100% found this document useful (1 vote)
312 views42 pages

Frobenius Series Solution

This document contains an assignment submission from students of Rajshahi University of Engineering & Technology on the topic of series solution and the Frobenius method for solving differential equations. It includes definitions of key concepts like power series, analytic functions, ordinary points, and singular points. It then outlines the Frobenius method and divides it into four cases, with each student assigned a sub-topic to explain the working rules and provide examples.

Uploaded by

fahim faisal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
312 views42 pages

Frobenius Series Solution

This document contains an assignment submission from students of Rajshahi University of Engineering & Technology on the topic of series solution and the Frobenius method for solving differential equations. It includes definitions of key concepts like power series, analytic functions, ordinary points, and singular points. It then outlines the Frobenius method and divides it into four cases, with each student assigned a sub-topic to explain the working rules and provide examples.

Uploaded by

fahim faisal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 42

“Heaven’s Light is Our Guide”

Rajshahi University of Engineering & Technology

Assignment

Course No: Math 2213


Course Name: Complex Variable, Differential Equation and Harmonic Analysis

Assignment Topic:
Series solution_ Frobenius method: Elementary discussion of series solution and solution of
Differential Equation using Frobenius method.

Submitted To:
Mst. Rupale Khatun
Assistant Professor
Dept. of Mathematics,
Rajshahi University of Engineering & Technology.

Submitted By:
Roll: 1703001-1703014, 1703016-1703030
Section : A
Dept. of Computer Science and Engineering.

Date of Submission: 22.12.2020


Submitted by
SI. Roll Name Topic
No
1. 1703001 Audity Ghosh Series solution.
2. 1703002 Oudarja Barman Tanmoy Analytic function.
3. 1703003 Kazi Mehrab Rashid Ordinary point.
4. 1703004 Tahsen Islam Sajon Singular point.
5. 1703005 Mahfujur Rahman Mahfuj Regular singular point.
6. 1703006 Mastura Khatun Irregular singular point.
7. 1703007 Shakil Mahmud Shuvo Frobenius method
8. 1703008 Mansura Naznine Case details of Frobenius method
9. 1703009 Maria Chowdhury Frobenius method: Case-1: Working rule
10. 1703010 M Mahbubul Alam Frobenius method: Case-1: Example
11. 1703011 Md. Sarowar Alam Sagor Frobenius method: Case-1: Example
12. 1703012 Farhan Fahmid Frobenius method: Case-1: Example
13. 1703013 Soumik Datta Frobenius method: Case-1: Example
14. 1703014 Kuldip Saha Frobenius method: Case-1: Example
15. 1703016 Kawshik Banerjee Frobenius method: Case-2: Working rule
& Example
16. 1703017 S M Nazmul Hasan Frobenius method: Case-2: Example
17. 1703018 Saifur Rahman Frobenius method: Case-2: Example
18. 1703019 Rizwan Ahmed Frobenius method: Case-2: Example
19. 1703020 Dilara Yasmin Disha Frobenius method: Case-2: Example
20. 1703021 Md. Shamsul Tahsin Zilani Frobenius method: Case-2: Example
21. 1703022 Mohammad Fakhrul Islam Frobenius method: Case-3: Working rule
22. 1703023 Wakil Mahmud Frobenius method: Case-3: Example
23. 1703024 Koushik Debnath Frobenius method: Case-3: Example
24. 1703025 Abdur Rahman Frobenius method: Case-3: Example
25. 1703026 Khaled Zinnurine Frobenius method: Case-3: Example
26. 1703027 Sheikh Tasnimul Hasan Thasin Frobenius method: Case-4: Working rule
27. 1703028 Md. Zihan Abedin Frobenius method: Case-4: Example
28. 1703029 S. M. Abdullah Ashik Frobenius method: Case-4: Example
29. 1703030 Md. Sonit Mia Frobenius method: Case-4: Example
Index
SI Sub-Topics Description Page
No. No.
1 Series Solution Definition 1-2
Example
Explanation
2 Analytic Function Definition 2-3
Example
Explanation
3 Ordinary Point Definition 4-5
Example
Explanation
4 Singular Point Definition 6
Example
Explanation
5. Regular Singular Point Definition 7
Example
Explanation
6 Irregular Singular Point Definition 8-9
Example
Explanation
7 Frobenius Method Trial Solution 10-15
Working steps
Case-Details
8 Case-1 Working Rule 15-21
Example -1
Example-2
Example-3
9 Case-2 Working Rule 22-29
Example-1
Example-2
Example-3
10 Case-3 Working Rule 29-35
Example-1
Example-2
11 Case-4 Working Rule 35-39
Example-1
Example-2
Page |1

Elementary discussion of series solution and solution of differential equations


using Frobenius method.

Series Solution: [1703001]


Power Series:
An infinite series of the form

∑𝑛=0 𝐶𝑛 (𝑥 − 𝑥0 )𝑛 = 𝐶0 + 𝐶1 (𝑥 − 𝑥0 ) + 𝐶2 (𝑥 − 𝑥0 )2+……….

is called a power series in (𝑥 − 𝑥0 ). In particular, a power series in x is an infinite series

∑∞ 𝑛 2
𝑛=0 𝐶𝑛 𝑥 = 𝐶0 + 𝐶1 𝑥 + 𝐶2 (𝑥) +…………..

Example of Power Series:

The exponential function



(𝑥𝑛 ) (𝑥2 ) (𝑥3 )
ex = ∑ = 1+𝑥+ + +………….
𝑛=0 𝑛! 2! 3!

A Property of Power Series:

The most important property of power series is Radius of convergence.



For any power series ∑𝑛=0 𝐶𝑛 (𝑥 − 𝑥0 )𝑛,

there is a number ρ ∈ [0, ∞] (meaning: ρ > 0 and can be infinity) such that

• the power series converges for all x such that |x – x0| < ρ;

• the power series diverges for all x such that |x – x0| > ρ.

This particular number ρ is called the radius of convergence.

Series Solution:

The power series method is used to seek a power series solution to certain differential equations.
In general, such a solution assumes a power series with unknown coefficients, then substitutes that
solution into the differential equation to find a recurrence relation for the coefficients.

The basic idea to finding a series solution to a differential equation is to assume that we can
write the solution as a power series in the form,

y(x)= ∑𝑛=0 𝐶𝑛 (𝑥 − 𝑥0 )𝑛 ………………(1)
Page |2

and then try to determine what the Cn’s need to be. We will only be able to do this if the
point x=x0, is an ordinary point. We will usually say that (1) is a series solution around x=x0.

Why should we try to solve the equation around singular points? If we want to know what happens
around a singular point, why couldn’t we just pick a regular point nearby, find the solution, and
then study the solution near the singular point? Afterall, this solution around the regular point
should converge for points near the singular point. Unfortunately, the above strategy will not give
us what we want. Because as x approaches the singular point, the convergence of the power series
(obtained by solving the equation at a nearby regular point) becomes worse and worse. And letting
x approach the singular point would not give us useful information.

Problem-Solving Strategy:

1. Assume the differential equation has a solution of the form



y(x)= ∑𝑛=0 𝐶𝑛 (𝑥)𝑛

2. Differentiate the power series by term to get


∞ ∞
y′(x) = ∑𝑛=0 𝑛𝐶𝑛 (𝑥)𝑛−1 and y′′(x) = ∑𝑛=0 𝑛(𝑛 − 1)𝐶𝑛 (𝑥)𝑛−2
3. Substitute the power series expressions into the differential equation.
4. Re-index sums as necessary to combine terms and simplify the expression.
5. Equate coefficients of like powers of x to determine values for the coefficients Cn
in the power series.

Analytic Function: [1703002]

1)

Definition 1: A function f(z) is said to be analytic in a region R of the complex plane if f(z) has
a derivative at each point of R and if f(z) is single valued.

Definition 2: A function f(z) is said to be analytic at a point z if z is an interior point of some


region where f(z) is analytic. Hence the concept of analytic function at a point implies that the
function is analytic in some circle with center at this point.

Theorem: If f(z) is analytic at a point z, then the derivative f’(z) is continuous at z.

Corollary: If f(z) is analytic at a point z, then f(z) has continuous derivatives of all order at
the point z.
Page |3

2) Conditions for a Complex Function to be Analytic:

a. A necessary condition for a complex function to be analytic:


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

be a complex function. Since x = (z + z̅)/2 and y = (z − z̅)/2i, substituting for x and y gives

f(z, z̅) = u(x, y) + iv(x, y)


𝑑𝑓
A necessary condition for f(z, z̅) to be analytic is, = 0. (1)
𝑑𝑧̅

Therefore, a necessary condition for f = u + iv to be analytic is that f depends only on z. In terms


of the of the real and imaginary parts u, v of f, condition (1) is equivalent to,
𝑑𝑢 𝑑𝑣
= (2)
𝑑𝑥 𝑑𝑦

𝑑𝑢 𝑑𝑣
= − 𝑑𝑥 (3)
𝑑𝑦

Equations (2, 3) are known as the Cauchy-Riemann equations. They are a necessary condition for
f = u + iv to be analytic.

b. Necessary and sufficient conditions for a function to be analytic:


The necessary and sufficient conditions for a function f = u + iv to be analytic are that:
𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑣
1. The four partial derivatives of its real and imaginary parts , 𝜕𝑦 , 𝜕𝑦 , 𝜕𝑥 satisfy the
𝜕𝑥
Cauchy-Riemann equations (2, 3).
𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑣
2. The four partial derivatives of its real and imaginary parts 𝜕𝑥 ,𝜕𝑦 , 𝜕𝑦 , 𝜕𝑧 are continuous.

Theorem:
𝜕2 𝑢 𝜕2 𝑢
If f(z) is analytic, then + =0 (4)
𝜕2 𝑥 𝜕𝑦2

𝜕2 𝑣 𝜕2 𝑣
+ =0 (5)
𝜕2 𝑥 𝜕𝑦2

The real and imaginary parts of an analytic function are harmonic conjugate functions, i.e.,
solutions to Laplace equation and satisfy the Cauchy Riemann equations (2,3).

3.Example:
𝑥−𝑖𝑦
𝑓(𝑥)=𝑥2 +𝑦2 𝑓(𝑧) =𝑧 3 , z = (x +i y)
Page |4

Ordinary Point: [1703003]


An ordinary point is a point in a domain in which a given function is analytic or defined.

Example:

1
f(x) = where x ∈ ℝ
1−𝑥

In this function all values except x = 1 are ordinary points cause the function is defined by all
those values except 1.

So here ordinary point of the function = {x : x ∈ ℝ & x ≠ 1}

Ordinary point of a differential equation:


Let , the differential equation of the form,

𝑑2𝑦 𝑑𝑦
+ P(x) + Q(x)y = 0 ……………(i)
𝑑𝑥2 𝑑𝑥

A point x = a is said to be an ordinary point of a differential equation (i) if the functions P(x) and
Q(x) are analytic at the point x = a,

i.e. they have power series expansions in non-negative powers of (x-a) with a positive radius of
convergence. Note that in the above definition the coefficient of the highest order derivative term
is taken as 1.

If x = a is not an ordinary point of the equation, then it is termed as a singular point of the equation.

Example:

𝑑2𝑦 1 𝑑𝑦 1
+ + y = 0
𝑑𝑥2 𝑥−1 𝑑𝑥 𝑥

Comparing with equation no (i) we get,


1 1
P(x) = and Q(x) =
𝑥−1 𝑥
Page |5

P(x) is undefined for x = 1 and Q(x) is undefined for x = 0. So the differential equation given in
example is defined for all values except 0 and 1.

