ME5011

ANALYTICAL METHODS IN
MECHANICAL ENGINEERING

1
 FOURIER SERIES
In solving certain boundary value problems involving linear differential equations, an f-
function defined in the interval [-L, L] can be expanded into an orthogonal series containing
trigonometric functions.
a 
 n n 
f ( x) = 0 +   an cos x + bn sin x
2 n =1  L L 
where, a0, a1, a2, …, b1, b2, b3,… are coefficients and can be found as in the extension of
orthogonal series.

Theorem: Fourier Series

The Fourier series of an f-function defined in the interval (-L, L) is:
a 
 n n 
f ( x) = 0 +   an cos( x) + bn sin( x) 
2 n =1  L L 
where, n is positive integer, the coefficients are calculated as follows,
1 L
a0 =  f ( x)dx
L −L

1 L n
an = 
L −L
f ( x) cos( x)dx
L

1 L n
bn = 
L −L
f ( x) sin(
L
x)dx

1
Example: Let f(x) = x for −   x   . Write the Fourier series of given f(x) function.
a 
 n n 
f ( x) = 0 +   an cos( x) + bn sin( x) 
2 n =1  L L 
L=

1  x 2

1   2 (− ) 2 
L
1 1
a0 =  f ( x)dx =  xdx = =  − =0
L −L  −   2 −   2 2 
 
n  cos(nx) x sin( nx) 
L
1 1 1
an = 
L −L
f ( x) cos( x)dx =  x cos(nx)dx =
L  −   n2 +
 n 
 −
1  cos(n ) x sin( n ) cos(−n ) (− ) sin( −n ) 
= + − −
  n 2 n n2 n 

1  cos(n ) x sin( n ) cos(n )  sin( n ) 
=  + − − =0
  n2 n n2 n 
cos(−n ) = cos(n )
NOTE:
sin( −n ) = − sin( n )
 
n 1  sin( nx) x cos(nx) 
L
1 1
bn =  f ( x) sin( x)dx =  x sin( nx)dx =  − 
L −L L  −   n2 n  −
1  sin( n ) (− ) cos(−n ) sin( −n ) (− ) cos(−n ) 
= − − +
  n 2 n n2 n 

1  sin( n )  cos(n ) sin( n )  cos(−n )  1  2 sin( n ) 2 cos(n ) 
=  − + −  =   n2 − 
  n2 n n2 n   n 
1  − 2 cos(n )  1 2 (−1) n +1 2(−1) n +1
= = =
  n  n n
cos(n ) = (−1) n
NOTE:
sin( n ) = 0
a0   n n 
f ( x) = +   an cos( x) + bn sin( x) 
2 n =1  L L 
 
2(−1) n+1 2 1
f ( x) =  bn sin( nx) =  sin( nx) = 2 sin( x) − sin( 2 x) + sin( 3x) − sin( 4 x) + 
n =1 n =1 n 3 2

2
 0 −π  x0
Example: Expand f(x) =  in Fourier series.
π − x 0  x  π

a0   n n 
f ( x) = +   an cos( x) + bn sin( x) 
2 n =1  L L 
L=
 
1  1 x2  
L 0
1
a0 =  f ( x)dx =   0dx +  ( − x)dx = x −  =
L −L  − 0   2 0 2

n 1 
L 0
1
an = 
L −L
f ( x ) cos(
L
x ) dx = 
 −
0 cos(nx) dx + 
0
( − x) cos(nx)dx


1
 
  sin( nx) 0 1  cos(nx) x sin( nx) 

  0 0
=  cos( nx ) dx − ( x cos(nx) dx =
  −  + 
  n −   n
2
n 0

sin( 0) − sin( −n ) 1  cos(n )  sin( n ) cos(0) 0 

= −  + − − 
n   n2 n n2 n
(−1) n 1 1 − (−1) n
=− + =
n 2 n 2 n 2

n 1 
L 0
1
bn =  f ( x) sin( x)dx =   0 sin( nx)dx +  ( − x) sin( nx)dx
L −L L  − 0 
   
1  1
  sin( nx)dx −  x sin( nx)dx =  sin( nx)dx −  x sin( nx)dx
 0 0  0 0
 − cos(nx)  1  sin( nx) x cos(nx)  
= −  − 
 n   n2 n
 0 0

− cos(n ) + cos(0) 1  sin( n )  cos(n ) sin( 0) 0 cos(0) 

= −  − − +
n   n2 n n2 n 
1 − (−1) n 1   (−1) n  1
= − − =
n  n  n
⸫
 
1 − (−1) n 1 
f ( x) = +  cos(nx) + sin( nx)
2 n=1  n 2
n 

3
 − x −5  x  0
Example: Expand f(x) =  in Fourier series.
1 + x 0  x 5
2

a0   n n 
f ( x) = +   an cos( x) + bn sin( x) 
2 n =1  L L 
L=5
1  1  355  71
L 0 5
1
a0 =  f ( x)dx =   − xdx +  (1 + x 2 )dx =  =
L −L 5  −5 0  5 6  6
n 1  nx   nx  
L 0 5
1
an = 
L −L
f ( x ) cos(
L
x ) dx = 
5  −5
− x cos
 5 
 dx + 
0
(1 + x 2 ) cos
 5
dx
 

=
n
2
5
2
11(−1) n
−1
n 1  nx   nx  
L 0 5
1
L −L 5 −5 0
bn = f ( x ) sin( x ) dx =  − x sin   dx + (1 + x 2
) sin  dx
L  5   5  

4
Even and Odd Functions
Even Function
If f (-x) = f (x) on -L ≤ x ≤ L interval, f-function on [-L, L] interval is called even function. In
this case, bn = 0.
The graph of the double function is as follows.

(Source: internet: google.com)

Properties of Even Functions
a) The product of two even functions is an even function.
b) The sum (difference) of two even functions is an even function.
a a
c) If the f-function is an even function, then  f ( x)dx =2 f ( x)dx
−a 0
Odd Function
If f (-x) = -f (x) on -L ≤ x ≤ L interval, f-function on [-L, L] interval is said to be an odd
function. In this case an = 0.
The graph of the double function is as follows.

(Source: Internet: google.com)

Properties of Odd Functions
a) The product of two odd functions is an even function.
b) The product of an even function and an odd function is an odd function.
c) The sum (difference) of two odd functions is odd function.
a
d) If f-function is odd function, then  f ( x)dx =0
−a
NOTE: A function can be written as the sum of even and odd functions.
f ( x) =  f ( x) + f ( − x ) +  f ( x) − f ( − x)
1 1
2 2

Example:
e x = e x + e − x  + e x − e x 
1 1
2 2
= cosh(x) + sinh( x)

5
Fourier Cosine and Sine Series
If the f-function is an even function on the interval [-L, L], then a0, an and bn coefficients are
as follows:
1 L 2L
a0 =  f ( x)dx =  f ( x)dx
L −L L0
1 L n 2L n
an = 
L −L
f ( x ) cos(
L
x ) dx = 
L0
f ( x) cos( x)dx
L

1 L n
bn = 
L −L
f ( x) sin(
L
x)dx = 0

Similarly, if the f-function is the odd function on the interval [L, -L],
a n = 0, n = 0,1,2,3,...

1 L n 2L n
bn =  f ( x) sin( x)dx =  f ( x) sin( x)dx
L −L L L0 L
i) The Fourier series of an even function on the interval [-L, L] is the cosine series.
In this case [-L, L]
a0 
n
f ( x) = +  a n cos( x)
2 n =1 L
2L
where a 0 =  f ( x)dx
L0
2L n
an = 
L0
f ( x) cos( x)dx
L

ii) The Fourier series of an odd function on the interval [-L, L] is the sine series. In
this case

n
f ( x) =  bn sin( x)
n =1 L
2L n
where bn =  f ( x) sin( x)dx
L0 L

Examlpe: Write the Fourier series of f(x) = x4 function on [-1,1].

On the interval -1 ≤ x ≤ 1, due to f(x) = x4 =(-x)4 =f(-x), function is an even function.

Since it is an even function, this is a Fourier cosine series and bn = 0.
1  x5 1  15 (−1) 5  2
a0 =  x dx = 
1 4
=  − =
1 −1  5  5 5  5
 − 1

6
 1
1 4  x 4 sin( nx) 4x3  12
1
an =  x cos(nx)dx =  + cos(nx) − x
2
cos(nx)dx
1 −1  n (n ) 2
 (n )
2
−1
1
12  2 x cos(nx)  x 2 
1
 x 4 sin( nx) 4x3  2 
= + cos(nx )  − 2 
+  −  sin( nx)
3 
 n (n ) 2
 −1 (n )  (n )
2
 n (n )   −1
4 24 cos(n ) 4(−1) 3 24 cos(−n )
= cos(n ) − − cos(−n ) +
(n ) 2
(n ) 4
(n ) 2
(n ) 4
48 cos(n )
=
4
(n ) 2
cos(n ) −
(n ) 4
=
8
(n ) 4
 
n 2 2 − 6 (−1) n

⸫
1 
 n 2 2 − 6 
f ( x) = + 8  4 4 (−1) n cos(nx)
5 n =1  n 

7
Örnek: Write the Fourier series of f(x) = x3 function on interval − 4  x  4 .

Since f(x) = x3 is an odd function in the interval, this is a Fourier sine series and a0= 0, an= 0.

n  nx 
L 4
1 1
bn =  f ( x) sin( x)dx =  x3 sin  dx
L −L L 4 −4  4 
4
    
1  3 x 2 6   nx   6 x x3   nx 
=  −  sin  + −  cos 
4  ( n ) 2 ( n ) 4   4   ( n )3 ( n )   4 
   
 4 4   4 4   − 4
     
 48
 6   24
 sin (n ) + 
64  

 cos(n )

 − − 
 ( n ) 2 ( n ) 4   ( n )3 ( n ) 
   
1 
 4 4   4 4  
 
=  
4      
   48 6   − 24
 sin (− n ) + 
64  
 cos(− n )
− 
  n 2 − + 
 n 4   ( n )3 ( n )  
  ( ) ( )    
  4 4   4 4  
 
1  48 128 
= 

−

 cos(n ) =  3 3 −
768 128  2 2
(
 cos(n ) = 3 3 6 − n  (−1)
128 n
)
4  ( )3 ( ) 

n n
 n  n  n
 4 4 
128 2 2
( )
= 3 3 n  − 6 (−1) n +1
n
⸫

 n 2 2 − 6  nx
f ( x) = 128 3 3 (−1) n +1 sin( )
n =1  n  4

8
Example: Write the Fourier series of f(x) = x on −2  x  2 .

f(x) = x function is an odd function, so this is a sine series expansion.


n
f ( x) =  bn sin( x)
n =1 L
L=2
n n 4(−1) n +1
L 2
2
bn =  f ( x) sin( x)dx =  x s in( x)dx =
L0 L 0
2 n
⸫
4 (−1) n+1

n
f ( x) = 
 n=1 n
sin(
2
x)

− 1 −π  x  0
Örnek: f(x) =  Write Fourier series.
1 0  x π

f(x) is an odd function on the given interval - ≤ x ≤ .


n
f ( x) =  bn sin( x)
n =1 L
L=
n 2 1 − (−1) n
L 2
2 2
L 0  0
bn = f ( x ) sin( x ) dx = (1) s in( n x ) dx =
L  n
1 − (−1) n
2 
⸫ f ( x) =  n sin( nx)
 n =1
HALF-RANGE EXPANSIONS
Throughout the preceding discussion it was understood that a f-function defined on an interval
with the origin as midpoint (i.e. [-p,p]). However, in many instances we are interested in
representing a function that is defined on an interval (0, p) by a trigonometric series. This can
be done in many different ways by supplying an arbitrary definition of the function on the
interval (-p, 0). For brevity we consider the three most important
cases. If y = f(x) is defined on the interval (0, L), then:
(i) reflect the graph of the function about the y-axis; the function is even on the interval [-p, p]

9
 Even function

(ii) reflect the graph of the function through the origin; the function is odd function on the
interval [-p, p]

Odd function

(iii) define f on (-p, 0) by f(x) = f(x + p)

In practice, there is no need for reflections in i and ii. If the f-function is defined on the interval
0 <x<p, then we can take L = p half period as the interval length. Series with coefficients a0,
an, and bn give either an even or odd periodic extension of the original function in period 2p.
The cosine and sine series obtained in this way are called half-range expansions.

