SMA 2323 Ordinary Differential Equations II

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 70

SMA 2323 ODE II LECTURE NOTES

DAVID CHEPKONGA-0716388191, MCS YEAR 3


Course Outline
- Wronskian.
- Reduction of n th order ordinary differential equations to first order systems and their
solutions.
- Transformation of the differential equations to the normal form.
- Total Legendre and Bessel’s differential equations.
- Existence and uniqueness.
- Non – linear equations.
- Power series solutions near singular points.
- Applications.

References
1. Mathematical methods for physics and engineering. A comprehensive guide, 2nd
edition by K. F. Riley, M.P Hobson and S.J. Bence.
2. Differential equations. Linear, non-linear, ordinary, partial. By A. C. King, J.
Billingham and S. R. Otto.
3. Differential equations, Schaum’s outlines, 3rd edition. 563 fully solved problems, by
Richard Bronson and Gabriel Costa.

Linear Second Order Differential Equations


A linear 2nd order differential equation is a differential equation of the form
d2y dy
a 2 ( x) 2  a1 ( x)  ao ( x) y  b( x)
dx dx

Where
a2 ( x), a1 ( x), ao ( x)andb( x) are functions of x alone.

Standard form of a Linear Second Order Differential Equations


d2y dy
2
 p ( x)  q ( x) y  g ( x)
dx dx
Where;
a ( x) a ( x) b( x )
p ( x)  1 , q( x)  0 and g ( x) 
a 2 ( x) a 2 ( x) a 2 ( x)
The linear 2nd order Differential Equations is said to be homogenous when g ( x)  0 .
If g ( x)  0 , the linear 2nd order equation is said to be non-homogenous.

Linear Differential Operators


Linear 2nd order Differential Equations are considered differential operators. For a
differential operator;
L[ y]  y' ' py'qy To be linear the following properties must hold;

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 1


SMA 2323 ODE II LECTURE NOTES

L[ y1  y2 ]  L[ y1 ]  L[ y2 ]

L[cy1 ]  cL[ y1 ]

Solutions to a Second Order Homogeneous Differential Equations


The solution to the 2nd Order Differential Equations y' ' py'qy  0 is given by
y( x)  c1 y1  c2 y2
Where c1 and c 2 are arbitrary constants and are two y1 and y 2 are two linearly
independent fundamental functions of x.
For two functions to be linearly independent, the Wronskian cannot equal zero. The
Wronskian is defined by
y1 ( x) y 2 ( x)
W [ y1 , y 2 ]( x) 
y '1 ( x) y ' 2 ( x)

= y1 ( x)  y' 2 ( x)  y2 ( x) y1 ' ( x)

Example 1;
Show that the differential operator
L[ y]  y' ' xy is linear.
Solution:
Show that property one holds by evaluating
L[ y1  y2 ]
L[ y1  y 2 ]  ( y1  y 2 )' ' x( y1  y 2 )

= y1 ' ' y2 ' ' xy1  xy 2


= ( y1 ' ' xy1 )  ( y2 ' ' xy 2 )
= L[ y1 ]  L[ y2 ]
Thus property one holds.
Next, we show that property two holds by evaluating L[cy1 ] .
L[cy1 ]  (cy1 )' ' x(cy1 )
= cy1 ' 'cxy1
= c( y1 ' ' xy1 )
= cL[ y1 ]

Thus, property two holds.


We can conclude that L[ y]  y' ' xy is indeed linear.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 2


SMA 2323 ODE II LECTURE NOTES

Example 2;
Determine if the following functions are linearly independent or linearly dependent.

a. e , e 
3x 4 x

b. tan x  sec
2 2
x,3
Solution;
a. Compute the Wronskian
e3x e 4 x
W [e 3 x , e  4 x ] 
3e 3 x  4e  4 x

= (e 3 x )(4e 4 x )  (e 4 x )(3e 3 x )
=  4e  x  3e  x  7e  x  0
Since the wronskian is not equal to zero, then the functions are linearly
independent.

b. compute the wronskian

Note: tan 2 x  sec 2 x  1

1 3

W tan x  sec 2 x,3  W  1,3   0 0

= (-1) (0)-(3) (0)


Since the wronskian =0=0 is equal to zero, then the functions are linearly dependent.

EXERCISE
1. Show that e x , e 2 x are linearly independent on all intervals
2. Show that x, x 2 are linearly independent on x | x  , x  0.
3. The function y1  sin x and y2  cos x are both solutions of
equation y' ' y  0. evaluate the wronskian of these two solutions, and hence
show that they are linearly dependent.

THE WRONSKIAN METHOD


If y1 and y 2 are two linearly independent solutions of the standard equation
y' ' p( z) y'q( z) y  0
Then the wronskian of these two solutions is given by W (z) = y1 y2 ' y2 y1 '
Dividing the wronskian by y12 we obtain

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 3


SMA 2323 ODE II LECTURE NOTES

w y 12 y1 ' y12  d  1 
  y2      y 2
y12 y1 y12 y1  dz  y1 
d  y2

= 
dz  y1

This integrates to give
z
W (u )
y 2 ( z )  y1 ( z )  du
y12 (u )
Now, using the alternative expression for W (z) given in
 x 
W(x) =C exp   p(u )du  with c=1(since we are not concern with the normalizing
 
factor), we find
z
1 

u 

y 2 ( z )  y1 ( z )  2 exp   p(v)dvdu …………………..………….. (1)
y1 (u ) 
 

Hence, given y1 , we can in principle compute y 2 . Note that the lower limits of
integration have been omitted. If constant lower limits are included then they merely
lead into a constant times the solution.

EXAMPLE
3 1
Find a second solution to y ' ' y ' y  0 ………………….. (2)
z 1 z ( z  1)
Using the wronskian method.

Solution
3
For the ODE (2) we have P (z) = and from
z 1

y1 ( z )  z  (n  1) z n  z (1  2 z  3z 2  .....)
n 0

z
=
(1  z ) 2

z
We see that one solution to (2) is y1  Substituting for p and y1 in (1) we
(1  z ) 2
have

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 4


SMA 2323 ODE II LECTURE NOTES

z
z
(1  u ) 4  u 3 
 du
(1  z ) 2  u 2   v 1 
y2 ( z)  exp dv
 
(1  u ) 4
z
exp  3 ln(u  1)du
z
(1  z ) 2  u 2

(1  u )
z
z
2 
 du
(1  z ) u2
z  1
 2 
ln z  
(1  z )  z
By calculating the wronskian of y1 and y 2 it is easily shown that, as expected, the two
solutions are linearly independent.
In fact as the wronskian has already been evaluated as W (u) =exp [-3ln (u-1)],
I.e. W (z) = ( z  1) 3 , no calculations needed.

EXERCISE
Verify that y  e x , y  xe x , y  x 2 e x and y  e 2 x are four linear solutions of the
d 4 y 3d 2 y dy d3y
problem   5  2 y   0 and write the primitive.
dx 4 dx 2 dx dx 3

REDUCTION OF ORDER

Hypothesis: let f be a non-trivial solution of the n th order homogeneous L.D.E


dny d n1 y d y
ao ( x) n  a1 ( x) n1  ............  an1 ( x)  an( x) y  0 …………. (1)
dx dx dx
Conclusion: The transformation y=f(x) v reduces, equation (1) to an (n  1) st order
homogeneous linear differential equation in the dependent variable
dv
w .
dx
This theorem states that if one non zero solution of the n th order homogeneous Linear
Differential Equation (1) is known, then by making the appropriate transformation we
may reduce the given equation to another homogenous linear equation that is one
order lower than the original. We shall consider the 2nd order equation. Suppose f is a
known non- trivial solution of the second order homogeneous linear equation.
d2y dy
a0 ( x) 2  a1 ( x)  a2 ( x) y  0 ……………………… (2)
dx dx
Let us make the transformation;

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 5


SMA 2323 ODE II LECTURE NOTES

y  f ( x)v ……………………………………………….. (3)


Where f is the known solution (2) and v is a function of x that will be determined.
Then, differentiating we obtain:

dy dv
 f ( x)  f ' ( x)v …………………………………… (4)
dx dx

d2y d 2v dv
2
 f ( x ) 2
 2 f ' ( x)  f ' ' ( x)v …………………….. (5)
dx dx dx

Substituting (3), (4) and (5) into (2) we obtain


 d 2v dv   dv 
a0 ( x)  f ( x) 2  2 f ' ( x)  f ' ' ( x)v  a1 ( x)  f ( x)  f ' ( x)v  a2 ( x) f ( x)v  0
 dx dx   dx 

Or

d 2v
 2a0 ( x) f ' ( x)  a1 ( x) f ( x)  a0 ( x) f ' ' ( x)  a1 ( xf ' ( x)  a2 ( x) f ( x)v  0
dv
a0 ( x) f ( x) 2
dx dx

Since f is a solution of (2), the coefficient of v is zero, and so the last equation reduces
d 2v
to a0 ( x) f ( x) 2  2a0 ( x) f ' ( x)  a1 ( x) f ( x)  0
dv
dx dx

dv
Letting w  , this becomes
dx
 2a0 ( x) f ' ( x)  a1 ( x) f ( x)w  0 ………………………….. (6)
dw
a0 ( x) f ( x)
dx

This is the first order homogeneous linear differential equation in the dependent
variable w. the equation is separable; thus assuming f ( x)  0 and a0 ( x)  0 , we may
write;

dw  f ' ( x) a1 ( x) 
  2   dx
w  f ( x) a0 ( w) 

Thus, integrating we obtain


ln w   ln f ( x)   1
a ( x)
dx  ln c
2

a0 ( x)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 6


SMA 2323 ODE II LECTURE NOTES

Or
 a ( x) 
C exp   1 dx 
W  a0 ( x) 
 f ( x)2
This is the general solution of equation (6). Choosing the particular solution for which
dv
c=1, recalling that  w , and integrating again, we now obtain
dt

 a ( x) 
exp   1 dx 
v  a 0 ( x )  dx
[ f ( x)]2

Finally from (3), we obtain

 a ( x) 
exp   1 dx 
y  f ( x)   a 0 ( x)  dx ……………………………………………………………
[ f ( x)]2
(7)

The function defined in the right member of (7) which we shall hence forth denoted by g is
actually a solution of the original 2nd order equation (2). Furthermore, this new solution g and
the original known solution f are linearly independent since,

f ( x) g ( x) f ( x) f ( x )v
W ( f , g )( x  =
f ' ( x) g ' ( x) f ' ( x) f ( x)v' f ' ( x)v

 a ( x) 
= [ f ( x)] 2 v'  exp   1 dx   0
 a 0 ( x) 

Thus the linear combination


c1 f  c 2 g , is the general solution of equation (2). We now summarize the discussion
in the following theorem (not necessary).