So here ordinary points of the differential equation = { x : x ∈ ℝ & x ≠ 0 & x ≠ 1}

If we take another example here ,

𝑑2𝑦 𝑑𝑦
𝑥3 + 𝑥 + y = 0
𝑑𝑥2 𝑑𝑥

𝑑2𝑦 1 𝑑𝑦 1
⇒ + + y = 0
𝑑𝑥2 𝑥2 𝑑𝑥 𝑥3

Here we divided both sides of the equation with 𝑥 3 just to form the shape of the differential
equation of the form (i).

So comparing with (i) we get ,


1
P(x) = 𝑥2
1
And Q(x) =
𝑥3

Both P(x) and Q(x) are undefined at x = 0. So that differential equation is defined for all values
except 0.

So here ordinary points of the differential equation = { x : x ∈ ℝ & x ≠ 0 }


Page |6

Singular Point [1703004]


Let, the differential equation of the form,

𝑑2 𝑦 𝑑𝑦
+ 𝑃(𝑥) + 𝑄(𝑥)𝑦 = 0 ……….. (i)
𝑑𝑥2 𝑑𝑥

Singular Point: If P(x) or Q(x) or both P(x) and Q(x) are non-analytic at the point x = a, then
x = a is called singular point of equation ( i )

Two types of Singular Point exist. They are:

i) Regular Singular Point

ii) Irregular Singular Point

Explanation: The function f (z) = ez/z is analytic throughout the complex plane—for all values
of z—except at the point z = 0, where the series expansion is not defined because it contains the
term 1/z. The series is:

1/z + 1 + z/2 + z2/6 +⋯+ z n/(n+1)! +⋯

where the factorial symbol (k!) indicates the product of the integers from k down to 1. When the
function is bounded in a neighborhood around a singularity, the function can be redefined at the
point to remove it; hence it is known as a removable singularity. In contrast, the above function
tends to infinity as z approaches 0; thus, it is not bounded and the singularity is not removable (in
this case, it is known as a simple pole).

In general, because a function behaves in an anomalous manner at singular points, singularities


must be treated separately when analyzing the function, or mathematical model, in which they
appear.

Example Math:
𝑑 2𝑦 𝑑𝑦
2𝑥 2 2 + 𝑥 + (𝑥 2 + 1)𝑦 = 0
𝑑𝑥 𝑑𝑥

Comparing the above equation with Standard Form,

1 (𝑥2 +1)
𝑃(𝑥) = 𝑄(𝑥) =
2𝑥 2𝑥2

Here, as for x = 0 both P(x) and Q(x) is infinite/undefined, it is called a Singular Point.
Page |7

Regular Singular Point: [1703005]

Definition:
A point x = x0 is called an ordinary point of the equation,

y" + P(x)y' + Q(x)y = 0 …………..(1)

if both the functions P(x) and Q(x) are analytic at x = x0

If the point x = x0 is not an ordinary point of the differential equation (1), then it is called a singular
point of the differential equation (1). And a singular point x = x0 of the equation (1) is called a
regular singular point of the differential equation (1) if both (x = x0) P(x) and (x = x0)2 Q(x) are
analytic at x = x0.

Example-1:
Let the equation, x2y" + x y' + (x2 – n2)y = 0

Re-writing the given equation,

y" + (1/x)y' + {(x2 – n2) / x2}y = 0 ……(1)

Comparing (1) with y" + P(x)y' + Q(x)y = 0, we get, P(x) = 1/x and Q(x) = (x2 – n2) / x2.

Here (x – 0)P(x) = 1 and (x – 0)2 Q(x) = x2 – n2, showing that both (x – 0)P(x) and (x – 0)2 Q(x)
are analytic at x = 0.

Hence, x = 0 is a regular singular point of the equation, x 2y" + x y' + (x2 – n2)y = 0.

Example-2:
Let the equation, 9x(1 – x)y" - 12y' + 4y = 0 ………(1)

Dividing by 9x(1 – x), equation (1) can be put in standard form,

4 4
𝑦 ′′ − 𝑦′ + 𝑦 =0
3𝑥(1 − 𝑥 ) 9𝑥(1 − 𝑥 )
4 4
Comparing it with y" + P(x)y' + Q(x)y = 0, we have P(x) = − and Q(x) = .
3𝑥(1−𝑥) 9𝑥 (1−𝑥)

Since P(x) and Q(x) are not both analytic at x = 0, so x = 0 is not ordinary point of equation (1).
−4 4𝑥
Again x P(x) = 3(1−𝑥) and x2Q(x) = 9(1−𝑥) , showing that both P(x) and Q(x) are analytic at x = 0.

Therefore, x = 0 is a regular singular point of equation (1).


Page |8

Irregular Singular Point: [1703006]


A concept of irregular singular point that arises in the analytic theory of linear ordinary differential
equations.

Let, a second order differential equation is,

y" + P(x)y' +Q(x)y=0 ……………………(1)

Now, from the above equation, there are two functions P(x) and Q(x) by which regular singular
point and irregular singular point can be determined.

Case 1: (x-x0)P(x) is analytic at point x0 .

Case 2: (x-x0)2Q(x) is analytic at point x0 .


If case 1 and case 2 both are finite then it is called regular singular point.

If case 1 is finite and case 2 is infinite then it is called irregular singular point.

If case 1 is infinite then it is called irregular singular point.

Let A(t) be an (n × n) -matrix that is holomorphic in a punctured neighborhood of t ≠ ∞ and that


has a singularity at t0.

The point t0 is then called a singular point of the system

x˙=A(t)x ………………………… (I)

There are two non-equivalent definitions of an irregular singular point.


Definition 1: According to the first one, t0 is called an irregular singular point of (I) if A(t) has a
pole of order greater than one at t0 .

Definition 2: According to the second definition, t0 is called an irregular singular point of (I) if
there is no number σ>0 such that every solution x(t) grows not faster than |t−t0|-σ as t→t0 . The
case t0=∞ can be reduced to the case t0=0 by the transformation t→t-1.

An irregular singular point is sometimes called a strongly-singular point. In a neighborhood of an


irregular singular point the solutions admit asymptotic expansions.
Page |9

At regular singular point singularity can be removed. But at irregular singular point singularity can
not be removed.

Example:
x3 y" + x2 y' +3y=0 where, x=0 is a singular point. Prove that, x=0 is irregular singular point.

Solution: Given equation, : x3 y" + x2 y' +3y = 0


1 3
=> y" + 𝑥 y' + 𝑥3 y = 0

1
Where, p(x)=
𝑥

3
Q(x)=
𝑥3

For, x0 = 0,

Case 1 : P(x)(x-x0)
1
= (x-0)
𝑥

=1

Case 1 is finite.

Case 2 : Q(x)(x-x0)2

= 3
(x-0)2
𝑥3

3
=x

3
=0

Case 2 is infinite.

So, x= 0 is a irregular singular point.

[Proved]
P a g e | 10

Frobenius Method: [1703007]


Definition:

Consider the differential equation of the form

y2+ (F(x)/x) y1 + (G(x)/x2)y = 0

where y1 = d y/dx , y2 = d2y/dx2 ---(1)

and the functions F(x) and G(x) are analytic at x=0. Then the following method for solving (1) is
called Frobenius Method

Working rule for Solution by Frobenius Method:

Consider linear differential equation of order two:

f(x)y” + g(x)y’ + r(x)v = 0 -------------------------(1)

Step 1: Suppose that a trial solution of (1) be of the form:

= x k (c0 + c1x + c2x2+…..+ cm x m +….)

i.e.: y = x k ∑∞
𝑚=0 cm x
m
-----------------------------------(2)

Thus we take

: y = ∑∞
𝑚=0 cm x
m+k
where, c0 ≠0 -------------------------(3)

Step 2: Differentiate (3) and obtain

y’ = ∑∞
𝑚=0 cm (m +k) xm+k-1

And y” = ∑∞
𝑚=0 cm (m +k)(m+k-1)x
m+k-2
-----------------------(4)

Using (3) and (4), (1) reduces to and identity


P a g e | 11

Step 3:

Equating to zero the coefficient of the smallest power of x in the identity obtained in step 2
above, we obtain a quadric equation is k. The quadric equation so obtained is called the indicial
equation

Step 4:

Solve the indicial equation. The following cases arise:

1. The roots of indicial equation unequal and not differing by an integer


2. The roots of indicial equation unequal and not differing by an integer and making
a coefficient of y indeterminate.
3. The roots of indicial equation unequal and not differing by an integer and making
a coefficient of y infinite.
4. The roots of indicial equation equal
Step 5:

We equate to zero the coefficient of general power in the identity obtained in step 2. The
equation so obtained will be called the recurrence relation

Step 6:

If the recurrence relation connects c m and cm-2 then in general determine c 1 by equation to zero
the coefficient of the next higher power. On the other hand, if the recurrence relation connects c m
and cm-1 this step may be omitted.

Step 7:

After getting various coefficients with the help of steps 5 and 6 above, solution is obtained by
substituting these in (2) and (3) above. Other necessary working will be shown in details.
P a g e | 12

Case details: [1703008]


Case 1: The roots of indicial equation are unequal and not differing by an integer.

Let, k1 and k2 be the roots of the indicial equation. If k 1 and k2 do not differ by an integer. Then
in general, two independent solutions u and v are obtained by putting k=k 1 and k2 in the series
for y. Then the general solution is y= au + b v , where a and b are arbitrary constants.

For k=k1, we obtain a solution,

u(x) = xk1 ( c0 + c1x + ………….)

Proceeding similarly by using k= k 2 , we obtain a second solution

v(x) = xk2 ( c’0 + c’1x + …………)

Where c’0,c’1,….are the new values of the coefficients corresponding to k=k 2. Since k1-k2 is not
an integer , u/v is never constant. Thus u and v are two independent solutions. The general
solution will be,

Y= au + bv , where a and b are arbitrary constants.

Case 2: Roots of indicial equation are equal.

If the indicial equation has two equal roots k 1=k2=k, we obtain two independent solutions by
substituting this value of k in y and d y/dk.

The indicial equation has double root k if, (a 0 - 1)2 - 4b0 = 0 giving, r = (1-a0)/2

We obtain a first solution,

u(x) = x k (c0 + c1x + ………….)

To determine another solution, we use method of variation of parameters, that is, we replace the
constant c in the solution cu by a function w(x) to be determined such that,

v(x) = u(x) w(x)

Differentiating this equation, we got,

v1 = u1w + uw1

v2 = u2w + 2u1w1 + uw2

Since v(x) is a solution, by assumption, x 2v2 + F(x) xv1 + G(x) v = 0


P a g e | 13

Putting the values of v, v1, v2 we get,

x2(u2w + 2u1w1 + uw2) + x F(x)[u1w + uw1] + G(x)u w = 0

[x2u2 + F(x)xu1 + G(x)u] w + x2(2u1w1 + uw2) + x F(x)uw1=0

Since u(x) is a solution, we have,

x2u2 + F(x)xu1 + G(x)u=0

So, the above equation reduces to,

x2(2u1w1 + uw2) + x F(x)uw1=0

Dividing by x2u and putting the value of F(x) gives,


𝑢1 𝑎0
𝑤2 + (2 + + ⋯ ) 𝑤1 = 0
𝑢 𝑥
Now from u(x) we obtain,

𝑢1 𝑥 𝑘−1 [𝑘𝑐0 + (𝑘 + 1) 𝑐1 𝑥 + ⋯ ] 𝑘
= = +⋯
𝑢 𝑥 𝑘[𝑐0 + 𝑐1 𝑥 + ⋯ ] 𝑥

Then,

2𝑘 + 𝑎0
𝑤2 + ( + ⋯ ) 𝑤1 = 0
𝑥

But from, 2r + a0 = 1. Hence the above equation reduces to,

1
𝑤2 + ( + ⋯ ) 𝑤1 = 0
𝑥
Or,
𝑤2 1
= −𝑥 + ⋯
𝑤1

1
Integrating, log 𝑤1 = − log 𝑥 + ⋯ or, 𝑤1 = 𝑥 𝑒 (𝑥)

Expanding the exponential function in powers of x and integrating once more, we obtain the
expression of w always in the following form,

𝑤 = log 𝑥 + 𝑘1 𝑥 + 𝑘2 𝑥 2 + ⋯

Putting this value of w in v(x), we obtained desired another independent solution v(x). Then
general solution will be, y = au + b v, where a and b are arbitrary constants.
P a g e | 14

Case 3: Roots of indicial equation unequal, differing by an integer and making a coefficient
of y infinite.