Finally, we define the value of the function on the interval –p<x<0 in iii the same as in the
range 0 <x<p. As in the first two cases, we practically do not need this situation. If the f-function
is defined in the range 0 <x <p, then 2L = p or L = p/2. Thus, the Fourier series gives the
periodic expansion of the f-function with the p period. The Fourier series on the interval 0 <x<p
is as follows:

2p
p 0
a0 = f ( x)dx

2p 2 n
an = 
p0
f ( x) cos(
p
x)dx

10
 2p 2n
bn = 
p0
f ( x) sin(
p
x)dx

a0   2 n 2n 
f ( x) = +   an cos( x) + bn sin( x) 
2 n =1  p p 

Example: Expand f(x) = x2 function on 0 < x < p a) in cosine series, b) in sine series, c) in a
Fourier series.

a) L=p
p p
2 2 2 2 2
p 0 p 0
a0 = f ( x ) dx = x dx = p
3

n n 4 p 2 (−1) n
p p
2 2
a n =  f ( x) cos( x)dx =  x 2 cos( x)dx =
p0 L p0 p n 2 2

a0   n  p 2 4 p 2  (−1) n n
f ( x) = +   an cos( x)  = + 2  2 cos( x)
2 n=1  L  3  n=1 n p

b) L=p
n n 2 p 2 (−1) n+1 4 p 2
( )
p p
2 2 2
bn =  f ( x) sin( x)dx =  x sin( x)dx = + 3 3 (−1) n − 1
p0 L p0 p n n
n  (−1) n+1 n 
( )
 
2 p2 2
f ( x) =  bn sin( x) =   + 3 2 (−1) n − 1 sin( x)
n =1 L  n =1  n n p 

11
 ve n = 2n
p 1 2
c) L= , =
2 L p L p

L p p
1 2 2 2
a0 =  f ( x)dx =  f ( x)dx =  x 2 dx = p 2
L −L p0 p0 3

n 2n
L p
1 2 2 p2
an =  f ( x) cos( x)dx =  x cos( x)dx = 2 2
L −L L p0 p n

n 2n
L p
1 2 2 p2
bn =  f ( x) sin( x)dx =  x sin( x)dx = −
L −L L p0 p n

a0   n n 
f ( x) = +   an cos( x) + bn sin( x) 
2 n =1  L L 
p2 p2 
 1 2n 1 2n 
f ( x) = +
3 
  n  cos(
2
x) − sin( x)
n =1  p n p 

12
INTEGRATION OF FOURIER SERIES
Fourier series expansion of an term by term integrable periodic function on the interval [-L, L]
was as follows:

a0   n n 
f ( x) = +   an cos( x) + bn sin( x) 
2 n =1  L L 

For –L ≤ x≤ L

x x
a0  x
 n n 
 f ( x)dx = 2 dx +    an cos( x) + bn sin( x) dx
−L −L n =1 − L L L 

(x + L ) + L  1  an sin( n x) + bn cos(n ) − cos(n x) 

x 
a0
 f ( x)dx =
2  n =1 n  L  L 
−L

2 
(−1) n
Example: Fourier series of f(x) = x2 on −  x   is x 2 = + 4 2
cos(nx) ,
3 n =1 n
integrate the series.

x x
2 
(−1) n
x

 x dx = 
2
dx + 4  2 cos(nx)dx
− −
3 n =1 − n

(−1) n 
x − x = 12 3 sin( nx)
3 2

n =1 n

13
Theorem: If the f-function is an integrable function on the interval [-L, L], then the Fourier
series coefficients on the interval [-L, L] satisfy the following condition.
a0 +  (an + bn2 )   f ( x) 2 dx
1 2  2 1 L
2 n=1 L −L

− 12  (−1)n nx
Example: Fourier series of f(x) = 2x on − 3  x  3 is f ( x) = 
 n =1 n
sin(
3
).
Using this expansion, find the Fourier series of f(x) = x2 , also expand f(x) = x2 as Fouirer
series.
− 12  (−1) n nx
x x

−3 2 xdx =  
 n =1 −3 n
sin(
3
)dx

12 
 (−1) n nx 1
2 
x2 − 9 = cos( )− 2  (1)
 n =1  n 2
3 n 
ve
1 2  2
2
( )
1 L
a0 +  an + bn2   f ( x) 2 dx
L −L
n=1

− 12 (−1)n
a0 = 0 an = 0 bn =
 n
 3
144 1 1

n =1  2
n 2
  (2 x) 2 dx
3 −3

144 1
 n =1
2
n2
 24

1 2

n =1 n
2

6
Substituting in Eq. (1),
36   (−1) n nx  36   1 
f ( x) = x 2 = 9 + 2   2 cos( ) − 
 n =1  n 3   2 n =1  n 2 
36 
 (−1) n nx  36  2
f ( x) = x 2 = 9 + 
 2 n =1  n 2
cos( ) −
3  2 6
(−1) n
36 
nx
f ( x) = x = 3 + 2  2 cos(
2
)
 n =1 n 3

DIFFERENTIATION OF FOURIER SERIES

Let f-function be a continuous function on the interval [-L, L]. The Fourier series of f-function
a 
 n n 
on the interval [-L, L] is as follows f ( x) = 0 +   an cos( x) + bn sin( x)  . So f (x) is
2 n =1  L L 
as follows:

a0  +  d an cos(n x) + bn sin( n x)


1 d
f ( x) =
2 dx n =1 dx  L L 

14
Example: Fourier series of f(x) = x2 on −   x   is
2 
(−1) n
x =
2
+ 4 2 cos(nx) . a) Find the series expansion of x using differentiation, b)
3 n =1 n
expand f(x) = x as Fourier series.
2 
(−1) n
f ( x) = x 2 = + 4 2 cos(nx)
3 n =1 n
d 2 d  2  
d  (−1) n 
f ( x) =  x  =   + 4  2 cos(nx)
dx dx  3  n =1 dx  n 

(−1) n
f ( x) = 2 x = 4 2 ( (−n) sin(nx) )
n =1 n

(−1) n +1
f ( x) = 2 x = 4 sin(nx)
n =1 n

(−1) n +1
f ( x) = x = 2 sin(nx)
n =1 n

15
 SPECIAL FUNCTIONS
In this section, special functions encountered in solving differential equations will be
discussed.
THE FACTORIAL FUNCTION


x e
n −x
dx = n!
0

n!= n(n −1)(n − 2)3.2.1 n > 0, n: integer


Example: Evaluate the integral  e dx .
−x

0


x e
0 −x
dx = 0!= 1
0

THE GAMMA FUNCTION


(n) =  x n −1e − x dx = (n − 1)! n > 0 n  -1, -2, -3, …
0
is called the Gamma function.

Properties of the Gamma Function

If n > 0
(n + 1) = n(n)
(1) = 1
If n is positive integer
(n + 1) = n! n = 1, 2, 3, …
(1 / 2) = 

Example:
(2) = 1!= 1
(5) 4!
(6) = 5!= 120 = = 12
(3) 2!

NOTE:
1 (2n)! 
( n + ) =
2 n!4n

(n + 1)
n < 0 ise (n) =
n

16
 7
Example: (− ) = ?
2
 7   5  5   3  3 
 − + 1  −   − + 1  −   − + 1
 7
 −  = 
2 
= 
2
= 
2   2 =  2 
=
 2  7  7  7  5   7  5   7  5  3 
−  −   −  −   −  −   −  −  − 
 2  2  2  2   2  2   2  2  2 
 1  1  1
 −   − + 1  
 2  2  2 16 
= = = =
 7  5  3   7  5  3  1   7  5  3  1  105
 −  −  −   −  −  −  −   −  −  −  − 
 2  2  2   2  2  2  2   2  2  2  2 

Example: (3.2) = ?
(n + 1) = n(n)
(3.2) = (2.2 + 1) = (2.2)(2.2) = (2.2)(1.2)(1.2) = (2.2)(1.2)(0.91817 ) = 2.42397

Example:


 xe
− x
dx = ? √𝑥 = 𝑢 olarak seçilirse,
0

𝑥 = 𝑢2
 ( n ) =  x n −1e− x dx = ( n − 1)!
0 𝑑𝑥 = 2𝑢𝑑𝑢
 

 xe dx =  u 2e−u 2udu
− x

0 0

= 2 u 3e−u du = 2 ( 4 ) = 2 ( 3!) = 12
0

THE BETA FUNCTION

1
B(m, n) =  x m −1 (1 − x) n −1 dx m > 0, n > 0
0
(m)(n)
B(m, n) =
 ( m + n)

Properties of the Beta Function

B(m, n) = B(n, m)
 /2
B(m, n) = 2  sin 2 m−1  cos2 n−1  d
0

x m−1

B(m, n) =  m+ n
dx
0 (1 + x )

17
 
x3
Example: Evaluate I =  dx .
0 ( x + 1)
5


x m−1
B(m, n) =  m+ n
dx
0 (1 + x )
m-1=3 → m=4
m+n=5→ n=1


x3  ( 4 )  (1) 3! 1
I = dx = B ( 4,1) = = =
0
( x + 1) 5
 ( 5) 4! 4
 /4

 sin
3
Example: Evaluate the integral (2 x) cos2 (2 x)dx .
0

If 2x = , then 2dx = d
 /4  /2
d
 sin (2 x) cos (2 x)dx =  sin  cos 
3 2 3 2

0 0
2
 /2
B(m, n) = 2  sin 2 m−1  cos2 n−1  d
0

2m - 1 = 3 → m = 2
2n - 1 = 2 → n = 3/2
 3
 /4  /2 B  2, 
1 1  2 1  3
 sin (2 x) cos (2 x)dx = 2  sin 3  cos 2  d = = B  2, 
3 2

0 0
2 2 4  2
3 1 1
 ( 2)     ( 2)     
 3 2 = 2 2 = 4
B  2,  =
 2 7  5  3  1   1  15
        
2  2  2  2   2 
 /4
1  3 1 4 1
 sin (2 x) cos (2 x)dx = 4 B  2, 2  = 4 15 = 15
3 2

1/ 4
1 
1
Example: Evaluate I =   − 1 dx .
0 x 
1
B(m, n) =  x m −1 (1 − x) n −1 dx
0

 1− x 
1/4 1/4
1  3 5
1 1 1
I =   − 1 dx =    dx =  (1 − x ) x dx = B  , 
1/4 −1/4

0
x  0
x  0 4 4
3 5
    
=    =
4 4 (1.22541333)( 0.906402 ) = 1.1107
 ( 2) 1

18
THE ERROR FUNCTION and THE COMPLEMENTARY ERROR FUNCTION

THE ERROR FUNCTION

2 x −u 2
erf ( x) =  e du
 0

THE COMPLEMENTARY ERROR FUNCTION

TAMAMLAYICI HATA FONKSİYONU

2  −u 2
erfc ( x) = 1 − erf ( x) =  e du
 x

e−x
2

erfc ( x) ~
x 

(Source: internet)

i) The error function is an odd function.

erf(-x) = -erf(x)

2  −u 2 2 (1 / 2)
ii) erf () = 
 0
e du =
 2
=1

iii) For very small x values

2  x3 x5 
erf ( x) = x − + + ... x  1
  3.1! 5.2! 
iv) For very big x values
e−x  
2
1 1.3 1.5
erfc ( x)  1 − 2 + − + ... x  1
 x  2 x 2x 2
2
( ) ( )
2x 2
3


19
Derivative of the error function
d
(erf ( x)) = 2 e − x
2

dx 

d2
(erf ( x)) = − 4 xe − x
2

dx 2



i n erfc ( x) =  i n −1erfc ( )d n = 0,1,2,...
0

1
ierfc ( x) = e − x − xerfc ( x)
2

1 2 2 
i 2 erfc ( x) = (1 + 2 x 2 )erfc ( x) − xe − x 
4  
Expansion in series

x 2 n +1
 (− 1) n!(2n + 1)
2
erf ( x) =
n

 n =0

2
Example: Evaluate I =  e dx .
−x 2

2
 
I =  e− x dx = erf (2) = (0.9953) = 0.88206
2

0
2 2

Example: u = k  u2
2
0 < x < L, t > 0
t x

Initial and boundary conditions: u(x,0) = T0 = constant

u(0, t )
u( L, t ) = =0 t 0
t
The solution is obtained as follows

  (2n + 1) L − x   (2n + 1) L + x  
u ( x, t ) = T0 − T0  (−1) n erfc   + erfc  
n =1   2 kt   2 kt 

20
THE DELTA FUNCTION (THE DIRAC DELTA FUNCTION)

Informally, an impulse is a force of extremely large magnitude applied over an extremely

short period of time (for example hitting on a nail with a hammer). We can model this idea
as follows. First, for any positive number  consider the pulse  defined by

1
 =  H (t ) − H (t −  ) H (t ) :Heaviside function


The Heaviside function is a function to define “off-on” physical process.

 (t ) = lim   (t ) Dirac delta function

 →0

 ( x) = 0  x  0

21


  ( x)dx =1
−

n
 n ( x) = e − nx n = 1,2,3,...
2
i)

ii)  n ( x) → 0 n → , x  0
iii)  n (0) →  n→

iv) Tüm n=1, 2, 3, … için   ( x)dx =1
−

v)  f ( x) ( x − a)dx = f (a)
−
Ayrıca,
vi) f ( x) ( x − a) = f (a)
vii) x ( x) = 0
viii)  ( x − a ) =  (a − x) .
ix)  < a < ;  ve  sabit sayı


 f ( x) ( x − a)dx = f (a)




 f ( x) ( x − a)dx = (−1) n f ( n) (a)

( n)
x) n. türevi
−


 f ( x) ( x − a)dx = − f (a)

−

a +

  ( x − a)dx = 1
− a +

22
CAUCHY-EULER DIFFERENTIAL EQUATION (EULER DIFFERENTIAL
EQUATION)

a0 xn y( n) + a1xn −1 y( n −1) + ... + an y = 0

where a0, a1, … , an are constant coefficients. To solve this differential equation x= et
transformation is performed. If we substitute x =et into y(x), we obtain a function of t as
Y (t)= y(et ).

Example: Solve x 2 y + xy − y = 0 .

Using x = et conversion

x = et
dx = et dt
dy dy dt 1 dY dY
= = = e− t = e−tY 
dx dt dx x dt dt

d 2 y d  dy  d  −t dY  d  −t dY  dt d  −t dY  1
=   = e  = e  = e 
dx 2 dx  dx  dx  dt  dt  dt  dx dt  dt  x
 dY d 2Y 
=  −e − t + e−t 2  e−t = e−2t (Y  − Y  )
 dt dt 

Substituting into the differential equation and rearranged

Y  − Y = 0
The general solution

Y (t ) = y (et ) = C1et + C2 e − t
1
y ( x) = C1 x + C2
x

23
LEGENDRE POLYNOMIALS
On the interval -1 ≤ x ≤ 1, n = 0, 1, 2, 3,… to be an integer

(1 − x 2 ) y − 2 xy + n(n + 1) y = 0

This differential equation is Legendre’s differential equation, and the solution of this equation
is written in terms of Legendre polynomials

1 dn 2
Pn ( x) = n n
( x − 1) n
2 n! dx

where Pn(x) is nth order Legendre function.