Example:
Given that y  x is a solution of
d2y dy
( x  1) 2  2 x  2 y  0 ……………………………………….. (1)
2

dx dx
Find a linearly independent solution by reducing the order.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 7


SMA 2323 ODE II LECTURE NOTES

Solution:
First observe that y  x does satisfy equation (1).
Then let y  xv
Then
dy dv
 x  v And
dx dx

d2y d 2v dv
2
 x 2
2
dx dx dx

Substituting into equation (1), we obtain

 d 2v dv   dv 
( x 2  1) x 2  2   2 x x  v   2 xv  0
 dx dx   dx 
Or

d 2v dv
x( x 2  1) 2
2 0
dx dx

dv
Letting w  we obtain the first order homogenous linear equation.
dx
dw
x( x 2  1)  2w  0
dx

The equation is separable


dw  2dx

w x( x 2  1)

Or

dw   2 2x 
  2 dx
w  x x 1

Note: use partial fractions

2 A BX  C
  2
x( x 2  1) x x 1

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 8


SMA 2323 ODE II LECTURE NOTES

 2  A( x 2  1)  x( Bx  C )
And compare the coefficients.

Integrating:
ln w  ln x 2  ln( x 2  1)  ln c
c( x 2  1
w )
x2

dv
Choosing c=1, and recalling  w we integrate to obtain.
dx
1
v( x)  x 
x
Now forming g=fv, we obtain

1
g ( x)  x( x  )  x 2  1
x
The general solution

y  c1 x  c 2 ( x 2  1) I.e. use y  c1 f  c 2 g  c1 x  c 2 ( x 2 1)

Example 2:
Find the general solution to 2t 2  y' 'ty '3 y  0 given that y1 (t )  t 1 is a
solution by the method of reducing of order.

Solution:
Let the 2nd solution be of the form
y 2 (t )  v(t ) y1 (t ) ………………………………………….. (1)
For a proper choice of v (t). To determine the proper choice, we plug the guess
into the differential equation and get a new differential equation that can be solved for
v (t). The 2nd form of the solution together with its derivatives is given by

y 2 (t )  t 1v

y' 2 (t )  t 2 v  t 1v'

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 9


SMA 2323 ODE II LECTURE NOTES

y' ' 2 (t )  2t 3 v  2t 2 v't 1v' '


Plugging these into the differential equation gives

2t 2 (2t 3 v  2t 2 v't 1v' ' )  t (t 2 v  t 1v' )  3(t 1v)  0


Rearranging and simplifying gives

2tv' '(4  1)v'(4t 1  t 1  3t 1 )v  0

2tv' '3v'  0

So, in order for (1) to be a solution then v must satisfy


2tv' '3v'  0 ………………………………………………………….. (2)
Making the change of variable
w  v' w'  v' '
With this change of variable (2) becomes
2tw'3w  0
And this is a linear; first order differential equation that we can solve. This also explains the
name of this method. We’ve managed to reduce a 2nd order differential equation down to first
order differential equation.
The solution to this differential equation is
3

w(t )  ct 2

But we are after a solution to equation (2). However, we can now find this. Recall our change of
variable v'  w .
Within this we can easily solve for v (t).
3 5
2
v(t )   wdt   ct dt  ct 2  k
2
5

5
Choosing c  , k=0.
2
Using these gives the following for v (t) and for the 2nd solution

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 10


SMA 2323 ODE II LECTURE NOTES
5 5 3
1
v(t ) t  y 2 (t )  t (t )  t
2 2 2

The general solution will then be


3

y(t )  c1t 1  c 2 t 2

Example 3:

Find the general solution to t 2 y' '2ty '2 y  0

Given that y1 (t )  t is a solution.

Solution:
The form for the 2nd solution as well as its derivatives is

y 2 (t )  tv

y 2' (t )  v  tv '

y 2'' (t )  2v'tv ' '

Plugging these into differential equation gives,

t 2 (2v'tv' ' )  2t (v  tv' )  2(tv)  0  0

Rearranging and simplifying gives the differential equation that we’ll need to solve in order to
determine the correct v that we’ll need for the 2nd solution.

t 3 v' '4t 2 v'  0


Next use the variable transformation as we did in the previous example.
w  v'
w'  v' '
With this change of variable the differential equation becomes

t 3 w'4t 2 w  0
And this is a linear; first order differential equation that we can solve

w(t )  ct 4

Now solve for v (t)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 11


SMA 2323 ODE II LECTURE NOTES

 1 3
v(t )   wdt   ct  4 dt  ct  k
3
Dropping the constants and using the following v (t).

v(t )  t 3

y 2 (t )  t (t 3 )  t 2

The general solution will then be

c2
y (t )  c1t 
t2

Exercise
1. Given that y=x is a solution of x 2 y' '4 xy '4 y  0 . Find a linearly independent solution by
reducing the order. Write the general solution. Answer; y  c1 x  c 2 x 4 .
2. Given that y=x+1 is a solution of ( x  1) 2 y' '3( x  1) y'3 y  0 . Find a LI solution by
reducing the order. Write the general solution.
3. Given that y  e 2 x is a solution to (2 x  1) y' '4( x  1) y'4 y  0 .find a LI solution by
reducing the order. Write the general solution. Answer; y  c1e 2 x  c 2 ( x  1) .
4. Given that y  x 2 is a solution of ( x 3  x 2 ) y' '( x 3  2 x 2  2 x) y'(2 x 2  2 x  2) y  0 .find
a linear independent solution and write the general solution.

Transformation of the Differential Equations to the Normal Form


Normal Form

Second Order Equations


Consider the second order equation
y' ' p( x) y'q( x) y  0 …………………………………………. (1)
Through a change of dependent variable, thus equation can be transformed to
u' ' I ( x) y  0
This is known as the normal form of equation (1). The function I (x) is known as the
invariant of the equation.
Now to find the change of variables that will accomplish this transformation, we make
the substitution y(x) =a(x) u(x) in equation (1),

au' '2a' u'a' ' u  p(au'a' u)  qau  0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 12


SMA 2323 ODE II LECTURE NOTES

 a'   a' ' pa' 


u ' ' 2  p u '   q u  0
 a   a a 

To eliminate the u ' term, a (x) must satisfy

a'
2  p0
a

1
a' pa  0
2

 1 
a  c exp   p( x)dx 
 2 
For this choice of a, our differential equation for u becomes

 p 2 p' 
u ' q   u  0
 4 2

Two differential equations having the same normal form are called equivalent.

Note: the change of variables

 1 
y( x)  exp   p( x)dx u ( x)
 2 

Transforms the differential equation

y' ' p( x) y'q( x) y  0


Into its normal form
u' ' I ( x)u  0
Where the invariant of the equation, I(x) is

p 2 p'
I ( x)  q  
4 2

Higher Order Differential Equations

Consider the third order differential equation y' ' ' p( x) y' 'q( x) y'r ( x) y  0

We can eliminate the y ' ' term. Making the change of dependent variable.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 13


SMA 2323 ODE II LECTURE NOTES

 1 
y  u exp   p( x)dx 
 3 

 1   1 
y'  u ' pu  exp   p( x)dx 
 3   3 

 2 1   1 
y' '  u ' ' pu ' ( p 2  3 p' )u  exp   p( x)dx 
 3 9   3 

 1

y' '  u ' ' ' pu' ' p 2  3 p' u '1

9 p'9 p' ' p 3 u exp  31  p( x)dx 
 3 27   

Yields the differential equation

1 1
u' ' ' (3q  3 p' p 2 )u' (27r  9 pq  9 p' '2 p 3 )u  0
3 27

Note: the change of variable

 1 
y( x)  exp   p n 1 ( x)dx u ( x)
 n 

Transforms the differential equation

y ( n)  p n 1 ( x) y n 1  p n 2 ( x) y n 2  ...  p0 ( x) y  0

Into the form

u ( n)  a n 2 ( x)u ( n 2)  a n 3 ( x)u ( n 3)  ...  a0 ( x)u  0

Transformation of the Independent Variable

Transformation to the form u' 'a( x)u  0

Consider the 2nd order LDE y' ' p( x) y'q( x) y  0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 14


SMA 2323 ODE II LECTURE NOTES

We make the change of independent variable


  f ( x), u( )  y( x)
The derivatives in terms of  are
d d d d
  f'
dx dx d d
2 2
d d d 2 d d
 f ' f '  ( f ' )  f ''
dx 2
d d d 2
d

The differential equation becomes

( f ' ) 2 u' ' f ' ' u' pf ' u'qu  0

In order to eliminate the u ' term, f must satisfy


f ' ' pf '  0
f '  exp(   p( x)dx)

f   exp(   p( x)dx)dx

The differential equation for u is then


q
u ' ' u0
( f ')2

u ' ' ( )  q( x) exp( 2 p( x)dx)u ( )  0

Result: the change of variables.

 
   exp   p( x)dx dx, u( )  y( x)

Transform the differential equation y' ' p( x) y'q( x) y  0 into

 
u' ' ( )  q( x) exp 2 p( x)dx u( )  0

Transformation to a constant Coefficient Equation:

Consider the 2nd order LDE y' ' p( x) y'q( x) y  0 with the change of independent
variable   f ( x), u( )  y( x)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 15


SMA 2323 ODE II LECTURE NOTES

The differential equation becomes

( f ' ) 2 u' '( f ' ' pf ' )u'qu  0

For this to be a constant coefficient equation we must have

( f ' ) 2  c1 q1
f ' ' pf '  c 2 q

For some constants c1 and c 2 .

