 8 [B]

4
 CHAPTER I

Series Solutions of Differential Equations

1. Taylor’s series method
2. Power series solution about an ordinary point
Method of undetermined coefficients
3. Frobenius Method

5
 1. Taylor’s series method

If f ( x) is defined in some interval a  x  b and if x0 is a point in this

interval and if all derivatives of f ( x) exist at x0 , then Taylor’s series

of f ( x) is
f ( x0 ) f ( x0 ) f ( x0 )
f ( x)  f ( x0 )  ( x  x0 )  ( x  x0 ) 2  ( x  x0 )3  ...
1! 2! 3!
If x0  0 , we obtain Maclaurin’s series
f (0) f (0) 2 f (0) 3
f ( x)  f (0)  x x  x  ...
1! 2! 3!
The series converges for all points in the interval.
We illustrate this method by the following examples:
Ex(1)
Find the first four non- zero terms in a power series solution for the
differential equation y  x 2  y 2 , when x  1, y  1
Solution
y (1)  1
y  x 2  y 2 y(1)  0
y  2 x  2 yy y(1)  2
y  2  2( y2  yy) y(1)  2
y iv  2(3 yy  yy) y iv (1)  4,
Then the first four non -zero terms of the power series solution is
2 2 4
y 1 ( x  1) 2  ( x  1)3  ( x  1) 4  ...
2! 3! 4!
Or

6
y  1  ( x  1) 2  13 ( x  1)3  16 ( x  1) 4  ...
Ex(2)
Find the first four non- zero terms in a power series solution for the
differential equation y  xy  yy , when x  0, y  0, y  1, y  2
Solution
y (0)  0
y(0)  1
y(0)  2
y  xy  yy y(0)  0
y iv  y  xy  y2  yy y iv (0)  1
y v  y  xy  y  2 yy  yy  yy
 2 y  xy  3 yy  yy y v (0)  8
Then the first four non -zero terms of the power series solution is
1 2 1 8
y x  x 2  x 4  x 5  ...
1! 2! 4! 5!
Or
1 4 1 5
y  x  x2  x  x  ...
24 15

7
 SHEET #1
Obtain the first four non- zero terms in a power series solution for each of the
following initial value problems:
1. y  x 2 y 2  1 , when x  1, y  1

2. y  sin( xy )  x 2 when x  0, y  3

3. y  x 2  y 2 when x  0, y  1, y  0

4. y  x  e y when x  0, y  1
5. y3  3 xy2  x  y  0 when x  0, y  1

6. 3y 2 y  y 3  x , when x  0, y  1

8
 2. Power series solution about an ordinary point
Method of undetermined coefficients
Consider the second order homogenous linear differential equation

a2 ( x) y  a1 ( x) y  a0 ( x) y  0 (2.1)

We assume that (2.1) has a solution expressible in the form

y  c0  c1 ( x  x0 )  c2 ( x  x0 ) 2  ...
 (2.2)
  ck ( x  x0 ) k ,
k 0

where c0 , c1 , c2 ,... are constants.

An expression of the form (2.2) is called a power series in ( x  x0 ) . We have

thus assumed that the differential equation (2.1) has a so- called power series
solution of the form (2.2). Assuming that this assumption is valid, we can
proceed to determine the coefficients c0 , c1 , c2 ,... in (2.2) in such a manner that
the expression (2.2) does indeed satisfy the equation (2.1).

But under what conditions is this assumption actually valid? That is , under
what conditions can we be certain that the differential equation (2.1)actually
does have a solution of the form (2.2)? In order to answer this question, we
shall first introduce certain basic definitions. For this purpose let us write the
differential equation (2.1) in the equivalent normalized form

y  P( x) y  Q( x) y  0 (2.3)

9
 a1 ( x) a ( x)
where P( x)  and Q( x)  0 .
a2 ( x) a2 ( x)

Definition 1:

If P( x) and Q ( x) are finite, single valued and differential at x0 , then x0 is

called an ordinary point. In this case

lim
x x
P( x) and lim
x x
Q( x)
0 0

both exist, i.e. are finite. If, on the other hand, either of the two limits is infinite,
we say that x0 is a singular point.

Definition 2:

Suppose that x0 is a singular point, but that

lim(
x x
x  x0
) P ( x ) and lim(
x x
x  x0
) 2
Q( x)
0 0

both finite, x0 is said to be a regular singular point. If either limit is infinite,

there is an essential singularity at x0 . In this case x0 is called irregular singular
point.

If x0 is an ordinary point of the differential equation (2.1), then we assume a

series solution of the form (2.2).

If x0 is a regular singular point of the differential equation (2.1), then we assume

a Frobenius series solution of the form

10
 
y   ck ( x  x0 ) k  , c0  0 (2.4)
k 0

Examples (Ordinary points)

Ex(1)

Find a power series solution for the equation y  xy  y  0 about x  0 .

It is clear that x  0 is an ordinary point; therefore assume a power series

solution of the form


y   ck x k , (2.5)
k 0

Differentiating (2.5) with respect to x twice, we obtain

 
y   kck x k 1   (k  1)ck 1 x k , (2.6)
k 1 k 0

 
y   k (k  1)ck x k  2   (k  1)(k  2)ck  2 x k , (2.7)
k 2 k 0

Substituting in the differential equation we obtain

  

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

k 0 k 0 k 0

Now, since the sum in the last equation must be identically zero then each
coefficient of various powers of x must be zero. In particular, equating the
coefficients of x k to zero, we obtain

11
 (k  1)(k  2)ck  2  kck  ck  0

ck
 ck  2   , k 0 (2.8)
k 2

If c0 and c1 are kept arbitrary, then all the other ck ' s will be in terms of these
arbitrary constants. Substitute in (2.8) for k  0,1,2,...

c0
k  0, c2  
2
c
k  1, c3   1
3
c c
k  2, c4   2  0
4 8
c c
k  3, c5   3  1
5 15
c c
k 4 c6   4   0
6 48

Therefore the series solution can be written as


y   ck x k  c0  c1 x  c2 x 2  c3 x 3  ....
k 0

c0 2 c1 3 c0 4 c1 5 c0 6
 c0  c1 x  x  x  x  x  x  ....
2 3 8 15 48
 1 1 1   1 1 
 c0 1  x 2  x 4  x 6  ....   c1  x  x 3  x 5  .... 
 2 8 48   3 15 

Ex(2)

Find a power series solution for the equation y  y  xy  0 , when x  0 ,

12
y  2, y  0 .

Solution

We have

  
y   ck x k , y   (k  1)ck 1 x k , y   (k  1)(k  2)ck  2 x k
k 0 k 0 k 0

To satisfy the initial condition, we have c0  2, c1  0

Substitute for y, y , y in differential equation, we get

  

 (k  1)(k  2)ck 2 x k   (k  1)ck 1 x k  ck x k 1  0,

k 0 k 0 k 0

equating the coefficients of x k to zero, we obtain

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

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

c1  c1 c
k  0, c2  , c1  0  c2   1  0
2 2
c  c 0  2 1
k  1, c3  2 0  
6 6 3
3c3  c1 c3 1
k  2, c4   
12 4 12
4c4  c2 c4 1
k  3, c5   
20 5 60

13
 5c5  c3 1
k 4 c6  
30 120

Therefore the series solution can be written as


y   ck x k  c0  c1 x  c2 x 2  c3 x 3  ....
k 0

1 1 1 1 6
 2  x 3  x 4  x 5  x  ....
3 12 60 120

14
 SHEET #2
a. Find the recurrence relation for the coefficients of the power series solution
for the following differential equations about x  0
1. y  y  0
2. y  2 y  xy  0

3. y  x 3 y  0
4. y  xy  2 xy  0
b. Find the first four non – zero terms in the power series solution for the
following differential equations about x  0
1. 2 y  4 xy  8 x 2 y  0

2. y  12 y  x 2 y  0
c. Find the first five non – zero terms of the solution of the initial value
problems
1. y  2 xy  ( x  1) y  0 when x  0 , y  1 and y  2
2. y  y  xy  0 when x  0 , y  1 and y  1
d. Solve in series the following differential equations near x  0
1. (1  x 2 ) y  2 xy  y  0

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

3. (1  x 2 ) y  xy  y  0

4. (1  x 2 ) y  xy  4 y  0

5. (1  x 2 ) y  2 xy  y  0

6. y  xy  x 2 y  0

7. (1  x 2 ) y  xy  xy  0

15
 3. The method of Frobenius
If x  0 is a regular singular point of the differential equation
y  P( x) y  Q( x) y  0 , then we assume a Frobenius series solution of the

form y   ck ( x  x0 ) k  , c0  0 .
k 0

Case 1:
The roots of the indicial equation are distinct and the difference is not an
integer.
Ex (1)
Find a series solution for the following differential equation about origin:
2 xy  (1  x) y  2 y  0
Solution

Assume a solution of the form y   ck x k  , c0  0 ,
k 0


y   (k  )ck x k 1 ,
k 0

y   (k  )(k    1)ck x k 2
k 0

Substitute in differential equation

  
2 x  (k  )(k    1)ck x k 2  (1  x) (k  )ck x k 1  2 ck x k   0
k 0 k 0 k 0

0r
   
2 (k  )(k    1)ck x k 1   (k  )ck x k 1   (k  )ck x k   2 ck x k   0
k 0 k 0 k 0 k 0