1 d0 2
P0 ( x) = 0 0
( x − 1) 0 = 1
2 0! dx
1 d 1
P1 ( x) = 1
( x 2 − 1) = (2 x) = x
2 1! dx 2

1 d2 1
P2 ( x) = 2 2
( x 2 − 1) 2 = (3x 2 − 1)
2 2! dx 2

1
P3 ( x) = (5 x 3 − 3x)
2
1
P4 ( x) = (35 x 4 − 30 x 2 + 3)
8

Recurrance relationships
2n + 1 n
Pn+1 ( x) = xPn ( x) − Pn−1 ( x)
n +1 n +1

Pn+1 ( x) − Pn−1 ( x) = (2n + 1) Pn ( x)

(1 − x 2 ) Pn( x) = −nxPn ( x) − nPn−1 ( x)

Ortogonality
1

 Pm ( x) Pn ( x)dx = 0 mn
−1

24
 BESSEL FUNCTIONS

The differential equation x y  + xy  + ( x −  ) y = 0 in which ν ≥ 0, is called Bessel’s

2 2 2

equation of order ν. This differential equation is second order, and the phrase “order ν”
refers to the parameter ν appearing in it.

Bessel Functions of the First Kind

x 2 y  + xy  + ( x 2 −  2 ) y = 0  0

The solution of Bessel’s equation of order ν leads to Bessel functions.


(−1) n
J n ( x) =  2 k +1
x 2k +n
k =0 2 k! ( n + k )!

Jn(x), is the Bessel function of the first kind of order n.

y 2
If Bessel equation is written as y + + (1 − 2 ) y = 0 , then there will be a singular point at 0.
x x

We therefore attempt a Frobenius solution y ( x) =  Cn x n + r
n =0

then

(−1) n ( + 1) 2 n +
y ( x) = C0  x
n =0 n!(1 + + n)

1
If C0  0 and C0 =  0 are selected, solution is obtained as
2 ( + 1)


2 n +

(−1) n  x
J ( x) =   
n =0 n!(1 +  + n)  2 

While to get a solution for y(x), recall that the indicial equation of Bessel’s differential
equation is r 2 − 2 = 0 , with roots  and -. Similarly for r2 = -

2 n −

(−1) n  x
J − ( x) =   
n =0 n!(1 − + n)  2 

Depending on the value of , J-(x)may contain negative powers of x and hence converge on
the interval (0, ).

Some point could be taken into account for the solution of the Bessel equation,
• If  = 0, then J(x) ve J-(x) are the same expressions.
• If  > 0 ve r1-r2 =  - (-) = 2 is not a positive intenger; the general solution for the
given interval is
y ( x) = C1 J ( x) + C2 J − ( x)

25
 • If r1-r2 = 2 is a positive integer then two possibilities are present

i. If  = m a positive integer Jm(x) and J-m(x) are not linearlu independent

solutions. In this case, J-m(x) is the multiplication of Jm(x) with a coefficient as

J − ( x) = (−1) m J m ( x)

This is the Bessel function of the second kind.

ii. If  is the half of an integer, r1-r2 = 2 may be a positive integer. In this case
J(x) and J-(x) are linearly independent solutions.The solution of the Bessel on
(0,) interval

y ( x) = C1 J ( x) + C2 J − ( x)   integer

1
Example: Find the general solution of the equation x y  + xy  + ( x − ) y = 0 .
2 2

9
x 2 y + xy + ( x 2 − 2 ) y = 0
1
=
3

y ( x) = C1 J 1 ( x) + C2 J 1 ( x)
−
3 3
1
Example: Find the general solution of the equation x y  + xy  + ( x − ) y = 0 on (0,)
2 2

4
interval.

Bessel Functions of the Second Kind

If   integer , the function defined by the linear combination

cos( ) J ( x) − J − ( x)
Y ( x) =
sin( )
and the function J(x) are linearly independent solutions of the Bessel equation. Thus another
form of the general solution of the Bessel equation is

y ( x) = C1 J ( x) + C2Y ( x)

26
 0
As  → m, (m is an integer), Y(x) has the indeterminate form . However, it can be shown
0
by L'Hopital's rule that lim Y ( x) exists. Moreover, the function
 →m

Ym ( x) = lim Y ( x)
 →m

and Jm(x) are linearly independent solutions of x 2 y + xy + ( x 2 − m2 ) y = 0 . Hence for any
value of , the general solution on the interval (0, ) can be written as

y( x) = C1J ( x) + C2Y ( x)

Y(x) is called the Bessel function of the second kind of order .

Example: Find the general solution of the equation x y + xy + ( x − 9) y = 0 on (0,)
2 2

interval.

x 2 y + xy + ( x 2 −  2 ) y = 0
 =3
y( x) = C1 J 3 ( x) + C2Y3 ( x)

Parametric Bessel Equation of order 

Sometimes it is possible to transform a differential equation into the Bessel equation by means
of a change of variable. We can then express the solution of the original equation in terms of
Bessel functions.

Example: Find the solution of the given differential equation

x 2 y + xy + ( 2 x 2 − 2 ) y = 0   0 .

If t =  x t = x
dt=dx

dy dy dt dy
= =
dx dt dx dt
d 2 y d  dy  dt d2y
=   = =2
dx 2 dt  dx  dx dt 2

Substituting into the differential equation

x 2 y + xy + ( 2 x 2 − 2 ) y = 0
2
 t  2 d y  t  dy
2

   +   + (t 2 − 2 ) y = 0
     dt
2
dt

27
 d 2 y dy 2 2
t 2 + t + (t − ) y = 0
2

dt dt
This equation is Bessel's equation of order  with solution

y(t ) = C1 J (t ) + C2Y (t )

Resubstituting t =  x

y( x) = C1 J ( x) + C2Y ( x)

NOTE:
x 2 y + xy + ( 2 x 2 − 2 ) y = 0   0

General solution of this equation is

y( x) = C1 J ( x) + C2Y ( x)

 may be equal to zero, a fractional number or an integer.

The Modified Bessel Equation

x 2 y + xy − ( x 2 + 2 ) y = 0 is called the modified Bessel equation. If t=ix is used (i2 = -1) the
given equation can be written as

d2y dy
t2 2
+t + (t 2 − 2 ) y = 0
dt dt

The solution of this differential equation is in terms of J(t) and Y(t). The complex number
valued solution of the modified Bessel differential equation is in the form of J(ix) and Y(ix).
The modified Bessel function of the first kind of order  is defined according to J(ix):

I ( x) = i − J (ix)

  integer, the modified Bessel function of the second kind of order  is

 I − ( x) − I ( x)
K ( x) =
2 sin( )

and  is an integer, for  = n

K n ( x) = lim K ( x)
 →n

For any value of , I(x) and K(x) are linearly independent on the interval (0,) and the solution
of the modified Bessel equation is

28
y( x) = C1 I ( x) + C 2 K ( x)

where

I(x) is the modified Bessel function of the first kind of order ,

K(x) is the modified Bessel function of the second kind of order .

Theorem:
(
d −
dx
)
x J ( x) = − x − J +1 ( x)

J (ix) = i I ( x)
 2 
Y (ix ) = i iI ( x) − (−1) K ( x)
  
 mx Z −1 (mx) Z = J ,Y , I
d 
(
x Z (mx) =  ) 
dx − mx Z −1 (mx) Z=K
− mx − Z +1 (mx) Z = J ,Y , K
(
d −
x Z (mx) =  )−
dx  mx Z +1 ( mx) Z=I

 = 0 special case
− mZ1 (mx) Z = J ,Y , K
d
(Z 0 (mx) ) = 
dx  mZ1 ( mx) Z=I

 x J n−1 ( x)dx = x J n ( x) + C
n n

Example:
Evaluate the integral I =  x J1 ( x)dx .
2

 x J n−1 ( x)dx = x J n ( x) + C
n n

n=2
I =  x 2 J1 ( x)dx = x 2 J 2 ( x) + C

29
GENERALIZED FORM OF THE BESSEL EQUATION

Some differential equations may be in the form of

1 − 2a  a 2 − p 2c 2 
y  + y  +  b 2 c 2 x 2 c − 2 +  y = 0 p0
x  x2 

where a,b, and c are constants. The general solution of this equation is

y ( x) = x a C1 J p (bx c ) + C2Yp (bx c )  .

If p is not an integer, Yp and J-p can be replaced.

Example: Find the general solution of the equation xy + 3 y + 9 y = 0 on (0,).

1 − 2a  a 2 − p 2c 2 
y + y +  b 2c 2 x 2 c −2 + y=0
x  x2 
xy + 3 y + 9 y = 0
3 9
y + y + y = 0
x x

1 − 2a = 3  2a = −2  a = −1

 2 2 2c−2 a 2 − p 2c 2  9
b c x + =
 x2  x
9 1
b2c 2 x 2c −2 =  b2c 2 = 9  c=
x 2
1
2c − 2 = −1 b2 = 9 → b = 6
4
a 2 − p 2c 2 = 0
2
1
(−1) − p   = 0
2 2
 p=2
2

General solution of xy + 3 y + 9 y = 0 is

y ( x) = x a C1 J p (bx c ) + C2Yp (bx c ) 

y ( x) = x −1 C1 J 2 (6 x1/2 ) + C2Y2 (6 x1/2 ) 

Example: Find the solution of the differential equation 9 x y − 27 xy + (9 x + 35) y = 0 .
2 2

30
Example: Suppose we have a thin rod of constant weight w per unit length, length L and circular
cross section of radius a. The rod is clamped in a vertical position. If the rod is “long enough,”
and the upper end is displaced and held in position until the rod is at rest, the rod will remain
bent. By contrast, if the rod is “short enough”, it will return to its vertical position after the end
has been displaced slightly. Derive an expression for the critical length LC.

From the theory of elasticity

d 2 g
EI +  gx = 0 2 = = =
w
, g = w
dx 2 EI EI
d 2
2
+  2 x = 0
dx
Boundary conditions   ( 0 ) = 0 ve  ( L ) = 0
1 − 2a  a 2 − p 2c 2 
y + y +  b 2c 2 x 2 c −2 + y=0
x  x2 
1
1 − 2a = 0  a=
2
3
b2c 2 x 2c−2 =  2 x  b 2c 2 =  2  c=
2
9 2
2c − 2 = 1 b2 =  2 → b = 
4 3
a 2 − p 2c 2 = 0
2 2
1 23 1
  − p   =0  p=
2 2 3
2  2 
The general solution,  ( x) = C1 x1/2 J 1   x3/2  + C2 x1/2 J 1   x3/2 
33  33 
−

Applying boundary conditions

  ( 0) = 0  C1 = 0
2 
 ( x) = C2 x1/2 J 1   x3/2 
− 3 
3

2 
 ( L) = 0  C2 L1/2 J 1   L3/2  = 0
− 3 
3

2 
C2  0 ve J 1   L3/2  = 0
− 3 
3

z =  L3/2 then J 1 ( z ) = 0
2
3 −
3

Using table or a software J

−
1 ( z ) = 0 is solved to get z  1.8663
3

2 w 3/2 EI
Lc  1.8663  Lc  1.9863 critical length
3 EI w

31
  2 3 −1 
Example: Find the solution of the differential equation y −   y +  784x 6 − 61  y = 0 .
  x2 
 x 

1 − 2a  a 2 − p 2c 2 
y + y +  b 2c 2 x 2 c −2 + y=0
x  x2 

( )
1 − 2a = − 2 3 − 1 = 1 − 2 3  a = 3

b2c 2 x 2c −2 = 784 x 6  b2c 2 = 784  c=4

784
2c − 2 = 6 b216 = 784 → b = =7
16
a 2 − p 2c 2 = −61
3 − p 216 = −61
− p 216 = −64  p2 = 4  p=2


y ( x) = x a C1 J p (bxc ) + C2Yp (bxc ) 
y( x) = x 3 C1 J 2 (7 x 4 ) + C2Y2 (7 x 4 ) 

STURM –LIOUVILLE THEOREM

y + R( x) y + Q( x) + P( x)y = 0 differential equation (ry) + (q + p) y = 0 …. (1)

can be written as
(ry) + (q +  p) y = 0
ry + r y + (q +  p ) y = 0
r q p
y + y + ( +  ) y = 0  y + R( x) y + Q( x) +  P ( x)  y = 0
r r r

When p, q, r , and r are continuous on (a, b), and r(x) > 0 and p(x) > 0 on (a, b), equation (1)
is called the Sturm-Liouville differential equation.
Three types of boundary conditions can be encountered in the Sturm-Liouville equation in the
interval (a, b). These are

1. B1 y(b) + B2 y(b) = 0 B1  0, B2  0 …. (2)

2. A1 y(a) + A2 y(a) = 0 A1  0, A2  0 …. (3)
3. r (a) = r (b) = 0 …. (4)

32
A1, A2, B1, B2 are real ve independent of .  is eigenvalue, obtained solutions are
eigenfunctions.
Equation (1) is linear and homogeneous. Equations (2) ve (3) are homogeneous.
A2 y(b) + B2 y(b) = C2 (C2  0, constant) boundary condition is not homogeneous. Generally,
the homogeneous differential equation has both equation and boundary conditions are
homogeneous, else differential equation is nonhomogeneous..

Teorem: Properties of Sturm-Liouville problem

1. There is an infinite sequence of eigenvalues n which can be ordered so that

1<2<3<…<n<…,
Further, with this ordering as an increasing sequence n→ as n→
2. For each eigenvalue there is only one eigenfunction (except for nonzero constant
multiples).
3. Eigenfunctions corresponding to different eigenvalues are linearly independent.
4. The set of eigenfunctions corresponding to the set of eigenvalues is orthogonal with respect
to the weight function p(x) on the interval [a, b ].
5. Every eigenvalue of the Sturm-Liouville problem is real.

33
Example: Solve the following Strum-Liouville problem for [0,L] (Find eigenvalues and
eigenfunctions)
y + y = 0 y(0) = y( L) = 0

with y = emx conversion

m2 +  = 0

Case1:  = 0 case
y = 0

Integrating twice
y ( x) = cx + d

Applying boundary conditions

1/ y(0) = 0  d =0

2/ y( L) = 0  c=0

∴  = 0 is not an eigenvalue.
Case 2:  < 0 case ( = -k2 , k > 0)
y + y = 0

y − k 2 y = 0

General solution

y( x) = C1ekx + C2e− kx
Applying boundary conditions
1/ y(0) = 0  C1e0 + C2e0 = 0  C1 = −C2

y ( L) = 0  C1e kL + C2e − kL = 0
2/ C1 ( e kL − e − kL ) = 0
C1 = 0 ve C2 = 0

∴ For  < 0 case there is no eigenvalues.