If indicial equation has two unequal roots k 1 and k2 differing by an integer, and if some of the
coefficients of y become infinite when k=k 2, we modify the form of y by replacing c 0 by do (k-
k2). We then obtain two independent solutions by putting k=k2 in the modified form of y and
dy/dk. The result of putting k=k1 in y gives a numerical multiple of that obtained by putting k=k 2
and hence we reject the solution obtained by putting k=k 1 in y.

Let the roots k1 and k2 of the indicial equation differ by an integer, say k 1=k and k2=k-p, where p
is a positive integer, then as before one solution is,

u(x) = xk1 (c0 + c1x + c2x2+………….)

For k2 it’s may not be possible to determine another independent solution v(x) . In such case, we
determine v(x) by using method outlined in case-2. Thus as before we first obtained,

2𝑘 + 𝑎0
𝑤2 + ( + ⋯ ) 𝑤1 = 0
𝑥

k1 + k2 = - (a0 - 1) , using theory of equation.

Since, k1=k and k2=k-p, this gives 2k+a0 = p+1. Thus,

𝑤2 𝑝+ 1
= −( + ⋯)
𝑤1 𝑥

Integrating, log 𝑤1 = −(𝑝 + 1) log 𝑥 + ⋯

Or, 𝑤1 = 𝑥 𝑝+1 𝑒 (𝑥)

Expanding the exponential function and simplifying, we get,

1 𝑟1 𝑟𝑝
𝑤1 = + + ⋯ + + 𝑟𝑝+1 + 𝑟𝑝+2 𝑥 + ⋯
𝑥 𝑝+1 𝑥 𝑝 𝑥
Integrating again,

1
𝑤=− − ⋯ + 𝑟𝑝 𝑙𝑜𝑔 𝑥 + 𝑟𝑝+1 𝑥 + ⋯
𝑝𝑥 𝑝

putting this value of w in, v(x) = u(x) w(x), we obtain another solution v(x) and as usual the
general solution y = au + b v, where a and b are arbitrary constants.
P a g e | 15

Case 4: Roots of indicial equation unequal, differing by an integer and making a coefficient
of y indeterminate.

If the indicial equation has two roots k1 and k2 (say k1<k2) and if one of the coefficients of y
becomes indeterminate when k=k2, the complete solution is given by putting k=k 2 in y, which
then contains two arbitrary constants. The result of putting k=k1 in y merely gives a numerical
multiple of one of the series contained in the first solution. Hence, we reject the solution obtained
by putting k=k1.

Working rules of each cases with examples

Case 1: [1703009]

Roots of indicial equation unequal and not differing by an integer.

Rule:
Let k1 and k2 be the roots of the indicial equation. It k1 and k2 do not differ by an integer. Then,
in general, two independent solutions u and v are obtained by putting k=k1 and k2 in the series
for y. Then the general solution is v = au + b v .Where a and b are arbitrary constants.

Example 1: [1703010]
Solve the series: 9x(1-x)y′′ - 12y′ + 4y = 0

Solution:
Given equation is 9x(1-x)y′′ - 12y′ + 4y = 0 ……………. (1)
Dividing by 9x(1-x), equation 1 can be written in standard form

𝑑2 𝑦 4 𝑑𝑦 4𝑦
- 3𝑥(1−𝑥) + 9𝑥(1−𝑥) = 0
𝑑𝑥2 𝑑𝑥

Comparing it with y′′ + P(x)y′ + Q(x)y = 0, we have P(x) = -4 / [3x(1-x)] and


Q(x) = 4 / [9x(1-x)]. Since P(x) and Q(x) are not both analytic at x = 0, so
x = 0 is not an ordinary point of (1). Again, x P(x) = -4 / [3(1-x)] and 𝑥 2 Q(x) = 4x / [(1-x)],
showing that both P(x) and Q(x) are analytic at x = 0. So, x=0 is a regular singular point of (1).
To find solution of (1) we take,

𝑦 = ∑∞ 𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚 , where 𝐶0 ≠ 0…………….. (2)


So, 𝑦′ = ∑∞
𝑚=0 𝑐𝑚 (𝑘 + 𝑚)𝑥 𝑘+𝑚−1
P a g e | 16

And, 𝑦′′ = ∑∞
𝑚 =0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2 ……….(3)

Substituting the series (2) and (3) in (1), we have

9x(1-x)∑∞
𝑚 =0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2 - 12∑∞
𝑚 =0 𝑐𝑚 (𝑘 + 𝑚)𝑥 𝑘+𝑚−1 +
𝛼 𝑘+𝑚
4∑𝑚 = 0 𝑐𝑚 𝑥 =0

Or, 9x ∑∞
𝑚 =0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2 - 9𝑥 2 ∑∞
𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 −
𝑘+𝑚−2 ∞ 𝑘+𝑚−1 ∞ 𝑘+𝑚
1)𝑥 -12∑𝑚 = 0 𝑐𝑚 (𝑘 + 𝑚)𝑥 + 4∑𝑚 = 0 𝑐𝑚 𝑥 =0

Or, ∑∞
𝑚 =0 𝑐𝑚 {9(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) − 12(𝑘 + 𝑚)}𝑥 𝑘+𝑚−1 +
∑∞
𝑚 =0 𝑐𝑚 {4 − 9(𝑘 + 𝑚)(𝑘 + 𝑚 − 1)}𝑥 𝑘+𝑚 = 0 ………… (4)

But, 9(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) − 12(𝑘 + 𝑚) = 3(𝑘 + 𝑚)(3𝑘 + 3𝑚 − 7) …(5)


And, 4 − 9(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) = 4 − 9(𝑘 + 𝑚)2 + 9(𝑘 + 𝑚)
= −[9(𝑘 + 𝑚)2 − 9(𝑘 + 𝑚) − 4]
= −[9(𝑘 + 𝑚)2 − 12(𝑘 + 𝑚) + 3(𝑘 + 𝑚) − 4]
= −[3(𝑘 + 𝑚){3(𝑘 + 𝑚) − 4} + 3(𝑘 + 𝑚) − 4]
= -{3(𝑘 + 𝑚) − 4}{3(𝑘 + 𝑚) + 1}
Thus, 4 − 9(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) = -{3(𝑘 + 𝑚) − 4}{3(𝑘 + 𝑚) + 1}……(6)

Using (5) and (6),(4) can be re-written as, [1703011]


3∑∞
𝑚=0 𝑐 m (k +m)(3k+3m-7)x
k+m-1 ∑∞
- 𝑚=0 𝑐 m (3k+m-4)(3k+3m+1)x k +m =0, ………(7)

Which is an identity in x. Equating to zero the coefficient of the smallest power of x, namely xk-
1
,(7) gives the indicial equation,

3c0k(3k-7)=0 or k(3k-7)=0 [∵c00]

Thus k=0 and 7/3, ……………..(8)

Which are unequal and not differing by an integer. To obtain the recurrence relation, we equate
to zero the coefficient of xk+m-1 and obtain,

3cm (k +m)(3k+3m-7)-cm-1[3k+3(m-1)-4]*[3k+3(m-1)+1]=0
3𝑘+3𝑚−2
Or cm = cm-1 ........................................(9)
3(𝑘+𝑚)

𝑐 3𝑘+1
Taking m=1 in (9) gives c1= 30 ∗ ……………….(10)
𝑘+1
P a g e | 17

3𝑘 +4 3𝑘 +4 𝑐0 3𝑘+1
Next, taking m=2 in (9) gives c2= c1=3(𝑘+2) ∗ ∗ …………….(11)
3(𝑘+2) 3 𝑘+1

And so on, Putting these values in (2) i.e., y=x k(c0+c1x+c2x2 +……..) gives
3𝑘+1 (3𝑘+1)(3𝑘+4)
y=c0xk[1+3(𝑘+1)x+ 𝑥 2 +………………..] ……(12)
32

1 1∗4
Putting k=0 and replacing 𝑐0 by b in (12),y=a(1+3x+3∗6 𝑥 2 +………..)=au,say

7
7 8 8∗11
Next, putting k=3 and replacing c0 by b in (12), y=b𝑥 3 (1+10x+10∗13 𝑥 2 +………..)=bv,say.

The required solution is given by y=a u +b v i.e.


7
1 1∗4 8 8∗11
y= a(1+ 𝑥+ 𝑥 2+………..)+ b𝑥 3 (1+ x+ 𝑥 2 +………..)
3 3∗6 10 10∗13

Example- 2: [1703012]
Solve the following differential equation in series:

2𝑥 2 𝑦 ′′ − 𝑥𝑦 ′ + (1 − 𝑥 2 )𝑦 = 0

Solution:

Given 2𝑥 2 𝑦 ′′ − 𝑥𝑦 ′ + (1 − 𝑥 2 )𝑦 = 0 (1)

1 1−𝑥2 1
Dividing by 𝑥 2 , (1) takes standard form 𝑦 ′′ − (2𝑥 ) 𝑦 ′ + { 2𝑥2 } y so that both 𝑥𝑃 (𝑥 ) = −(2)
and 𝑥 2 𝑄(𝑥 ) = (1 − 𝑥 2 )/2. Thus both P(x) and Q(x) are analytic at x=0 and so x=0 is a
regular singular point. Let the series solution of (1) be of the form

(𝑦) 𝑛 = ∑𝑚=0 𝐶𝑚 𝑥 𝑘+𝑚 where 𝐶0 ≠ 0 (2)


Therefore, (𝑦) ′ = ∑𝑚=0 𝐶𝑚 (𝑘 + 𝑚)𝑥 𝑘+𝑚−1 and

(𝑦) ′′ = ∑ 𝐶 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2 (3)
𝑚=0 𝑚
P a g e | 18

Substitution for y, 𝑦 ′ , 𝑦 ′′ in (1), we have


2𝑥 2 ∑ 𝐶𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2


𝑚=0
∞ ∞
𝑘+𝑚−1
− 𝑥 ∑ 𝐶𝑚 (𝑘 + 𝑚)𝑥 + (1 − 𝑥 ) ∑ 𝐶𝑚 𝑥 𝑘+𝑚 = 0
2

𝑚=0 𝑚=0

or

∞ ∞

2 ∑ 𝐶𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚 − ∑ 𝐶𝑚 (𝑘 + 𝑚)𝑥 𝑘+𝑚 + ∑ 𝐶𝑚 𝑥 𝑘+𝑚


𝑚=0 𝑚=0
𝑚=0

− ∑ 𝐶𝑚 𝑥 𝑘+𝑚+2 = 0
𝑚=0

or
∞ ∞
𝑘+𝑚
∑ 𝐶𝑚 {2(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) − (𝑘 + 𝑚) + 1}𝑥 − ∑ 𝐶𝑚 𝑥 𝑘+𝑚+2 = 0
𝑚=0 𝑚=0

or
∞ ∞
∑𝑚=0 𝐶𝑚 {2(𝑘 + 𝑚)2 − 3(𝑘 + 𝑚) + 1}𝑥 𝑘+𝑚 − ∑𝑚=0 𝐶𝑚 𝑥 𝑘+𝑚+2 = 0

∞ ∞
or ∑𝑚=0 𝐶𝑚 (2𝑘 + 2𝑚 − 1)(𝑘 + 𝑚 − 1)𝑥 2 − ∑𝑚=0 𝐶𝑚 𝑥 𝑘+𝑚+2 = 0 (4)