Solving the first condition,

f '  c q1

f  c  q( x)dx

The 2nd constraint becomes

f ' ' pf '


 const
q

1 1
1 2
cq q ' pcq 2
2  const
q

q '2 pq
3
 const
2
q

Result:
Consider the differential equation y' ' p( x) y'q( x) y  0

If the expression
q '2 pq
3
2
q

Is a constant then the change of variables:

  c  q( x)dx , u( )  y( x)

Will yield a constant coefficient differential equation. (Here c is an arbitrary constant).

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 16


SMA 2323 ODE II LECTURE NOTES

Example 1:
Solve the differential equation
4 1
y' '(2  x) y' (24  12 x  4 x 2 ) y  0
3 9
Solution:
Transform the equation to normal form. To transform the equation to normal form we make the
substitution;

 1  4  
y  exp    2  x dx u
 2  3  

 e  x x u
3

The invariant of the equation is


1 1 4 1 d 4
I ( x)  (24  12 x  4 x 2 )  (2  x) 2  (2  x)
9 4 3 2 dx 3

1
The normal form of the differential equation is then
u' 'u  0
Which has the general solution

u  c1 cos x  c 2 sin x

Thus the general solution for y has the general solution

2 2
 x x 3  x x 3
y  c1 e cos x  c 2 e sin x

Example 2:
Show that the solution of the differential equation
y' '2(a  bx) y'(c  dx  ex 2 ) y  0

Can be written in terms of one of the following canonical forms:

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 17


SMA 2323 ODE II LECTURE NOTES

v' '( 2  A)v  0

v' '  v

v' 'v  0
v' '  0

Solution:
Hint: transform the equation to the normal form and then apply the scale transformation
x    
The substitution that will transform to normal form is

 1 
y  exp 
 2
 2(a  bx)dx u

bx2
 ax
e 2
u
The invariant of the equation is

1 1 d
I ( x)  c  dx  ex 2  (2(a  bx)) 2  (2(a  bx))
4 2 dx

 c  b  a 2  (d  2ab) x  (e  b 2 ) x 2

   x  rx 2

The normal form of the differential equation is:

u' '(  x  rx 2 )u  0

We consider the following cases:


r0
 0
 0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 18


SMA 2323 ODE II LECTURE NOTES

We immediately have the equation u' '  0


  0 With the change of variables
1

v( )  u ( x), x   2 

We obtain
v' 'v  0
  0 . We have the equation

y' '(  x) y  0

The scale transformation x     yields

v' '2 (   (   )) y  0


 
v' '  3  2 (    ) v Choosing

1

  (  ) ,   
3

Yields the differential equation
v' '  v
  0 .the scale translation x     yields

 
v' '2 (   (   ))  y (   ) 2 v  0
v' ' 2

    2   (   2 )  2  2 v  0 Choosing

1
   4,  
2

Yields the differential equation

v' '( 2  A)v  0

Where

EXAMPLES 3
Show that the solution of the differential equation

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 19


SMA 2323 ODE II LECTURE NOTES

b d e
y' '2(a  ) y'(c   2 ) y  0
x x x
Can be written in terms of the following canonical terms.

A B
v' '(1   )v  0
 2

1 A
v' '(  2 )v  0
 
A
v' ' v0
2
Solution
Hint: transform the equation to normal form and then apply the scale transformation
x  
The substitution that will transform the equation to normal form is

 1 b 
y  exp    2(a  )du u
 2 x 
b ax
x e u
The invariant of the equation is
2
d e 1 b  1 d  b 
I ( x)  c   2   2(a  )    2(a  ) 
x x 4 x  2 dx  x 
d  2ab e  b  b 2
 c  a2  
x x2
 
  
x x2

The invariant form of the differential equation is

  
u ' '   u  0
 x x

We consider the following cases


 0
We immediately have the equation
 0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 20


SMA 2323 ODE II LECTURE NOTES


u ' ' u0
x2
  o We have the equation

 
u ' '(  )u  0
x x2
The scale transformation u( x)  v( ), x   yields

 
v' '(  )u  0
 2

Choosing    1 we obtain

1 
v' '(  2 )u  0
 
  0 The scale transformation x   yields
 
v' '(2   )v  0
 2

Choosing

1

  2
, we obtain
1

   2
v' '(1   2 )v  0
 

EXAMPLE 4
Show that the second order Euler equation

d2y dy
x2
2
 a1 x  a0 y  0
dx dx

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 21


SMA 2323 ODE II LECTURE NOTES

Can be transformed to a constant coefficient equation


Solution

Hint: make the change of variable x  e t , y( x)  u(t )

Write the derivative with the respect to x in terms of t

x  et
dx  e dt
t
d d
e
dx dt
d d
x 
dx dt

d2
Now we express x 2 in terms of
dx 2

d2y d  d  d d2 d
x2  x  x   x  
dx 2 dx  dx  dx dt 2 dt

Thus under the change of variables x  e t , y( x)  u(t ) the Euler equation becomes

u ' 'u ' a1u 1  a0 u  0


u ' '(a1  1)u ' a0 u  0

EXAMPLE 5
Solve Bessel’s’ equation of order ½,
1 1
y' ' y'(1  2 ) y  0
x 4x

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 22


SMA 2323 ODE II LECTURE NOTES

SOLUTION: transform the equation to the normal form.

The transformation

 1 1 
1

y  exp    du u  x 2 u
 2 x 
Will put the equation in normal form. The invariant is

 1  1 1  1 1
I (u )  1  2   2    2 1
 4x  4 x  2 x
Thus we have the differential equation

u' 'u  0
Write the solution

u  c1 cos x  c2 sin x

The solution of Bessel’s` equation of order ½ is

1 1
 
y  c1 x 2
cos x  c2 x 2
sin x

THE LEGENDRE’S AND BASSELS DIFRENTIAL EQUATIONS


THE LEGENDRES DIFRENTIAL EQUATION
The differential equation

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 23


SMA 2323 ODE II LECTURE NOTES

(1  x 2 ) y' '2 xy  p( p  1)  0 ……………………………………………….. (1)

Where p is a constant, is called Legendre`s differential equation

Now I n (1) p 0 ( x)  1  x 2 . p0 (0)  0 And hence x =0 is an ordinary point. We assume


the power series solution


y   c n x n ............................................................................(2)
n 0


y '   c n nx n 1
n 1

and


y ' '   c n n(n  1) x n  2
n2

Substituting the value of y, y' andy' ' in (1),

We obtain
   

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


n2 n2 n 1 n 0

or
   

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


n 0 n2 n 1 n 0

or
 
(2)(1)c 2 x 0  (3)(2)c3 x'2 (n  2)(n  1)c n  2 x n   n(n  1)c n x n 2c1 x'
n2 n2
 
2 nc n x n  p( p  1)c0 x 0  p( p  1)c1 x' p( p  1) c n x n  0
n2 n 0

Or

(2)(1)c2  p( p  1)c0 x 0  (3)(2)c3  2c1  p( p  1)c1 x'


 (n  2)(n  1)c


n2
n2  n(n  1)c n  2nc n  p( p  1)c n x n  0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 24


SMA 2323 ODE II LECTURE NOTES

Equating to zero the coefficients


of each power of x in the above
(2)(1)c2  p ( p  1)c0  0,
equation, we find
(3)(2)c3  2c1  P ( p  1)c1  0
 p ( p  1)c0
c2  ,
(1)(2)
2  p ( p  1)
c3  c1
(2)(3)
and
(n  2)(n  1)cn  2  n(n  1  2)cn  p ( p  1)cn  0
therefore
n(n  1)  p ( p  1)
cn  2  cn , n  2
(n  2)(n  1)
whichyield s
(2)(3)  p ( p  1)
c4 
(3)(4)
(2)(3)  p ( p  1)  p ( p  1)
c2   c0
(3)(4) (1)(2)
p ( p  1)(2)(3)  p 2 ( p  1) 2
c4  c0
(4)(3)(2)(1)
(3)(4)  p ( p  1) (3)(4)  p ( p  1) 2  p ( p  1)
c5  c3   c1
(5)(4) (5)(4) (2)(3)
(2)(3)(4)  2 p ( p  1)  (3)(4) p ( p  1)  p 2 ( p  1) 2
 c1
(2)(3)(4)(5)

And so on.

Now we substitute the values of c2 , c3,...................1

In the equation (2) it becomes,

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 25


SMA 2323 ODE II LECTURE NOTES

 p( p  1) 2  p( p  1)
y  c0  c1 x  c0 x 2  c1 x 3 
(1)(2) (2)(3)
p ( p  1)(2)(3)  p 2 ( p  1) 2
 c0 x 4 
(2)(3)(4)(5)
(2)(3)(4)  2 p( p  1)  (3)(4) p( p  1)  p 2 ( p  1) 2
c1 x 5
(2)(3)(4)(5)
 ...
 p ( p  1) 2 p 2 ( p  1) 2  (2)(3) p( p  1) 4 
 c0 1  x  x ...
 2 (2)(3)(4 
 2  p( p  1) 3 (2)(3)(4)  2 p( p  1)  (3)(4) p ( p  1)  p 2 ( p  1) 2 5 
 c1  x  x  x  ...
 (2)(3) (2)(3)(4)(5) 
Which is the required solution of the given Legendre equation.

LEGENDRE POLYNOMIALS
In many applications, the parameter in the Legendre’s equation (1  x 2 ) y' '2 xy 'n(n  1) y  0

Will be non-negative integer. From the recurrence relation


(n  s)(n  s  1)
as2   as
( s  2)( s  1) ……………………………………………………… (1)
( s  0,1...)

The RHS is zero when s=n, and therefore an2  0, an 4  0, an6  0,...