16
Now, to get the indicial equation, we equate the coefficient of the lowest power
of x , ( x 1 ) to zero, we obtain
 2(  1)    c 0
 0  (2  1)c0  0, c0  0
 (2  1)  0
Therefore the roots of the indicial equation are   0 and   12 . The difference
between the two roots is not an integer then for each value of  corresponds a
solution.
Equating the coefficients of x k  , we obtain
2(k    1)(k  )ck 1  (k    1)ck 1  (k  )ck  2ck  0
or
(k    1)(2k  2  1)ck 1  (k    2)ck  0 
(k    2)
ck 1   ck
(k    1)(2k  2  1)
The first solution corresponds to   0
The recurrence relation becomes:
(k  2)
ck 1   ck
(k  1)(2k  1)
We have, for various values of k :
k  0, c1  2c0
c1 c0
k  1, c2  
6 3
k  2, c3  0
Then for k  2 , ck  0 , the solution corresponding to   0 is

 1 
y1   ck x k    c0  c1 x  c2 x 2   c0 1  2 x  x 2 
k 0  3 

17
For   12 : the recurrence relation becomes
(k  32 ) (2k  3)
ck 1   ck   ck
(k  2 )(2k  2)
3
2(2k  3)(k  1)
For different values of k :
c0
k  0, c1  
2
c c
k  1, c2  1   0
20 40
c c
k  2, c3   2  0
42 40.42
The solution corresponding to   12 is

 1 
 x  c0  c1 x  c2 x 2  ... c0 x 1  x  x 2 
1 1
y2   ck x
k  12
x 3  ... 
k 0  2 40 40.42 
The general solution is
 1   1 1 1 
y  A 1  2 x  x 2   B x 1  x  x 2  x 3  ...  .
 3   2 40 40.42 
Case 2:
The two roots of the indicial equation are equal
In this case, 1   2   and we obtain only one Frobenius series solution.
Ex (2)
Find the series solution for the following differential equation about x  0 :

x  x  y  (2 x  1) y 
1
2
y0
x
Solution
The coefficient functions of the differential equation are

18
 2x  1 1
P( x)  , Q( x)  
x2  x x( x  x)
2

P( x) and Q ( x) are infinite as x  0 but since

lim xP ( x)  1 and lim x 2Q ( x )  1 ,

x 0 x 0

the two limits are finite, therefore the point x  0 is a regular singular point.
Assume a Frobenius series solution

y   ck x k  , c0  0
k 0

Then
 
y   (k   )ck x k 1 , y   (k  )(k    1)ck x k 2
k 0 k 0

Substitute for y, y, y into the differential equation, we obtain

 
1 
x 2
 x   (k  )(k    1)ck x k 2  (2 x  1) (k  )ck x k 1 
k 0 k 0

x k 0
ck x k   0

or
 

 (k  )(k    1)c x
k 0
k
k 
  (k  )(k    1)ck x k 1
k 0
  
2 (k  )ck x k    (k  )ck x k 1   ck x k 1  0
k 0 k 0 k 0

To get the indicial equation, we equate the coefficient of the lower power of x
( x 1 ) to zero:
(  1)c0  c0  c0  0  ( 2  2  1)c0  0, c0  0
(  1) 2  0    1,1
We have in this case one solution corresponding to   1
Now, equating the coefficients of x k  to zero, we obtain

19
(k  )(k    1)ck  ( k    1)(k  )ck 1  2(k  )ck  (k    1)ck 1  ck 1  0
Or
 (k  )(k    1)  2(k  ) c   (k    1)(k  )  (k    1)  1 c
k k 1
0
(k  )(k    1)c   (k    1)(k  )  (k  )  c  0
k k 1

(k    1)ck
(k  )(k    1)ck  (k  ) 2 ck 1  0  ck 1 
(k  )
k2
For   1 , ck 1  ck and for various values of k , we obtain
k 1
k  0, c1  2c0
3
k  1, c2  c1  3c0
2
4
k  2, c3  c2  4c0
3
5
k  3, c4  c3  5c0
4
The solution corresponding to   1 is

y1   ck x k 1  x  c0  c1 x  c2 x 2  ... c0 x 1  2 x  3 x 2  4 x 3  5 x 4  ...
k 0

It can be shown that, the second solution is given by


y2  y1 ln x   k x k 
k 1


y2  c0 x ln x 1  2 x  3 x  ...   k x k 1 ,
2

k 1

The coefficients k are obtained by substituting y2 and its derivatives into the
differential equation.

20
Case 3:
The roots of the indicial equation are distinct by an integer
Theorem
If x  0 is a regular singular point of the differential equation
y  P ( x) y  Q( x) y  0 , and if 1 and  2 are the two roots of the indicial
equation and if 1   2 is a positive integer, then there is a Frobenius series
solution of the form

y1   ck x k  , 1
c0  0
k 0

Also the second linearly independent solution is given by


y2  my1 ln x   k x k  2

k 0

The constants m and k are obtained by substituting y2 and its derivatives into
the differential equation and equating the coefficients of different powers of
x to zero.

Ex (3)
Find the series solution for the following differential equation about x  0 :
xy  (4  x) y  2 y  0
Solution
The point x  0 is a regular singular point of this differential equation
Assume a Frobenius series solution

y   ck x k  , c0  0
k 0

Then

21
  
y   (k  )ck x k 1 , y   (k  )(k    1)ck x k 2
k 0 k 0

Substitute for y, y, y into the differential equation, we obtain

  
x  (k  )(k    1)ck x k  2
 (4  x) (k  )ck x k 1
 2 ck x k   0
k 0 k 0 k 0

or
   

 (k  )(k    1)ck x k 1  4 (k  )ck x k 1   (k  )ck x k   2 ck x k   0

k 0 k 0 k 0 k 0

To get the indicial equation, we equate the coefficient of the lower power of x
( x 1 ) to zero:
   

 (k  )(k    1)ck x k 1  4 (k  )ck x k 1   (k  )ck x k   2 ck x k   0

k 0 k 0 k 0 k 0

(  1)c0  4c0  0  (  5)c0  0, c0  0

 (  5)  0
Then 1  5 and  2  0 . The roots are distinct and the difference between them
is an integer.
Equating the coefficients of x k  to zero, we obtain
(k    1)(k  )ck 1  4(k    1)ck 1  (k  )ck  2ck  0
(k    1)(k    4)ck 1  (k    2)ck
(k    2)ck
 ck 1 
(k    1)(k    4)
(k  3)ck
For 1  5 , ck 1 
(k  6)(k  1)
For various values of k , we have
c0
k  0, c1 
2
2c c
k  1, c2  1  0
7 7

22
 5c2 5c0
k  2, c3  
8.3 8.7.3
The solution corresponding to 1  5 is

 1 
y1   ck x k 5   c0  c1 x  c2 x 2  ... c0 x5 1  x  x 2 
1 5 3
x  ... 
k 0  2 7 8.7.3 
The second linearly independent solution is given by

y2  my1 ln x   k x k
k 0

The constants m and k are obtained by substituting y2 and its derivatives into
the differential equation and equating the coefficients of different powers of
x to zero.

23
 SHEET#3
Case1
Obtain the series solution of the following differential equations about x  0 .
1. 2 x( x  1) y  3( x  1) y  y  0
2. 2 xy  (2  x) y  2 y  0

3. 2 xy  (1  2 x 2 ) y  4 xy  0
4. 2 xy  (1  2 x) y  5 y  0

5. 9 x 2 y  3 x( x  3) y  (1  4 x) y  0
Case2
Obtain the two linearly independent solutions of the following differential
equations about x  0 .
1. x( x  1) y  (3x  1) y  y  0
2. xy  y  xy  0

3. x 2 y  x( x  1) y  (1  x) y  0
4. x( x  2) y  2( x  1) y  2 y  0
Case3
Obtain series solutions for the following differential equations about x  0 .
1. xy  (3  x) y  2 y  0 2. x(1  x) y  3 y  2 y  0
3. xy  (4  3x) y  3 y  0 4. x( x  3) y  9 y  6 y  0
5. xy  ( x  1) y  2 y  0 6. x(1  x) y  4 xy  2 y  0
7. xy  (1  x) y  3 y  0

24
 CHAPTER II

SPECIAL FUNCTIONS
1. Gamma function
2. Beta function
3. Bessel Functions
4. Legendre Polynomials

25
 1. Gamma function

The Gamma function, denoted by (n) , is defined by the improper integral

(n)   e  x x n1dx (1.1)

0

The integral is convergent for n  0 .

Properties of Gamma function

(1) (1)  1

We have, from definition (1.1),

 

(1)   e x dx   e  x dx  e  x 0  1 .
x 11

0 0

(1)  1 (1.2)

(2) (n  1)  n(n) , (n  0)

We have also, from definition (1.1),

(n  1)   e  x x n dx , on integrating by parts

0

 

(n  1)  e  x x n 0    e  x  nx n1dx  0  n  e  x x n 1dx ,



0 0

where the first term vanishes at the upper limit since the exponential dominates
the power, and at lower term since n  0 . Therefore

26
 (n  1)  n(n) , (n  0) (1.3)

(3) If n is a positive integer, then ( n  1)  n! . By repeated use of property (2)

and remembering that n is integral, we have

(n  1)  n(n)  n(n  1)(n  1)

 n(n  1)(n  2)(n  2)
 n(n  1)(n  2)(n  3)....3.2.1.(1)
 n!(1)  n!.