[∴ 2(𝑘 + 𝑚)2 − 3(𝑘 + 𝑚) + 1 = 2(𝑘 + 𝑚)2 − 2(𝑘 + 𝑚) − (𝑘 + 𝑚) + 1 =


2(𝑘 + 𝑚)[(𝑘 + 𝑚) − 1] − [(𝑘 + 𝑚) − 1] = (𝑘 + 𝑚 − 1)[(2𝑘 + 2𝑚 − 1) ]

(4) is an identity. Equating to zero the coefficient of the smallest power of x namely 𝑥 𝑘 , (4)
gives the indical equation 𝐶0 (2𝑘 − 1)(𝑘 − 1) = 0 or (2k-1)(k-1)=0 [ ∴ 𝐶0 ≠ 0 ]

so that k=1 and k=1/2 (5)


P a g e | 19

[1703013]
Which are unequal and not differing by an integer. To obtain the recurrence

relation, we equate to zero the co-efficient of 𝑥 𝑘+𝑚 and obtain

𝑐𝑚 (2𝑘 + 2𝑚 − 1)(𝑘 + 𝑚 − 1) − 𝑐𝑚−2 = 0

𝑐𝑚−2
Giving 𝑐𝑚 = ------------------------------------(6)
(2𝑘 +2𝑚−1)(𝑘+𝑚−1)

To obtain 𝑐1, we now equate to zero the co-efficient of 𝑥 𝑘+1 and get
1
𝑐1(2k+1)k = 0, so that 𝑐1 = 0 for both roots k = 1 and k = of the indicial
2

equation. Then from (6) and 𝑐1 = 0

We have, 𝑐1 = 𝑐3 = 𝑐5 = ⋯ … … … . = 0 ---------------------------(7)

Further taking n = 2 in (6) gives


1
𝑐2 = 𝑐 ---------------------------------------(8)
(2𝑘+3)(𝑘+1) 0

Next taking n=4 in (6) and using (8) gives


1 1
𝑐4 = 𝑐 = 𝑐 -------(9)
(2𝑘+5)(𝑘+3) 2 (𝑘+1)(𝑘+3)(2𝑘+3)(2𝑘+5) 0

and so on, putting these values in (2) i.e 𝑦 = 𝑥 𝑘(𝑐0 + 𝑐1 𝑥 + 𝑐2 𝑥 2 + 𝑐3 𝑥 3 +. . )

gives
𝑥2 𝑥4
x = 𝑐0 𝑥 𝑘 [1 + ( + + ⋯ ] --------(10)
𝑘+1)(2𝑘+3 ) (𝑘+1)(𝑘+3)(2𝑘+3)(2𝑘+5)

Putting k = 1 and replace 𝑐0 by a in (10)


𝑥2 𝑥4
Gives y = ax[1 + 2.5 + 2.4.5.9 +…..] = au, say

1
Next putting k = 2 and replacing 𝑐0 by b in (10) gives

1
𝑥2 𝑥4
y = b𝑥 2 [1 + + + ⋯ … ] = bv, say (Ans)
2.3 2.3.4.7

The required series solution is given by y = au + b v where a and b are arbitrary


constants.
P a g e | 20

Example-3:
Solve the following differential equation in series.

3𝑥𝑦 ′′ + 2𝑦 ′ + 𝑦 = 0

Solution:

Given, 3𝑥𝑦 ′′ + 2𝑦 ′ + 𝑦 = 0 ----------(1)

Since x = 0 is a singular point, let



𝑦 = ∑𝑘=0 𝑎𝑘 𝑥 𝑚+𝑘 -------------------(2)

𝑦 ′ = ∑𝑘=0 𝑎𝑘 (𝑚 + 𝑘 )𝑥 𝑚+𝑘−1 ---------(3)

𝑦 ′′ = ∑𝑘=0 𝑎𝑘 (𝑚 + 𝑘 )(𝑚 + 𝑘 − 1 )𝑥 𝑚+𝑘−2 -------(4)

From we substitute equation (1) by equation (2) (3) (4) and we found

3x∑∞
𝑘=0 𝑎𝑘 (𝑚 + 𝑘 )(𝑚 + 𝑘 − 1)𝑥
𝑚+𝑘−2 + 2 ∑∞ 𝑎 (𝑚 + 𝑘 )𝑥 𝑚+𝑘−1 + ∑∞ 𝑎 𝑥 𝑚+𝑘 = 0
𝑘=0 𝑘 𝑘=0 𝑘

➔ 3∑∞
𝑘=0 𝑎𝑘 (𝑚 + 𝑘 )(𝑚 + 𝑘 − 1)𝑥
𝑚+𝑘−1
+ 2∑∞
𝑘=0 𝑎𝑘 (𝑚 + 𝑘 )𝑥
𝑚+𝑘−1
+ ∑∞
𝑘=0 𝑎𝑘 𝑥
𝑚+𝑘
=0

➔ ∑∞
𝑘=0 𝑎𝑘 (𝑚 + 𝑘 )𝑥
𝑚+𝑘−1 [ (
3 𝑚 + 𝑘 − 1) + 2] + ∑∞
𝑘=0 𝑎𝑘 𝑥
𝑚+𝑘
=0

➔ ∑∞
𝑘=0 𝑎𝑘 (𝑚 + 𝑘 )𝑥
𝑚+𝑘−1 [3𝑚 + 3𝑘 − 1] + ∑∞ 𝑎 𝑥 𝑚+𝑘 = 0 -----(5)
𝑘=0 𝑘

Equating the coefficient of the lowest degree term of x

i.e. 𝑥 𝑚−1

𝑎0 𝑚(3𝑚 − 1) = 0[𝑘 = 0]

-> m(3m – 1) = 0 [𝑎0 ≠ 0]

-> m = 0,1/3

y = ∑∞
𝑘=0 𝑎𝑘 𝑥
𝑚+𝑘
= 𝑥 𝑚 (𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + ⋯ )
P a g e | 21

[1703014]
Equating the coefficients of 𝑥 𝑚 ,

𝑎1 (𝑚 + 1)(3𝑚 + 2) + 𝑎0 = 0 [𝑘 = 1]
𝑎0
∴ 𝑎1 = −
(𝑚 + 1)(3𝑚 + 2)

Equating the coefficient of 𝑥 𝑚+𝑘 ,

𝑎𝑘+1 (𝑚 + 𝑘 + 1)(3𝑚 + 3𝑘 + 2) + 𝑎𝑘 = 0
𝑎𝑘
∴ 𝑎𝑘+1 =
(𝑚 + 𝑘 + 1)(3𝑚 + 3𝑘 + 2)

Putting 𝑘 = 0, 1, 2, 3, … here, we find:


𝑎1 𝑎0
𝑎2 = − =
(𝑚 + 2)(3𝑚 + 5) (𝑚 + 1)(𝑚 + 2)(3𝑚 + 2)(3𝑚 + 5)
𝑎2 𝑎0
𝑎3 = − =−
(𝑚 + 3)(3𝑚 + 8) (𝑚 + 1)(𝑚 + 2)(𝑚 + 3)(3𝑚 + 2)(3𝑚 + 5)(3𝑚 + 8)

and so on.
𝑎0 𝑎 𝑎
For 𝑚 = 0, 𝑎1 = − 2
; 𝑎2 = 200 ; 𝑎3 = − 480
0

1 𝑎0 𝑎 𝑎
For 𝑚 = 3, 𝑎1 = − ; 𝑎2 = 560 ; 𝑎3 = − 1680
0
4

Here, 𝑦 = 𝑥 𝑚 (𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + 𝑎3 𝑥 3 + ⋯ )
𝑎0 𝑎0 2 𝑎0 3
∴ (𝑦) 𝑚=0 = 𝑥 0 (𝑎0 − 𝑥+ 𝑥 − 𝑥 +⋯)
2 20 480
1 1 2 1 3
∴ (𝑦)𝑚=0 = 𝑎0 (1 − 𝑥 + 𝑥 − 𝑥 + ⋯)
2 20 480
1 𝑎0 𝑎0 2 𝑎0 3
∴ (𝑦)𝑚= 1 = 𝑥 3 (𝑎0 − 𝑥+ 𝑥 − 𝑥 + ⋯)
3 4 56 1680
1 1 4 1 7 1 10
∴ (𝑦) 1 = 𝑎0 (𝑥 3 − 𝑥 3 + 𝑥3 − 𝑥 3 +⋯)
𝑚=
3 4 56 1680

General solution: 𝑦 = 𝑐1 (𝑦)𝑚1 + 𝑐2 (𝑦) 𝑚2 (Ans) where c1 and c2 are arbitrary


constants.
P a g e | 22

Case-2: [1703016]

Roots of indicial equation are equal.

Rule:
If the indicial equation has two equal roots (i.e., 𝑘 = 𝑘1 ), we obtain two independent solutions
𝛿𝑦
by substituting this value of 𝑘 in 𝑦 and 𝛿𝑘 .

𝛿𝑦
𝑦 = 𝑐1 (𝑦)𝑚1 + 𝑐2 ( )𝑚1
𝛿𝑚

Example 1:

Solve the Bessel’s equation of order zero 𝑥𝑦" + 𝑦′ + 𝑥𝑦 = 0 in series.

Solution: Given, 𝑥𝑦" + 𝑦′ + 𝑥𝑦 = 0 … (1)


1
Dividing this by x, 𝑦" + ( )𝑦′ + 𝑦 = 0
𝑥

1
Comparing it with 𝑦" + 𝑃(𝑥)𝑦′ + 𝑄(𝑥)𝑦 = 0, here we find: 𝑃(𝑥 ) = , 𝑄(𝑥 ) = 1; so that
𝑥
𝑥𝑃 (𝑥 ) = 1 and 𝑥 2 𝑄(𝑥 ) = 𝑥 2 ; which are analytic at 𝑥 = 0. So, 𝑥 = 0 is a regular singular point.
Let the series solution of the given equation be of the form,

𝑦 = ∑∞
𝑚=0 𝐶𝑚 𝑥
𝑘+𝑚
… (2)

∴ 𝑦′ = ∑∞
𝑚=0 𝐶𝑚 (𝑘 + 𝑚)𝑥
𝑘+𝑚−1 and 𝑦" = ∑∞
𝑚=0 𝐶𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥
𝑘+𝑚−2 … (3)

Where, 𝐶0 ≠ 0. Putting the above values of 𝑦, 𝑦 ′ , 𝑦" into equation (1) gives,
∞ ∞ ∞

∑ 𝐶𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−1 + ∑ 𝐶𝑚 (𝑘 + 𝑚)𝑥 𝑘+𝑚−1 + ∑ 𝐶𝑚 𝑥 𝑘+𝑚+1 = 0


𝑚=0 𝑚=0 𝑚=0

∞ ∞
𝑘+𝑚−1
∴ ∑ 𝐶𝑚 {(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) + (𝑘 + 𝑚)}𝑥 + ∑ 𝐶𝑚 𝑥 𝑘+𝑚+1 = 0
𝑚=0 𝑚=0

∴ ∑∞ 2 𝑘+𝑚−1 + ∑∞
𝑚=0 𝐶𝑚 (𝑘 + 𝑚) 𝑥 𝑚=0 𝐶𝑚 𝑥
𝑘+𝑚+1 = 0 … (4)

Which is an identity. Equating to zero, the coefficient of the smallest power of x, namely 𝑥 𝑘−1
gives the indicial equation,