Hence if n is even
n(n  1) 2 (n  2)n(n  1)(n  3) 4
y1 ( x)  1  x  x ……………………………………. (2)
2! 4!
Reduce to a polynomial of degree n. if n is odd, the same is true for
(n  1)(n  2) 3 (n  3)(n  1)(n  2)(n  4) 5
y 2 ( x)  x  x  x  ... ……………….. (3)
3! 5!

These polynomials multiplied by some constants are called Legendre polynomials. Since they
are of great practical importance, let us consider them in more detail.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 26


SMA 2323 ODE II LECTURE NOTES

For this purpose we solve (1) for a s , obtaining

( s  2)(s  1)
as   a s  2 , s  n  2 …………………………………………. (4)
(n  s)(n  s  1)

And may then express all the non-vanishing coefficients in terms of the coefficient a n of the
highest power of x of the polynomial. The coefficient a n is at first still arbitrary. It is customary
to choose a n  1 when n  0 and

(2n)! 1  3  5...(2n  1
an  n 2
 ) ……………………………………………….. (5)
2 (n!) n!

Where n=1, 2 …

The reason is that for this choice of a n all those polynomials will have the value 1 when x=1 (As
seen from the generating function below). We then obtain from (4) and (5).

 n(n  1)
a n2  an
2(2n  1)

 n(n  1)(2n)!

2(2n  1)2 n (n! ) 2

 n(n  1)2n(2n  1)(2n  2)!



2(2n  1)2 n n(n  1)! n(n  1)(n  2! )

(2n  2)!

2 (n  1)!(n  2)!
n

Similarly,
 (n  2)(n  3)
a n4  a n2
4(2n  3)

(2n  4)!

2 n! (n  2)!(n  4)!
n

etc

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 27


SMA 2323 ODE II LECTURE NOTES

In general, when n  2m  0

(1) m (2n  2m)!


a n2m  ………………………………………………….. (6)
2 n m! (n  m)!(n  2m)!

The resulting solution of Legendre polynomial of degree n and is denoted by p n (x) . From (6) we
obtain
M
(1) m (2n  2m)!
p n ( x)   x m2m
m  0 2 m! ( n  m)! ( n  2m)!
n

……………………………………………… (7)
(2n)! n (2n  2)!
 x  n x n  2  ...
n
2 (n!) 2
2 1!(n  1)!(n  2)!

n (n  1)
Where M  or
2 2
Whichever is an integer.
In particular
p 0 ( x)  1
p1 ( x)  x
1
p 2 ( x)  (3x 2  1)
2
1
p 3 ( x)  (5 x 3  3x)
2
1
p 4 ( x)  (35 x 4  30 x 2  3)
8
1
p 5 ( x)  (63x 5  70 x 3  15 x)
8
etc

An easy method for computing the successive Legendre polynomials is the Rodrigues formula;

1 dn
p n ( x)  n n
( x 2  1) n
2 n! dx
Proof:

We apply the Binomial theorem to ( x 2  1) n and differentiate the result n times term by term

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 28


SMA 2323 ODE II LECTURE NOTES

1 dn
p n ( x)  n n
(1)(1  x 2 ) n
2 n! dx

1 dn  n(n  1) 4 n(n  1)(n  2) 6 


 (1) n 1  nx 2  x  x  ...  x 2 n 
 
n n
2 n! dx 2! 3!

1 n d
n
 n n! 
n 
 ( 1) x 2n2m 
n
2 n! dx m 0 m! (n  m)! 


d n 2n2m

n
1 n!

2 n n!
( 1) n

m  0 m! ( n  m)! dx
n
x

1 n
n! (2n  2m)! x n  2 m

2 n n!
( 1) n

m  0 m! ( n  m)! ( n  2m)!

Extending the range of m to 0 to n changes nothing since the new terms of degree < n their
derivatives are zero.

Orthogonality:
The most important property of the Legendre polynomials is the fact that

0, ifm  n 
 
1

1 p m ( x) p n ( x)dx   2 , ifm  n



 2n  1 

This often expressed by saying that p0 ( x), p1 ( x),... is a sequence of orthogonal functions on the
interval  1  x  1.

Proof:
Let f(x) be any function with at least n combinations derivatives on the interval  1  x  1
And consider the integral

1
I   f ( x) p n ( x)dx
1

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 29


SMA 2323 ODE II LECTURE NOTES

Rodrigues formula enables us to write this as

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

And on integration by parts gives


1
1  d n 1 2 
I  n  f ( x) n 1 ( x  1) n 
2 n!  dx  1

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

The expression in the brackets vanishes at both limits

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

And continuing to integrate by parts, we obtain

1
1

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

If f ( x)  p m ( x) with m<n, then f ( n) ( x)  0 and consequently I=0, which proves the first part.

To establish the second part, we put f ( x)  p n ( x)

(2n)!
Since p n( n ) ( x)  , it follows that
2 n n!
1 n
(2n)!
I  2n  (1  x
2
) dx
2 (n!) 2 1

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

If we change the variable by writing x  sin  , and recall the formula (proved by integration by
parts)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 30


SMA 2323 ODE II LECTURE NOTES

1 2n
 cos d  cos 2 n  sin    cos 2 n 1 d
2 n 1

2n  1 2n  1
Then the definite integral becomes
 
2
2n 2
 cos d   cos 2 n 1 d
2 n 1

0
2n  1 0

2n 2n  2 22
  .......  cos d
2n  1 2n  1 30

2 n n! 2 2 n (n! ) 2
 
1  3...(2n  1)(2n  1) (2n)!(2n  1)

And

2(2n)! 2 2 n (n! ) 2 2
I  2n  
2 (n! ) (2n)! (2n  1) 2n  1
2

consider
1
 p 0 ( x)  p1 ( x)t  p 2 ( x)t 2  ..............
1 2 t  t 2

The function on the LHS is called the generating function of the Legendre polynomials.
Proof:
1
 1  t (2 x  t ) 2
1
1 2t  t2

1 1 3 2 1  3...(2n  3) n 1
1 t (2 x  t )  2 t (2 x  t ) 2  ...  n 1
t (2 x  t ) n 1 
2 2  2! 2 (n  1)!
1  3...(2n  1) n
n
t (2 x  t ) n  ....
2 n!
Or
1
Start from the binomial expansion of ,
1 v

Then set v  2 xt  t 2 ,

Multiply the powers of 2 xt  t 2 out,

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 31


SMA 2323 ODE II LECTURE NOTES

Collect all the terms involving t n and verify that the sum of these terms is

p n ( x)t n i.e.

1 3 1 3 5
1   
1 2 2 2 2 2 v 3  ....
(1  v)  1  v 
2
v 
2

2 2! 3!
1 3 15
 1  v  v 2  v 3  ...
2 8 48
1 3 15
 1  (2 xt  t 2 )  (4 x 2 t 2  4 xt 3  t 4 )  (8 x 3 t 3  .....)  .......
2 8 48
3 1 5 3 
 1  xt   x 2  t 2   x 3  x t 3  ....
2 2 2 2 

  p n ( x)t n
n0

 p 0 ( x)  p1 ( x)t  p 2 ( x)t 2  p 3 ( x)t 3  .....

Exercise
1. Using the generating function for the Legendre polynomials, verify that
a. p n (1)  1 and
b. p n (1)  (1) n

Solution:
1



a. n 0
t p n (1)  (1  2t  t )
n 2 2
 (1  t ) 1  1  t  t 2  ...

p n (1)  1
n
1 2


 1
b. n 0
t p n (1)  (1  2t  t )
n 2 2
 (1  t )  1  t  t  t 3  ....

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 32


SMA 2323 ODE II LECTURE NOTES

p n (1)  (1) n

2. Consider the generating function


1
  p n ( x)t n
1  2 xt  t 2 n 0

a. By differentiating both sides w.r.t t, show that

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

b. Equate the coefficients of t n in (a) to obtain the recurrence formula

(n  1) p n1 ( x)  (2n  1) xp n ( x)  np n1 ( x)

c. Assume that p 0 ( x)  1 and p1 ( x)  x are known, and use the recursion formula in (b)
to calculate p 2 ( x), p3 ( x), p 4 ( x)andp5 ( x)

Solution:
xt 
a. 3
  np n ( x)t n 1
n 0
(1  2 xt  t ) 2 2

Substituting the original function


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

b. Expanding the above

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 33


SMA 2323 ODE II LECTURE NOTES
  

 np n ( x)t n1   2 xnp n ( x)t n   np n ( x)t n1


n 0 n0 n 0
 
  p n ( x)t n 1   xp n ( x)t n  0
n 0 n0

or
  

 (n  1) p n1 ( x)t n   2 xnp n ( x)t n   (n  1) p n1 ( x)t n


n 1 n 0 n 1

 
  p n 1 ( x)t n   xp n ( x)t n  0.
n 1 n0

  
p1 ( x)   (n  1) p n 1 ( x)t n  2 xnp n ( x)t n   (n  1) p n 1 ( x)t n
n 1 n 1 n 1
 
  p n 1 ( x)t n  xp 0 ( x)   xp n ( x)t n  0.
n 1 n 1

(n  1) p n 1 ( x)  2 xnp n ( x)  (n  1) p n 1 ( x)  p n 1 ( x)  xp n ( x)  0
(n  1) p n 1 ( x)  np n 1 ( x)  (2n  1) xp n ( x)

c. When n=1
2 p2 ( x)  3 xp1 ( x)  p0 ( x)
 3x( x)  1 When n=2

2 p2 ( x )  3 x 2  1
1
p2 ( x )  (3 x 2  1)
2

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 34


SMA 2323 ODE II LECTURE NOTES

3 p 3 ( x)  5 xp 2 ( x)  2 p1 ( x)
1
 5 x( (3x 2  1))  2( x)
2
5
 x(3 x 2  1)  2( x)
2
15 5 15 9
 x 3  x  2x  x 3  x
2 2 2 2
15 3 9 5 3
p 3 ( x)  x  x  x3  x
3.2 3.2 2 2

when
n3

4 p 4 ( x)  7 xp 3 ( x)  3 p 2 ( x)
 5x 3 3 3 
 7 x  x  (3x 2  1) 
 2 2 2 
 5x 3 3  9 x 2 3
 7 x  x    )
 2 2  2 2
35 x 4 21 2 9 x 2 3
  x  
2 2 2 2
4
35 x 3
4 p 4 ( x)   15 x 2 
2 2
1 1 1
p 4 ( x)   (35 x 4  30 x 2  3)  (35 x 4  30 x 2  3)
4 2 8
when
n4
5 p 5 ( x)  9 xp 4 ( x)  4 p 3 ( x)
1 5 3 
 9  (35 x 4  30 x 2  3)  4 x 3  x 
8 2 2 
1 5 3 
 (315 x 5  270 x 3  27 x)  32 x 3  x 
8 2 2 
1
 (315 x 5  270 x 3  27 x  80 x 3  48 x)
8
1
5 p 5 ( x)  (315 x 5  350 x 3  75 x)
8

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 35


SMA 2323 ODE II LECTURE NOTES

1
p5 ( x)  (63x 5  70 x 3  15 x)
8

3. Using M
M
(2n  2m)!
p n ( x)   (1) m x n2m
m0 2 m!(n  m)!(n  2m)!
n

Show that;

p n ( x)  (1) n p m ( x)
and
p' n ( x)  (1) n 1 p' m ( x)

Exercise
1. Legendre polynomials satisfy the condition
1

 p ( x)  dx 
2 2
2n  1
n
1

Show that this is true for p3 ( x) .