Case 3: For  > 0 ( = k2 , k > 0)

y + y = 0

y + k 2 y = 0

34
General solution

y( x) = C1 cos(kx) + C2 sin(kx)
Applying boundary conditions
1/ y (0) = 0  C1 cos(0) + C2 sin(0) = 0  C1 = 0
C1 (1) + C2 (0) = 0
2 / y ( L) = 0  C2 sin(kL) = 0

sin(kL) = 0 sin(n ) = 0 n = 1, 2,3,...

kL = n
n
k= n = 1, 2,3,...
L

n2 2
∴ n = 2 n = 1, 2,3,... eigenvalues
L
n x
yn ( x) = Cn sin( ) n = 1, 2,3,... , Cn ≠ 0 a real number. Eigenfunctions.
L

Example: Solve the following Strum-Liouville problem for [0,L] (Find eigenvalues and
eigenfunctions)
y + y = 0 y(0) = y( L) = 0

Case 1:  = 0 case
y = 0

Integrating twice

y( x) = c1 x + c2
Applying boundary conditions
1/ y (0) = 0  c2 = 0

2/ y( L) = 0  c1 = 0

∴  = 0 case has no solutions.

Case 2:  < 0 case ( = -2 ,  > 0)

y + y = 0

y −  2 y = 0

General solution

35
 y( x) = C1e x + C2e− x
Applying boundary conditions
1/ y(0) = 0  C1e0 + C2e0 = 0  C1 = −C2

2 / y( L) = 0
y( x) = C1 e x − C2 e − x = C1 ( e x + e − x )
y( L) = 0  C1 ( e L + e − L ) = 0

C1 = 0 ve C2 = 0

∴  < 0 case have no solution.

Case 3:  > 0 case ( = 2 ,  > 0)
y + y = 0

y +  2 y = 0

general solution

y( x) = C1 cos( x) + C2 sin( x)
applying boundary conditions

1 / y (0) = 0  C1 cos(0) + C2 sin(0) = 0  C1 = 0

2 / y( L) = 0
y( x) = −C1 sin( x) + C2 cos( x) = C2 cos( x)

y( L) = C2 cos( L) = 0

 1 
cos( L) = 0 cos  (n − )  = 0
 2 
1
 L = (n − )
2
1 
 n = (n − ) n = 1, 2,3,...
2 L

1 2 2
∴ n =  = (n − ) 2 n = 1, 2,3,... eigenvalues
2
n
2 L
 1 x
yn ( x) = Cn sin  (n − )  n = 1, 2,3,... eigenfunctions.
 2 L 

36
Example: Solve the following Strum-Liouville problem for [0,/2] (Find eigenvalues and
eigenfunctions)
y + y = 0 y(0) = y( / 2) = 0

37
Example: Solve the following Strum-Liouville problem for [0,/2] (Find eigenvalues and
eigenfunctions)
y + y = 0 y(0) = y( / 2) = 0

Case 1:  = 0 case
y = 0

y( x) = c

y ( x) = cx + d

Applying B.C.s
1/ y(0) = 0  c=0


2/ y( ) = 0  c=0
2

y( x) = d , d = constant,

∴  = 0 eigenvalue, and y( x) = d eigenfunction

Case 2:  > 0 case ( = 2 ,  > 0)

y + y = 0

y +  2 y = 0

General solution of the differential equation

y( x) = C1 cos( x) + C2 sin( x)
Applying B.C.s

y( x) = −C1 sin( x) + C2 cos( x)

1 / y(0) = 0  −C1 sin(0) + C2 cos(0) = 0  C2 = 0

y( x) = −C1 sin( x)

 
2 / y( ) = 0 = −C1 sin( ) = 0
2 2

sin( )=0 sin(n ) = 0
2

 = n
2
 n = 2n

∴ n =  n = 4n n = 1,2,3,... eigenvalues
2 2

38
 yn ( x) = Cn cos(2nx) n = 1, 2,3,... eigenfunctions.

Case 3:  < 0 case ( = -2 ,  > 0)

y + y = 0

y −  2 y = 0

General solution of the differential equation

y( x) = b1e x + b2e− x
Applying B.C.s

y( x) = b1 e x − b2 e− x

1/ y(0) = 0  b1 e0 − b2 e0 = 0  b1 = b2

y( x) = b1 ( e x − e − x )

   2 
− 
2 / y( ) = b1  e − e 2  = 0
2  

b1 = 0 ve b2 = 0

∴  < 0 case does not have any eigenvalue.

39
Example: Find the eigenvalues and eigenfunctions of the given problem
y + y + y = 0 y(0) = 0 , y(1) = 0

−1  12 − 4
m +m+ = 0 →
2
m1,2 =
2
Case 1: 1 − 4  0 case
1 − 4  0
1  4
1

4

1 − 4  0 then m1,2 roots are real and distinct

−1 + 1 − 4 −1 − 1 − 4
m1 = ve m2 =
2 2
general solution

y( x) = C1em1x + C2e−m2 x
applying boundary conditions

1/ y(0) = 0  C1e0 + C2e0 = 0  C1 = −C2

y ( x) = C1 ( e m1x − e − m2 x )

2 / y (1) = 0  C1 ( e m1 − e − m2 ) = 0

C1 = 0 ve C2 = 0

1
∴ For   case, there is no solution.
4
1
Case 2:  = case
4
−1
1 − 4 = 0 then m1,2 =
2
General solution of y + y + y = 0
−1 −1
x x
y ( x) = C1e 2 + C2 xe 2

BCs

1/ y(0) = 0  C1e0 + C2 (0)e0 = 0  C1 = 0

40
 1
− x
y ( x) = C2 xe 2

1
−
2 / y (1) = 0  C2 (1)e 2
=0

C2 = 0

1
∴ = is not an eigenvalue.
4
1
Case 3: 1 − 4  0 case (   case)
4

−1  1 − 4 −1 4 − 1
m2 + m +  = 0 → m1,2 = = i
2 2 2
General solution of y + y + y = 0
1
− x  4 − 1 4 − 1 
y ( x) = e 2
 C1 cos( x) + C2 sin( x) 
 2 2 

Boundary conditions

1/ y (0) = 0  e0 ( C1 cos(0) + C2 sin(0) ) = 0  C1 = 0

1
− x 4 − 1
y ( x) = e 2 C2 sin( x)
2
−
1
4 − 1
2 / y (1) = 0  2
e C2 sin( )=0
2

4 − 1
sin( )=0 sin(n ) = 0
2
4 − 1
= n
2
1
n = + n 2 2 n = 1, 2,3,...
4
1
− x 4 − 1
yn ( x) = e Cn sin(
2
x) n = 1, 2,3,... eigenfunctions
2

41
EXPANSION OF EIGENFUNCTIONS

(ry) + (q + p) y = 0 is the Strum-Liouville differential equation.


f ( x) =  Cm X (m , x) a≤x≤b
m =1

Cm is evaluated as:

- Multiply both sides of the equation with p( x) X (m , x) then

- Integrate between a to b.
b b

 P( x) X (n , x) f ( x)dx = Cm  P( x) X (n , x)dx

2

a a

 P( x) X (n , x) f ( x)dx
Cm = a
b

 P( x) X
2
(n , x)dx
a

Example: Expand f ( x) = L − x in a series of the eigenfunctions of the Sturm-Liouville problem

𝑦 ′′ + 𝜆𝑦 = 0 y(0) = 0 , y′(L) = 0

First we must generate the eigenvalues and eigenfunctions.

For  > 0 y( x) = C1 cos( x) + C2 sin( x)

Applying boundary conditions

1/ y(0) = 0  C1 cos(0) + C2 sin(0) = 0  C1 = 0

2 / y( L) = 0
y( x) = −C1 sin( x) + C2 cos( x) = C2 cos( x)

y( L) = C2 cos( L) = 0

42
  1 
cos( L) = 0 cos  (n − )  = 0
 2 
1
 L = (n − )
2
1 
 n = (n − )
2 L

1 2 2
∴ n =  = (n − ) 2 n = 1, 2,3,... eigenvalues
2
n
2 L
 1 x
yn ( x) = Cn sin  (n − )  n = 1, 2,3,... eigenfunctions.
 2 L 

f ( x) = L − x =  Cn sin( n x)
n =1

If we compare with (ry) + (q + p) y = 0 , p(x) = 1.

L L

 ( L − x ) sin( x)dx =  C sin ( x)dx

2
n n n
0 0
L

 ( L − x ) sin( x)dx n

Cn = 0
L

 sin ( x)dx
2
n
0

L
L
 sin ( n x)dx = ... =
2

0
2
L L L
− sin( n L) L
 ( L − x ) sin( n x)dx = L  sin( n x)dx −  x sin( n x)dx =
0 0 0
 2
n
+
n
 1 
− sin  (n − ) 
=  2 
+
L
 1 
2
 1 
 (n − )   (n − ) 
 2 L  2 L
 2n − 1 
sin  ( ) 
=−  2 + L
( 2n − 1) 
2 2
( 2n − 1) 
4 L2 2L

substituting

43
 4 L2  2n − 1  2 L 
L 2

 ( L − x ) sin( x)dx n
− sin
( 2n − 1)  2  2
2  ( )  +
 ( 2n − 1)
Cn = 0
L
=
L
 sin ( n x)dx
2
2
0

8L  2n − 1  4 L
=− sin  ( )  +
( 2n − 1)   2  ( 2n − 1)
2 2

   2n − 1     2 ( −1) 
n
2
= 4L  − sin  ( )   = 4 L  − 
 ( 2n − 1) ( 2n − 1)   2    ( 2n − 1) ( 2n − 1)  
2 2 2 2

  2 ( −1) 
n

Cn = 4 L  −  n = 1, 2,3,...
 ( 2n − 1) ( 2n − 1)  
2 2


f ( x) = L − x =  Cn sin( n x)
n =1

  2 ( −1)   (2n − 1)  x 
n

f ( x) = L − x = 4 L  −  sin  
n =1  ( 2n − 1) ( − )    2 L 
2 2
 2 n 1

Strum-Liouville Systems and Bessel Functions

d 2 R 1 dR  2  2 
+ + + 2  R = 0 b
dr 2 r dr  r 
r
or
d  dR   2  2 
r  + + 2  R = 0
dr  dr   r 

and on the interval (0,b)

Bessel functions J ( r ) are orthogonal to each other according to w(r ) = r on the interval

0 < r < b, where w(r ) = r is the weight function.

44
Example: Expand f(r) function in a series of the eigenfunctions of the Sturm-Liouville
problem below.

d  dR   2  2 
r  + + 2  R = 0 0r b
dr  dr   r 

B.C. R(0) = finite

R(b) = 0

R(r ) = AJ ( r ) + BY ( r )

Applying boundary conditions

1/ R(0) = finite  B=0

2/ R(b) = 0  J (nb) = 0

ns are the roots of this equation.


f (r ) =  Cn J (n r )
n =1

Above equation is multiplied by rJ (n r ) and integratin 0 to b

b b

 rf (r ) J ( r )dr = C  rJ ( r )dr

2
n n n
0 0

 rf (r ) J ( r )dr n

Cn = 0
b

 rJ ( r )dr
2
n
0

b
b
 r2  2  2 r2  d  
2

0 rJ (n r )dr =  2 1 − r 2n2  J (n r ) + 2  dr J (n r )  

2

 0
b2  2 
2
b2  d 
= 1 − 2 2  J2 (nb) +  J (nb) 
2  b n  2  dr 
b2  
2

=  n J (nb) − J +1 (nb) 
2 b 
b2 2
= J +1 (nb)
2

45
 b b

 rf (r ) J ( r )dr  rf (r ) J ( r )dr
n n
2
b

b J +1 (nb) 0
Cn = 0
b
= 0
= rf (r ) J (n r )dr
b2 2 2 2

 rJ (n r )dr J +1 (nb)

2

0
2


f (r ) =  Cn J (n r )
n =1
  
2
b

f (r ) =   2 2  rf (r ) J (n r )dr  J (n r )
n =1  b J +1 (n b) 0 

FOURIER – LEGENDRE SERIES

𝑑 𝑑𝑦
[(1 − 𝑋 2 ) 𝑑𝑥 ] + 𝑛(𝑛 + 1)𝑦 = 0
𝑑𝑥

𝑓(𝑥) = ∑ 𝐶𝑛 𝑃𝑛 (𝑥)
𝑛=0

Both sides of the above equation is multiplied by Pn(x) and integrating -1 to 1

1 1
∫ 𝑓(𝑥)𝑃𝑛 (𝑥)𝑑𝑥 = 𝐶𝑛 ∫ 𝑃𝑛2 (𝑥)𝑑𝑥
−1 −1

1 1
𝐶𝑛 = 1 ∫−1 𝑓(𝑥)𝑃𝑛 (𝑥)𝑑𝑥
∫−1 𝑃𝑛2 (𝑥)𝑑𝑥

1
2
∫ 𝑃𝑛2 (𝑥)𝑑𝑥 =
−1 2𝑛 + 1

2𝑛 + 1 1
𝐶𝑛 = ∫ 𝑓(𝑥)𝑃𝑛 (𝑥)𝑑𝑥
2 −1

where
1 𝑑𝑛
𝑃𝑛 (𝑥) = 2𝑛 𝑛! 𝑑𝑥 𝑛 (𝑥 2 − 1)𝑛 n = 0, 1, 2, …..

