Partial Dierential Equations

Module 1: Introduction

Dr.rer.nat. Narni Nageswara Rao

January 2012

1 Series solutions of dierential equations

In the rst year rst semester, we have studied the methods for solving gen-
eral linear rst order dierential equations, constant coecient second and
higher order dierential equations and a special case of variable coecient
dierential equation (Euler-Cauchy type equation). The solutions of these
equations were all closed form solutions in terms of standard functions. How-
ever, it is often not possible to express the solutions of variable coecient
equations in closed form using the standard functions. In such cases, we seek
the solution as an innite series in terms of the independent variable.

Many of the important physical problems can be described by second

order variable coecient equations. Solutions of such equations can be ob-
tained in terms of innite series. The series solution methods can be classied
into two categories: power series method and general series solution method
(Frobenius method).

If the solution is not just twice, but innitely dierentiable, we call it

an \analytic" function. According to Taylor's theorem, any function that
is analytic at a point x = x can be expanded in a power series y (x) =
0
P1 y (k) x0
ak (x x ) , where ak = .
k ( )
k
=0 0 !k

 nnrao maths@yahoo.co.in

1
 Results from Calculus

A function f (x) is said to be P

analytic at a point x = x if a
power series expansion f (x) = 1
0

ak (x x ) converges in some
k
k =0 0

open neighbourhood of x . 0

A function is said to be analytic in an open interval I if it is

analytic at every point t 2 I .

Taylor's theorem: If f (x) is innitely dierentiable at x then 0

the power series converges and ak = f k (x )=k! and we generally ( )

0

write 1 k
X f (x ) ( )

f (x) = (x x )k : 0

k!
0

k =0

Linear Independence of f1 ; x; x ; x ;
2 3



g.
1
X
ak x
k
=0 8 , x ak =0 8 k:

k =0

Ordinary and singular points of an equation Consider the variable

coecient second order, linear homogeneous equation
a (x) 6= 0
00 0
a (x)y + a (x)y + a (x)y = 0; (1)
0 1 2 0

Dividing the equation by a (x), we can write it in the standard form (normal
0

form or canonical form) as

00 0
y + p(x)y + q (x)y = 0 (2)
where p(x) = a (x)=a (x) and q (x) = a (x)=a (x). Let x be a point in the
1 0 2 0 0

interval I and a (x); a (x) be analytic (i.e. Taylor series expansion exists)
1 2

at x (that is a (x); a (x) are dierentiable at x and at every point in its

0 1 2 0

neighbourhood). We dene the following.

Ordinary point: A point x 2 I is said to be an ordinary point of equa-

6 0. Ordinary point is also called a regular point of the
0

tion (1) if a0 (x ) =
0

equation.

Singular point: A point x 0 2 I is said to be a singular point of equation

(1) if a (x ) = 0.
0 0

2
 Example I: Find the regular and singular points of the dierential equa-
tions
1.
(1 x
2
)y 00 2xy 0 + n(n + 1)y = 0
2.
x y
2 00 + axy 0 + by = 0; a; b are constants
Solution:

1. Setting a (x) = 1 x = 0, we get x = 1. Hence, x = 1 are the

0
2

singular points of the equation, while all other points are regular points.
Note that a (x) = 2x and a (x) = n(n + 1) are analytic at x = 1.
1 2

2. Setting a (x) = x = 0; we get x = 0, Hence, x = 0 is the singular

0
2

point of the equation, while all other points are regular points.
Let x 2 I be a singular point of equation (1), where a (x) and a (x) are
0 1 2

analytic at x . Consider now the standard form of equation (1), which is

0

given by
00 0
y + p(x)y + q (x)y = 0

where p(x) = a (x)=a (x) and q (x) = a (x)=a (x). Now, write this equation
1 0 2 0