𝐶0 𝑘 2 = 0

Since, 𝐶0 ≠ 0, 𝑘2 = 0 ∴ 𝑘 = 0, 0
P a g e | 23

[1703017]
For recurrence relation, equating to zero the coefficient of 𝑥 𝑘+𝑚+ 1 in (4) gives
−𝐶𝑚−2
𝐶𝑚 (𝑘 + 𝑚)2 +𝐶𝑚−2 = 0 or, 𝐶𝑚 = … (5)
(𝑘+𝑚) 2

To compute 𝐶1 , equating to zero the coefficient of 𝑥 𝑘 gives 𝐶1 (𝑘 + 1) 2 = 0 so that 𝐶1=0 (for


k=0 is root of indicial equation). Using 𝐶1=0 and (5), we get

𝐶1 =𝐶3 =𝐶5 = 𝐶7 = … = 0 … (6)

Putting m = 2, 4, 6, … in turn in (5) and simplifying, we get


0 𝐶
𝐶2 =− (𝑘+2) 2

𝐶2 𝐶0
𝐶4 = − =
(𝑘+4) 2 (𝑘+2)2 (𝑘+4) 2

𝐶4 𝐶0
𝐶6 = − =
(𝑘+6) 2 (𝑘+2)2 (𝑘+2) 2 (𝑘+6) 2

and so on. Putting these values in (2)


𝑥2 𝑥4 𝑥6
𝑦 = 𝐶0 𝑥 𝑘 {1 − (𝑘+2) 2 + (𝑘+2) 2 (𝑘+4) 2 − (𝑘+2) 2(𝑘+4) 2(𝑘+6) 2 + ⋯ } … (7)

Putting k=0 and replacing 𝐶0 by 𝑎 in (7) gives


𝑥2 𝑥4 𝑥6
𝑦 = 𝑎 (1 − 22 + 22 .42 − 22 .42 .62 + ⋯ ) = 𝑎𝑢, say … (8)

To get another independent solution, substituting, (7) into the L.H.S of (1) and simplifying, we
find

𝑥 2 𝑦" + 𝑦′ + 𝑥𝑦 = 𝐶0 𝑘 2 𝑥 𝑘−1 … (9)

Differentiating both sides of (9) partially w.r.t. ‘k’ gives

𝜕 2 𝑑 2 𝑦 𝑑𝑦 𝐶0 𝜕 2 𝑘
[𝑥 + + 𝑥𝑦] = (𝑘 𝑥 )
𝜕𝑘 𝑑𝑥 2 𝑑𝑥 𝑥 𝜕𝑘

𝑑2 𝑑 𝜕𝑦 𝐶0
∴ [𝑥 2 + + 𝑥] = [2𝑘𝑥 𝑘 + 𝑘 2 𝑥 𝑘 log 𝑥] … (10)
𝑑𝑥2 𝑑𝑥 𝜕𝑘 𝑥

The presence of the factor k in each term on R.H.S of (10) shows that a second solution is
𝜕𝑦
( ) , where y is given by (7) . Differentiating partially w.r.t ‘k’ (7) gives
𝜕𝑘 𝑘=0
P a g e | 24

𝜕𝑦 𝑥2 𝑥4 𝑥6 𝑥2 −2
= 𝐶0 𝑥 𝑘 log 𝑥 {1 − (𝑘+2) 2 + (𝑘+2) 2 (𝑘+4) 2 − (𝑘+2) 2(𝑘+4) 2 (𝑘+6) 2 + ⋯ } + 𝐶0 𝑥 𝑘 {− (𝑘+2) 2 × 𝑘+2 +
𝜕𝑘
𝑥4 2 2 𝑥6 2 2 2
(𝑘+2) 2 (𝑘+4 )2
[− − ]−( [− − − ] + ⋯} … (11)
𝑘+2 𝑘+4 𝑘+2) 2(𝑘+4) 2 (𝑘+6) 2 𝑘+2 𝑘+4 𝑘+6

𝑑 1
[To find 𝑑𝑥 [ (𝑘+2) 2(𝑘+4) 2(𝑘+6) 2 ], we have proceeded as follows,

1
𝑧=[ ]
(𝑘 + 2)2 (𝑘 + 4)2 (𝑘 + 6)2

∴ log 𝑧 = −2 log(𝑘 + 2) − 2 log(𝑘 + 4) − 2 log(𝑘 + 6)

Differentiating w.r.t. ‘k’, we have

1 𝑑𝑧 2 2 2
=− − −
𝑧 𝑑𝑘 𝑘+2 𝑘+4 𝑘+6
𝑑𝑧 2 2 2
∴ = 𝑧 (− − − )
𝑑𝑘 𝑘+2 𝑘+4 𝑘+6
1
Where, 𝑧 = [ (𝑘+2) 2(𝑘+4) 2(𝑘+6) 2 ]; etc. Other terms can be similarly differentiated easily]

Putting k=0 and replacing 𝐶0 by 𝑏 in (11) gives


𝜕𝑦 𝑥2 𝑥4 𝑥6 𝑥2 𝑥4 1
(𝜕𝑘) 𝑘=0 = 𝑏 log 𝑥 (1 − 22 + 22 42 − 22 42 62 + ⋯ ) + 𝑏 {22 − 22 42 (1 + 2) +
𝑥6 1 1
(1 + + ) + ⋯ }
22 42 62 2 3

𝑥2 𝑥4 1
= 𝑏 [𝑢 log 𝑥 + {22 − 22 42 (1 + 2) + ⋯ }]

= 𝑏𝑣 , (say) [using (8)]

The required solution is

𝑦 = 𝑎𝑢 + 𝑏𝑣, where 𝑎 and 𝑏 are arbitrary constants.


P a g e | 25

Example 2: [1703018]
Solve in series (𝑥 − 𝑥 2 )𝑦 ′′ + (1 − 5𝑥 )𝑦 ′ − 4𝑦 = 0.

Solution: Given, (𝑥 − 𝑥 2 )𝑦 ′′ + (1 − 5𝑥 )𝑦 ′ − 4𝑦 = 0 … (1)

Dividing by (𝑥 − 𝑥 2 ), (1) gives

(1 − 5𝑥 ) ′ 4
𝑦 ′′ + { } 𝑦 − { } 𝑦 = 0.
(𝑥 − 𝑥 2 ) (𝑥 − 𝑥 2 )

(1−5𝑥)
Comparing it with 𝑦 ′′ + 𝑃(𝑥 )𝑦 ′ + 𝑄(𝑥 )𝑦 = 0, here 𝑃 (𝑥 ) = { (𝑥−𝑥2 )} and

4 (1−5𝑥) 4𝑥
𝑄(𝑥 ) = − {( } so that 𝑥𝑃(𝑥 ) = and 𝑥 2 𝑄(𝑥 ) = − ( . Since 𝑥𝑃(𝑥 ) and 𝑥 2 𝑄(𝑥 )
𝑥−𝑥2 ) (1−𝑥) 1−𝑥)
are analytic at 𝑥 = 0, hence 𝑥 = 0 is a regular singular point of (1).

Let the series solution of (1) be of the form



𝑦= ∑ 𝐶𝑚 𝑥 𝑘 + 𝑚 , where 𝐶0 ≠ 0. … (2)
𝑚= 0

∞ ∞

∴ 𝑦′ = ∑ 𝐶𝑚 (𝑘 + 𝑚) 𝑥 𝑘 + 𝑚−1 and 𝑦 ′′ = ∑ 𝐶𝑚 (𝑘 + 𝑚) (𝑘 + 𝑚 − 1) 𝑥 𝑘 + 𝑚−2 … (3)


𝑚=0 𝑚 =0

Putting above values of 𝑦, 𝑦 ′ , 𝑦 ′′ into (1) gives


∞ ∞

(𝑥 − 𝑥 2 ) ∑ 𝐶𝑚 (𝑘 + 𝑚) (𝑘 + 𝑚 − 1)𝑥 𝑘 + 𝑚−2 + (1 − 5𝑥 ) ∑ 𝐶𝑚 (𝑘 + 𝑚) 𝑥 𝑘 + 𝑚−1


𝑚 =0 𝑚= 0

−4 ∑ 𝐶𝑚 𝑥 𝑘 + 𝑚 = 0
𝑚=0

or

∑ 𝐶𝑚 {(𝑘 + 𝑚) (𝑘 + 𝑚 − 1) + (𝑘 + 𝑚)}𝑥 𝑘 + 𝑚−1


𝑚= 0

− ∑ 𝐶𝑚 {(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) + 5(𝑘 + 𝑚) + 4} 𝑥 𝑘 + 𝑚 = 0


𝑚= 0

or
∞ ∞

∑ 𝐶𝑚 (𝑘 + 𝑚)2 𝑥 𝑘 + 𝑚−1 − ∑ 𝐶𝑚 {(𝑘 + 𝑚) 2 + 4(𝑘 + 𝑚) + 4} 𝑥 𝑘 + 𝑚 = 0


𝑚 =0 𝑚=0
P a g e | 26

or
∞ ∞

∑ 𝐶𝑚 (𝑘 + 𝑚) 2 𝑥 𝑘 + 𝑚−1 − ∑ 𝐶𝑚 (𝑘 + 𝑚 + 2)2 𝑥 𝑘 + 𝑚 = 0 … (4)


𝑚=0 𝑚=0

which is an identity. Equating to zero the coefficient of the smallest power of 𝑥, namely 𝑥 𝑘−1 ,
(4) gives the identical equation

𝐶0 𝑘 2 = 0 or 𝑘 2 = 0. as 𝐶0 ≠ 0. Thus 𝑘 = 0, 0 (equal roots).

Finally, equating to zero the coefficient of x k +m in (4), we get [1703019]


(𝑘+𝑚+1) 2
Cm (k +m)2 – Cm-1(k+m+1)2 = 0 or 𝐶𝑚 = Cm-1 ... (5)
(𝑘+𝑚) 2

Putting m = 1, 2, 3, … in turn in (5) and simplifying


(𝑘+2) 2 (𝑘+3) 2 (𝑘+3) 2 (𝑘+4) 2 (𝑘+4) 2
𝐶1 = (𝑘+1) 2 𝐶0 𝐶2 = (𝑘+2) 2 𝐶1 = (𝑘+1) 2 𝐶0 𝐶3= (𝑘+3) 2 𝐶2 = (𝑘+1) 2 𝐶0

And so on. Putting these in (2),i.e. y = xk(C0 + C1x + C2x2 + C3x3 + … ), gives
(𝑘+2) 2 (𝑘+3) 2 (𝑘+4) 2
y = xkC0[1 + (𝑘+1) 2 x + (𝑘+1) 2x2 + (𝑘+1) 2x3 + … ] …(6)

putting k = 0 and replacing C0 by a in (6) gives,

y = a(1 + 22x + 32x2 + 42x3 + … ) = au, say … (7)

Differentiating partially w.r.t ‘k’, (6) gives,


𝜕𝑦 (𝑘+2) 2 (𝑘+3) 2 2 (𝑘+4) 2 3
= C0 xk log x[ 1 + x+ x + x + ……}
𝜕𝑘 (𝑘+1) 2 (𝑘+1) 2 (𝑘+1) 2

(𝑘+2) 2 2 2 (𝑘+3) 2 2 2 2
+ C0xk [(𝑘+1) 2 x {𝑘+2 –𝑘+1} + x {𝑘+3 – 𝑘+1} + …]
(𝑘+1) 2

Putting k = 0 and replacing C0 by b gives


𝜕𝑦 2
(𝜕𝑘) 𝑘=0 = b log x (1 + 22x + 32x2 + 42x3 + ………) + b{22 x(1-2) + 32. x2. 2(3 -2) + ........}

= b[u log x – 2(1.2x + 2.3x2 + ……..)]

=bv, say

The required solution is y = au + bv, where a and b are arbitrary constants.