(n  1)  n! , n is a positive integer (1.4)


 ( n)
e x n1dx 
 ax
(4) (1.5)
0 an

To prove this property, replace x with ax in equation (1.1),

   
 ( n)
(n)   e  ax (ax) n1d (ax)   e  ax x n1a n1adx  a n  e ax x n 1dx   e  ax x n1dx  .
0 0 0 0 an

(5) (n)  2 e  t t 2 n 1dt

2
(1.6)
0

In definition (1.1) substitute t 2 for x : x  t 2 , so that dx  2tdt ; when

x  0, t  0 and when x  , t   , so that we have

 

(n)   e  t (t 2 ) n 1 2tdt  2  e  t t 2 n1dt

2 2

0 0

27
 n 1
 1
1

(6) (n)    ln  dy (1.7)

0 y

1 1 1
Let x  ln   e x  y  e  x  dy  e  x dx ; when y  0, x  ln   and
y y 0
when y  1, x  ln1  0 , so that we have

n 1
1
 1 0 

0  ln y  dy   x  e dx   0 e x dx  (n) .
n 1 x  x n 1

 


(7) (n)(n  1)  (1.8)
sin n

(8) If n is zero or a negative integer, then (n)  

(n  1)
From equation (1.2), we have (n)  , then for
n

(1)
n  0; (0)   ,
0
(0)
n  1; (1)   ,
1
(1)
n  2; (2)   , etc.
2

It is now possible to draw a graph of (n) . Plotting of points combined with

information from various properties leads to a graph as shown in Fig.1.

28
 Fig.1 Graph of Gamma Function ( n)

n ( n) n ( n) n ( n)
1 1.00000 1.55 0.88887 1/5 4.5909
1.05 0.97350 1.60 0.89352 1/4 3.6256
1.10 0.95135 1.65 0.90012 1/3 2.6789
1.15 0.93304 1.70 0.90864 2/5 2.2182
1.20 0.91817 1.75 0.91906 3/5 1.4892
1.25 0.90640 1.80 0.93138 2/3 1.3541
1.30 0.89747 1.85 0.96180 3/4 1.2254
1.35 0.89115 1.90 0.96177 4/5 1.1642
1.40 0.88726 1.95 0.97988
1.45 0.88566 2.00 1.00000
1.50 0.88623
Table1. Some values of ( n)

29
 2.The Beta function

The Beta function denoted by ( p, q) , is defined by the integral

( p, q )   x p 1 (1  x) q 1dx (2.1)
0

This integral is convergent for p, q  0 .

Properties of Beta function

(1) The symmetry property:

( p, q )  (q, p ) (2.2)

In definition (2.1) substitute z for1  x : z  1  x , so that dz  dx ; when

x  0, z  1 and when x  1, z  0 , we have

0 1

( p, q )   1  z  z (dz )   z q 1 1  z  dz  (q, p )
p 1 q 1 p 1

1 0

/ 2

(2) ( p, q )  2  sin 2 p 1  cos 2 q 1 d  (2.3)

0

In definition (1) substitute sin 2  for x : x  sin 2  , so that dx  2sin  cos d  ;

when x  0,   0 and when x  1,    / 2 , we have

30
 / 2

( p, q )   sin (1  sin 2 ) q 1 2sin  cos d 

2 p 2

0
/ 2

  sin  cos 2 q 2 2sin  cos d 

2 p2

0
/ 2

 2  sin 2 p 1  cos 2 q 1 d 
0

 /2

 sin  cos 2 q 1 d   12 ( p, q )
2 p 1
Or
0

 ( p ) ( q )
(3) ( p, q )  (2.4)
( p  q )

To prove this relation, we have from property (5) of Gamma function:

 

( p )  2  e x  x2 2 p 1
(q )  2  e  y y 2 q 1dy
2
dx and
0 0

  

 ( p ) ( q )  2  e x  x2 2 p 1
dx.2  e  y2 2 q 1
dy  4   e  ( x  y ) x 2 p 1 y 2 q 1dxdy
2 2
y
0 0 0 0

To evaluate the last double integral, we change to polar coordinates, where

x  r cos , y  r sin  and the element of area dxdy  rdrd  . The values of r
and  that cover the area of integration are: r from 0 to  and  from 0 to  / 2 .
Therefore

/ 2 

( p )(q )  4   e  r  r cos    r sin  

2 p 1 2 q 1
rdrd 
2

0 0

From property (2) and from property (5) of Gamma function, we get

31
  /2
 ( p ) ( q )
 sin  cos 2 q 1 d   12
2 p 1

0 ( p  q)

Putting p  q  12 in the last relation, we obtain

 /2
( 12 )( 12 )  1
 d     ( 12 )   ( 12 )  
2
1

( 12  12 )
2
0 2 2

Examples

Ex(1)

Evaluate the following integrals:

1  /2

(i)  x (1  x) dx
4 3
(ii)  sin
6
 cos 7 d 
0 0

 /2  /2

 cos d   tan d 
8
(iii) (iv)
0 0

Solution

(i) Compare with the definition of ( p, q )   x p 1 (1  x) q 1dx

0

p  1  4, q  1  3  p  5, q  4 ,
1
(5)(4) 4!3! 1
 x (1  x) dx  (5,4)   
4 3

0 (9) 8! 280

32
  /2
 ( p ) ( q )
 sin  cos 2 q 1 d   12
2 p 1
(ii) Compare with
0 ( p  q )

2 p  1  6,2q  1  7  p  72 , q  4

 /2
( 72 )(4) ( 72 ) ( 72 ) 48 16
0 sin  cos d   ( , 4)  ( 7  4)  3 ( 15 )  3 13 11 9 7 ( 7 )  13.11.9.7  3003
6 7 1 7 1
2 2 2
2 2 2 2 2 2 2

 /2
 ( p ) ( q )
 sin  cos 2 q 1 d   12
2 p 1
(iii) Compare with
0 ( p  q)

2 p  1  0,2q  1  8  p  12 , q  92

 ( ) 
2
 /2
( 12 )( 92 ) 1 7 5 3 1 1
105 105
0      2   
8 2 2 2 2 2
cos d 1
( ,
1 9
) 1

( 12  92 ) (5)
2 2 2 2
32.4! 768

 /2 / 2

  sin    cos  
1/2
 tan d   d 
1/2
(iv)
0 0

2 p  1  12 ,2q  1   12  p  34 , q  14

 /2
( 14 )( 34 ) 1 ( 14 )( 34 ) 1 1

0
tan d   12 ( 14 , 34 )  12
( 4  4 )
1 3
2
(1)
 2 ( 4 )( 34 )


Putting n  14 , in the relation (n)(1  n)  , we get
sin n
  /2

 ( ) ( ) 
1
4
sin 4
 2 ,
3
4 Therefore 
0
tan d   12 2 
2

33
Ex(2)

1
Evaluate 1 t
0
4
dt

Solution

Let t 2  tan   1  t 4  1  tan 2  and

sec 2  sec 2 
2tdt  sec 2  d   dt  d  d
2t 2 tan 

When t  0,   0 and when t  ,    / 2 , therefore


1  /2
1 sec 2   /2  /2

   
1/ 2
0 1  t 4  0 sec2  2 tan    0    0   d
1/ 2
dt d 1
2 cot d 1
2 cos sin

2 p  1  12 , 2q  1   12  p  34 , q  14


1 ( 14 )( 34 ) 1 ( 14 )( 34 ) 1 1 2
0 1  t 4 dt  1 1
2 2 ( ,
1 3
4 4 )  1
4
( 14  34 )
 4
(1)
 4  ( 4 )  ( 3
4 ) 
4

Ex(3)
n
 1
1
1
Show that  x  ln  dx 
m
(n  1) and hence deduce that
0  x (m  1) n1
1/ 2 1/ 2
1
 1 1
 1
(i)   ln  dx   / 2 (ii)   ln  dx  
0 x 0 x

Solution

34
 1 1 1
Let z  ln   e z  x  e  z  dx  e  z dz ; when x  0, z  ln   and
x x 0

when x  1, z  ln1  0 , so that we have

n
1
 1 0 

0 x  ln x  dx    e  z  e dz   0 e
m z m n z  m 1 z n
z dz


 ( n)
e x n1dx  with a  m  1 and replace n  1
 ax
But using property (IV),
0 an
by n , we get

(n  1)
n
1
 1
0 x  ln x  dx   m  1n1
m

To find the value of the integral in (i), compare with the result just obtained,

m  0, n  12

( 12  1)
1/2
1
 1
0  x 
ln dx   ( 32 )  12 ( 12 )  12  .
1
1 1
2

And for the integral in (ii), m  0, n   12 ,

1/ 2
1
 1 ( 12  1)
0  ln x  dx   ( 12 )   .
1
 12 1

35
 SHEET #4

From exercise 1 to 15, evaluate the following integrals

1  /2  /2
d
 x (1  x) dx  sin  cos d  
7 8 4 6
1. 2. 3.
0 0 0 cos 

1
dx 1 2
x 2 dx
  x (1  x )dx 
3
4. 5. 6.
0 1  x3 0 0 2x

2 3 

 sin d  x 27  x 3 dx 9.  x 4e  x dx
8 3
7. 8.
0 0 0

  

10.  x 3e 2 x dx  xe  x dx 2
3 x 2
3
11. 12. dx
0 0 0

3
 1
dx  11

13.  e 3 y
dy 14.  15.   ln  dx
0 0  ln x 0 x

16. Show that ( p, q).( p  q, m)  (q, m).(q  m, p)

1
(1) n n!
17. Show that  x ln xdx  m n

 m  1
n 1
0

1.3.5....(2n  1)
18. Show that (n  12 )  
2n

1
dx
19. Show that   2
0 x ln(1 / x)

36
 3. Bessel Functions

The standard form of Bessel differential equation is given by

x 2 y  xy  ( x 2   2 ) y  0, (3.1)

where  is a nonnegative real number.

1 x2  2
Here P( x)  , Q( x)  .
x x2

1 x2  2
Since lim P ( x )  lim   and lim Q( x)  lim   and
x 0 x 0
x x 0 x 0
x2

2 x 
2 2
1
lim
x 0
xP ( x )  lim
x 0
x  1   and lim
x 0
x 2
Q ( x )  lim
x 0
x 2
  2
 ,
x x

therefore the point x  0 is a regular singular point. A solution is taken in

accordance with the Frobenius method to be


y   ck x k  , c0  0 (3.2)
k 0

The indicial equation gives 1   and  2   ,   0.