𝑃0 (𝑥) = 1

𝑃1 (𝑥) = 𝑥

46
 1
𝑃2 (𝑥) = (3𝑥 2 − 1)
2
1
𝑃3 (𝑥) = (5𝑥 3 − 3𝑥)
2
𝜋𝑥
Example: Expand 𝑓(𝑥) = cos ( 2 ) in Fourier-Legendre series on the interval -1 ≤ x ≤ 1.


f ( x) =  Cn Pn ( x)
n =0

1 1

 f ( x) P ( x)dx = C  P ( x) P ( x)dx
−1
n n
−1
n n

 f ( x) P ( x)dx n
2n + 1
1

2 −1
Cn = −1
1
= f ( x) Pn ( x)dx
P
2
n ( x)dx
−1

x 
f ( x) = cos( ) =  Cn Pn ( x) = C0 P0 ( x) + C1P1 ( x) + C2 P2 ( x) + C3 P3 ( x) + C4 P4 ( x) + ...
2 n =0

2n + 1 x 1
Cn =
2 −1  cos( ) Pn ( x)dx
2

1
1
x x 1
1
C0 = 
2 −1
cos( ) P0 ( x)dx =  cos( )dx =
2 −1
2 
3
1
x x 1
C1 = 
2 −1
cos( ) P1 ( x)dx =  cos( )( x)dx = 0
2 −1
2
5 x
1
x 1  10( 2 − 12)
1
C2 =  cos( ) P2 ( x)dx =  cos( )  (3 x 2 − 1)  dx =
2 −1 2 −1
2 2  3
7
1
x x 1 1

C3 = 
2 −1
cos( ) P3 ( x)dx =  cos( )  (5 x 2 − 3 X )  dx = 0
2 −1
2 2 
9 x
1
x 1 1
 18( 4 + 1680 − 180 3 )
C4 =  cos( ) P4 ( x)dx =  cos( )  (35 x 4 − 30 x 2 + 3)  dx =
2 −1 2 −1
2 8  5

x
f ( x) = cos( ) = C0 P0 + C1 P1 + C2 P2 + C3 P3 + C4 P4 + ...
2
 x 1 10( 2 − 12)  1 2  18( + 1680 − 180 )  1
4 3

f ( x) = cos( ) = +  (3 x − 1)  +  (35 x − 30 x + 3)  + ...
4 2

2   3
2   5
8 
x 1 5( 2 − 12) 9 ( 4 + 1680 − 180 3 )
f ( x) = cos( )= + (3x 2 − 1) + (35 x 4 − 30 x 2 + 3) + ...
2  3 4 5

47
Example: Expand 𝑓(𝑥) = 7𝑥 4 − 3𝑥 + 1 function in Fourier-Legendre series on the interval
-1 ≤ x ≤ 1.

48
BOUNDARY VALUE PROBLEM IN RECTANGULAR COORDINATE SYSTEM

SEPARABLE PARTIAL DIFFERENTIAL EQUATIONS

In this section we will deal only with the solution of linear partial differential equations.
THEOREM:
u is the dependent variable, x and y are the independent variables, then the general form of the
linear second order partial differential equation is as follows:

𝜕2𝑢 𝜕2𝑢 𝜕2𝑢 𝜕𝑢 𝜕𝑢

𝐴 𝜕𝑥 2 + 𝐵 𝜕𝑥𝜕𝑦 + 𝐶 𝜕𝑦 2 + 𝐷 𝜕𝑥 + 𝐸 𝜕𝑦 + 𝐹𝑢 = 𝐺 (1)

where A, B, C,…., G coefficients are either the constant numbers or the functions x and y.
When G(x,y) = 0, Equation (1) is homogeneous equation, else non-homogeneous equation.

For example,
𝜕2𝑢 𝜕2𝑢 𝜕2 𝑢 𝜕2𝑢
+ 𝜕𝑦 2 = 0 ve − 𝜕𝑦 2 = 𝑥𝑦
𝜕𝑥 2 𝜕𝑥 2

above equations are homogeneous and non-homogeneous equations, respectively.

Classification of Equations
A second order linear partial differential equation is

𝜕 2𝑢 𝜕 2𝑢 𝜕 2𝑢 𝜕𝑢 𝜕𝑢
𝐴 2+𝐵 +𝐶 2+𝐷 +𝐸 + 𝐹𝑢 = 0
𝜕𝑥 𝜕𝑥𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑦

where A, B, C, D, E and F are real number constants (at least one of A, B, C coefficients are
not equal to zero.)
B2 –4AC > 0 hyperbolic
B2 – 4AC = 0 parabolic
B2 – 4AC < 0 elliptic
partial differential equations.

49
SOLUTION OF PARTIAL DIFFERENTIAL EQUATIONS
The solution of a linear partial differential equation (Eq. 1) is a function of two independent
variables and it satisfies the equation on a given xy plane.

Our aim is not to find general solutions to linear partial differential equations. Solving a linear
second order partial differential equation is both difficult and often not useful in applications.
Therefore, in this section, we will find particular solutions of some important linear partial
differential equations that are widely used in practice.

The Separation of Variables Method

Although there are many methods to find the particular solution of a linear partial differential
equation, we are looking for particular solutions in the form of the product of x and y
functions in the separation of variables method.
u(x,y) = X(x)Y(y)

With this definition, sometimes a linear partial differential equation with two independent
variables turns into two ordinary differential equations. To do this

𝜕𝑢 𝜕𝑢 𝜕 2𝑢 𝜕 2𝑢
= 𝑋 ′ 𝑌, = 𝑋𝑌 ′ , = 𝑋 ′′ 𝑌 , = 𝑋𝑌′′
𝜕𝑥 𝜕𝑦 𝜕𝑥 2 𝜕𝑦 2

where ′ shows the derivation.

u dX ( x) dY ( y) dX ( x)
= Y+X = Y = X Y
x dx dx dx
Separation of variables is not a general method. Some partial differential equations cannot be
separated into variables. For example,

𝜕 2 𝑢 𝜕𝑢
− =𝑥
𝜕𝑥 2 𝜕𝑦

Let's consider the second order partial differential equation. For the steady conditions

𝑎1 (𝑥)𝑇𝑥𝑥 + 𝑎2 (𝑥)𝑇𝑥 + 𝑎3 (𝑥)𝑇 + 𝑏1 (𝑦)𝑇𝑦𝑦 + 𝑏2 (𝑦)𝑇𝑦 + 𝑏3 (𝑦)𝑇 = 0

T = T(x,y)

50
 2T
= Txx
x 2
T
= Tx
x
 2T
= Tyy
y 2

It can be written as 𝑇(𝑥, 𝑦) = 𝑋̅(𝑥)𝑌(𝑦).

𝑎1 ̅̅̅̅ ̅ 𝑌 + 𝑎3 𝑋̅𝑌 + 𝑏1 𝑋̅𝑌′′ + 𝑏2 𝑋̅𝑌′ + 𝑏3 𝑋̅𝑌 = 0

𝑋′′𝑌 + 𝑎2 𝑋′

dividing by 𝑋̅𝑌
1 1
(𝑎1 ̅̅̅̅ ̅ + 𝑎3 𝑋̅)
𝑋′′ + 𝑎2 𝑋′ = −(𝑏1 𝑌 ′′ + 𝑏2 𝑌 ′ + 𝑏3 𝑌) = ±𝜆2
𝑋̅ 𝑌
If we write equation for 𝑋̅ and Y separately.

𝑎1 ̅̅̅̅ ̅ + [𝑎3 ± 𝜆2 ]𝑋̅ = 0

𝑋′′ + 𝑎2 𝑋′

𝑏1 𝑌 ′′ + 𝑏2 𝑌 ′ + [𝑏3 ± 𝜆2 ]𝑌 = 0

Separation of variables can be applied if the following conditions are satisfied:

1. The equation in one direction must be homogeneous and the boundary conditions in that
direction must also be homogeneous. The differential equation in the other direction must also
be homogeneous.
2. Accordingly, the sign of λ2 should be selected in the direction where the Sturm-Liouville
equation is formed in the direction that is homogeneous.
The Superposition Theorem
If T1, T2, …., Tk are solutions of a homogeneous linear partial differential equation then

T=C1T1 + C2T2 + …. + CkTk where C’s are constants,

or it can be written as
∞

𝑇(𝑥, 𝑦) = ∑ 𝐶𝑖 𝑇𝑖
𝑖=1

51
Example: Find the solution of the differential equation below

Solution:
If we write as
u(x,y) = X(x)Y(y)
and use this relation in the differential equation

 2u
= X  Y and
x 2
u
= X Y
y

Substituting into the partial differential equation,

 2u u
=4
x 2
y

X  Y = 4 XY 

Dividing by 4XY
X  Y 
=
4X Y

In order to achieve such an equality, each side must equal the same constant in this equation.
If the real constant is denoted by λ
X  Y 
= = −
4X Y
Thus, two ordinary differential equations are obtained.
X  + 4X = 0 and Y  + Y = 0
There are three cases of λ. λ = 0, λ < 0 and λ > 0. In other words, λ = 0, λ =-α2< 0 and
λ = α2 > 0, where α > 0.
Case 1: λ = 0
Two ordinary differential equations can be written and solved
X  = 0 ve Y  = 0

X ( x) = C1 + C2 x and Y ( y ) = C3

52
Thus, the product solution

u ( x, y) = X ( x)Y ( y) = (C1 + C2 x )C3 = A1 + B1 x

u( x, y) = A1 + B1x (1)

where A1 = C1C3 ve B1 = C2C3

Case 2: λ < 0, λ =-α2 < 0

X  + 4X = 0 ve Y  + Y = 0

X  − 4 2 X = 0 ve Y  − 4 2Y = 0

X = C4 cosh(2x) + C5 sinh( 2x) ve Y = C6e

2
y

⸫ u( x, y) = X ( x)Y ( y) = (C4 cosh(2x) + C5 sinh( 2x) )C6e

 2
y

u ( x, y ) = A2 e y cosh(2x) + B2 e y sinh( 2x)

2 2
(2)

where A2 = C4C6 ve B2 = C5C6

Case 3: λ > 0, λ = α2 > 0

X  + 4X = 0 ve Y  + Y = 0

X  + 4 2 X = 0 ve Y  + 4 2Y = 0
the solutions are,

X = C7 cos(2x) + C8 sin( 2x) ve Y = C9e−

2
y

⸫ u( x, y) = XY = (C7 cos(2x) + C8 sin( 2x))C9e

− 2
y

u( x, y) = A3e− y cos(2x) + B3e− y sin( 2x)

2 2
(3)

where A3 = C7C9 ve B3 = C8C9

Equations (1),(2) ve (3) satisfy uxx = 4u y given differential equation.

53
Example: The vertical displacement of the rope is u(x,t) as seen on the figure below. Length
of the rope is L.

 2 u  2u
a2 = 0 < x < L , t >0
x 2 t 2

u(0,t) = 0 u(L,t) = 0 t>0

u
u(x,0) = f(x) = g ( x) 0<x<L
t t = 0

Solve the given boundary value problem.

Çözüm:
If it is written as
u(x,t) = X(x)T(t)

 2u
= X T and
x 2

 2u
= X T 
t 2

Substituting into the partial differential equation

 2 u  2u
a 2
=
x 2 t 2

a 2 X T = XT 
Dividing both sides of the equation with XT
X T XT 
a2 =
XT XT

X  1 T 
= 2 = −
X a T

Thus, two ordinary differential equation is obtained.

54
X  + X = 0 (1)

ve T  + a 2 T = 0 (2)
Boundary conditions:
u(x,t) = X(x)T(t)

1) u(0,t) = X(0)T(t) = 0 → X(0) = 0

2) u(L,t) = X(L)T(t) = 0 → X(L) = 0

Equation (1) with boundary conditions is a Sturm-Liouville problem.

X  + X = 0 , X(0) = 0, X(L) = 0.

λ is checked for three cases as λ = 0, λ < 0 ve λ > 0, in other words, λ = 0, λ =-α2 < 0 ve λ =
α2 > 0, where α > 0.

Case 1: λ = 0
X  = 0
Integrating,

X ( x) = C1 + C2 x
Using B.Cs,

1) X(0) = 0 → X(0) = C1 = 0 → C1 = 0

2) X(L) = 0 → X(L) = C2L = 0 → C2 = 0

⸫ λ = 0 is not an eigenvalue.
Case 2: λ < 0, λ =-α2 < 0, α > 0
X  + X = 0

X  −  2 X = 0
General solution: X ( x) = C3e x + C4e− x

Using B.Cs,

1) X(0) = 0 → X(0) = C3 + C4 = 0 → C4 = - C3

2) X(L) = 0 → X(L) = C3eL + C4 e-L = 0

C3 [eL - e-L]= 0

C3 = 0

55
 C3 = C4 = 0
⸫ For λ < 0 there is no solution
Case 3: λ > 0, λ = α2 > 0, α > 0
X  + X = 0

X  +  2 X = 0
General solution

X ( x) = C5 cos(x) + C6 sin(x)

Using B.Cs,

1) X (0) = 0 → X (0) = C5 = 0 → C5 = 0

2) X ( L) = 0 → X ( L) = C5 cos(L) + C6 sin(L) = 0

C6 sin(L) = 0

sin(L) = 0

L = n

n
n = n = 1,2,3,...
L

n 2 2
Thus, n =  n2 = n = 1,2,3,...
L2

and X ( x) = C6 sin(x)

n
X ( x) = C6 sin( x)
L
n
X n ( x) = C6 n sin( x) n = 1, 2,3,...
L

Solution for T(t)

T  + a 2 T = 0

 n 2 2 
T  + a  2  T = 0
2

 L 
General solution
an an
Tn (t ) = C7 cos( t ) + C8 sin( t) n = 1, 2, 3, ...
L L

56
⸫ u ( x, t ) = X ( x)T (t )

 an an  n
un ( x, t ) =  An cos( t ) + Bn sin( t )  sin( x) n = 1, 2, 3, ...
 L L  L

where An = C6C7 ve Bn = C6C8

and

 an an  n
u ( x, t ) =   An cos( t ) + Bn sin( t ) sin( x)
n =1  L L  L

Using initial conditions,


n
u ( x,0) = f ( x)=  An sin( x)
n =1 L

n n n
L L


0
f ( x) sin(
L
x)dx = An  sin(
0
L
x) sin(
L
x)dx

n
L
2
An =
L0 f ( x) sin(
L
x)dx

u
= g ( x)
t t = 0

u 
  an  n  n  an  n
( x, t ) =   An  −  sin( t ) + Bn   cos( t ) sin( x)
t n =1   L  L  L  L  L

u 
  an  n 
( x,0) = g ( x) =   Bn   sin( x) 
t n =1   L  L 

n  an  n n
L L

 g ( x)sin(
0
L
x)dx = Bn 
 L 0
  sin( x)sin( x)dx
L L

n
L
2
Bn = 
an 0
g ( x) sin(
L
x)dx

Thus,
 
n  an  2 L n  an  n
L
2
u ( x, t ) =    f ( x) sin( x)dx cos( t) +   g ( x ) sin( x ) dx sin( t ) sin( x)
n=1 
 0
L L  L  an 0
L  L 
 L

57
 2 y 2 y
Example: = 4 0 < x < 4 , t >0
t 2 x 2

y(0,t) = 0 y(4,t) = 0 t>0

y( x,0)
y( x,0) = x(x − 2) =1 0 < x <4
t

Solve the given differential equation.