BESSEL FUNCTIONS
The differential equation
x 2 y 2  xy '( x 2  p 2 ) y  0 ……………………………………………………….. (1)
Where p is a non- negative constant is called Bessel’s equation and its solutions are
known as Bessel function.

Definition of the function J p (x)


Dividing (1) by x 2 , coefficient of y ' and y are

1
p( x) 
x
and
x2  p2
Q( x) 
x2

So that x p (x) = 1

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 36


SMA 2323 ODE II LECTURE NOTES

And x 2 Q( x)   p 2  x 2
The origin is therefore a regular singular point and equation (1) has a solution of the
form

y   C n X n  r , C 0  0 …………………………………………………………….. (2)
n 0

Substituting this equation together with its derivatives into equation (1) we get

 
x 2  (n  r )(n  r  1)C n X n  r  2   (n  r )C n X n  r 1
n 0 n 0

 ( x 2  p 2 ) C n X n  r  0
n 0

or
 

 (n  r )(n  r  1)C n X n r  (n  r )C n X n r 
n 0 n0
 

C
n 0
n X nr 2  p 2  C n X nr  0
n 0

or

r (r  1)C 0 X  (1  r )rC1 X
r r 1
  (n  r )(n  r  1)C n X n  r 
n2

rC0 X r  (1  r )rC1 X r 1   (n  r )C n X n  r 
n2

2 
 


n2
C n2 X nr
 p  0

C X  C 1 X 1 r
 
n2
C n X nr   0

Equating the sum of the coefficients of like powers to zero.

r (r  1)c0  rc0  p 2 c0  0, (n  0) ……………………………………………………..


(3(a)).

(1  r )rc1  (1  r )c1  p 2 c1  0, (n  1) ………………………………………………… (3b.)

(n  r )(n  r  1)cn  (n  r )cn  cn2  p 2 cn  0 ……………………………….. (3c.)

From (3a.) we obtain the indicial equation (r  p)(r  p)  0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 37


SMA 2323 ODE II LECTURE NOTES

The roots are r1  p  0 and r2   p

Let us now determine a solution corresponding to the root r1 .for this value of
r1 equation (3b) yields c1  0 .equation (3c) may be written

(n  r  p)(n  r  p)cn  cn2  0 ………………………………………. (5)

cn2
cn  
( p  n  r )(n  r  p)
 c0  c0
c2  
(2  r  p)(2  r  p) (2  2 p)(2)
 c1
c3  0
(3  2 p)(3)

Since c1  0 and p  0 , it follows that c3  0, c5  0,.......................... successively.


If we set n  2m in (5), we obtain for other coefficients

1
c2 m  c2 m2 , m  1,2,3... ………………………………………………. (6)
2 m( p  m)
2

And may determine c2, c4 ,c6 ……………


Successively;
C
c2  2 0
2 ( p  1)

 c2 c0
c4  
2 2( p  2) 242!( p  1)( p  2)
2

And so on, and in general

(1) m c0
c2 m  , m  1,2.... ……………………………….. (7)
2 2 m m!( p  1)( p  2)...( p  m)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 38


SMA 2323 ODE II LECTURE NOTES

Bessel function J n (x) for integer p=n.

Integer values of p are denoted by n. this is standard for p=n the relation (7) becomes

(1) m c0
c2 m  , m  1,2,............. (8)
2 2 m m!(n  1)(n  2)...(n  m)

c0 Is still arbitrary, so that the series (2) with these coefficients would contain this arbitrary
factor c0 , a highly impractical situation for developing formulas or computing. Verify by the
ratio test.
Accordingly, we have to make a choice;

c0  1Would be possible, but more practical is

1
c0  n
………………………………………………………. (9)
2 n!
Because then n!(n  1)...(n  m)  (M  n)! in (8),

So that
(1)
c2 m  2 m 2 n
, m  1,2,......................................................... (10)
2 m!(n  m)!

With these coefficients and r1  p  n we get from (2) a particular solution of (1), denoted by J n (x)

Given by;


(1) m X 2 m
J n ( x)  X n  2 m n
........................................................... (11)
m0 2 m!(n  m)!

Are called the Bessel function of the first kind of order n. this series converges for all x,
as the ratio test shows, and in fact converges very rapidly because of the factorials in the
denominator.

Bessel functions J 0 ( x) and J1 ( x) for n=0, we obtain from (11) the Bessel function of
order 0.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 39


SMA 2323 ODE II LECTURE NOTES

 (1) m
x 2m
J 0 ( x)  n 0

2 2 m (m!) 2
x2 x4 x6
 1    ...
2 2 (1!) 2 2 4 (2!) 2 2 6 (3!) 2

Which looks similar to a cosine function. For n=1, we obtain the Bessel function of order
1.

(1) m X 2 m1
J 1 ( x)   2 m 1
m 0 2 m!(m  1)!
x x3 x5 x7
  3  5  7  ...
2 2 1!2! 2 2!3! 2 3!4!

Which looks similar to a sine function


Bessel functions for any p  0

Gamma function
All we need for extension from p=n to any p  0 is an extension of the factorials in (9) and (11)
to any p. this is done by the Gamma function ( ) defined by the integral


( )   e t t  dt
0

 e t |   e t t  1dt
t  
0
0

The first expression on the right is zero, and the integral on the right is ( ) . This yields the
basic relation

(  1)   ( )
sin ce

(1)   e t dt  1
0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 40


SMA 2323 ODE II LECTURE NOTES

We conclude that

(2)  (1)  1!,


(3)  2 (2)  2!.....

And in general (n  1)  n!(n  0,1,2,...............)

1
Now in (9) we have c0  n
2 n!
Which is
1
By above
2 n (n  1)

1
Therefore, for any p, c0  . Then (7) becomes
2 p
( p  1)

(1) m
c2m 
2 2 m p m!( p  1)( p  2)...( p  m) ( p  1)

But we have

p  1 ( p  1)  ( p  2)

( p  2) ( p  2)  ( p  3) , and so on, so that

( p  1)( p  2)...( p  m) ( p  1)  ( p  m  1)

Hence the coefficients are

(1) m
c 2m 
2 2 m  p m! ( p  m  1)

With these coefficients and r  r1  p we get from (2) a particular solution of (1), donated
by J p (x) , given by


(1) m X 2 m
J p ( x)  x p
2
m 0
2m p
m! ( p  m  1)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 41


SMA 2323 ODE II LECTURE NOTES

And are called the Bessel function of the first kind of order p. this series converges for all x, as
one can verify by the ratio test. Replacing p by –p in the above, we get


(1) m X 2 m
J  p ( x)  x  p 
m 0 2 2 m p m! (m  p  1)

Since Bessel equation involves p 2 , the formulas J p and J  p

Are solutions of the equation for the same p. if p is not an integer, they are linearly independent,
because the first terms in J p and J  p are finite non-zero multiples of x p and x  p respectively.
This yields the following result.

Theorem 1 (General Solution)


If p is not an integer, a general solution of Bessel’s equation for all x  0
Is y( x)  a1 J p ( x)  a2 J  p ( x)

Theorem 2 (Linear dependence)

For integer p=n, the Bessel functions J p and J  p (x) are linearly dependent, because
J  n ( x)  (1) n J n ( x), n  1,2,3..........
proof
2 mn
 x

(1)  
m

J  n ( x)   2
m  0 m!(  n  m)!
2 mn
 x

(1)  
m

J  n ( x)   2
m  n m!(  n  m)!