If  is not an integer the first solution corresponding to 1   is denoted by

J  ( x) and is given by


(1) k ( x / 2) 2 k 
J  ( x)   (3.3)
k  0 k ! (  k  1)

37
where J  ( x) is called Bessel function of the first kind of order  . For
 2   , we get the second solution as


(1) k ( x / 2) 2 k 
J  ( x)   (3.4)
k  0 k ! (   k  1)

It can be shown that, for  is not integer, the two functions J  ( x) and J  ( x) are
linearly independent, then a general solution of the Bessel differential equation
is

y  AJ  ( x)  BJ  ( x) (3.5)

If  is an integer,   n : the first solution in this case is


(1) k ( x / 2) 2 k  n
J n ( x)   (3.6)
k  0 k ! ( n  k  1)

For the other root   n , we have


(1) k ( x / 2) 2 k  n
J  n ( x)   (3.7)
k  0 k ! ( k  n  1)

Now we show that, for n the two solutions (3.6) and (3.7) are linearly
dependent

Ex (1)

If n is an integer, show that J  n ( x)  (1) n J n ( x) .

38
 1
Since for k  0,1,2,..., n  1,  0 that is the first n coefficients
(k  n  1)
become zero in the series (3.7), i.e.


(1) k ( x / 2) 2 k  n
J  n ( x)   ,
k  n k ! ( k  n  1)

by putting k  m  n , we get


(1) m n ( x / 2) 2 m n 
(1) m ( x / 2) 2 m  n
J  n ( x)    (1) n 
m  0 ( m  n )! ( m  1) m 0 m!(m  n)!

Therefore

J  n ( x)  (1) n J n ( x) (3.8)

Relation (3.8) means that J n ( x) and J  n ( x) are not linearly independent and
hence there exist another solution is denoted by Yn ( x) . Thus the general
solution of the Bessel differential equation is

yG  AJ n ( x)  BYn ( x) (3.9)

Yn ( x) is called Bessel function of the second kind of order n .

Parametric Bessel differential equation

The Bessel differential equation is

x 2 y  xy  ( 2 x 2   2 ) y  0, (3.10)

39
Has the same solution as before, but with replacing x by x

1
J0 (x)
J1 (x)

0.5

-0.5
0 1 2 3 4 5 6 7 8 9 10
x

Fig.2 Graphs of J 0 ( x ) and J1 ( x) Functions

Recurrence Relations for Bessel functions

d n d n
(I)  x J n ( x)   x n J n1 ( x) (II)  x J n ( x)    x  n J n 1 ( x)
dx dx

n n
(III) J n ( x)  J n 1 ( x)  J n ( x) (IV) J n ( x)  J n ( x)  J n1 ( x)
x x

40
 1 2n
(V) J n ( x)   J n1 ( x)  J n1 ( x) (VI) J n 1 ( x)  J n1 ( x)  J n ( x)
2 x

Proof

To prove I, we have

d  (1) k x 2 k  2 n 
(1) k 2(k  n) x 2 k  2 n 1
 x J n ( x)   dx 
d n

dx k 0 2
2 k n
k !(  n  1) k 0 22 k  n k !(  n  1)

(1) k (k  n) x 2 k  n1 
(1) k ( x / 2) 2 k  n1
 xn   x n

k 0 2
2 k  n 1
k !(k  n)(  n) k 0 k !(  n)
 x n J n1 ( x).

That is

 x J n ( x)   x n J n 1 ( x).
d n
(3.11)
dx

Similarly

 x J n ( x)   x n J n 1 ( x).
d n
(3.12)
dx

Differentiating the left sides of (3.11) and (3.12) and divide the results by the
factors x n and x  n respectively, we deduce that

n
J n ( x)  J n 1 ( x)  J n ( x) (3.13)
x

n
J n ( x)  J n ( x)  J n1 ( x) (3.14)
x

41
The sum and the difference of the above two equations give, respectively

1
J n ( x)   J n1 ( x)  J n1 ( x) (3.15)
2

2n
J n 1 ( x)  J n1 ( x)  J n ( x) (3.16)
x

Important relations

Letting    12 in (3.3) , we get respectively

2 2
J1/2 ( x)  sin x, J 1/2 ( x)  cos x, (3.17)
x x

Ex (2)

2
2
Show that J1/2 ( x)  J 21/2 ( x) 
x

sin 2 x  cos 2 x   sin 2 x  cos 2 x   .

2 2 2 2
2
We have J1/2 ( x)  J 21/2 ( x) 
x x x x

Ex (3)

Show that J n ( x)   1 J n ( x)

n

(1) k ( x / 2) 2 k  n

Replacing x by  x in the series, J n ( x)   we get
k  0 k ! ( n  k  1)


(1) k ( x / 2) 2 k  n  (1) k (1) 2 k  n ( x / 2) 2 k  n 
(1) k ( x / 2) 2 k  n
J n ( x)     (1) n 
k  0 k ! ( n  k  1) k 0 k !(n  k  1) k  0 k ! ( n  k  1)

42
That is J n ( x)  (1) n J n ( x)

which means that

J n ( x) is an odd function for n  1,3,5,...

J n ( x) is an even function for n  0,2,4,...

We note that J 0 (0)  1, J n (0)  0, n  1 .

Ex (4)

Show that
2  sin x  2  cos x 
J 3/2 ( x)    cos x  , J 3/2 ( x)     sin x  ,
x  x  x  x 

2n
Letting n  12 in the recurrence relation (3.16), J n 1 ( x)  J n 1 ( x)  J n ( x) :
x

2. 12 1
J 1/2 ( x)  J 3/2 ( x)  J1/2 ( x)  J 3/2 ( x)   J 1/2 ( x)  J1/2 ( x) ,
x x

Substitute for J 1/2 ( x) and J1/2 ( x) from (3.17), we get

2 1 2 2  sin x 
J 3/2 ( x)   cos x  sin x    cos x 
x x x x  x 

Again letting n   12 in the recurrence relation (3.16), we get the formula for
J 3/2 ( x) .

Integrals involving Bessel functions

Integrating (3.9) and (3.10), we get

x J n1 ( x)dx  x n J n ( x)  C
n
(3.18)

43
 x J n1 ( x)dx   x  n J n ( x)  C
n
(3.19)

Ex (5)

Evaluate the following integrals

(a)  x 4 J1 ( x)dx
(b)  J 3 ( x)dx

Solution

(a) I   x 4 J1 ( x)dx   x 2  x 2 J1 dx , using integration by parts

u  x2 , dv  x 2 J1dx
du  2 xdx, v  x2 J 2

I  x 4 J 2  2  x 3 J 2 dx  x 4 J 2  2 x 3 J 3  C

(b) I   J 3 ( x)dx   x 2 x 2 J 3 ( x)dx ), using integration by parts

u  x2 , dv  x 2 J 3dx
du  2 xdx, v   x 2 J 2

I   x 2 x 2 J 2    x 2 J 2 2 xdx   J 2  2  x 1 J 2 dx
  J 2  2 J1  C

2n
Another solution, using the recurrence relation J n 1 ( x)  J n ( x)  J n 1 ( x) ,
x

4
Let n  2  J 3 ( x)  J 2 ( x)  J1 ( x)  4 x 1 J 2 ( x)  J1 ( x) , then
x

44
I    4 x 1 J 2 ( x)  J1 ( x) dx    4 x 1 J 2 ( x)  x 0 J1 ( x) dx
 4 x 1 J 1 ( x )  J 0 ( x )  C .

Ex (6)

Evaluate the following integrals

x x

(a)  J1 (t )dt (b)  tJ 0 (t )dt

0 0

Solution
x x

(a)  J1 (t )dt   t 0 J1 (t )dt   J 0 (t ) 0   J 0 ( x)  J 0 (0)  1  J 0 ( x)

x

0 0
x

(b)  tJ 0 (t )dt  tJ1 (t ) 0  xJ1 ( x).

x

Ex (7)

Solve the following equations in terms of Bessel functions

d  2 dy 
  x 2  14  y  0, y ( / 2)  0, y( / 2)  1
dy
(a)  x   x
dx  dx  dx

(b) xy  y  5 xy  0

Solution

d2y
  x 2  14  y  0,
dy dy
(a) x 2
2
 2 x  x
dx dx dx

x 2 y  xy   x 2  14  y  0

This is a Bessel equation with  2  14     12 , that is  is not an integer, then

y  AJ1/2 ( x)  BJ 1/2 ( x)

45
 2 2
yA sin x  B cos x,
x x

From the condition y (  / 2)  0 , we have

2 2
0 A sin  / 2  B cos  / 2  A  0,
x x

2 2 cos x
yB cos x  B ,
x  x

cos x
 x sin x 
2 2 x
y  B
 x

From the condition y( / 2)  1, we get

2  /2 2 
1  B  1  B  B  , hence
 /2  2


y cos x
2x

(b) xy  y  5 xy  0 , multiply the equation by x , we get

 
2
x 2 y  xy  5 x 2 y  0  x 2 y  xy  5 x y,

which is a Bessel differential equation with   0 , that is  is an integer, then

yG  AJ 0 ( 5 x)  BY0 ( 5 x)