as
00 p (x) y 0 + q (x) y = 0
y +
1
(3) 1

x x (x x ) 0 0
2

where
(x x )a (x)
p (x) = (x x )p(x) =
0 1
;
a (x)
1 0
0

and
(x x ) a (x) 2

q (x) = (x x ) q (x) =
2 0 2
;
a (x)
1 0
0

we dene the following

Regular singular point A singular point x of equation (1) is said to 0

be a regular singular point if and only if the functions p (x) and q (x) de- 1 1

ned in (3) have removable discontinuities at x and become analytic when 0

these discontinuities are removed. In other words, after the discontinuity at

x is removed; the functions p (x) and q (x) have Taylor series expansions
0 1 1

about the point x = x . What is intended in the above is that the denitions
0

of p (x) and q (x) given in equation (3) may be continuously extended to

1 1

include the point x and these continuous extensions will lead to analytic
0

3
functions at the point x . 0

Irregular singular point A singular point x of equation (1) is said to 0

be an irregular singular point if and only if x is not a singular point. In other 0

words, if either p (x) or q (x) or both p (x) and q (x) do not have Taylor
1 1 1 1

series expansions about the point x = x , then x is irregular singular point. 0 0

Example II Classify the singular points of the following equations.

1.
2
x y
00 + axy 0 + by = 0; a; b are constants
2.
2
x y
00 + xy 0 + (x 2
n
2
)y = 0; n constant
3.
(1 x
2
)y 00 2xy 0 + n(n + 1)y = 0; n constant
4.
x
3
(x 2)y 00 + x y 0 + 6y = 0 3

Solution

1. Setting a (x) = x = 0; we get the singular point of the equation as

0
2

x = 0. Dividing the equation by x , we get

2

y
00 + ax y 0 + b
y =0
2 2
x x

y
00 + a

x
y
0+
x
b
2
y = 0; x 6= 0 (4)
comparing with equation (3), we get p (x) = a and q (x) = b (Note 1 1

that x = 0 is a removable discontinuity of p (x) = ax =x ) and q (x) = 1

2 2
1

(bx )=x :) Since a and b are constants, the expressions of p (x) and
2 2
1

q (x) are the Taylor series expansions. Hence, x = 0 is regular singular

1

point of the given equation.

2. The singular point of the equation is x = 0. Dividing by x , we get 2

y
00 + x
y
0 + 1 (x 2 2
n )y = 0
2 2
x x

y
00 + 1 y 0 + 1 (x 2
n
2
)y = 0; x 6= 0
2
x x

4
 comparing with equation (3), we get p (x) = 1; q (x) = x n . (Note 1 1
2 2

that x = 0 is a removable discontinuity of p (x) = x =x ; and q (x) = 1

2 2
1

x (x n )=x ). The expressions of p (x) and q (x) are Taylor series

2 2 2 2
1 1

expansions about x = 0. Hence, x = 0 is a regular singular point of the

given equation.
3. Setting a (x) = 1 x = 0, we get the singular point of the equation
2

as x = 1. Dividing the equation by (1 x ) and writing it in the

0
2

required form, we get

2x 0 n(n + 1)
x 6= 1
y
00 y + y = 0; (5)
1 x 1 x 2 2

   
00 1 2x 0 1 (1 x)n(n + 1)
y y + y = 0
(1 x) 1 + x (1 x) 1+x 2

where the singularity at x = 1 is being considered. Now p (x) = x

x 1
2
1+

and q (x) =
1
x n n
x
(1
are both analytic at x = 1 and hence Taylor
)

1+
( +1)

series expansions of these functions about x = 1 exist. For example,

2(x 1 + 1)
p (x) =
1
(x 1 + 2)
 
(x 1) 1

= [1 + (x 1)] 1 +
2
 
1 (x 1)




2
x
= [1 + (x 1)] 1 +
2 4
 
1 (x 1)
+


:
2
x
= 1+
2 4
Similarly, Taylor series expansion for q (x) can be written. Write now, 1

the equation (5) as

   
00 1 2x 1 (1 + x)n(n + 1)
y + y = 0
1+x 1 x (1 + x) 1 x 2

where the singularity at x = 1 is being considered. Now p (x) = x

x 1
2
1

and q (x) = x n xn are both analytic at x = 1 and hence have

(1+ ) ( +1)

Taylor series expansions about x = 1. Therefore, x = 1 are regular

1 1

singular points of the given equation.