P a g e | 27

Example 3: [1703020]
Solve the equation (𝑥 − 𝑥 2 )𝑦 ′′ + (1 − 𝑥 )𝑦 ′ − 𝑦 = 0 in series near x=0.

Solution:

Given, (𝑥 − 𝑥 2 )𝑦 ′′ + (1 − 𝑥 )𝑦 ′ − 𝑦 = 0 … (1)

Since x=0 is a regular singular point then let the series solution of equation (1) be of form,

y=𝑥 𝑘 (𝐶0 + 𝐶1 𝑥 + 𝐶2 𝑥 2 + ⋯ ) = ∑𝑚=0 𝐶𝑚 𝑥 𝑘+𝑚 , 𝐶0 ≠ 0 … (2)

𝑦 = ∑∞
𝑚=0(𝑘 + 𝑚)𝐶𝑚 𝑥
𝑘+𝑚−1
and 𝑦" = ∑∞
𝑚=0(𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝐶𝑚 𝑥
𝑘+𝑚−2
… (3)

Now, putting the value of 𝑦, 𝑦 ′ , 𝑦" in the equation (1) we have,

(𝑥 − 𝑥 2 ) ∑∞
𝑚=0(𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝐶𝑚 𝑥
𝑘+𝑚−2 +(1-x) ∑∞ (𝑘 + 𝑚) 𝐶 𝑥 𝑘+𝑚−1 -∑∞
𝑚=0 𝑚 𝑚=0 𝐶𝑚 𝑥
𝑘+𝑚 =0

or, ∑∞𝑚=0(𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝐶𝑚 𝑥


𝑘+𝑚−1 -∑∞ (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝐶 𝑥 𝑘+𝑚
𝑚=0 𝑚
∑ ∞ 𝑘+𝑚−1 ∑∞ 𝑘+𝑚 ∑∞ 𝑘+𝑚
+ 𝑚=0(𝑘 + 𝑚)𝐶𝑚 𝑥 - 𝑚=0(𝑘 + 𝑚)𝐶𝑚 𝑥 - 𝑚=0 𝐶𝑚 𝑥 =0

or, ∑∞
𝑚=0[(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) + (𝑘 + 𝑚)]𝐶𝑚 𝑥
𝑘+𝑚−1 -∑∞ [(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) +
𝑚=0
+(𝑘 + 𝑚)1]𝐶𝑚 𝑥 𝑘+𝑚 =0

or, ∑∞ 2
𝑚=0(𝑘 + 𝑚) 𝐶𝑚 𝑥
𝑘+𝑚−1 ∑∞
- 𝑚=0[(𝑘 + 𝑚)2 + 1]𝐶𝑚 𝑥 𝑘+𝑚 =0…………….(4)

Equation (4) is an identity. Equating to zero the coefficient of the smallest power of x, namely
𝑥 𝑘−1 , we obtain the indicial equation –

𝐶0 𝑘 2 = 0

So that, k= 0, 0 as 𝐶0 ≠ 0

Equating to zero the coefficient of 𝑥 𝑘+𝑚−1 in (4), we get:

(𝑘 + 𝑚)2 𝐶𝑚 − {(𝑘 + 𝑚 − 1)2 + 1}𝐶𝑚−1 = 0, for all m≥1


(𝑘+𝑚−1) 2 +1
𝐶𝑚 = 𝐶𝑚−1 , for all m≥1 …(5)
(𝑘+𝑚)
P a g e | 28

[1703021]
Putting m= 1, 2, 3… in (5) we have
𝑘2 +1
C1 = (k+1)2 C0 … (6)

(𝑘+1) 2+1 (𝑘2 +1) { (𝑘+1) 2+1}


C2 = C1 = C1, by (6) …(7)
(𝑘+2) 2 (𝑘+1) 2+(𝑘+2) 2

(𝑘+2) 2 +1 (𝑘2 +1) {(𝑘+1) 2 +1}+{ (𝑘+2) 2+1}


C3 = (𝑘+3)2
C2 = (𝑘+1) 1(𝑘+2)2 (𝑘+3) 3
C0

And so on, putting these values in (2), we have


(𝑘2 +1) (𝑘2 +1){(𝑘+1) 2 +1} (𝑘2 +1){(𝑘+1) 2+1}{ (𝑘+1) 2+1}
𝑦 = 𝐶0 𝑥^𝑘[1 + ( )2
+ (𝑘+1) 2 +(𝑘+2) 2
𝑥2 + (𝑘+1)+(𝑘+1) 2+(𝑘+3) 2
𝑥3 + ⋯ ] …. (8)
𝑘+1

Differentiating (8) partially w.r.t ‘k’ we have

𝜕𝑦 (𝑘2 +1) ( 𝑘2 +1){(𝑘+1) 2 +1} 2 (𝑘2 +1) {(𝑘+1) 2+1} { (𝑘+2) 2+1}
= C0 xk log x { 1+ x+ x + x3 + ⋯ } +
𝜕𝑥 (𝑘+1) 2 (𝑘+1) 2 (𝑘+2) 2 (𝑘+1) 2(𝑘+2) 2(𝑘+3) 2

2k 2(𝑘2 +1) 2𝑘({(𝑘+1) 2 +1}+2(𝑘+1)(𝑘2 +1) 2 (𝑘2 +1) {(𝑘+1) 2 +1}


C0 x k {[ 2- ]x+[ - -
(𝑘+1) (𝑘+1) 3 (𝑘+1) 2 (𝑘+2) 2 (𝑘+1) 3 (𝑘+2) 2
( 2 ) ( )
2 𝑘 +1 { 𝑘+1 +1}2
] x2 + ⋯ } … (9)
(𝑘+1) 2(𝑘+2)3

Putting k=0 and replacing C0 by a in (8)


2 2.5
y = a( 1+x+ 4 x 2 + 4.9 x 3 + ⋯)= au say,

Putting k=0 and replacing C0 by b in (8)


𝜕𝑦 2 2.5
(𝜕𝑥 ) 𝑘=0 = b logx (1+x+ 4 x 2 + 4.9 x 3 + ⋯)+ b(-2x-x2-…)

2 2.5
= b{ log x ((1+x+ 4 x 2 + 4.9 x 3 + ⋯)+ (-2x-x2-…)}

= b v, say

The required solution is y = au + b v, a being arbitrary constants.


P a g e | 29

Case 3: [1703022]

Roots of indicial equation, unequal, differing by an integer and making a


coefficient of y infinite.

Rule:
If the indicial equation has two unequal roots 𝑘1 and 𝑘2 (say 𝑘1 > 𝑘2) differing by an integer, and
if some of the coefficients of y become infinite when 𝑘 = 𝑘1 .We modify the form y by replacing
𝐶0 by 𝐷0 (𝑘 − 𝑘2 ) . We then obtain two independent solutions by putting 𝑘 = 𝑘2 in the modified
𝜕𝑦
form of y and . The result of putting 𝑘 = 𝑘1gives a numerical multiple of that obtained by
𝜕𝑘
putting 𝑘 = 𝑘2 and hence we reject the solution obtained by putting 𝑘 = 𝑘1in y.

Example-1: [1703023]
Solve in series 𝑥(1 − 𝑥 )𝑦 ′′ − 3𝑥𝑦 ′ − 𝑦 = 0 near x = 0.

Solution:

Given 𝑥(1 − 𝑥 )𝑦 ′′ − 3𝑥𝑦 ′ − 𝑦 = 0 ……………. (1)


𝑑2 𝑦 3 𝑑𝑦 1
Dividing by x (1-x), (1) yields − − 𝑦=0
𝑑𝑥2 1−𝑥 𝑑𝑥 𝑥 (1−𝑥)

Comparing it with 𝑦 ′′ + 𝑃(𝑥 )𝑦 ′ + 𝑄(𝑥 )𝑦 = 0,


3 1
Hence 𝑃 (𝑥 ) = − 1−𝑥 𝑎𝑛𝑑 𝑄(𝑥 ) = − 𝑥 (1−𝑥)

3𝑥 𝑥
So that 𝑥𝑃 (𝑥 ) = − 1−𝑥 𝑎𝑛𝑑 𝑥 2 𝑄(𝑥 ) = − 1−𝑥 .

Since 𝑥𝑃 (𝑥 ) and 𝑥 2 𝑄(𝑥 ) are both analytic at x = 0, so x = 0 is a regular singular point of (1).

Let the series solution of (1) be


𝑦 = ∑ 𝑐𝑚 𝑥 𝑘+𝑚 𝑤ℎ𝑒𝑟𝑒 𝑐0 ≠ 0 … … … … … … … … … … … … … . . (2)


𝑚=0

𝑦′ = ∑ 𝑐𝑚 (𝑘 + 𝑚)𝑘+𝑚+1
𝑚=0
P a g e | 30


′′
𝑦 = ∑ 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2 … … … … … … … … … . (3)
𝑚=0

Putting the above values of 𝑦, 𝑦 ′ , 𝑦′′ into (1) gives


∝ ∝ ∝
2) 𝑘+𝑚−2 𝑘+𝑚−1
(𝑥 − 𝑥 ∑ 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1) 𝑥 − 3𝑥 ∑ 𝑐𝑚 (𝑘 + 𝑚)𝑥 − ∑ 𝑐𝑚 𝑥 𝑘+𝑚
𝑚=0 𝑚=0 𝑚=0
=0

Or, ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−1 − ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚 −


3 ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚) 𝑥 𝑘+𝑚 − ∑∝𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚 = 0

Or, ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−1 − ∑∝𝑚=0 𝑐𝑚 {(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) + 3(𝑘 + 𝑚) +


1}𝑥 𝑘+𝑚 = 0

Or, ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−1 − ∑∝𝑚=0 𝑐𝑚 {(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) + 3(𝑘 + 𝑚) +


1}𝑥 𝑘+𝑚 = 0

Or, ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−1 − ∑∝𝑚=0 𝑐𝑚 {(𝑘 + 𝑚) 2 + 2(𝑘 + 𝑚) + 1}𝑥 𝑘+𝑚 = 0

Or, ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−1 − ∑∝𝑚=0 𝑐𝑚 (𝑘 + 𝑚 + 1)2 𝑥 𝑘+𝑚 = 0 … … … … (4)

Equating to zero the coefficient of the lowest power of x, namely 𝑥 𝑘+𝑚−1

Equation (4) gives the indicial equation

𝑐0 𝑘(𝑘 − 1) = 0

Or, 𝑘 (𝑘 − 1) = 0 since 𝑐0 ≠ 0

Which gives k = 0 and k = 1.