46
Ex (8)

Evaluate the following integrals  J 2 ( x)dx

Solution

1
I   J 2 ( x)dx , from the recurrence relation (3.15), J n ( x)   J n1 ( x)  J n1 ( x)
2

We have J n 1 ( x)  2 J n ( x)  J n 1 ( x) ,

Let n  1 , J 2 ( x)  2 J1( x)  J 0 ( x) , then

I    2 J1( x)  J 0 ( x) dx  2 J1 ( x)   J 0 ( x)dx

47
 SHEET #5

1. Show that y  x n J n ( x) is a solution of the differential equation

d2y dy
x 2  (2n  1)  xy  0
dx dx

2. Show that y  x  n /2 J n (2 x ) satisfies: xy  (n  1) y  y  0

3. Show that

2  3  x2 3 
(a) J 5/2 ( x)   cos x  sin x 
x  x 2
x 

2  3  x2 3 
(b) J 5/2 ( x)   sin x  cos x 
x  x 2 x 

4. Express J 3 ( x) in terms of J 0 ( x) and J1 ( x)

x
3
5. Express J 4 ( x)dx in terms of J 0 ( x) and J1 ( x)

6. Show that  x 1 J 4 ( x)dx   x 1 J 3 ( x)  2 x 2 J 2 ( x)  C

7. Evaluate the following integrals: (a)  x 3 J 0 ( x)dx , (b)  x 3 J 3 ( x)dx

 /2

8. Show that 
0
xJ1/2 ( x)dx  1

9. Solve the following equations in terms of Bessel functions:

(a) x 2 y  xy  25 x 2 y  19 y

48
 (b) 4 x 2 y  4 xy  4 x 2 y  y, y ( / 2)  0, y( / 2)  1

(c) x 2 y  xy  13 12 x 2  13  y  0,

( xy )   x 2  54  y  0, y ( / 2)  0, y( / 2)  1
d
(d) x 2 y 
dx

10. Evaluate the following integral: (a)  x 2 J 2 ( x)dx , (b)  x 2 J 2 ( x)dx .

49
 4.Legendre Polynomials

The Legendre differential equation is defined by the equation

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

2
(4.1)

where n is a nonnegative integer number. Here the point x  0 is an ordinary

point. The first solution of this differential equation is a polynomial of degree
n . This polynomial is denoted by Pn ( x) and is given by

N
(1) k (2n  2k )!
Pn ( x)   x n2 k (4.2)
k  0 2 k !( n  k )!( n  2k )!
n

where N  n / 2 when n is even and N  n  1 / 2 when n is odd. Pn ( x) known

as the Legendre polynomial of degree n . The first few Legendre polynomials
are

P0 ( x)  1
P1 ( x)  x
P2 ( x)  12 (3 x 2  2)
P3 ( x)  12 (5 x 3  3x)
P4 ( x)  18 (35 x 4  30 x 2  3)

These polynomials are plotted in Fig.2 for values of x between 1 .

50
 1

0.8

0.6

0.4

0.2

-0.2 P0 (x)
P1 (x)
-0.4
P2 (x)
-0.6 P3 (x)

-0.8 P4 (x)

-1
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
x

Fig.3 Graphs of some Legendre polynomials

Generating function for Legendre polnomails

The function 1  2xt  t 2 

1/2
is called the generating function for Legendre

polynomails. In fact we have

1  2 xt  t 
1/2
2
  t n Pn ( x), x  1, t  1 (4.3)
n0

Differential form for Legendre polynamials (Rodrigue’s formula)

51
One of the fundamental identities involving Legendre polynomails is
Rodrigue’s formula. This formula is given by

1 dn 2
n 
x  1
n
Pn ( x)  n (4.4)
2 n! dx

Recurrence Relations for Legendre polynamials

I. (n  1) Pn1 ( x)  (2n  1) xPn ( x)  nPn1 ( x)  0

II. nPn ( x)  xPn( x)  Pn1 ( x)
III. (2n  1) Pn ( x)  Pn1 ( x)  Pn1 ( x)

IV. (n  1) Pn ( x)  Pn1 ( x)  xPn( x)

V. (1  x 2 ) Pn( x)  n  Pn1 ( x)  xPn ( x) 

VI. (1  x 2 ) Pn( x)  (n  1)  xPn ( x)  Pn 1 ( x) 

Proof of the recurrence relation I:

From the generating functuin

1  2 xt  t 
1/2
2
  t n Pn ( x)
n0

Differentiating withrespect to t , we get


 12 1  2 xt  t 2 
3/2
(2 x  2t )   nt n1 Pn ( x)
n0

52
 Multiplying both sides by 1  2 xt  t 2  , we obtain

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

Equating the coefficients of t n in both sides we get

xPn ( x)  Pn 1 ( x)  (n  1) Pn 1 ( x)  2nxPn ( x)  (n  1) Pn 1 ( x) , or

(n  1) Pn1 ( x)  (2n  1) xPn ( x)  nPn1 ( x)  0 .

Orthogonal properties of Legendre polynamials

The most important property of Legendre polynamials is the orthogonality

property. This given by:

 0 if m  n
1

 Pm ( x)Pn ( x)dx   2 if m  n (4.5)
1  2n  1

Proof of the case m  n :

Since Pm ( x) and Pn ( x) are solutions of the Legendre differential equation, then

1  x 2  Pm   m(m  1) Pm  0,

1  x 2  Pn  n(n  1) Pn  0,

53
Multiplying the 1st equation by Pn ( x) and the 2nd by Pm ( x) and subtracting, we
obtain

1  x 2  Pm  Pn  1  x 2  Pn Pm   m(m  1)  n(n  1)  Pm Pn  0,

Integrating both sides with respect to x from 1 to 1

1
   
  1  x  Pm  Pn  1  x  Pn Pm dx
2 2

1  
1

  n(n  1)  m(m  1)   Pm Pn dx
1

Use integration by parts in the first integral:

1
   
  1  x  Pm  Pn  1  x  Pn Pm dx
2 2

1  
1 1

 1  x  Pm Pn 1   1  x 2  Pm Pndx  1  x 2  PnPm 1   1  x 2  PnPmdx  0

1 1
2

1 1

  n(n  1)  m(m  1)  Pm Pn dx  0

1

Since m  n , we get  P P dx  0 .
1
m n

Examples

Ex (1)

Using the generating function, 1  2 xt  t 2 
1/2
  t n Pn ( x)
n 0

54
 n(n  1)
(i) Pn (1)  1 (ii) Pn(1)  (iii) Pn ( x)  (1) n Pn ( x)
2

Solution

(i) Let x  1 in generating function,



1  2t  t 2    (1  t ) 2 
1/2 1/2
 (1  t ) 1   t n Pn (1)
n 0

Using the binomail expansion


(1  t ) 1  1  t  t 2  ...  t n  ...   t n 
n 0

 

 t n   t n Pn (1)
n 0 n 0

Equating the coefficients of t n in both sides, we obtain Pn (1)  1

(ii) Pn ( x) satisfies Legendre differential equation

1  x  P( x)  2 xP( x)  n(n  1) P ( x)  0,

2
n n n

Letting x  1 , we get 2 Pn(1)  n(n  1) Pn (1)  0 and since ,from (i), Pn (1)  1

n(n  1)
2 Pn(1)  n(n  1)  0  Pn(1) 
2

(iii) Replacing x by  x in the generating function, we get



1  2 xt  t 
1/2
2
  t n Pn ( x)
n 0

Now replacing t by t , we get

55
 

1  2 xt  t     t  Pn ( x) or
2 1/2 n

n 0

 

 t n Pn ( x)    1 t n Pn ( x)
n

n 0 n 0

Equating the coefficients of t n in both sides, we obtain Pn ( x)  (1) n Pn ( x)

This means if n is even, then Pn ( x) is even function and if if n is odd, then

Pn ( x) is odd function.

Ex (2)

Show that
1
2n
(i)  xP ( x) P
1
n n 1
( x)dx 
4n 2  1

 0 if m  n
1

(ii)  1  x  Pm ( x) Pn( x)dx   2n(n  1)
2

1  2n  1 if m  n

1
2n(n  1)
x P P dx 
2
(iii) n 1 n 1
1 (2n  3)(4n 2  1)

Solution

(i) From recurrence relation I, we have

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

56
 1 1
 n 1 n 
1 n n1
xP ( x ) P ( x ) dx  1  2n  1 Pn1 ( x)  2n  1 Pn1 ( x)  Pn1 ( x)dx
n 1 1 n 1

2n  1 1
 n1 n1
P ( x )P ( x ) dx 
2n  1 1
 Pn1 ( x)Pn1 ( x)dx
n 2 n 2 2n
   2
2n  1 2(n  1)  1 2n  1 2n  1 4n  1
1

(ii) I   1  x 2  Pm ( x) Pn( x)dx , integrating by parts:

1

u  1  x 2  Pn( x), dv  Pm ( x)dx

du  1  x 2  Pn( x)  2 xPn( x)  dx, v  Pm ( x)

I  1  x  Pn( x) Pm ( x) 1   Pm ( x) 1  x 2  Pn( x)  2 xPn( x)  dx

1
2

1

The first term to the right vanishes at both sides, and since Pn ( x) satisfies
Legendre differential equation, we get

1  x  P( x)  2 xP( x)  n(n  1) P ( x) , then

2
n n n

I  n(n  1)  Pm ( x) Pn ( x)dx
1

For m  n , I  0

2n 2n(n  1)
For m  n , I  (n  1).  .
2n  1 2n  1
1 1

(iii) I   x Pn 1 Pn1dx    xPn 1  xPn 1  dx

2

1 1

From recurrence relation I, we have

57
 n 1 n
xPn ( x)  Pn1 ( x)  Pn1 ( x) ,
2n  1 2n  1

replacing n once by n  1 and a second time by n  1 , we get

n2 n 1
xPn 1 ( x)  Pn  2 ( x)  Pn ( x)
2n  3 2n  3

n n 1
xPn 1 ( x)  Pn ( x)  Pn 2 ( x) , therefore
2n  1 2n  1
1
 n2 n  1  n n 1 
I   Pn 2  Pn  Pn  Pn 2  dx
1  2n  3 2n  3  2n  1 2n  1 
n 1 n 1
n 1 n 2 2n(n  1)