Solution:
y(x,t) = X(x)T(t) is assumed,

2 y
= X  T and
x 2

2 y
= X T 
t 2

Substituting into partial differential equation, and arranging,

X  1 T 
= = −2
X 4 T

Thus, two ordinary differential equation is obtained.

X  + 2 X = 0 (1)

and T  + 42T = 0 (2)

Boundary Conditions:
y(x,t) = X(x)T(t)

1) y(0,t) = X(0)T(t) = 0 → X(0) = 0

2) y(4,t) = X(4)T(t) = 0 → X(4) = 0

Equation (1) with boundary conditions is a Sturm-Liouville problem.
Solution of X(x)
Case 1: λ > 0

X  + 2 X = 0

General solution X ( x) = C1 cos(x) + C2 sin( x)

Using B.Cs,

1) X (0) = 0 → X (0) = C1 = 0 → C1 = 0

2) X (4) = 0 → X (4) = C2 sin( 4 ) = 0

58
 C2 sin( 4 ) = 0
n
n = n = 1, 2, 3,...
4

X ( x) = C2 sin(x)
n
X n ( x) = C2 sin( x) n = 1, 2, 3,...
4

Case 2: λ = 0
X  = 0
Integrating,

X ( x) = C + Dx

Using B.Cs,

1) X(0) = 0 → X(0) = C = 0 → C= 0

2) X(4) = 0 → X(4) = 4D = 0 → D= 0

⸫ λ = 0 , there is no solution.

Solution of T(t)

T  + 42T = 0

 n 
2

T  + 4  T =0
 4 
General solution
n n
Tn (t ) = C3 cos( t ) + C4 sin( t) n = 1, 2, 3,...
2 2
⸫ General solution of y(x,t)
y( x, t ) = X ( x)T (t )

n  n n 
yn ( x, t ) = C2 sin( x) C3 cos( t ) + C4 sin( t )  n = 1, 2, 3, ...
4  2 2 

n n 
n n
y ( x, t ) =  An sin( x) cos( t ) +  Bn sin( x) sin( t )
n =1 4 2 n =1 4 2

where An = C2 C3 ve Bn = C2C4

59
Using I.Cs,
y( x,0)
=1
t

y 
n  n  n 
n  n  n
( x, t ) =  An sin( x)  − sin( t ) +  Bn sin( x)  cos( t )
t n =1 4  2  2 n =1 4  2  2

y 
n  n 
( x,0) = 1 =  Bn sin( x) 
t n =1 4  2 

n  n  n n
4 4

 sin(
0
4
x)dx = Bn    sin(
 2 0 4
x) sin(
4
x)dx

n
4
4 − 4 cos(n )
 sin( 4
x)dx
n 1 − (− 1)
n
 
Bn = 0
= = 4 n = 1, 2, 3, ...
 n 
4
n n  n  sin( 2n )  n 2 2
  sin( x ) sin( x ) dx   2 − 
 2 0 4 4  2  n 

First initial condition y( x,0) = x(x − 2)


n
y ( x,0) = x(x − 2) =  An sin( x)
n =1 4

n n n
4 4

 x(x − 2)sin(
0
4
x)dx = An  sin(
0
4
x) sin(
4
x)dx

4
n  ( )
32 4 + − 4 + ( n ) cos(n ) − 3n sin( n )
2

 x(x − 2)sin( 4
x)dx −
(n )3
An = 0
= ... =
4
n n sin( 2n )
2−
0 sin( 4 x) sin( 4 x)dx n

An = −
 (
16 4 + − 4 + (n ) (− 1)
2
) n

(n )3
Substituting An ve Bn into y(x,t)

y ( x, t ) = − 3 
 (
16  4 + − 4 + ( n ) (− 1) n n
2
) n

4  1 − (− 1) n n  n

2 
sin( x ) cos( t ) + sin( x) sin( t )
 n =1 n 3
4 2  n =1 n 2
4 2

60
Example:

2 y 2 y
= + Ax 0 < x < L , t >0
t 2 x 2

y(0,t) = 0 y(L,t) = 0 t>0

y( x,0)
y( x,0) = 0 =1 0 < x <4
t

For the given boundary value problem A is a positive number and Ax is a force. Find the
solution of the given boundary value problem.
Solution:
y(x,t) = X(x)T(t)

2 y
= X T  and
t 2

2 y
= X  T
x 2
If we put into the differential equation,
XT  = X T + Ax
Dividing by XT
T  X  Ax
= +
T X XT
There is no way to separate the t dependency on one side of an equation and the x dependency
on the other. One strategy in such a case is to try to transform this problem into one to which
separation of variables applies.

y( x, t ) = Y ( x, t ) + ( x)
 2 y  2Y  2 y  2Y  2
= and = +
t 2 t 2 x 2 x 2 x 2

The idea is to choose  ( x) to obtain a problem for Y(x,t) that we can solve. Substitute y(x,t)
into the partial differential equation to get

 2Y  2Y  2
= 2 + 2 + Ax
t 2 x x
This equation can be written as two parts:

61
  2Y  2Y  2
= 2 ve + Ax = 0
t 2 x x 2

Also, initial and boundary conditions are needed to be transformed

Boundary conditions

1) y(0, t ) = 0 → y(0, t ) = Y (0, t ) + (0) = 0 → Y (0, t ) = 0 ve  (0) = 0

2) y( L, t ) = 0 → y( L, t ) = Y ( L, t ) + ( L) = 0 → Y ( L, t ) = 0 ve  ( L) = 0
Initial conditions

3) y( x,0) = 0 → y( x,0) = Y ( x,0) + ( x) = 0 → Y ( x,0) = − ( x)

y y Y
4) ( x,0) = 1 → ( x,0) = ( x,0) = 1
t t t

Two differential equations with boundary conditions

 2Y  2Y d 2
= 2 + Ax = 0
t 2 x dx2
Y (0, t ) = 0  (0) = 0

Y ( L, t ) = 0  ( L) = 0
Y ( x,0) = − ( x)

Y ( x,0)
=1
t

First (x) should be solved .

Solution of

d 2 Ax 3
+ Ax = 0 is  ( x) = − + cx + d
dx2 6
Applying B.C.s,
 (0) = 0  d = 0 d = 0,

AL2
 ( L) = 0  c=
6

 ( x) = Ax(L2 − x 2 )
1
6

62
Solution of Y(x,t)

 2Y  2Y
= 2
t 2 x

Boundary and initial conditions:

Y ( x,0)
Y (0, t ) = 0, Y ( L, t ) = 0, Y ( x,0) =
1
6
( )
Ax x 2 − L2 ,
t
=1

Y(x,t) = X(x)T(t), using separation of variables

X  + 2 X = 0 and T  + 2T = 0
Boundary conditions:

1) Y(0,t) = X(0)T(t) = 0 → X(0) = 0

2) Y(L,t) = X(L)T(t) = 0 → X(L) = 0

Solution for X(x)

Case 1: λ = 0
X  = 0
Integrating twice,

X ( x) = C + Dx

B.Cs,

1) X(0) = 0 → X(0) = C = 0 → C= 0

2) X(L) = 0 → X(L) = DL = 0 → D= 0

⸫ λ = 0 is not a solution.

Case 2: λ > 0

X  + 2 X = 0

X ( x) = C1 cos(x) + C2 sin( x)

B.C.s

1) X (0) = 0 → X (0) = C1 = 0 → C1 = 0

63
2) X ( L) = 0 → X ( L) = C2 sin( L) = 0

C2 sin( L) = 0
thus
n
n = n = 1, 2, 3,...
L

and X ( x) = C2 sin(x)
n
X n ( x) = C2 sin( x) n = 1, 2, 3,...
L

Solution for T(t)

T  + 2T = 0

 n 
2

T  +   T =0
 L 
General solution
n n
Tn (t ) = C3 cos( t ) + C4 sin( t) n = 1, 2, 3,...
L L
⸫ Y ( x, t ) = X ( x)T (t )

n  n n 
Yn ( x, t ) = C2 sin( x) C3 cos( t ) + C4 sin( t )  n = 1, 2, 3, ...
L  L L 

n n 
n n
Y ( x, t ) =  Cn sin( x) cos( t ) +  Dn sin( x) sin( t )
n =1 L L n =1 L L

where Cn = C2 C3 ve Dn = C2C4

Applying I.C.s

Y ( x,0) = − ( x) = Y ( x,0) =
1
6
(
Ax x 2 − L2 )
n
( )

1
Y ( x,0) = Ax x 2 − L2 =  Cn sin( x)
6 n =1 L

n n n
( )
L L
1
0 6 Ax x − L sin( L x)dx = Cn 0 sin( L x) sin( L x)dx
2 2

64
 n
 6 Ax(x )
L
1 2
− L2 sin( x)dx
2 1 n 
( )
L
L
Cn = =   Ax x − L sin(
2 2
0
x)dx
n
4
L 0 6 L 
0 sin ( 4 x)dx
2

Y ( x,0)
Other I.C. =1
t

Y 
 n  n n 
 n  n n
( x, t ) =  C n  −  sin( x ) sin( t ) +  Dn   sin( x) cos( t )
t n=1  L  L L n=1  L  L L

Y 
 n  n
( x,0) = 1 =  Dn   sin( x)
t n=1  L  L

n  n  2 n
L L

 sin(
0
L
x)dx = Dn    sin (
 L 0 L
x)dx

 2  n
L
Dn =    sin( x)dx
 n  0 L

Substituting Cn and Dn into Y(x,t)

Y ( x, t ) =
2 AL3  (− 1)
 sin(
n
n
x ) cos(
n
t ) +
2 L  1 − (− 1)


sin(
n
x
n
)

sin(
n
t)
 3 n=1 n3 L L  2 n=1 n 2 L L

 y( x, t ) = Y ( x, t ) + ( x)

2 AL3  (− 1)
n
n n 
2 L  1 − (− 1) n
n
 n 1
( )
3  2 
y ( x, t ) = sin( x ) cos( t ) + sin( x) sin( t ) + Ax L2 − x 2
 n=1 n 3
L L  n=1 n 2
L L 6

65
Example: Consider a rod of length L with on both sides for t > 0 temperature is zero and
initial temperature is f(x) along the rod. Calculate the temperature distribution inside the rod.
Use the separation of variables methor in the solution. (Note: One dimensional transient
conduction problem)

 2u u
k = 0 < x < L , t >0
x 2 t

u(0,t) = 0 u(L,t) = 0 t>0

u(x,0) = f(x) 0<x<L

2   n   −kn 2 2  n
L
u ( x, t ) =    f ( x) sin( x)dx  exp  2
t  sin( x)
L n =1  0 L   L  L

Numerical values:

If u(x,0) = 100, L = , k = 1

 1 − ( −1)n 
 exp ( −n2t ) sin(nx)
200
u ( x, t ) =  
 n=1  n 

Example:
Consider the two-dimensional temperature distribution u(x,y) in a rectangular plate under
steady state conditions. If the plate is insulated at x = 0 and x = a, and on the upper and lower
surfaces (y = b and y = 0) temperatures be f(x) and 0, respectively.

The boundary value problem for this plate is as follows:

66
 2u  2u
+ =0
x 2 y 2

u u
=0 =0 0 < y <b
x x =0 x x=a

u(x,0) = 0 u(x,b) = f(x) 0<x<a

Find the expression that gives the temperature distribution on the plate for the given boundary
value problem.

Solution:
Using u(x,y) = X(x)Y(y)

 2u
= X  Y and
x 2

 2u
= X Y 
y 2

If subtitution and arrangements are done

X  + X = 0 and Y  − Y = 0
Boundary conditions:
u
1) (0, y) = 0  X (0)Y ( y) = 0  X (0) = 0
x

u
2) (a, y) = 0  X (a)Y ( y) = 0  X (a) = 0
x

3) u(x,0) = X(x)Y(0) = 0 → Y(0) = 0

solution for X(x)

The first differential equation with these boundary conditions is the Sturm-Liouville problem.
So,
X  + X = 0 , X’(0) = 0, X’(a) = 0.

There are three cases of λ. λ = 0, λ < 0 and λ > 0. In other words, λ = 0, λ =-α2 < 0 ve λ = α2
> 0, where α > 0.

Case 1: λ = 0
67
X  = 0
Integrating,

X ( x) = C2

X ( x) = C1 + C2 x

Using boundary conditions,

1) X’(0) = 0 → X’(0) = C2 = 0 → C2 = 0

2) X’ (a) = 0 → X’(a) = C2 = 0 → C2 = 0
⸫ X(x) = C1 then λ = 0 is a solution.

Case 2: λ < 0, λ =-α2 < 0

X  + X = 0

X  −  2 X = 0
General solution: X ( x) = C3ex + C4 e −x

Applying boundary conditions

X ( x) = C3ex − C4e −x

1) X (0) = 0 = C3 − C4  C3 = C4

(
2) X (a) = 0 = C3 ea − e −a )  C3 = 0

C3 = C4 = 0
⸫ λ < 0 is not solution.