P a g e | 31

[1703024]
These are unequal and differ by an integer. Next, to find the recurrence relation we equate to
zero the coefficient of 𝑥 𝑘+𝑚−1 and obtain

𝐶𝑚 (𝑘 + 1)(𝑘 + 𝑚 − 1) − 𝐶𝑚−1 (𝑘 + 𝑚)2 = 0


𝑘 +𝑚
Or 𝐶𝑚 = 𝐶𝑚−1 ………………………………… (5)
𝑘+𝑚−1

( )
Putting m = 1 in (5) gives 𝐶1 = { 𝑘 + 1 ⁄𝑘 } 𝐶0 ………………………. (6)

𝑘+2 𝑘+2
Putting m = 2 in (5) and using (6) gives 𝐶2 = 𝐶1 = 𝐶0……… (7)
𝑘+1 𝑘

𝑘+3 𝑘+3
Putting m = 3 in (5) and using (7) gives 𝐶3 = 𝐶2 = 𝐶0 ……….. (8)
𝑘+2 𝑘

Putting these values in (2). i.e. 𝑦 = 𝑥 4 (𝐶0 + 𝐶1 𝑥 + 𝐶2 𝑥 2 + ⋯ ) Gives


𝑘+1 𝑘+2 𝑘+3
𝑦 = 𝐶0 𝑥 𝑘[1 + 𝑥+ 𝑥2 + 𝑥 3 + ⋯] …………………………. (9)
𝑘 𝑘 𝑘

If we put k = 0 in (9), we find that due to presence of the factor k in their denominators, the
coefficients becomes infinite. To remove this difficulty, we write 𝐶0 = 𝑘 do in (9). Then (9)
becomes

𝑦 = 𝑑0 𝑥 𝑘[𝑘 + (𝑘 + 1) 𝑥 + (𝑘 + 2)𝑥 2 + (𝑘 + 3) 𝑘 3 + ⋯ ] …………..……….. (10)

Putting k = 0 and replacing do by a in (10) gives

𝑦 = 𝑎(𝑥 + 2𝑥 + 3𝑥 2 +. . . ) = au, say …………………………………..…………… (11)

To obtain a second solution, if we put k = 1 in (9) we obtain

𝑦 = 𝐶0 (𝑥 + 2𝑥 + 3𝑥 2 + ⋯ ) ……………………………………………….… (12)

Which is not distinct (i.e. not linearly because ratio of the two series in (11) and (12) is a
constant from (11). Hence (12) will not serve the purpose of a second solution. In such a case the
𝜕𝑦
second independent solution is given by (𝜕𝑥 )𝑘=0 Differentiating (10) partially w.r.t ‘k’

𝜕𝑦
= 𝑑0 𝑥 𝑘𝑙𝑜𝑔𝑥 [𝑘 + (𝑘 + 1)𝑥 + (𝑘 + 2) 𝑥 2 + ⋯ ] + 𝑑0 𝑥 𝑘[1 + 𝑥 + 𝑥 2 + ⋯ ]……….. (13)
𝜕𝑘
P a g e | 32

Putting k = 0 and replacing 𝑑0 by b (13) gives

𝜕𝑦
( ) = 𝑏𝑙𝑜𝑔𝑥(𝑥 + 2𝑥 2 + 3𝑥 3 + ⋯ ) + 𝑏(1 + 𝑥 + 𝑥 2 + ⋯ )
𝜕𝑘 𝑘=0
𝜕𝑦
(𝜕𝑘) 𝑘=0 = 𝑏[𝑢𝑙𝑜𝑔𝑥 + (1 + 𝑥 + 𝑥 2 + ⋯ )] = 𝑏𝑣, 𝑏𝑦(11) …………. (14)

The required solution is 𝑦 = 𝑎𝑢 + 𝑏𝑣 , where a and b are arbitrary constants.

[1703025]

Example: 02
Solve in series the Bessel’s equation of order 2, near x=0, 𝑥 2 𝑦 ′′ + 𝑥𝑦 ′ + (𝑥 2 − 4)𝑦 = 0.

Solution:

Given 𝑥 2 𝑦 ′′ + 𝑥𝑦 ′ + (𝑥 2 − 4)𝑦 = 0.
1 (𝑥2 −4)
Dividing by 𝑥 2 , 𝑦 ′′ + (𝑥 ) 𝑦 ′ + [( )]y=0.
𝑥2

1
Comparing it with 𝑦 ′′ + 𝑃(𝑥 )𝑦 ′ + 𝑄(𝑥 )𝑦 = 0. ℎ𝑒𝑟𝑒 𝑃(𝑥 ) = and Q(x) = (𝑥 2 − 4)/𝑥 2 so
𝑥
that xP(x) =1 and 𝑥 2 𝑄(𝑥 ) = 𝑥 2 − 4. 𝑆𝑖𝑛𝑐𝑒 𝑥𝑃(𝑥 )𝑎𝑛𝑑 𝑥 2 𝑄(𝑥 )𝑎𝑟𝑒 𝑏𝑜𝑡ℎ 𝑎𝑛𝑎𝑙𝑦𝑡𝑖𝑐 𝑎𝑡 𝑥 = 0.
Hence x=0 is a regular singular point of (1). Let the series solution of (1) be


y = ∑𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚 , where 𝑐0 ≠ 0.

∞ ∞
y′ = ∑𝑚=0 𝑐𝑚 (𝑘 + 𝑚)𝑥 𝑘+𝑚−1 and y′′ = ∑𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2 …
(3)

Putting the above values of y, y’, 𝑦 ∞ into (1) gives,


∞ ∞ ∞
𝑥 2 ∑𝑚=0 𝑐𝑚 (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝑥 𝑘+𝑚−2 +x∑𝑚=0 𝑐𝑚 (𝑘 + 𝑚)𝑥 𝑘+𝑚−1 + 𝑥 2 ∑𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚 −

4 ∑𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚 = 0.
P a g e | 33



Or ∑ 𝑐𝑚 {(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) + (𝑘 + 𝑚) − 4}𝑥 𝑘+𝑚 + ∑𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚+2 =
𝑚=0
∞ ∞
0 or∑𝑚=0 𝑐𝑚 {(𝑘 + 𝑚)2 − 4}𝑥 𝑘+𝑚 + ∑𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚+2 = 0.

∞ ∞
Or ∑𝑚=0 𝑐𝑚 (𝑘 + 𝑚 + 2)(𝑘 + 𝑚 − 2)𝑥 𝑘+𝑚 + ∑𝑚=0 𝑐𝑚 𝑥 𝑘+𝑚+2 = 0 …. (4)

Which is an identify Equation to zero the coefficient of the smallest power of x. namely 𝑥 𝑘 ,
given the indicial equation 𝑐𝑜 (k+2) (k-2) =0 or (k+2) (k-2) =0 [𝑠𝑖𝑛𝑐𝑒 , 𝑐𝑜 ≠ 0]

[1703026]
This gives k = 2 and k = -2. This are unequal and differ by an integer. For the recurrence relation
we equate to zero the coefficient of 𝑥 𝑘+𝑚 and get
1
𝐶𝑚 (k+m+2) (k+m-2) + 𝐶𝑚−2 = 0 so 𝐶𝑚 = - 𝐶𝑚−2 …………………………….
(𝑘+𝑚+2)(𝑘+𝑚−2)
(5)

To determine𝐶1, we equate to zero the coefficient of xk+1 and get 𝐶1 (k+3) (k-1) = 0 giving C1 = 0
for both the roots k = 2 and k = -2 of the identical equation. Now using C1 = 0 and (5), we get

𝐶1 = 𝐶3 = 𝐶5 = 𝐶7 = ⋯ = 0................................................. (6)

Next, putting m = 2, 4, 6, in (5) and simplifying. We have

𝐶2 = − 𝐶0 ⁄𝑘(𝑘 + 4) 𝐶4 = 𝐶2 ⁄(𝑘 + 2)(𝑘 + 6) = 𝐶0 ⁄𝑘(𝑘 + 2)(𝑘 + 4)(𝑘 + 6) .

𝐶6 = − 𝐶4 ⁄(𝑘 + 4)(𝑘 + 8) = − 𝐶0 ⁄𝑘(𝑘 + 2)(𝑘 + 4)2 (𝑘 + 6)(𝑘 + 8)

And so on. Putting this value in (2), i.e.

y = 𝑥 𝑘 (𝐶0 + 𝐶1 𝑥 + 𝐶2 𝑥 2 + 𝐶3 𝑥 3 + 𝐶4 𝑥 4 + 𝐶5 𝑥 5 + 𝐶6 𝑥 6 + 𝐶7 𝑥 7 + 𝐶8 𝑥 8 + …………), we get


𝑥2 𝑥4
y =𝐶0 𝑥 𝑘 {1- + 𝑘(𝑘+2)(𝑘+4)(𝑘+6) −
𝑘 (𝑘+4)
𝑥6
+………}……………………….. (7)
𝑘 (𝑘+2)(𝑘+4) 2 (𝑘+6)(𝑘+8)

𝑥2 𝑥4 𝑥6
Putting k = 2 in (7) yields, y = 𝐶0 𝑥 2 {1- 2.6 + 2.4.8.6 − 2.4.62 8.10 +……} = 𝐶0 𝜔 , say …………….
(8)
P a g e | 34

Next if we putt k = -2 in (7), the coefficient of 𝑥 4 , 𝑥 6 , becomes infinite. To get rids of this
difficulty, we put 𝐶0 = 𝐷0 (𝑘 + 2) in (7) and obtained the modified solution as,
(𝑘+2)𝑥2 𝑥4
y = 𝐷0 𝑥 𝑘{(𝑘 + 2) − + 𝑘(𝑘+4)(𝑘+6) −
𝑘(𝑘+4)
𝑥6
2 +………}………………………. (9)
𝑘 (𝑘+4) (𝑘+6)(𝑘+8)

Putting k = -2 and replacing 𝐷0 by a in (9) gives,


𝑥4 𝑥6 −𝑎 𝑥2 𝑥2 𝑥4
y = 𝑎𝑥 −2 {0 − 0𝑥 2 + (−2)(2)(4) − (−2)(2) 3(4)(6) +…} or y = {1− 2.6 + 2.4.8.6 −…} = au, say...
16
(10)

Now (8) and (10) =>


− 𝑥2 𝑥2 𝑥4 −𝜔
u= {1− 2.6 + 2.4.8.6 −…} = 16 ,
16

Showing that 𝜔 and u are dependent solutions. Hence we must find one more independent solution
in order to obtain the required general solution.

To get another independent solution, substituting (9) into the L.H.S. of (1) and simplifying. We
find 𝑥 2 𝑦 ′′ + 𝑥𝑦 ′ + (𝑥 2 − 4)𝑦 = 𝐷0 (𝑘 − 2)(𝑘 + 2)2 𝑥 𝑘 ………….. (11)

Differentiating both side of (11) partially w.r.t. k, we get,

𝜕 𝑑 2𝑦 𝑑𝑦 𝜕
[ 𝑥2 2 + 𝑥 + (𝑥 2 − 4)𝑦] = 𝐷0 [(𝑘 + 2)2 . 𝑥 𝑘 . (𝑘 − 2)]
𝜕𝑘 𝑑𝑥 𝑑𝑥 𝜕𝑘

𝑑2 𝑑 𝜕𝑦
Or [ 𝑥 2 𝑑𝑥2 + 𝑥 𝑑𝑥 + (𝑥 2 − 4)] 𝜕𝑘 = 2𝐷0 (𝑘 + 2). 𝑥 𝑘 . (𝑘 − 2) + 𝐷0 (𝑘 + 2)2 [𝑥 𝑘 log 𝑥 . (𝑘 −
2) + 𝑥 𝑘 − 1 ]………………………………….. (12)

The presence of the factor (𝑘 + 2) in each term on R.H.S. of (12) shows that a second solution is
𝜕𝑦
(𝜕𝑘) 𝑘=2 where y is given by (9). Differentiating (9) partially w.r.t. ‘k’ we get

𝜕𝑦 (𝑘 + 2)𝑥 2 𝑥4 𝑥6
= 𝐷0 𝑥 𝑘 log 𝑥. {(𝑘 + 2) − + − + ⋯}
𝜕𝑘 𝑘 (𝑘 + 4) 𝑘 (𝑘 + 4)(𝑘 + 6) 𝑘(𝑘 + 4)2 (𝑘 + 6)(𝑘 + 8)
+

(𝑘 + 2) 𝑥 2 1 1 1 𝑥4 1 1 1
𝐷0 𝑥𝑘 {1 − ( − − )+ (− − − )
𝑘 (𝑘 + 4) 𝑘 + 2 𝑘 𝑘 + 4 𝑘 (𝑘 + 4)(𝑘 + 6) 𝑘 𝑘+4 𝑘+6

𝑥6 1 2 1 1
− 2 × (− − − − ) + ⋯… ……… .. }
𝑘 (𝑘 + 4) (𝑘 + 6)(𝑘 + 8) 𝑘 𝑘+4 𝑘+6 𝑘 +8
P a g e | 35