2n  3 2n  1 1
 Pn2 dx  
2n  3 2n  1 2n  1 (2n  3)(4n 2  1)

Ex (3)

Evaluate the following integrals

1
x 2 dx 1
P1 ( x) P5 ( x)dx
(i) 
1 5  4x
(ii) 
1 2  2x

Solution
1
x 2 dx
(i) I  
1 5  4x

Letting t  12 in the generating function, we get

 
1 1
  x    
1/2 1/2
5
4 P
n n
( x ) or 5  4 x  P ( x)
n 1 n
n 0 2 n 0 2

Also , since P2 ( x)  12 (3 x 2  2), P0 ( x)  1  x 2  13 (2 P2  P0 ) , then

58
 1 
1
I  13  (2 P2  P0 ) P dx
n 1 n
1 n 0 2

From the orthogonality relations, all terms in the integral vanish except for n  0
and n  2 , we obtain

 1 2 1 2  11
I  13  2 21  01 
 2 2.2  1 2 0  1  30
1
P1 ( x) P5 ( x)dx
(ii) I  
1 2  2x

Letting t  1 in the generating function, we get


1
 2  2x   Pn ( x) also
1/2

2  2 x n 0

P1 ( x) P5 ( x)  xP5 ( x) and from recurrence relation I, xP5  116 P6  115 P4 , we get

1  1

 P62  115 P42  dx 

238
I    116 P6  115 P4   Pn dx   6
11 .
1 n 0 16 1287

59
 SHEET #6

(1) If P0 ( x)  1 and P1 ( x)  x , use recurrence relation I, find P2 ( x) and P3 ( x) .

(2)In Legendre differential equation, use the substitution x  cos  to show that

d2y dy 1 d  dy 
(a)  cot   n(n  1) y  0 , (b)  sin    n(n  1) y  0 .
d 2
d sin  d   d 

(1  x 2 ) Pn( x)
1

(3) Using Legendre differential equation, show that  Pn ( x)dx  .

x n(n  1)

(4) Using generating function or otherwise, show that:

n(n  1)
(a) Pn (1)  (1) n (b) Pn(1)  (1) n1
2

(2n)!
(c) P2 n (0)  (1) n (d) P2 n1 (0)  0 .
2n
2 (n!) 2

(5) Show that g ( x, t )  1  2 xt  t 2 

1/2
satifies:

g g
(a) 1  2 xt  t 2   (x  t)g (b) 1  2 xt  t 2   tg
t x

(6) If Pn ( x) is Legendre polynomial of degree n , and if a is such that Pn (a)  0 ,

i.e. a is a root of Pn ( x)  0 , show that Pn 1 (a) and Pn1 (a) are of opposite

signs. (Hint: use recurrence relation I).

60
 1
n(n  1)
(7) Show that  x 2 Pn 1 Pn 1dx  .
0 (2n  3)(4n 2  1)
1
2t n
 1  2 xt  t 
1/2
(8) If n is a positive integer, show that 2
Pn ( x)dx  .
1 2n  1

(9)Evaluate the following integrals:

1 1 1
(1  x 3 )dx
 xP  x P ( x)dx (c) 
2 3
(a) ( x)dx (b)
1 (1  x )
n 4 3/2
1 1

(10) Show that

1
 0 if n  1
(i)  nxP ( x ) dx  
1 2 / 3 if n  1
1

(ii)  P ( x ) P
1
n n 1
( x)dx  2, n  0,1, 2,...

1
2n
(iii)
1
 xP ( x) P( x)dx  2n  1 ,
n n
n  0,1,2,...

61
 CHAPTER III

Boundary Value Problems

in Partial Differential Equations

2. Heat Equation
(a) Ends with fixed temperature
(b) Insulated ends
3. Wave Equation

62
 1. Heat Equation

Mathematical models of various phenomena may lead to partial differential

equations because often more than one independent variable is involved. A
partial differential equation is an equation which contains one or more partial
derivatives. In physical applications, the partial differential equation
associated by other conditions which may:
Initial conditions - specifying the state of the system at some initial time
and or
Boundary conditions - specifying the state of the system at ends of an
interval over which the solution is to be defined.

Here we study a boundary value problem which models the temperature

distribution in a straight, thin rod under simple circumstances.
Let u ( x, t ) represents the temperature at location x and timet . Using the law
of conservation of heat energy, we can show that the temperature satisfies the
partial differential equation:

u 2  u
2

a (3.1)
t x 2

where a 2 is a constant called thermal diffusivity . Equation (3.1) is called heat

conduction equation or simply heat equation.

Let the rod initially heated with temperature f ( x ) , that is

63
 u ( x,0)  f ( x), 0 x (3.2)

where  is the length of the rod.

For boundary conditions, we consider two cases.

Case I

The ends of the rod are kept at constant temperatures,

u (0, t )  c1 , u (, t )  c2 t0 (3.3)

Case II

The ends of the rod are insulated,

 
u (0, t )  0, u ( , t )  0 t0 (3.4)
x x

Examples

Ex (1)

Find the temperature of a thin rod of length  , if the ends of the rod are kept at
zero temperature. The initial temperature is given by

1, 0  x  12 
f ( x)  
0, 2   x  
1

Take a 2  1.

64
Solution

The boundary value problem is

u  2u
 , (0  x  , t  0)
t x 2

u (0, t )  0, u ( , t )  0 t 0

1, 0  x  12 
u ( x,0)  f ( x)  
0, 2   x  
1

We will apply the method of separation of variables.

We seek a solution of the form
u ( x, t )  X ( x)T (t )

u  2u
We have 
 XT ,  X T
t x 2
Substitute in differential equation, we get
XT   X T
T  X 
or 
T X
Since x and t are independent, both sides of this equation must equal the same
constant.
 0
T  X   2
  C    (ve)
T X  2
 (ve)
For C   2 , we have
X    2 X  0 and T    2T  0

65
The condition that u (0, t )  0 implies that X (0)T (t )  0 for t  0 and hence
X (0)  0 ,
assuming that T (t ) is not identically zero. Similarly, u (, t )  0  X ()T (t )
implies that
X ()  0 .

The differential equation for X is X    2 X  0 with X (0)  X ()  0

The general solution of this equation is
X  C1 cos x  C2 sin x
X (0)  0  0  C1
X ()  0  0  C2 sin   0, C2  0  sin   0 , therefore
  n, n  1, 2,3,... Corresponding to each positive integer n , we a solution
for X :
nx
X n   n sin

Now, the differential equation for T is T    2T  0 , its solution is T  C3e  t .
2

For each positive integer n , a solution for T :

Tn  n e  ( n /  ) t
2

Superposition solutions, we get


nx  ( n /  ) t
u ( x, t )   bn sin
2
e
n 1 
Applying the initial condition:

nx
f ( x)   bn sin
n 1 
This is a Fourier sine series, therefore

66
 2 nx
bn 
0
 f ( x)sin

dx

2   /2 nx 
nx  2  /2 nx
   1.sin dx   0.sin dx    sin dx
0   /2    0 
 /2
2  nx  2  n 
   cos    cos  1 .
  n  0 n  2 
thus

2  n  nx  ( n /  ) t
u ( x, t )    cos  1 sin
2
e .
n 1 n  2  
Note that the cases C   2 and C  0 implies the trivial solution ( u ( x, t )  0 ).

Ex (2)

Solve the problem in Ex (1) but with    and f ( x)  5sin 3 x  2sin 7 x .

Solution

As in example (1) we have


u ( x, t )   bn sin nx e  ( n /  ) t
2

n 1

Applying the initial condition, u ( x,0)  5sin 3x  2sin 7 x


5sin 3 x  2sin 7 x   bn sin nx  b3 sin 3x  b7 sin 7 x
n 1

Equating coefficients, we have

b3  5, b7  2 and bn  0 for all n except n  3 and n  7 . Therefore the

solution is u ( x, t )  5sin 3 x e 9t  2sin 7 x e 49 t .

67
Ex (3)

Find the temperature of a thin rod of length  , if the ends of the rod are
insulated. The initial temperature is given by

 x, 0  x  12 
f ( x)  
 0, 2   x  
1

Take a 2  1.

Solution

The boundary value problem is

u  2u
 , (0  x  , t  0)
t x 2

 
u (0, t )  0, u ( , t )  0 t 0
x x

 x, 0  x  12 
u ( x,0)  f ( x)  
 0, 2   x  
1

We will apply the method of separation of variables.

We seek a solution of the form
u ( x, t )  X ( x)T (t )

u  2u
We have  XT ,  X T
t x 2
Substitute in differential equation, we get

68
 XT   X T
T  X 
or 
T X
Since x and t are independent, both sides of this equation must equal the same
constant.
 0
T  X   2
  C    (ve)
T X  2 (ve)

For C   2 , we have
X    2 X  0 and T    2T  0

The condition that u (0, t )  0 implies that X (0)T (t )  0 for t  0 and hence
x
X (0)  0 ,

assuming that T (t ) is not identically zero. Similarly, u (, t )  0  X ()T (t )
x
implies that
X ()  0 .