1
Since the factors are zero when m=0, 1 …n-1.
(n  m)!

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 42


SMA 2323 ODE II LECTURE NOTES

On replacing the dummy variable m by n+m, and compensating by beginning the summation at
m=0, we obtain
2( n m)n
 x
m n

(1)  
J  n ( x)   2
m 0 m!(n  m)!
 (1) n J n ( x)

Further properties of J p (x)

Multiplying J p by X P we have


(1) m X 2 m 2 p
X p J p ( x)  
m 0 2 2 m p m! ( p  m  1)

Differentiating this and using

(  1)   ( )

We have

(1) m 2(m  p) X 2 m  2 p 1
( X p J p )'  
mn 2 2 m  p m! ( p  m  1)

(1) m X 2 m
( X p J p )'  X p X p 1
2
m 0
2 m  p 1
m! ( p  m)
therefore
d
( X p J p ( x))  ( X p ) J p 1 ( x)...................................................(1)
dx

Similarly by multiplying by X  p ,
We get by differentiation and an index shift

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 43


SMA 2323 ODE II LECTURE NOTES

(1) m X 2 m 1
( X  p J p ( x))'  
m 1 2 2 m  p 1 m! ( p  m  2)

(1) m 1 X 2 m 1
( X  p J p ( x))'  
m 1 2 2 m  p 1 m! ( p  m  2)

(1) m X 2 m  p 1
 X p 
m 0 2 2 m  p 1 m! ( p  m  2)
  X  p J p 1 ( x)...............................................................................(2)

Writing out formulas (1) and (2) and multiplying (2) by x 2 p


We get

px p 1 J p  x p J ' p  x p J p 1 ..............................................................(1*)
and
 px p 1 J p  x p J ' p   x p J p 1 ..........................................................(2*)

Subtracting (2*) from (1*) and dividing the result by x p , we obtain the first recurrence relation

2p
J p 1 ( x)  J p 1 ( x)  J p ( x)..........................................................(3)
x

Adding (1*) and (2*) and dividing the result by x p ,


We obtain the 2nd recurrence relation
J p 1 ( x)  J p 1 ( x)  2 J ' p ( x)..........................................................(4)

Example 1:
Find formulae for computing
J 3 (1)
and
J 3 ( 2)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 44


SMA 2323 ODE II LECTURE NOTES

Solution
Formulae (3) with p=1 and p=2, gives
2
J 2 ( x)  J 1 ( x)  J 0 ( x)
x
4
J 3 ( x)  J 2 ( x)  J 1 ( x)
x
 8  4
  2  1 J 1 ( x)  J 0 ( x)
x  x
hence
J 3 (1)  7 J 1 (1)  4 J 0 (1)  0.0199
J 3 (2)  J 1 (2)  2 J 0 (2)  0.1289

Example 2
2
Integrate I   x 3 J 4 ( x)dx
1

Solution
From (2) with p=3, and example 1 we get

I   x 3 J 3 ( x) |12
1
  J 3 (2)  J 3 (1)  0.0038
8

Example 3
Express the following integral in terms of Bessel functions

x
2
J 2 ( x)dx

Solution
Integrating by parts and using (1) with p=2 we get

x J 2 ( x)dx   x 4 ( x 2 J 2 ( x))dx
2

 x 3 2  x 3 2
 x J2   x J 1dx
3 3
J 1
  2   x 1 J 1dx
3x 3

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 45


SMA 2323 ODE II LECTURE NOTES

On the right, by another integration by parts and


use of (1) with p=1, we obtain
x J 1 dx   x 2 xJ 1dx
1

  x 1 xJ 1   x 1 xJ 0 dx
together
 J 2 ( x) J 1 ( x) 1
x   J 0 dx
2
xJ 2 ( x)dx  
3x 3 3
J p ( x) with
1 3 5
p   , , ,.........................
2 2 2
It often happens that, in special cases, higher functions become functions known from calculus.
Consider

(1) m X 2 m
J p ( x)  x p 
m 0 2 2 m p m! ( p  m  1)

1
When p  ,
2

(1) m X 2 m
J 1 ( x)  x  1
m 0 2 m 3
2
2 2
m! (m  )
2
2  (1) m X 2 m1
 
x m 0 3
..................................................................(1)
2 m 1
2 m! (m  )
2
now
 1
1
    t 2 e dt
t

2 0
let
t  s2
dt  2 sds

e s
2
1
( )  2 s  ds
2 0
s

 2  e  s ds
2

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 46


SMA 2323 ODE II LECTURE NOTES

Since s is a dummy in (a) we can write


2
 1    2   2 
 ( )   4  e  x dx   e  y dy 
 2    
  0  0 
 
 4 e
( x 2  y 2 )
dxdy
0 0

By changing this double integral to polar coordinates, we get



2
 1  2 
 ( )   4 e
( r 2 )
rdrd
 2 
  0 0

2
1
 4  d
0
2

1 1
 4   |02  4    
2 4
 1 
 ( )  
 2 
 
Now

 3  1 1 3 1 1 
 (m  )   (m  )(m  )...   ( ) 
 2  2 2 2 2  2 

 2 ( m1) (2m  1)(2m  1)....3  1 

In the denominator, we further have

2 2 m1 m! 2 2 m1 m(m  1)...2  1


 2 m12m(2m  2)...4  2

Together the denominator becomes

(2m  1)! 
sothat
2  (1) m x 2 m1
J 1 ( x)  
x m0 (2m  1)!
2

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 47


SMA 2323 ODE II LECTURE NOTES

This is the familiar maclaurian Series of sinx. Thus;

2
J 1 ( x)  sin x................................................................(6)
2
x

1
By differentiating and by (1) with p  ,
2
We get
1
2
x J 1 ( x)  cos x  x J 1 ( x)
2

2
 2

and
2
J 1 ( x)  cos x...........................................(7)

2
x

Exercise

1. Show that
d p
dx
 
x J p ( x)  x p J p 1 ( x)
Solution
We know that

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 48


SMA 2323 ODE II LECTURE NOTES

x
 (1) n ( ) 2 n  p
J p ( x)   2
n 0 n!(n  p )!


(1) n x 2 n  2 p
x p J p ( x)   2n p
n  0 n!( n  p )!2

  (1) n (2n  2 p ) x 2 n  2 p 1

d p
x J p ( x)  
dx n 0 n!(n  p )!2 2 n  p


(1) n x 2 n  2 p 1
 2 n p
n  0 n!( n  p )!2

(1) n x 2 n  2 p 1
 2 n  p 1
n  0 n!( n  p  1)!2

(1) n x 2 n  p 1 xp
  2 n  p 1
n  0 n!( n  p  1) 2
n x 2 n  p 1
 ( 1) ( )
 2 xp
n 0 n !( n  p  1 )!
n x 2 n  p 1
 ( 1) ( )
 xp 2
n 0 n !( n  p  1)!

therefore
d p
dx
 
x J p ( x)  x p J p 1 ( x)

note
J 0 (0)  1,
but
J 0 (0)  0
for
p 1

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 49


SMA 2323 ODE II LECTURE NOTES

2. Prove that

(1) k (2k ) x 2 k 1 
(1) k x 2 k 1
2
k 0
2k  p
k! ( p  k  1)
 
k 0 2
2 k  p 1
k! ( p  k  2)
Solution
Writing the k=0 term separately, we have

(1) k (2k ) x 2 k 1 
(1) k (2k ) x 2 k 1

k 0 2
2k  p
k! ( p  k  1)
 0   2k  p
k 1 2 k! ( p  k  1)

Which under the change of variables j=k-1, becomes



(1) j 1 (2 j  1) x 2( j 1)1 
(1)(1) j 2( j  1) x 2 j 1
2
j 0
2 ( j 1)  p
( j  1)! ( p  j  1  1)

j 0 2 2 j  p  2 ( j  1)! ( p  j  2)


(1) j 2( j  1) x 2 j 1
 
j 0 2 2 j  p 12( j  1)( j!) ( p  j  2)


(1) j x 2 j 1
 
j 0 2 2 j  p 1 ( j!) ( p  j  2)

The desired result follows by changing the dummy variable in the last summation from I
to k.

3. Prove that

(1) k x 2 k  p  2 
(1) k (2k ) x 2 k  p
 
k 0 2 2 k  p 1 (k!) ( p  j  2) k 0 2 2 k 1 (k!) ( p  j  1)

Solution:
Make the change of variable j=k+1:

(1) k x 2 k  p  2 
(1) j 1 x 2( j 1) p  2
  
k 0 2 2 k  p 12(k!) ( p  j  2) j 1 2 2( j 1) p 12( j  1)! ( p  j  1  2)


(1) j x 2 j  p

j 1 2 2 j  p 1 ( j  1)! ( p  j  1)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 50


SMA 2323 ODE II LECTURE NOTES

Now, multiply the numerator and denominator in the last summation by 2j, noting that j
(j-1)! =j! And

2 2 j  p1 (2)  2 2 j  p

The result is


(1) j (2 j ) x 2 j  p
2
j 1
2 j p
( j )! ( p  j  1)

Owing to the factor j in the numerator, the last infinite series is not altered if the lowered
limit in the sum is changed from j=1 to j=0. Once this is done, the desired result is achieved by
simply changing to the dummy index from j to k.

4. Prove that
d p 1
dx
 
x J p 1 ( x)  x p 1 J p ( x)

Solution

We may differentiate the series for the Bessel function term by term. Thus,
d  p 1  
d p 1
 
x J p 1 ( x)   x  2 k  p 1
(1) k x 2 k  p 1

dx dx 
 k 0 2 k ! ( k  p  1  1) 

d  (1) k x 2 k  2 p  2 
  2 k  p 
dx  k 0 2
 ( 2) k ! ( k  p  2) 

(1) k (2k  2 p  2) x 2 k  2 p 1


k 0 2 2 k  p k!2 (k  p  2)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 51


SMA 2323 ODE II LECTURE NOTES

Noting that 2 k  p  2  2k  p  1 k  p  1


And that the factor 2k  p  2 

Cancels, we have

  (1) k x 2 k  2 p 1

d p 1
x J p 1 ( x)   2 k  p x p 1 J p ( x)
dx k 0 2 k! (k  p  1)

For the particular case p=0, it follows that


d
xJ 1 ( x)  xJ 0 ( x)
dx

5. Prove that xJ ' p ( x)  pJ p ( x)  xJ p1 ( x)


Solution
We have

(1) k x 2 k  2 p 
(1) k x 2 k  2 p 1
pJ p ( x)  xJ p 1 ( x)  p  2k  p
x 
k 0 2 k! (k  p  1) k 0 2 2 k  p 1 k! (k  p  2)


(1) k px 2 k  p 
(1) k x 2 k  2 p  2
 2k  p
 
k 0 2 k! (k  p  1) k 0 2 2 k  p 1 k! (k  p  2)

Using the problem (3) on the last summation, we find



(1) k px 2 k  p 
(1) k (2k ) x 2 k  2 p
PJ p (k )  XJ p 1 ( x)   2k  p
x 
k 0 2 k! (k  p  1) k 0 2 2 k  p k! (k  p  1)