Case 3: λ > 0, λ = α2 > 0

X  + X = 0
X  +  2 X = 0
General solution
X ( x) = C5 cos(x) + C6 sin(x)

Applying boundary conditions

X ( x) = −C5 sin(x) + C6 cos(x)

68
1) X (0) = −C5 sin( 0) + C6 cos(0) = 0  C6 = 0

2) X (a) = −C5 sin( a ) = 0

C5 sin( a ) = 0

sin(a) = 0

a = n
n
n = n = 1,2,3,...
a

n2 2
Therefore, n =  n2 = n = 1,2,3,...
a2
n
and X n ( x) = C5 cos( x) n = 1, 2, 3,...
a

solution for Y(y):

Y  − Y = 0 3) Y(0) = 0

Case 1: λ = 0
Y  = 0
Integrating,

Y ( y) = C1y + C2
Applying boundary conditions,

3) Y (0) = 0 = C10 + C2  C2 = 0

Y ( y) = C1y
Case 3: λ > 0, λ = α2 > 0
Y  − Y = 0
Y  −  2Y = 0

General solution

Y ( y) = C3 sinh(y) + C4 cosh(y)
Applying boundary conditions

3) Y (0) = C3 sinh( 0) + C4 cosh(0) = 0  C4 = 0

69
  n 
Yn ( y ) = C3 sinh( n y ) = C3 sinh  y n = 1, 2, 3, ...
 a 

Thus,

un ( x, y) = X n ( x)Yn ( y)

u0 ( x, y) = C2 C1 y  = A0 y

  n    n   n   n 
un ( x, y ) = C5 cos x  C3 sinh  y  = An cos x  sinh  y
  a    a   a   a 


 n   n 
u ( x, y ) = A0 y +  An cos x  sinh  y
n =1  a   a 

Applying the nonhomeogeneous 4th B.C. u ( x, b) = f ( x)


 n   n 
u ( x, b) = f ( x) = A0b +  An cos x  sinh  b 
n=1  a   a 
The above equation is the half-range expansion of the f(x) function in the Fourier cosine
series.
a0  n 
A0b = and An sinh  b  = an n = 1, 2,3,...
2  a 

2p
f ( x)dx and an = 2  f ( x) cos n 
p
using a0 = 
p0
x dx
p0  p 
a a
2 1
a 0 ab 0
2 A0b = f ( x)dx  A0 = f ( x)dx

 n  2  n   n 
a a
2
An sinh  b  =  f ( x) cos x dx  An =  f ( x) cos x dx
 a  a0  a   n  0  a 
a sinh  b
 a 

 
1 a
  
2
a
 n    n   n 
u ( x, y ) =   f ( x)dx y +    f ( x) cos x dx cos x  sinh  y
 ab 0  n =1   n   a    a   a 
a sinh  b  0
  a  

70
Example: Superposition Theorem

 2u  2u
+ =0
x 2 y 2

u(0,y) = F(y) u(a,y) = G(y) 0 < y <b

u(x,0) = f(x) u(x,b) = g(x) 0<x<a
Solve the given boundary value problem.
When two boundary conditions, which are parallel, are homogeneous, the problem can be
solved by the separation of variables method. In this case, the problem can be divided into
two problem parts.
Problem 1:

 2 u1  2 u1
+ =0 0  x  a, 0 yb
x 2 y 2

u1 (0, y ) = 0, u1 (a, y ) = 0 0 yb

u1 (x,0) = f ( x), u1 (x, b ) = g ( x) 0 xa

Problem 2:

 2u2  2u 2
+ 2 =0 0  x  a, 0 yb
x 2 y

u 2 (0, y ) = F ( y ), u 2 (a, y ) = G ( y ) 0 yb

u 2 (x,0) = 0, u 2 ( x, b ) = 0 0 xa

In this case, let us assume that the solutions of Problem 1 and Problem 2 are u1 and u2. In this
case, u(x,y) = u1(x,y) + u2(x,y). All boundary conditions satisfy the boundary conditions of
the main problem. For example,

u (0, y ) = u1 (0, y ) + u2 (0, y ) = 0 + F ( y ) = F ( y ),

u (x, b ) = u1 (x, b ) + u2 (x, b ) = g ( x) + 0 = g ( x)

This method is called the superposition method.

71
Solution of Problem 1

 2 u1  2u1
+ =0 0  x  a, 0 yb
x 2 y 2

u1 (0, y ) = 0, u1 (a, y ) = 0 0 yb

u1 (x,0) = f ( x), u1 (x, b ) = g ( x) 0 xa


  n   n   n   n 
u1 ( x, y ) =   An sin  x  sinh  y  + Bn sin  x  cosh y 
n=1   a   a   a   a 

 n 
a
2
Bn =  f ( x) sin  x dx
a0  a 

1 2 a  n   n 
An =   g ( x) sin  x dx − Bn cosh b 
 n  a 0  a   a 
sinh  b
 a 

Solution of Problem 2

 2u2  2u 2
+ 2 =0 0  x  a, 0 yb
x 2 y

u 2 (0, y ) = F ( y ), u 2 (a, y ) = G ( y ) 0 yb

u 2 (x,0) = 0, u 2 ( x, b ) = 0 0 xa


  n   n   n   n 
u2 ( x, y ) =  Cn cosh x  sin  y  + Dn sinh  x  sin  y 
n=1   b   b   b   b 

 n 
b
2
Cn =
b0 F ( y ) sin  y dy
 b 

1 2 b  n   n 
Dn =   G( y) sin  y dy − Cn cosh b 
 n   b 0  b   a 
sinh  a
 b 

 u ( x, y) is obtained as u( x, y) = u1 ( x, y) + u2 ( x, y)

72
Example: Fourier Series Solutions of Heat Transfer Equations
Let us consider the boundary value problem for the temperature distribution in the insulated
rod with no internal energy generation. The differential equation and the boundary conditions
that give the temperature distribution inside the rod are as follows:

T  2T
=a 2 0 < x < L , t >0
t x

Tx (0, t ) = 0 Tx ( L, t ) = 0 T ( x,0) = x

T ( x, t ) = + 
 e

L  2 L (− 1) − 1 − an2 2t / L  nx 
n
cos 
2 n =1 n 2 2  L 

73
Example: Fourier Series Solutions of Heat Transfer Equations

 2u  2u
+ =0 0 < x < , 0 < y < 
x 2 y 2

u(0,y) = 0 u(,y) = 0 0<y <

u(x,0) = f(x) u(x, ) = 0 0<x<

a 
  f ( x) sin (nx)dx 
2   
u ( x, y ) =   0
 n =1  (e − 2 n
−1)  (
sin (nx) e − 2 n e ny − e − ny )

 
 

Example: Fourier Series Solutions of Heat Transfer Equations

Solve the following differential equation where u(x,t) is temperature.

u  2u
=a 2 0 < x < L , t >0
t x

u x (0, t ) = 0
u ( L, t )
k + hu( L, t ) = 0
x
u ( x,0) = f ( x) = 100 (1 − x)

Solution:
If u(x,t) is writtten as u(x,t) = X ( x)(t )

X = X 
X  1 T 
= = −2
X aT

Solution for X(x)

X  + 2 X = 0
1) X (0) = 0
2) kX ( L) + hX ( L) = 0
Case 1: λ > 0

X  + 2 X = 0

74
General solution
X ( x) = A cos(x) + B sin( x)

X ( x) = − A sin( x) + B cos(x)

Applying boundary conditions

1) X (0) = 0
X (0) = 0 = − A sin( 0) + B cos(0)  B=0

X ( x) = A cos(x)

2) kX ( L) + hX ( L) = 0

−kA sin( L) + hA cos( L) = 0

kA sin( L) = hA cos( L)
h hL 1
tan(n L) = =
k n k n L
H hL
tan(n L) = where H =
n L k
ns are the roots of the above equation.

Case 2: λ = 0
X  = 0

Integrating twice
X ( x) = C + Dx

Applying boundary conditions,

1) X (0) = 0 D=0

2) kX ( L) + hX ( L) = 0  hC = 0 C =0
⸫λ=0 There is no solution.

Solution for (t)

 + a2n  = 0

(t ) = C exp(−a2nt )

75
un (x,t) = X n ( x) n (t ) = ( A cos(n x) ) C exp(−an2t )

u(x,t) =  an e− an t cos(n x)
2

n =1

Initial condition
u(x,0 ) = f ( x) = 100(1 − x)

100 (1 − x) =  an e − an t cos(n x)
2

n =1

L
100 (1 − x ) cos(n x )dx
2
L 0
an =

200  sin (n L )

(1 − L ) − cos(2 n L ) + 12 

an = 
L  n n n 

200  sin (n L )

(1 − L ) − cos(2 n L ) + 12 e − ant cos(n x)


 u(x,t) = 
2


n =1 L  n n n 

Example: Solve the following differential equation.

u  2u
=k 2 0 < x < L , t >0
t x

u (0, t ) = T1 t 0
u ( L, t ) = T2 t 0
u ( x,0) = f ( x) 0 x0

Solution:
u(x,t) = X(x)T(t),
y
= X T and
t

2 y
= X  T
x 2
Substituting into the partial differential equation,
XT  = kX T
Dividing by XT

1 T  X 
= = −
k T X
X  + X = 0 and T  + kT = 0

76
Initial and boundary conditions:

1) u(0,t) = X(0)T(t) = T1 → X(0) = T(t) / T1

2) u(L,t) = X(L)T(t) = T2 → X(L) = T(t) / T2

3) u(x,0) = X(x)T(0) = f(x) → T(0) = f(x)/X(x)

There is no homogeneous boundary condition in the x direction. Transformation is required to

apply the separation of variables method.

u( x, t ) = U ( x, t ) + ( x)

u U  2u  2U  2
= and = +
t t x 2 x 2 x 2

U   2U  2 
= k  + 
t  x
2
x 2 

This equation can be written as two parts:

U  2U  2
=k ve =0
t x 2 x 2

Boundary and initial conditions are needed to be transformed

1) u(0, t ) = T1 → u(0, t ) = U (0, t ) + (0) = T1 → U (0, t ) = 0 ve  (0) = T1

2) u( L, t ) = T2 → u( L, t ) = U ( L, t ) + ( L) = T2 → U ( L, t ) = 0 ve  (L) = T2

3) u ( x,0) = f ( x) → u( x,0) = U ( x,0) + ( x) = f ( x) → U ( x,0) = f ( x) − ( x)

Solution for (x):

d 2
=0  (0) = T1 , ( L) = T2
dx2

Integrating and applying B.Cs

T −T 
 ( x) =  2 1  x + T1
 L 

77
Solution for U(x,t)

U  2U
=k
t x 2

1) U (0, t ) = 0
2)U ( L, t ) = 0
 T −T 
3) U ( x,0) = f ( x) − ( x) = f ( x) −  2 1  x − T1
 L 
U(x,t) is solved using separation of varibles

n  − kn 2 2t 
U ( x, t ) =  Cn sin( x) exp  2

n =1 L  L 

2   T2 − T1   n 
L
Cn =    f ( x) −   x − T1  sin( x)dx
L 0   L   L 


2  T −T   n  n  − kn 2 2t 
L
U ( x, t ) =     f ( x) −  2 1  x − T1  sin( x)dx sin( x) exp 
 L 
2
n =1 L  0   L  L  L 


u ( x , t ) = U ( x, t ) +  ( x )
  2  L   T2 − T1   n  n  − kn 2 2t   T2 − T1 
u ( x, t ) =     f ( x) −   x − T1  sin( x)dx sin( x) exp  +   x + T1
n =1   L    L 
2
L 0   L  L  L

78
 COMPLEX NUMBERS

A complex number is a symbol x + iy, or x + yi, where x and y are real numbers and i2=−1.

Arithmetic of complex numbers:

Equality a +ib=c+id exactly when a =c and b=d.

Addition (a +ib)+(c+id)=(a +c)+i (b+d).
Multiplication (a +ib)(c+id)=(ac−bd)+i (ad +bc).

The magnitude of complex numbers:

The magnitude of a + bi is denoted by a + bi and is calculated as a + bi = a 2 + b2

Example:

2 − 5i = (2) 2 + (−5) 2 = 4 + 25 = 29

Conjugate of complex numbers

The complex conjugate, or just conjugate of x +iy is the complex number x −iy with the sign
of the imaginary part reversed. Denote the conjugate of a + bi as a + bi

Example: 3 − 8i = 3 + 8i

i = −i

25 = 25

Division operation of complex numbers:

a + bi a + bi ac − adi + bci − bdi2 ac + bd + i (bc − ad )

= = =
c + di c + di c 2 − d 2i 2 c2 + d 2
(c − di)

Example:

2 − 7i 2 − 7i 16 − 6i − 56i + 21i 2 −5 − 62i 5 62

= = = =− − i
8 + 3i 8 + 3i 64 − 9i 2
73 73 73
(8 − 3i )

79
Argument and polar form of complex numbers
a = r cos
b = r sin 

r = a 2 + b2
b
 = tan −1 
a

z = a + ib = r(cos + i sin  )

ei = cos + i sin 

z = re i (Polar form of z complex number)

Example:
Write the complex number z = −1 − i in polar form

5
z = 2 exp( i)
4

80
 COMPLEX FUNCTIONS
A complex function is a function that acts on complex numbers and produces complex
numbers.

Limit:

lim f ( z ) = L and lim f ( z) = lim− f ( z) = lim+ f ( z) is to be satisfied.

z → z0 z → z0 z → z0 z →z0

z 2 zi
Example: f ( z ) =  lim f ( z ) = ?
0 z=i z →i

If lim f ( z ) = lim− f ( z ) = lim+ f ( z ) is satisfied, there is limit.

z →i z →i z →i

lim− f ( z ) = lim− z 2 = i 2 = −1
z →i z →i

lim+ f ( z ) = lim+ z 2 = i 2 = −1
z →i z →i

⸫ lim f ( z ) = −1
z →i

Continuity:

If lim f ( z ) = f ( z 0 ) then f(z) is continuous at z0.

z → z0

Differentiability:
The complex derivative is modeled after the real derivative. We say that f is differentiable
at z0 if for some number L
f ( z 0 + h) − f ( z 0 )
f ( z ) = lim =L
h →0 h

81
Example: Find the derivative of f ( z ) = z 2 ( f (z ) =?).

f ( z 0 + h) − f ( z 0 )
f ( z ) = lim
h→0 h
( z + h) − ( z ) 2
2
z 2 + 2 zh + h 2 − z 2

f ( z ) = lim = lim = lim (2 z + h ) = 2 z
h→0 h h→0 h h→0

ANALYTICAL FUNCTIONS
In a complex plane, if there is a derivative at a point, this function f(z) is called an analytic
function.