The differential equation for X is X    2 X  0 with X (0)  X ()  0

The general solution of this equation is
X  C1 cos x  C2 sin x

X     C1 sin x  C2 cos x 

X (0)  0  0  C2 ,   0  0  C2
X ()  0  0  C1 sin , C1  0,   0  sin   0 , therefore
  n, n  1, 2,3,... Corresponding to each positive integer n , we a solution
for X :

69
 nx
X n   n cos , n  1,2,3,...

Now, the differential equation for T is T    2T  0 , its solution is T  C3e t .
2

For each positive integer n , a solution forT :

Tn  n e  ( n /  ) t
2

For C  0 , the differential equation for X is X   0 with X (0)  X ()  0

The solution is X  C1 x  C2 , X   C1
The conditions give C1  0 , then X  C2 , combine this solution with the above
X n , we have
nx
X n   n cos , n  0,1, 2,3,...

Superposition solutions, we get

nx  ( n /  ) t
u ( x, t )   an cos
2
e
n 0 

nx  ( n /  ) t
 a0   an cos
2
e
n 1 
Applying the initial condition:

nx
f ( x)  a0   an cos
n 1 
This is a Fourier cosine series, therefore
1 1  /2 1 1 2  /2 
a0  
0
f ( x )dx 
 0
 xdx 

2 x 0  8
2 nx
an 
0
 f ( x)cos

dx

2   /2 nx 
nx  2  /2 nx
   x.cos dx   0.cos dx    x cos dx
0   /2    0 

70
Use integration by parts:
nx
ux dv  cos dx

 nx
du  dx v  sin
n 

2    /2 nx 
 /2
nx
an   x sin
  n  0

n 0
 sin  dx 

2  2 nx 
 /2
n   
2

an   sin    cos 
  2n 2  n   0 
 n 2  n 
an  sin  2 2  cos  1
n 2 n  2 
Thus

    n 2  n   nx  ( n /  ) t
u ( x, t )     sin  2 2  cos  1  cos
2
e .
8 n1  n 2 n  2  

Note that the case C   2 implies the trivial solution ( u ( x, t )  0 ).

Ex (4)

Solve the problem in Ex (3) but with    and f ( x)  cos3 x .

Solution

As in example (3) we have


nx  ( n /  ) t
u ( x, t )  a0   an cos
2
e
n 1 
Applying the initial condition, u ( x,0)  cos3 x

71
 
cos3 x   an cos nx  a3 cos3 x
n 0

Equating coefficients, we have

a3  1, and an  0 for all n except n  3 . Therefore the solution is

u ( x, t )  cos3x e 9t .

72
 2. Wave Equation

The wave equation is an important second order linear partial differential

equation for the description of waves – as they occur in physics – such as
sound waves, light waves and water waves. It arises in the fields like acoustics,
electromagnetics and fluid mechanics.
The simplest system to visualize and describe are waves on a stretched elastic
string of length  . The differential equation which models this system in a
simplified case is

 2u 2  u
2

c , (3.5)
t 2 x 2

u ( x, t )

x x

where u ( x, t ) represents the vertical displacement of the string at location x and

time t . c 2 is the speed squared.

The motion of the string will be determined by both the initial position and
initial velocity of the string. Therefore we must specify initial conditions

u ( x,0)  f ( x), 0  x   (initial displacememt) (3.6)

73
 u
( x,0)  g ( x), 0  x   (initial velocity) (3.7)
t

In addition, we must include into the model the information that the ends of the
string are fixed. Certainly the motion would be different if one end were free.
Thus we must state boundary conditions at the ends of the string.

u (0, t )  u (, t )  0, t0 (3.8)

Note: when the string starts vibrating from rest, we have

u
( x,0)  0, 0  x   (initial velocity). (3.9)
t

Examples

Ex (1)

Find the displacement of a vibrating string of length  , if both ends are fixed.
5
The string starts vibrating with velocity g ( x)  5sin x and initial

displacement f ( x)  0 . Take a 2  1.

Solution

The boundary value problem is

 2 u  2u
 , (0  x  , t  0)
t 2 x 2

74
B.C: u (0, t )  0, u ( , t )  0 t 0

I.C: Initial displacement u ( x,0)  0

 5
Initial velocity u ( x,0)  5sin x
t 

We will apply the method of separation of variables.

We seek a solution of the form
u ( x, t )  X ( x)T (t )

 2u  2u
We have 2  XT ,  X T
t x 2
Substitute in differential equation, we get
XT   X T
T  X 
or 
T X
Since x and t are independent, both sides of this equation must equal the same
constant.
 0
T  X   2
  C    (ve)
T X  2
 (ve)
For C   2 , we have
X    2 X  0 and T    2T  0
The condition that u (0, t )  0 implies that X (0)T (t )  0 for t  0 and hence
X (0)  0 ,

75
assuming that T (t ) is not identically zero. Similarly, u (, t )  0  X ()T (t )
implies that
X ()  0 .

The differential equation for X is X    2 X  0 with X (0)  X ()  0

The general solution of this equation is
X  C1 cos x  C2 sin x
X (0)  0  0  C1
X ()  0  0  C2 sin   0, C2  0  sin   0 , therefore
  n, n  1, 2,3,... Corresponding to each positive integer n , we a solution
for X :
nx
X n   n sin

Now, the differential equation for T is T    2T  0 , with T (0)  0 ; its solution is
T  C3 cos t  C4 sin t .
T (0)  0  0  C3 , therefore For each positive integer n , a solution for T :
n
Tn  n sin t

Superposition solutions, we get

nx n
u ( x, t )   bn sin sin t
n 1  
 
n nx n
u ( x, t )   bn sin cos t
t n 1   
Applying the condition of initial velocity:
5 
n nx 5 5x
5sin x   bn sin  b5 sin
 n 1    

76
Equating coefficients, we have

b5  , and bn  0 for all n except n  5 . Therefore the solution is

 5x 5
u ( x, t )  sin sin t
  
Note that the cases C   2 and C  0 implies the trivial solution ( u ( x, t )  0 ).

Ex (2)

Solve the problem in Ex (1) but with the initial displacement

 x, 0  x  12 
f ( x)  
 0,
1
2   x  ,

and the string starts vibrating from rest.

Solution

The boundary value problem is

 2 u  2u
 , (0  x  , t  0)
t 2 x 2

B.C: u (0, t )  0, u ( , t )  0 t 0

 x, 0  x  12 
I.C: Initial displacement u ( x,0)  f ( x)  
0, 2   x  ,
1


Initial velocity u ( x,0)  0
t

77
The differential equation for T is T    2T  0 , with T (0)  0 ; its solution is
T  C3 cos t  C4 sin t
T     C3 sin t  C4 cos t  .

T (0)  0  0  C4 ,   0  0  C4 , therefore for each positive integer n , a

solution for T :
n
Tn  n cos t

Superposition solutions, we get

nx n
u ( x, t )   an sin cos t
n 1  
Applying the condition for initial displacement , we get

nx
f ( x)   an sin
n 1 
This a Fourier sine series, therefore
2 nx
an 
0
 f ( x)sin

dx

2   /2 nx 
nx  2  /2 nx
   x.sin dx   0.sin dx    x sin dx
0   /2    0 
Use integration by parts:
nx
u  x, dv  sin
dx

 nx
du  dx, v   cos
n 

78
 2  nx 
 /2
nx   /2
an    x cos
  n  0

n 0
 cos

dx 

2  2 nx 
 /2
n   
2

an    cos    sin 
  2n 2  n   0 
 n 2 n
an   cos  2 2 sin
n 2 n 2
Thus

  n 2 n  nx n
u ( x, t )     cos  2 2 sin  sin cos t .
n 1  n 2 n 2  

79
 SHEET #7
[A] Heat Equation (Heat transfer in a bar)
1. Solve the heat equation subject to the given conditions. Assume a bar of
length  and a 2  1
1, 0  x  12 
a. u (0, t )  0, u (, t )  0, u ( x,0)  
0, 2   x  ,
1

b. u (0, t )  0, u (, t )  0, u ( x,0)  x(  x)

2. Find the temperature u ( x, t ) in a rod of length  , if the initial temperature is
f ( x) throughout and if the ends x  0 and x   are insulated and the initial

 x, 0  x  12 
condition is f ( x)  
 0,
1
2   x  ,

u  2u u
3. Solve the heat equation  2 where    and (0, t )  0,
t x x
u
(, t )  0 and u ( x,0)  cos3 x, 0  x   .
x
[B] Wave Equation (Vibration of a string)
 2 u  2u
1.Solve the wave equation 2  2 subject to the following boundary
t x
Conditions:
u
a. u (0, t )  0, u (, t )  0, u ( x,0)  14 x(  x) and ( x,0)  0
t
u
b. u (0, t )  0, u (, t )  0, u ( x,0)  0 and ( x,0)  x(  x)
t
u
c. u (0, t )  0, u (, t )  0, u ( x,0)  0 and ( x,0)  5sin 5 x
t

80
 u
d. u (0, t )  0, u (, t )  0, ( x,0)  0 and u ( x,0) is given by
t

1
3  x
1
2  2
3 
2.Find the displacement u ( x, t ) of a vibrating string of length  , if both ends
of the string are fixed at x  0 and x   , and the string starts vibrating
from rest. The initial displacement is given by
 x, 0  x  12 
u ( x,0)  
   x, 2   x  ,
1

take c 2  1 .
[C] Boundary Value Problem
Show that the solution of the boundary value problem:
 2u  2 u
 2  u , 0  x  12 , t  0
t 2
x
u (0, t )  0, u (, t )  0,

 x, 0  x  12 
u ( x,0)   t 0
   x, 2   x  ,
1

u
and  0, t  0
t t 0
4  (1) k 1
is u ( x, t )   cos (2k  1) 2  1 t sin(2k  1) x .
 k 1 (2k  1) 2

81
 CHAPTER IV

Conformal Mapping

1. Bilinear Transformation
2. Schwarz Christoffel Transformation

82
 Conformal Mapping

Definition
A complex function w  f ( z ) of a complex variable z gives a mapping of its
domain of definition D in the z  plane into the complex w  plane or onto its
range of values in that plane.