(1) k cp (2k ) x 2 k  p
 2k  p
 xJ ' p ( x)
k 0 2 k! (k  p  1)

For the particular case p=0, it follows that xJ '0 ( x)   xJ 1 ( x)

Or J '0 ( x)   J1 ( x)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 52


SMA 2323 ODE II LECTURE NOTES

6. Prove that xJ ' p ( x)   pJ p ( x)  xJ p1 ( x)

Solution

(1) k px 2 k  p 
(1) k x 2 k  p 1
 pJ p ( x)  xJ p 1 ( x)   p x
k 0 2 2 k  p k! (k  p  1) k 0 2 2 k  p 1 k! (k  p)

Multiplying the numerator and denominator in the second summation by 2(p+k) and noting that

( p  k)  p  k    p  k  1
We find 

(1) k ( p) x 2 k  p 
(1) k 2( p  k ) x 2 k  p
 pJ p ( x)  xJ p 1 ( x)   p  2k  p
 
k 0 2 k! (k  p  1) k 0 2 2 k  p k! (k  p  1)


(1) k  p  2( p  k )x 2 k  p

k 0 2 2 k  p k! (k  p  1)


(1) k (2k  p) x 2 k  p
 2k  p
 xJ ' p ( x)
k 0 2 k! (k  p  1)

7. Use problems (5) and (6) to derive the recurrence relation


2p
J p 1 ( x)  J p ( x)  J p1 ( x)
x
Solution
Subtracting the results of problem (6) from the results of problem (5), we find that
0  2 pJ p ( x)  xJ p1 ( x)  xJ p1 ( x)

Upon solving for J p1 ( x) we obtain the desired result.

8. Show that y  xJ 1 ( x) is a solution of xy ' ' y' x 2 J '0 ( x)  0

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 53


SMA 2323 ODE II LECTURE NOTES

Solution

First note that J 1 ( x) is a solution of Bessel’s equation of order one;

x 2 J1 ' ' ( x)  xJ 1 ' ( x)  ( x 2  1) J1 ( x)  0

Now substitute y  xJ 1 ( x)

Into the left side of the given differential equation.

x[ xJ 1 ( x)]' '[ xJ 1 ( x)]' x 2 J '0 ( x)  x[2 J '1 ( x)  xJ ' '1 ( x)]  [ J 1 ( x)  xJ '1 ( x)]  x 2 J '0 ( x)
but
J '0 ( x)   J 1 ( x)

By problem (5), so that the R.H.S becomes

x 2 J ' '1 ( x)  2 xJ '1 ( x)  J1 ( x)  xJ '1 ( x)  x 2 J1 ( x)  x 2 J ' '1 ( x)  xJ '1 ( x)  ( x 2  1) J1 ( x)  0

The last equality following from (*).

9. Show that y  x J 3 ( x) is a solution of x 2 y' '( x 2  2) y  0


2

Solution
Observe that J 3 ( x) is a solution of Bessel’s equation of order 3
2 2

x 2 J ' ' 3 ( x)  xJ ' 3 ( x)  ( x 2  9 ) J 3 ( x)  0


2 2 4 2

Now substitute y  x J 3 ( x) into the L.H.S of the given differential equation, obtaining
2

x 2 [ x J 3 ( x)]' '( x 2  2) x J 3 ( x)
2 2
3 1 1 1
1  
 x 2 [ x J 3 ( x)  x J ' 3 ( x)  x 2 J ' ' 3 ( x)]  ( x 2  2) x 2 J 3 ( x)
2 2
4 2 2 2 2

9
 x [ x 2 J ' ' 3 ( x)  xJ ' 3 ( x)  ( x 2  ) J 3 ( x)]  0
2 2
4 2

The last equality following from (**). Thus x J 3 ( x) satisfies the given differential equation.
2

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 54


SMA 2323 ODE II LECTURE NOTES

FIRST ORDER ODE


The general first order ODE can be written as
dy
f ( x, y , )0
dx
We normally expect that the above first order ODE will have a solution f and solution contain
one arbitrary constant.

Example
We consider the following examples
2
 dy 
  1  0
 dx 
Has no real valued solutions at all.
2
 dy 
  y 0
2

 dx 
Has only the solution y  0 , which contains no arbitrary constant.

Situations of this kind raise difficult theoretical questions about the existence and nature of
solutions of different equations’

EXISTENCE OF A SOLUTION

We consider the first order ODE

dy
 f ( x, y)...........................................................................(1)
dx

We assume that f (x, y) is a continuous function throughout some rectangle R in the xy- plane.
The geometric meaning of a solution of (1) is as follows;

If p0  ( x0 , y0 ) is a point in R then the number

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 55


SMA 2323 ODE II LECTURE NOTES

 dy 
  p 0  f ( x0 , y 0 )
 dx 

Determines a direction at p 0 . Now let p1  ( x1 , y1 )

Be a point near to p 0 in this direction, then

 dy 
  p1  f ( x1 , y1 )
 dx 

It determines a new direction at p1 . If we continue like this, we obtain a broken line.

P2

P1

P0
R

If we assume that the successive points move closer to one another then the broken line
approaches a smooth curve through the initial point Po .

This curve is a solution y  y  x  of equation (1). Note that the different initial point produces a
different curve. Thus, the solution of (1) forms a family of curves called Integral Curves.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 56


SMA 2323 ODE II LECTURE NOTES

LINEAR AND NON-LINEAR DE’S


A differential equation written as D (y (t)) =0 is linear if

D( y1  y2 )  D( y1 )  D( y2 )

NOTE 1:
A differential equation is linear if all the derivative terms and the unknown function y are not
raised to any power or part of a function.
Example 1

d2y
 y  0 Is linear
dt 2

d2y
 ty  t 3 Is linear
dt 2

d2y
2
 y 2  0 Is non-linear
dt

Bernoulli’s Equation:
The non-linear differential equation:
dy
 p( x) y  Q( x) y n ..............................................................(1)
dx
Where n is a constant but not necessary an integer, known as Bernoulli’s equation.
If n=0, n=1, the equation is linear and we know the solution methods for n  2

We use the substitution v  y1n to reduce Bernoulli’s equation to linear equation. The above
equation (1), we rewrite as

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 57


SMA 2323 ODE II LECTURE NOTES

dy
y n  p ( x) y 1 n  Q( x)
dx
taking
v  y 1 n
weget
dv dy
 (1  n) y  n
dx dx

dv dy
 (1  n) y  n
dx dx
or
dy 1 dv
y n 
dx (1  n) dx

Finally we get the linear equation of the form;


1 dv
 p ( x )v  Q ( x )
(1  n) dx

We solve the linear equation for v and we replace v  y1n to get the required general solution for
equation (1).

EXAMPLE:
Solve the differential equation

x 2 y'2 xy  y 3  0

Solution
The general solution is
2 1
y 2  x  cx 4 , where c is an arbitrary constant.
5

Power series solution and solutions near singular points; applications ;( see
the printed copies)

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 58


SMA 2323 ODE II LECTURE NOTES

Differential Equations Reducible to linear Equations with constant


Coefficients
We will now study two types of differential equations with variable coefficients which can be
transformed into differential equations with constant coefficients by appropriate substitutions.
a.) The Cauchy homogeneous linear equations
An equation of the form
dny d n1 y d n2 y
x n n  a1 x n1 n1  a2 x n2 n2  ...  an y  x...............................................(1)
dx dx dx
(Where a’s are all constants and x is a function of x) is called The Cauchy homogeneous
linear equations.
Equation (1) can be solved by changing it into a linear equation with constant coefficients
by the following transformations.
Put

x  ez
or
z  log x
dy dy dz 1 dy
 
dx dz dx x dx
therefore
dy dy
x   Dy.
dx dz
where
d
D
dx
also
d 2 y d  dy  d  1 dy 
    
dx 2 dx  dx  dx  x dx 
1 dy 1 d  dy 
 2   
x dz x dx  dz 
1 dy 1 d  dy  dz
   
x 2 dz x dz  dz  dx
1 dy 1 d 2 y
 2 
x dz x 2 dz 2
therefore
d 2 y d 2 y dy
x2  2   D( D  1) y
dx 2 dz dz

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 59


SMA 2323 ODE II LECTURE NOTES

Similarly;
d3y
x 3 3  D( D  1)( D  2)
dx
And so on.
Substituting for
dy d2y
x , x 2 2 ...,
dx dx
Equation (1) is reduced to one with constant coefficients and thus can be solved by the
methods already discussed.
Example 1
Solve the differential equation

d2y dy
x2 2
 x  3 y  x 2 log x
dx dx
Solution
d
Put x  e z or z  log x and defining D by D  the given equation is changed to
dz

{D( D  1)  D  3} y  ze 2 z
or
( D 2  2 D  3) y  ze 2 z

This is linear differential equation with constant coefficients. The A.E is


(  2  3)  0
2

giving
  3,1
 C.F  1e 3 z  2 e  z

The particular integral P.I is given by

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 60


SMA 2323 ODE II LECTURE NOTES

1 1
P.I  y p  ze 2 z  e 2 z Z
D  2D  3
2
( D  2)  2( D  2)  3
2

1
1 e2z  2D  D 2 
e 2z
z   1   z
D 2  2D  3 3  3 

e 2 z  2D 
 1  z
3  3 

 ze 2 z 2 2 z  e2z

   e  (3 z  2)
 3 9  9

 C.S
is
e2z
y  1e 3 z  2 e  z  (3 z  2)
9
or
2 x2
 1 x 3   (3 log x  2)
x 9

b.) Legendre Linear Equation;


This equation is of the form;
dny d n1 y
(a  bx) n n  a1 (a  bx) n1 n1  ...  an y  x..............................................(2)
dx dx
We will reduce this equation to Cauchy form by the substitution t=a+bx and the resulting
equation will be solved by the method discussed above.
Let t=a+bx then

dy dy dt dy
 . b
dx dt dx dt
2 2
d y d y
2
 b2 2
dx dt

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 61


SMA 2323 ODE II LECTURE NOTES

Equation (2) in terms of the new independent variable t is given by;


n 1
dny n 1 n 1 d y
b nt n n
 a1b t n 1
 ...  an y  x
dt dt
Note: equation (2) can straightway be reduced to one with constant coefficients by putting
a  bx  e z
or
z  log( a  bx).