4. Setting a (x) = x (x 2) = 0, we get the singular points as
0
3
x = 0 ; 2:
Dividing the equation throughout by x (x 2), we get 3

6
00 + 0+ 6= 0 2
3
x
y y y = 0; x ; (6)
x
3
(x 2) x
3
(x 2)
5
 Consider now, the singularity at x = 2. Comparing equation (6) with
00 p (x) y 0 + q (x) y = 0
y +
1 1

x 2 (x 2) 2

we get p (x) = 1 and q (x) = 6(x 2)=x . Since p (x) and q (x) are
1 1
3
1 1

both analytic at x = 2, Their Taylor series expansions about x = 2

exist. Hence x = 2 is regular singular point of the given equation.
Consider now, the singularity at x = 0. Comparing equation (6) with
00 p (x) y 0 + q (x) y = 0:
y +
1 1

2
x x

We get p (x) = x=(x 2), and q (x) = 6=[x(x 2)]. Now q (x) is not
1 1 1

analytic at x = 0, and Taylor series expansion of q (x) about x = 0 does

1

not exist. Hence, x = 0 is an irregular singular point of the equation.

2 Appendix
2.1 Review of convergence of series

The Polynomials are functions of the form

a + a x + a x +


+ an + an xn
2 n 1
0 1 2 x 1

where a ; a ;


; an are constants. This is a polynomial of degree n. A

0 1

rational function is quotient of polynomials. For example,

3x x + 4 3

r (x) =
x + 5x + 1
2

is a rational function.

A transcendental function y = f (x) is one that is not solution of a poly-

nomial equation in x and y . The elementary transcendental functions are the
ones that we encounter in calculus: sine, cosine, logarithm, exponential, and
their inverses and combinations using arithmetic/algebraic operations.

The higher transcendental functions are ones that are dened using power
series. These are often discovered as solution of dierential equations. These
functions a bit dicult to understand, just because they are not given by
elementary formulas. Higher transcendental functions are frequently termed
as Special functions. These functions were studied extensively in the eigh-
teenth and nineteenth centuries-by Gauss, Euler, Abel, Jacobi, Weirstrass,
Riemann, Hermite, Poincare and other leading mathematicians of the day.

6
2.2 Review of power series

A series of the form

1
X
j
aj x =a +a0 1x +a 2x
2
+


(7)
j =0

is called a "power series" in x. Slightly more general is the series

X1
aj (x a)
j

j =0

which is a power series in x a. The series in the equation (7) is said to

converge at a point x if the limit
k
X
lim aj x
j
k!1 j =0

exists.

2.3 Problems

1. Calculate the radius of convergence of the series

X1 j
x
2
j
j =0

Solution We apply ratio test:

   
( + 1)
 xj +1 = j   

= jxj:
2 2
j
lim   = lim x
j!1  j
x =j
2  !1 (j + 1)
j 2 

We know that the series will converge when this limit is less than 1, or
jxj < 1. Likewise it diverges when jxj > 1. Thus the radius of conver-
gence is R = 1.

In practice, one has to check the end points of the interval of conver-
gence for each case. In this example, we see immediately that the series
converges at 1: Thus we may say that the full interval of convergence
is [ 1; 1].

7
2. Calculate the radius of convergence of the series
X1 j
x

j
j =0

Solution We apply ration test:

   
( + 1) 
 xj +1 = j  
lim
!1
j

 j
x =j

= lim

!1 (j + 1)
j

j
x

= jxj:

We know that the series will converge when this limit is less than 1, or
jxj < 1. Likewise it diverges when jxj > 1. Thus the radius of conver-
gence is R = 1.

In this example, we see immediately that the series converges at 1 (by

alternating series test) and diverges at +1 (since this gives harmonic
series). Thus we may say that the interval of convergence is [ 1; 1).
3. Calculate the radius of convergence of the series
X1 j
x
j
j
j =0

Solution We use root test:

 1  
 xj  j x
lim   = lim   =0
j!1  j
j j !1  
j

Of course 0 < 1, regardless of the value of x. So the series converges for

all x. The radius of convergence is +1 and the interval of convergence
is [ 1; +1]. There is no need to check the end points of the interval
of convergence because there is none.