For any point z0 in D , the point w0  f ( z0 ) is called the image of z0 with

respect to f .

Conformal Mapping:
1. We say that a mapping preserves angles if curves intersecting at an angle
 in z  plane map to curves intersecting at the same angle  in the w  plane.

C2
z0  C1 C2*

w0 C1*

z  Plane w  Plane

2. A mapping preserves orientation if a counterclockwise rotation in the

z  plane is mapped to a counterclockwise rotation in the w  plane. This
means that, if a line L is rotated counterclockwise in the z  plane to form a

83
 new line L* , the image f ( L) is rotated counterclockwise in the w  plane to

the image f ( L* ) as shown in figure.

L* L f ( L* )
z0
w0 f ( L)

z  Plane w  Plane

3. A conformal mapping is one which preserves both angles and orientation.

The following theorem enables us to produce many examples of conformal

mapping.

Theorem 1
Suppose f ( z ) is analytic on a domain D in the z  plane and that f ( z )  0 for
z in D .Then w  f ( z ) is a conformal mapping of D onto f ( D) in the
w  plane.

84
 1. Linear Fractional Transformation (LFT)
(Bilinear Transformation)
An important class of mappings called linear fractional transformations, which
are functions of the form

az  b
w  f ( z)  (4.1)
cz  d

where a, b, c and d are constants and ad  bc  0 . Such functions are also

called bilinear transformation. Differentiation (4.1) gives

ad  bc
w  (4.2)
(cz  d ) 2

Since ad  bc  0 by assumption, w( z )  0 on any domain in which w is

defined (that is, z  d / c ). From theorem 1, the transformation is conformal
on any such domain.

Theorem 2
Let z1 , z2 and z3 be distinct points in the z  plane and let w1 , w2 and w3 be distinct
points in the w  plane. Then there is a linear fractional transformation which
maps
z1  w1 , z2  w2 and z3  w3
It can be shown this transformation has the form

w  w1 w2  w3 z  z1 z2  z3
.  . (4.3)
w  w3 w2  w1 z  z3 z2  z1

85
If one of the points, say z3   , the transformation takes the form

w  w1 z  z1 z2  z3
 . (4.3)
w2  w1 z  z3 z2  z1

Examples

Ex (1)

Find a bilinear transformation that maps z  1 to w  3

z1  1 z3  1
z3  3 w1  3

z2  i

w2  3i

z  Plane w  Plane

1  3,  i  3i and 1  3
We have
w  w1 w2  w3 z  z1 z2  z3
.  .
w  w3 w2  w1 z  z3 z2  z1

86
Substitute for z1 , z2 , z3 , w1w2 , w3 in the transformation
w  3 3i  3 z  1 i  1
.  .
w  3 3i  3 z  1 i  1
w  3 z  1 (1  i ) 2
 .
w  3 z  1 (1  i ) 2
w  3 z 1

w  3 1 z
( w  3)  ( w  3) ( z  1)  (1  z )

( w  3)  ( w  3) ( z  1)  (1  z )
w 1 3
  w
3 z z

Ex (2)

Find a linear fractional transformation that maps the upper half of the plane
onto the unit disk in the w  plane.

z1  1 z2  0 z3  1 w1  1 w3  1

w2  i
z  Plane w  Plane
Substitute for z1 , z2 , z3 , w1w2 , w3 in the transformation,

87
 w  w1 w2  w3 z  z1 z2  z3
.  . ,
w  w3 w2  w1 z  z3 z2  z1
we get
w  1 i  1 z  1 0  1
.  .
w  1 i  1 z  1 0  1
w 1 z 1 1 i
 .
w 1 z 1 i 1
w  1 z  1  iz  i

w  1 z  1  iz  i
( w  1)  ( w  1) ( z  1  iz  i )  ( z  1  iz  i )

( w  1)  ( w  1) ( z  1  iz  i )  ( z  1  iz  i )
2 w 2 z  2i z i
  w
2 2  2iz 1  iz

Ex (3)

Find a linear fractional transformation that maps z  2 in the z  plane onto the
right half plane Re( w)  0 .

w1  i
z1  2 z3  2
w2  0

z2  2i w2  i

z  Plane w  Plane

88
Substitute for z1 , z2 , z3 , w1w2 , w3 in the transformation,
w  w1 w2  w3 z  z1 z2  z3
.  . ,
w  w3 w2  w1 z  z3 z2  z1
we get

w  i 0  i z  2 2i  2
.  .
w  i 0  i z  2 2i  2
w  i z  2 i  1 iz  z  2i  2
 . 
w  i z  2 1  i z  iz  2  2i
( w  i )  ( w  i ) (iz  z  2i  2)  ( z  iz  2  2i )

( w  i )  ( w  i ) (iz  z  2i  2)  ( z  iz  2  2i )
2 w 2 z  4i z  2i 2  iz
  w  i  w
2i 2iz  4 iz  2 2  iz

89
 2. Schwarz Christoffel Transformation (SC)

Let P be a polygon in the w  plane with vertices w1 , w2 ,..., wn . Let the

interior angles of P be 1 ,  2 ,...,  n , as shown in figure.

 3 w3

1 2
w1 w2
z1 z2 z3 z 4

z  Plane w  Plane

Then there are constants A and B and real numbers z1 , z2 ,..., zn that the function
1 2 n
1 1 1
f ( z )  A ( z  z1 ) 
( z  z2 ) 
....( z  zn ) 
dz  B (4.4)

map the domain Im( z )  0 onto the domain bounded by P .

In differential form this is

1 2 n
1 1 1
f ( z )  A( z  z1 )  ( z  z2 )  ....( z  zn )  (4.5)

If we choose zn   , so that z1 , z2 ,..., zn1 ,  are mapped to the vertices of the

polygon, Schwarz Christoffel Transformation takes the form

90
 1 2  n1
1 1 1
f ( z )  A ( z  z1 )  ( z  z2 )  ....( z  zn 1 ) 
dz  B , (4.6)

and its differential form is

1 2  n1
1 1 1
f ( z )  A( z  z1 ) 
( z  z2 ) 
....( z  zn 1 ) 
(4.7)

Important integrals

dz
1.   sinh 1 z  C
z 1
2

dz
2.  z 1
2
 cosh 1 z  C

dz
3.  1 z 2
 sin 1 z  C

4.  
z 2  1 dz  12 z z 2  1  cosh 1 z  C 
Note: cosh 1 1  0, cosh 1 (1)  i

Examples

Ex (1)

Find SC transformation that maps the upper half pane of the z  plane onto the
region of the first quadrant of the w  plane the real axis and the line w  i .

91
 w1  i 1
2
z1  1 z2  1 w2  0

z  Plane w  Plane

1   2  12 

Substitute in (4.7)

1 1
1 1 1
f ( z )  A( z  1) ( z  1)
2 2
A
z2 1
dz
f ( z )  A B
z2 1
Therefore
w  A cosh 1 z  B
At z1  1  w1  i  i  A cosh 1 (1)  B  i  iA  B ,

and z2  1  w2  0  0  A cosh 1 (1)  B 00 B B 0

then A  1
Therefore, the mapping is w  cosh 1 z

92
Ex (2)

Find a conformal mapping that maps the lower half of the z  plane onto the
outside of the region in the first quadrant of the w  plane bounded by the
imaginary axis and the line w  2 .

z1  1 z2  1

3 3
2 2

w1  0 w1  2

z  Plane w  Plane

3 3
1 1
1   2    f ( z )  A( z  1) 2 ( z  1) 2  A z 2  1
3
2

f ( z )  A z 2  1dz  B

Therefore


w  12 A z z 2  1  cosh 1 z  B 
At z2  1  w0  i 2B ,

4
and z1  1  w1  0  0  12 A(i)  2  A
i

Therefore, the mapping is w

2
i

z z 2  1  cosh 1 z  2 

93
 SHEET #8

[A] Linear Fractional (Bilinear) Transformation

1. Find a linear fractional transformation that maps z  1 into w  3 .

2. Find a linear fractional transformation that maps the upper half of
the z  plane onto in the unit disk in the w  plane.
3. Find a linear fractional mapping from half Re( z )  0 onto the disk w  2 .

4. Find a linear fractional transformation that maps z  2 in the z  plane

onto in the right half plane Re( w)  0 .

5. Find a bilinear mapping from the half plane Re( z )  0 onto the disk w  1 .
6. Find a linear fractional transformation that maps the lower half plane
Im( z )  0 into the right half plane Re( w)  0 .

[B] Schwarz - Christoffel formual

1. Using Schwarz - Christoffel formula, find a conformal mapping that maps

the lower of the z  plane onto the region in the first quadrant of the
w  plane bounded by the real axis and the line w  i .

2. Using Schwarz - Christoffel formula, find a conformal mapping that maps

the upper half of the z  plane onto the region in the first quadrant of the
w  plane bounded by the imaginary axis and the line w   .

94
3. Find a conformal mapping that maps the upper half of the z  plane onto the
region in the second quadrant of the w  plane bounded by the imaginary axis
and the line w  1 .

4. Find a conformal mapping that maps the lower half of the z  plane onto the
outside of the region in the first quadrant of the w  plane bounded by the
imaginary axis and the line w  2 .

5. Find a Schwarz - Christoffel transformation that maps an equilateral triangle

ABC whose base BC has the coordinates (0,0) and (1,0) in the w  plane
onto the upper half of the z  plane.

95