Example:
Solve the differential equation:

d2y dy
(3x  2) 2 2
 3(3x  2)  36 y  3x 2  4 x  1
dx dx

Solution

3x  2  e z
or
ez  2
x
3
or
z  log( 3x  2)
then
dy dy dz 3 dy
 . 
dx dz dx 3 x  2 dz
or
dy
(3 x  2)  3D y ,
dx
d 2 y d  3 dy  32 dy 32 d2y
     
dx 2 dx  3 x  2 dz  (3 x  2) 2 dz (3 x  2) 2 dz 2

2
Put  (3 x  2) 2 d y  3 2 [ D( D  1) y ]
dx 2

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 62


SMA 2323 ODE II LECTURE NOTES

Therefore the given equation can be written as


2
 ez  2  4
[3 D( D  1)  3 D  36] y  
2 2
  (e z  2)  1
 3  3
or
2
 ez  2  4
9( D  4) y  
2
  (e z  2)  1
 3  3

The A.E is

2 40
giving
  2,2
 C.F  1e 2 z  2 e 2 z

The right side

(e z  2) 2 4 z (e 2 z  1)
  (e  2)  1 
3 3 3

The particular integral is


1 1 2z
yp  (e  1)
9( D  4) 3
2

1 e2z 1 1 1
 z  .  ( ze 2 z  1)
27 4 27 4 108

Hence the complete solution is


1
y  1e 2 z  2 e 2 z  ( ze 2 z  1)
108
or

y  1 (3x  2) 2  2 (3x  2) 2 
1
108
 
(3x  2) 2 log( 3x  2)  1

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 63


SMA 2323 ODE II LECTURE NOTES

EXERCISE
Solve the following equations.

d2y 1
x2  2y  x2 
1. dx 2
x

2 log x 2 1
Answer; y  1 x 2   (x  )
x 3 x

d2y dy
2. 2
 x  2 y  x log x
x2
dx dx
Answer y  1 cos(log x)  2 sin(log x)x  x log x

d3y 2
2 d y dy
3. x3 3
 2 x 2
 x  y  3x 2  log x  8x
dx dx dx
 7
Answer y  (1  2 log x) x  3  x 2 log x  x 2  2 x(log x) 2
x 3

d2y dy log x sin(log x)  1


4. x2 2
 3x  y 
dx dx x
Answer


y  x 2 1 x 3
 2 x  3
 1x log x 5sin(log x) 61 6 cos(log x)  54 sin(log x)3721
 382 cos(log x) 

d2y dy
5. x2 2
 2 x  20 y  ( x  1) 2
dx dx

 x2 x 1 
Answer y  1 x 5  2 x 4     
 14 9 20 

d2y dy
6. x2 2
 2 x  4 y  x 2  2 log x
dx dx

1 x2 1 3
Answer y   2 x 4   log x 
x 6 2 8

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 64


SMA 2323 ODE II LECTURE NOTES

d2y dy
7. (2 x  1) 2 2
 4(2 x  1)  8 y  8 x
dx dx
1
Answer y  1 (2 x  1)  2 (2 x  1) 2  (2 x  1) log( 2 x  1) 
2

d2y dy sin(log x)
8. x2 2
 3x  y 
dx dx x
1
Answer y  x 2 (1 x 3
 2 x  3 )  (5 sin log x  6 cos log x)
61x

d2y dy
9. ( x  2) 2 2
 ( x  2)  y  3x  4
dx dx
3
Answer y  ( x  2)[1  2 log( x  2)  log 2 ( x  2)]  2
2

d2y dy
10. ( x  1) 2 2
 ( x  1)  (2 x  3)(2 x  4)
dx dx
Answer y  1  2 log( x  1)  [log( x  1)]2  x 2  8x

APPLICATIONS OF DIFFERENTIAL EQUATIONS


EXAMPLE 1:
An electric circuit consists of an inductance of 0.1 Henry, a resistance of 20 ohms and a
condenser of capacitance 25 micro farads. Find the charge q and the current I at time t, given the
initial conditions.
dq
a. q=0.05 coulombs, i   0 when t=0.
dt
b. q=0.05 coulomb, I = -0.2 ampere when t=o.

Solution
Here
d 2q dq q
0.1 2  20   E (t )  0,
dt dt 25  10 5
or
d 2q dq
2
 200  400000q  0
dt dt

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 65


SMA 2323 ODE II LECTURE NOTES

Integrating q  e 100t ( A cos100 39t  B sin100 39t )


Differentiating w.r.tt,
dq
i  100e 100t [( 39 B  A) cos 100 39t ]  ( 39 A  B) sin 100 32t )
dt
0.05
a) using the initial conditions q=0.05, I=0 when t=0, A=0.05 and B=  0.008
39
Hence, q=

e 100t (0.05 cos 624.5t  0.008 sin 624.5t )


and
i  0.32e 100t sin 624.5t
b) using the initial conditions q=0, i=-0.2 when t=0, A=0.05 and B=0.0077
Hence q=
e 100t (0.05 cos 624.5t  0.0077 sin 624.5t )
and
i  e 100t (0.2 cos 624.5t  32 sin 624.5t

Note:
Here q and I are both transients, each becoming negligible very quickly.

EXAMPLE 2:
The differential equation of a shaft which is whirling with the line bearings horizontal is given by

d 4 y Ww2
EI  y W
dx 4 g

Where W is the weight of the shaft and w is the whirling speed. Taking the shaft of length 2l
d2y
with the origin at its center and short bearings at both ends (i.e. for x  l , y  2  0 )
dx
Show that

g  cos mx cosh mx 
y  cos ml  cosh ml  2
2u 2
where
Ww 2
m 
4

gEI

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 66


SMA 2323 ODE II LECTURE NOTES

g
And max, deflection is (sec ml  sec ml  2)
2w 2
Solution
The given differential equation may be written in an equivalent form, by substituting

Ww 2
m 
4
,
gEI
as
d4y W
 m4 y  .......................................................................(1)
dx4 EI
C.F  A cosh mx  B sinh mx  cos mx  D sin mx
And the particular integral is

1
1 W 1  D4  W W
.  4 1  4  .  4
D  m EI m  m  EI
4 4
m EI

g
Substituting for m 4 , the P.I is -
w2
The complete solution of the equation (1) is
g
y  A cosh mx  B sinh mx  C cos mx  D sin mx  ..........................................(2)
w2
Differentiating equation (2) successively twice, we get

1 dy
.  A sinh mx  B cosh mx  C sin mx  D cos mx......................................................(3)
m dx
1 d2y
.  A cosh mx  B sinh mx  C cos mx  D sin mx......................................................(4)
m 2 dx 2
Applying the given boundary conditions, i.e. when

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 67


SMA 2323 ODE II LECTURE NOTES

d2y
x  l , y  0, 0
dx 2
on
equations (2)and (4)
wehave
g ………………………… (5)
0  A cosh ml  B sinh ml  C cos ml  D sin ml 
w2
g
0  A cosh ml  B sinh ml  C cos ml  D sin ml  2
w
0  A cosh ml  B sinh ml  C cos ml  D sin ml
0  A cosh ml  B sinh ml  C cos ml  D sin ml
Adding the first, and 2nd equations, and third and fourth equations in (5) we get
g
A cosh ml  C cos ml 
w2
and
A cosh ml  C cos ml  0......................................................................(6)

And by further addition and subtraction, we’ve

g
A cosh ml 
w2
or
g
A 2
2 w cosh ml
and
g
C cos ml 
2w 2
or
g
C 2
2 w cos ml
Now taking the difference of the first and 2nd and also of the third and fourth of the equations in
(5), we have

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 68


SMA 2323 ODE II LECTURE NOTES

B sinh ml  D sin ml  0
B sinh ml  D sin ml  0
Solving these we have B sinh ml  0 and D sin ml  0 but as ml  o we have B=D=0. Substituting
the values of A, B, C and D in equation (2), we’ve

g  cos mx cosh mx 
y 2 
  2
2w  cos ml cosh ml 
The required result.
The max. Deflection will occur at the middle point i.e. when x=0.
g
 y max  (sec ml  sec hml  2)
2w 2

EXERCISE
1. A cylindrical tank 3m long and 2m in diameter is placed with its axis in a horizontal
position. Water initially filling the tank flows through a circular orifice of diameter 3cm
located in the bottom of the tank. How much time will be required for all the water to
flow out?
2. A body falling from rest is subjected to a force of gravity and an air resistance of
n2
times the square of velocity. Show that the distance travelled by the body in t seconds
g
g
is 2 log cosh nt.
n
3. A tank contains 5 cubic meters of fresh water. Salt water containing 10kg of salt cubic
meter runs in at the rate of 0.1 cubic meters per minute. The mixture kept uniform by
stirring runs out at the rate of 0.05 cubic meters per minute. Find the amount of salt in the
tank at time t.
4. A mass M suspended from the end of a helical spring is subjected to a periodic force
f  F sin wt in the direction of its length. The force f is measured positive vertically
downwards and at zero time M is at rest. If the spring stiffness is S, prove that the
displacement of M at time t from the commencement of motion is given by
F  w 
x 2 
sin wt  sin pt 
M(p  w ) 
2
p 
where
S
p2 
M
And damping effects are neglected.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 69


SMA 2323 ODE II LECTURE NOTES

5. A spring for which k=700 Newton per meter hangs in a vertical position with its upper
end fixed. A mass of 7kg is attached to the lower end. After coming to rest, the mass is
v
pulled down 0.05m and released. The medium offers a resistance equal to and the
4
support of the spring is given a motion x  0.3 cos 4t meters. Discuss the resulting motion of
the mass.

SMA 2323 ORDINARY DIFFERENTIAL EQUATIONS II Page 70

You might also like