Cauchy-Reimann Equations
Let f (z)=u(x, y)+iv(x, y) be differentiable at z =x +iy. Then, at (x, y)
u v ve v u
= =−
x y x y

Example: Check if the function f(z)=z 2 is analytical or not.

f(z)=z 2 = (x + iy ) = x 2 − y 2 + 2 xyi = u( x, y) + iv( x, y)

2

u ( x, y) = x 2 − y 2 v( x, y) = 2 xy

u 
= 2x 
x
 u v
 =
v  x y
= 2x 
y 

v 
= 2y 
x
 v u
 =−
u  x y
= −2 y 
y 

u v and v u conditions are satisfied therefore f(z) is an analytical function.

= =−
x y x y

82
Exponential and Trigonometric Functions

e z = e x+iy = e x cos y + i sin y 

e i = cos + i sin 

cos =
2

1 −i
e + ei 

cos z =
2

1 −iz
e + eiz 

sin z =
2i

1 iz −iz
e −e 

Example: e 2−i = ?

e 2−i = e 2e −i = e 2 cos(− ) + i sin( − )

cos(− ) = cos
sin( − ) = − sin 

e 2−i = e 2 cos(− ) + i sin( − ) = e2 cos − i sin   = e 2 − 1 = −e 2

Example: cosi = ?

cosi =
2

1 −i 2 1
 
e + ei = e + e−1 = 1.543
2

2


 
Example: sin  + i ln 2  = ?
2 
 
  1  i − ln 2 − i + ln 2  1  2 i − ln 2 
− i 
sin  + i ln 2  = e 2 −e 2  = e e − e 2 eln 2 
2  2i   2i  

i  
e 2 = cos + i sin =i
2 2

− i    
e 2
= cos(− ) + i sin( − ) = cos − i sin = −i
2 2 2 2
eln 2 =2

83
   1 i  1 5
sin  + i ln 2  =  − (−i )2 = + 1 =
2  2i  2  4 4

Hyperbolic Functions

sinh z =
2

1 z
e − e−z 

cosh z =
2

1 z
e + e− z 
sinh( iz ) = i sin z

cosh(iz ) = cos z

sinh z
tanh z =
cosh z

d
sinh z = cosh z
dz
d
cosh z = sinh z
dz

Complex Logarithm:

ln z = ln( re i ) = ln r + i

Example: ln (1 + i) = ?

z = 1+ i
r= 2

= + 2n
4

( + 2 n ) i
z = rei = 2e 4


 ( + 2 n ) i  
ln (1 + i ) = ln  2e 4  = ln 2 +  + 2n i
 4 
 

84
COMPLEX INTEGRATION

The Integral of a Complex Function

On the interval [a,b] integral of f(x)=u(x)+iv(x ) is evaluated as

b b b

 f ( x)dx =  u ( x)dx + i  v( x)dx

a a a

2
Example: f(x)=x − ix 2 1 x  2 Evaluate  f ( x)dx .
1

2 2 2 2
3 7
 f ( x)dx =  ( x − ix )dx =  xdx − i  x 2 dx = − i
2

1 1 1 1
2 3

 /4
Example: f(x)= cos(2 x) + i sin( 2 x) 0  x   / 4 , evaluate  f ( x)dx .
0

 /4  /4  /4  /4

 f ( x)dx =  (cos(2 x) + i sin( 2 x))dx =  cos(2 x)dx + i  sin( 2 x)dx = 2 + 2 i

1 1
0 0 0 0

Cauchy Theorem
If a function f(z) is continuous at every point in and above the C-closed region and is a
differentiable function then

 f ( z)dz = 0
C

where  the oval on the integral is a reminder that the path is closed.
C

Contour Integral

I =  f ( z )dz
C

f(z)=u(x, y)+iv(x, y )

85
 z = x + iy
dz = dx + idy

I =  f ( z )dz =  u(x, y)+iv(x, y )dx + idy

C C

I =  u(x, y)dx − v(x, y)dy + i  v(x, y)dx + u(x, y)dy

C C

Example: The graph of z(t)=(5-t)+(t-2)i 2 ≤ t ≤ 3 is as

given in the figure. Evaluate  zdz .
L

z(2) = 3 ve z(3) = 2+i

I =  zdz =  ( 5 - t)+(t - 2 )i) −1 + i  dt

L L
3
=   −3 + 7i − 2ti  dt = −3 + 2i
2

Example:

C curve consists of lines. AB + BD . Evaulate  z dz .

2

z 2 = ( x + iy ) 2 = x 2 − y 2 + i 2 xy
dz = dx + idy

 z dz =  x    
− y 2 + i 2 xy dx + idy =  ( x 2 − y 2 )dx − 2 xydy + i  2 xydx + ( x 2 − y 2 )dy 
2 2

C C C C

 z dz = I + I BD
2
AB
C

Along AB y = 1 and dy = 0

5
 x3 
 
5 5
=  ( x − 1)dx + i  2 xdx =  − x + i x 2
5
I AB 2
−1 = 36 + i 24
−1 −1 3  −1

86
Along BD x = 5 and dx = 0

3
 y3 
  ( 
3 3
124
I BD =  − 10 ydy + i  25 − y dy = − 5 y
2 2 3
1 + i − 40 +  = −40 + i
1 1  3 1 3

124 196
 z dz = I + I BD = 36 + i 24 − 40 + i = −4 + i
2
AB
C
3 3

Example:

Let C be the region (closed path) that defines the

square.

Evaluate  ( z + 1)dz .
C

 ( z + 1)dz =  ( x + iy + 1)(dx + idy) =  ( x + 1)dx − ydy + i  ydx + ( x + 1)dy

C C C C

 ( z + 1)dz =  ( z + 1)dz +  ( z + 1)dz +  ( z + 1)dz +  ( z + 1)dz = I

C OA AB BD DO
OA + I AB + I BD + I DO

Along OA y = 0 and dy = 0
1 1 1

 ( x + 1)dx − ydy+ i   ydx + ( x + 1)dy =  ( x + 1)dx =

3
 ( z + 1)dz =
OA 0 0 0
2

Along AB x = 1 and dx = 0
1 1
1
 ( z + 1)dz =  − ydy + i  2dy = −
AB 0 0
2
+ 2i

Along BD y = 1 and dy = 0
0 0
3
 ( z + 1)dz =  ( x + 1)dx + i  dx = −
BD 1 1
2
−i

87
Along DO x = 0 and dx = 0
0 0
1
 ( z + 1)dz =  − ydy + i  dy = 2 − i
DO 1 1

3 1 3 1
 ( z + 1)dz = I
C
OA + I AB + I BD + I DO = − + 2i − − i + − i = 0
2 2 2 2

Cauchy Integral Formula

Let z0 be on the C-closed curve. Thus

f ( z)
 z−z 0
dz = 2if ( z0 )

sin( z)
Example: Evaluate 
C
z
dz .

z0 = 0
f ( z) sin( z )
z−z
C 0
dz = 2if ( z0 ) = 
C
z
dz = 2i sin( 0) = 0

88
 z 2 − 4z + 4
Example: Since the C-region is the circle with z = 2 , calculate the integral  dz
C z+i
using the Cauchy integral formula.

z 2 − 4z + 4
 
C z + i dz = 2if (−i) = 2i (−i) − 4(−i) + 4 = 2i(3 + 4i) = −8 + 6i
2

2
eZ
Example: Evaluate the integral  dz
C
z −i

a) if C closed path does not enclose i,

b) if C closed path encloses i.
a) If the C region does not enclose i, according to the Cauchy theorem, the integral must be
differentiable on and within the C- closed path. Therefore,

2
eZ

C
z −i
dz = 0

c) b) if C closed path encloses i.

f ( z) = e z
2

2
eZ
C z − i dz = 2if (i) =2ie = 2ie
i2 −1

89
 2z + 1
Example: Evaluate  z 3 − iz 2 + 6 z dz , C region is the circle with r= 1/3 radius center at
C

z=3i.

2z + 1 2z + 1
z
C
3
− iz + 6 z
2
dz = 
C
z ( z + 2i )( z − 3i )
dz

z =0, z = −2i, z = 3i

2z + 1 A B C
= + +
z − iz + 6 z z z + 2i z − 3i
3 2

1 −6−i 4i − 1
 A= , C = , B=
6 15 10
2z +1 A B C
= + +
z − iz + 6 z z z + 2i z − 3i
3 2

2z +1 A ( z + 2i )( z − 3i ) + B ( z )( z − 3i ) + C ( z )( z + 2i )
=
z − iz + 6 z
3 2
z 3 − iz 2 + 6 z

2 z + 1 = A ( z + 2i )( z − 3i ) + B ( z )( z − 3i ) + C ( z )( z + 2i )

z = 0 için 1 = −6i 2 A
z = 3i için 6i + 1 = C (3i )(5i )
z = −2i için −4i + 1 = B(−2i)(−5i)
1 −6 − i 4i − 1
 A= , C = , B=
6 15 10
1 4i − 1 −6 − i
2z +1

C
z − iz 2 + 6 z
3
dz = 
C
6 dz +
z 
C
10 dz +
z + 2i 
C
15 dz
z − 3i
1 dz 4i − 1 dz −6 − i dz
=
6 
C
z
+ 
10 C z + 2i
+
15  z − 3i
C

−6 − i dz  −6 − i 
=
15  z − 3i dz = 
C
15 
 2 i

dz
f ( z ) = 1 → f (3i ) = 1  z − 3i dz = 2 i(1)
C

dz dz

C
z
=0 and  z + 2i = 0, because
C
z = 0 and z = −2i are out of C-region.

90
THEOREM:
If z0 point is enclosed by C-closed path

f ( n ) ( z0 ) =
n! f ( z)

2i C ( z − z0 )n +1
dz

or

2i ( n )
f (z0 )
f ( z)
 (z − z )
C 0
n+1
dz =
n!

sin( 2 z )
Example:  (6 z −  )3 dz = ?
C

sin( 2 z ) sin( 2 z ) 1 sin( 2 z )

 (6 z −  ) dz = 
3
 
3
dz =
63 C  
3
dz
C C
63  z −  z − 
 6  6

z0 = , n +1 = 3 → n = 2
6
f ( z ) = sin( 2 z )
f ( z ) = 2 cos(2 z )
f ( z ) = −4 sin( 2 z )

1 sin( 2 z ) 1 2i    i     3i
63  
3
dz = 3
6 2!
f   = 3  − 4 sin( 2 )  = −
6 6  6  108
z − 
C

 6

z4
Example: C contour encloses z = 1. Evaluate  ( z − 1)3 dz .
C

z0 = 1, n +1 = 3 → n = 2
f ( z) = z 4
f ( z ) = 4 z 3
f ( z ) = 12 z 2

2i 2i
z4
C ( z − 1)3 dz = 2! f (1) = 2! 12(1) = 12i
2
( )

91
Example: C-closed pathe encloses -2 and 4i points, evaluate
z
 ( z + 2)( z − 4i)dz .
C

z  A B 
 ( z + 2)( z − 4i)dz =   ( z + 2) + ( z − 4i) dz
C C

z A B
= +
( z + 2)( z − 4i ) ( z + 2) ( z − 4i )
A( z − 4i ) + B( z + 2) = z
2i
z = 4i  B=
1 + 2i
1
z = −2  A=
1 + 2i
z A B 1 dz 2i dz
 ( z + 2)( z − 4i)dz =  ( z + 2)dz +  ( z − 4i)dz = 1 + 2i  ( z + 2) + 1 + 2i  ( z − 4i)
C C C C C

1 2i 2i − 4 2 (i − 2)
= 2i(1) + 2i (1) = =
1 + 2i 1 + 2i 1 + 2i 1 + 2i

z
Örnek: C-closed pathe encloses -2 and 4i points, evaluate  ( z + 2) 2 ( z − 4i)dz .
C

z A B C
= + +
( z + 2) ( z − 4i) ( z + 2) ( z + 2) ( z − 4i)
2 2

92
 SERIES REPRESENTATIONS OF COMPLEX FUNCTIONS

ONE VARIABLE TAYLOR SERIES

Taylor series expansion of function f(x) at point x = a is written as

f (a)( x − a)2 f (n −1) (a)( x − a)n −1

f ( x) = f (a) + f (a)( x − a) + ++
2! (n − 1)!

Example: Expand Taylor series of f(x) = ex function about x = 0.

f(x) = e x f (0) = 1
f (x) = e x f (0) = 1
f (x) = e x
f (0) = 1
f (x) = e x
f (0) = 1

f (0)( x − 0) 2
f ( x) = e = f (0) + f (0)( x − 0) +
x
+
2!
x 2 x3
f ( x) = e x = 1 + x + + +
2! 3!

Example: Expand Taylor series of f(x) = sin (x) function about x = 0.

f(x) = sin( x) f (0) = 0

f (x) = cos(x) f (0) = 1
f (x) = − sin( x) f (0) = 0
f (x) = − cos(x) f (0) = −1
f ( 4) (x) = sin( x) f ( 4 ) (0) = 0
f (5) (x) = cos(x) f (5) (0) = 1

x x2 x3 x4 x5
f ( x) = sin( x) = 0 + 1 + 0 + (−1) + 0 + 1 + 
1! 2! 3! 4! 5!

x3 x5 x7
sin( x) = x − +1 − +
3! 5! 7!

93