𝑑 𝑘 +2 𝑘 +2 𝑘 +2
[To find𝑑𝑘 {𝑘(𝑘+4)}, we proceed as follows: Take z = 𝑘(𝑘+4) so that log 𝑧 = log[𝑘(𝑘+4)]

Or log 𝑧 = log(𝑘 + 4) − log 𝑘 − log(𝑘 + 4)

Differentiating it w.r.t. ,‘k’, we get


1 𝑑𝑧 1 1 1 𝑑 𝑘+2 𝑘+2 1 1 1
= − 𝑘 − 𝑘+4 =≫ { }= ( − 𝑘 − 𝑘+4) Etc.
𝑧 𝑑𝑘 𝑘+4 𝑑𝑘 𝑘 (𝑘+4) 𝑘 (𝑘+4) 𝑘+4

Other terms can be similarly differentiated easily. ]

Putting k = -2 and replacing 𝐷0 by b in (13) gives

𝜕𝑦 𝑥4 𝑥6 𝑥2
= 𝑏𝑥 −2 log{0 − 0𝑥 2 + − + ⋯ } + 𝑏𝑥 −2 {1 +
𝜕𝑘 𝑘=−2 (−2)(2)(4) ( −2)(2) 2 (4)(6) 22
𝑥4 1 𝑥6 1 1 1
+ 2 ( )+ 3 ( −1+ − )+⋯
2 .4 4 2 . 4.6 2 4 6
𝑥 2 𝑥4
= 𝑏 [𝑢 log 𝑥 + 𝑥 −2 {1 + 2 + 2 2 − ⋯ }] = 𝑏𝑣
2 2 .4

by (10)

The required solution is 𝑦 = 𝑎𝑢 + 𝑏𝑣, where ‘a’ and ‘b’ are arbitrary constants.

Case 4: [1703027]

Roots of indicial equation unequal, differing by an integer and making a coefficient


of y indeterminate.

Rule:
If the indicial equation has two roots k1 and k2 (say k1<k2) and if the one of the coefficients of y
becomes indeterminate when k=k2, the complete solution is given by putting k=k 2 in y, which
then contains two arbitrary constants. The result of putting k=k 1 in y merely gives a numerial
multiple of one of the series contained in the first solution. Hence, we reject the solution obtained
by putting k=k1.
P a g e | 36

Example-1: [1703029]
Find solution near x = 0 of 𝑥 2 𝑦 ′′ +(𝑥 + 𝑥 2 )𝑦 ′ +(𝑥 − 9)𝑦 = 0

Solution:

Given 𝑥 2 𝑦 ′′ +(𝑥 + 𝑥 2 )𝑦 ′ +(𝑥 − 9) 𝑦 = 0 …...(1)

Dividing by 𝑥 2 , (1) can be put in standard form as

𝑦 ′′ + {(1 + 𝑥 ) ∕ 𝑥 }𝑦 ′ + {(𝑥 − 9) ∕ 𝑥 2 }𝑦 = 0

Comparing it with 𝑦 ′′ + 𝑃(𝑥 )𝑦 ′ + 𝑄(𝑥 )𝑦 = 0,

here 𝑃(𝑥 ) = (1 + 𝑥 )⁄𝑥 and 𝑄(𝑥 ) = (𝑥 − 9) ∕ 𝑥 2

so that 𝑥𝑃(𝑥 ) = 1 + 𝑥 and 𝑥 2 𝑄(𝑥 ) = 𝑥 − 9.

Since both 𝑥𝑃 (𝑥 ) and 𝑥 2 𝑄(𝑥 ) are analytic at x = 0, hence x = 0 is a regular singular point of (1).
Let the series solution of (1) be

𝑦 = 𝑥 𝑘(𝐶0 + 𝐶1 𝑥 + 𝐶2 𝑥 2 + . . 𝑎𝑑. 𝑖𝑛𝑓. ) = ∑𝑚=0 𝐶𝑚 𝑥 𝑘+𝑚 , where 𝐶0 ≠ 0 …(2)

So 𝑦 ′ = ∑𝑚=0(𝑘 + 𝑚)𝐶𝑚 𝑥 𝑘+𝑚−1

and 𝑦 ′′ = ∑𝑚=0(𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝐶𝑚 𝑥 𝑘+𝑚−2 …(3)

Putting the above values of 𝑦, 𝑦 ′ 𝑎𝑛𝑑 𝑦 ′′ in (1), we have


∞ ∞

𝑥 2 ∑ (𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝐶𝑚 𝑥 𝑘+𝑚−2 + (𝑥 + 𝑥 2 ) ∑ (𝑘 + 𝑚)𝐶𝑚 𝑥 𝑘+𝑚−1


𝑚=0 𝑚=0

+ (𝑥 − 9) ∑ 𝐶𝑚 𝑥 𝑚+𝑘 = 0
𝑚=0

∞ ∞
Or ∑𝑚=0(𝑘 + 𝑚)(𝑘 + 𝑚 − 1)𝐶𝑚 𝑥 𝑘+𝑚 + ∑𝑚=0(𝑘 + 𝑚) 𝐶𝑚 𝑥 𝑘+𝑚 +

∞ ∞ ∞

∑ (𝑘 + 𝑚) 𝐶𝑚 𝑥 𝑘+𝑚+1 + ∑ 𝐶𝑚 𝑥 𝑘+𝑚+1 − 9 ∑ 𝐶𝑚 𝑥 𝑘+𝑚 = 0


𝑚=0 𝑚=0 𝑚=0

∞ ∞
Or ∑𝑚=0{(𝑘 + 𝑚)(𝑘 + 𝑚 − 1) + (𝑘 + 𝑚) − 9}𝐶𝑚 𝑥 𝑘+𝑚 + ∑𝑚=0(𝑘 + 𝑚 + 1)𝐶𝑚 𝑥 𝑘+𝑚+1 = 0
P a g e | 37

∞ ∞
Or ∑𝑚=0 {(𝑘 + 𝑚)2 − 32 } 𝐶𝑚 𝑥 𝑘+𝑚 + ∑𝑚=0(𝑘 + 𝑚 + 1) 𝐶𝑚 𝑥 𝑚+𝑘+1 = 0

Or ∑∞
𝑚=0(𝑘 + 𝑚 + 3)(𝑘 + 𝑚 − 3) 𝐶𝑚 𝑥
𝑘+𝑚
+ ∑𝑚=0 (𝑘 + 𝑚 + 1) 𝐶𝑚 𝑥 𝑚+𝑘+1 = 0…….(4)

Equating to zero coefficient of the smallest power of x namely 𝑥 𝑘 , [1703028]


the above identity (4) gives the identical equation

(𝑘 + 3)(𝑘 − 3)𝐶0 so that k=3, -3 as 𝐶0 ≠ 0.

Next equating to zero the coefficient of 𝑥 𝑘+𝑚 in (4) , we get

(𝑘 + 𝑚 + 3)(𝑘 + 𝑚 − 3)𝐶𝑚 +(𝑘 + 𝑚) 𝐶𝑚−1 = 0


𝑘+𝑚
𝐶𝑚 =− ( 𝐶𝑚−1 ….(5) ,
𝑘+𝑚+3)(𝑘+𝑚−3)

Putting m=1,2,3 ….. in (5), we get


(𝑘+1)
𝐶1 = − (𝑘+4)(𝑘−2)
𝐶0 ,

(𝑘+2) (𝑘+1)(𝑘+2)
𝐶2 = - (𝑘+5)(𝑘−1) 𝐶1=(𝑘+4)(𝑘−2)(𝑘+5)(𝑘−1) 𝐶0,

−(𝑘+3) (𝑘+1)(𝑘+2)(𝑘+3)
𝐶3 = (𝑘+6)k 𝐶2 = - 𝑘(𝑘+4)(𝑘−2)(𝑘+5)(𝑘−1)(𝑘+6) 𝐶0 ,And so on.

Putting these values in (2), we have


(𝑘+1) (𝑘+1)(𝑘+2) (𝑘+1)(𝑘+2)(𝑘+3)
𝑦=𝐶0 𝑥 𝑘[1-(𝑘+4)(𝑘−2) 𝑥+(𝑘+4)(𝑘−2)(𝑘+5)(𝑘−1) 𝑥 2 +𝑘(𝑘+4)(𝑘−2)(𝑘+5)(𝑘−1)(𝑘+6) 𝑥 3 + ….ad Inf ] ….(6)

Putting k=3 and replacing 𝐶0 by a in (6), we have


4 4.5 4⋅5.6
y = a𝑥 3 [1- 𝑥+ 𝑥 2+ 𝑥 3 - …..] = au, say
2.7 2.1⋅7⋅8 3⋅1.7⋅8⋅9.2

putting k=-3 and replacing 𝐶0 by b in (6), we have

y=b𝑥 −3 {1-(2/5) 𝑥+(1/20) 𝑥 2} = b𝑣, say,

The required solution is 𝑦 = 𝑎𝑢 + 𝑏𝑣 ,where a and b are arbitrary constants.


P a g e | 38

Example-2. [1703030]
Solve in series (1-x2) 𝑦 ′′ -x𝑦 ′ +4y=0

Solution:

Given,

(1-x2) 𝑦 ′′ -xy′+4y=0 ….(1)

Clearly x=0 is an ordinary point of (1). Let the series solution of (1) be of the form

y =∑𝛼𝑚=0 𝐶𝑚 x k+ m where C0≠ 0 ….(2)

y′ =∑𝛼𝑚=0 𝐶𝑚 xk+m-1

and y n =∑𝛼𝑚=0 𝐶𝑚 (k + m)(k + m − 1)xk+m-2 …(3)

putting the above values of y, y′, 𝑦 ′′ in (1) gives

or (1-x2) ∑𝛼𝑚=0 𝐶𝑚 (k + m)(k + m − 1)xk+m-2 -x∑𝛼𝑚=0 𝐶𝑚 (k + m)xk+m-1 + 4∑𝛼𝑚=0 𝐶𝑚 x k+ m =0

or ∑𝛼𝑚=0 𝐶𝑚 (k + m)(k + m − 1)xk+m-2 - ∑𝛼𝑚=0 𝐶𝑚 (k + m)(k + m − 1)x k+ m -


∑𝛼𝑚=0 𝐶𝑚 (k + m)x k+ m + 4∑𝛼𝑚=0 𝐶𝑚 x k+ m =0

or ∑𝛼𝑚=0 𝐶𝑚 (k + m)(k + m − 1)xk+m-2 -

∑𝛼𝑚=0 𝐶𝑚 [(k + m)(k + m − 1) + (k + m) − 4]xk+ m =0

or ∑𝛼𝑚=0 𝐶𝑚 (k + m)(k + m − 1)xk+m-2 -∑𝛼𝑚=0 𝐶𝑚 [(k+m)2−4]xk+ m =0

or ∑𝛼𝑚=0 𝐶𝑚 (k + m)(k + m − 1)xk+m-2 - ∑𝛼𝑚=0 𝐶𝑚 (k + m + 2)(k + m − 2)x k+ m =0 …(4)

which is an identity. Equating to zero the coefficient of the smallest power of x, namely xk-2

(4) gives the indicial equation

C0k(k-1)=0 or k(k-1)=0 (As zero is not equal to C0)

giving k=1 and k=0.These are unequal and differ by an integer .

To get the recurrence relation, we equate to zero the coefficient of x k +m- 2 .Thus

Cm (k +m) (k+m-1)-Cm-2 (k +m) (k+m-4) =0


𝑘+𝑚−4
giving Cm = Cm-2 …(5)
𝑘+𝑚−1
P a g e | 39

Next, we equate to zero the coefficient of xk-1 and get C1(k+1) k=0 ...(6)

If we take k=0, (6) shows that C1 is indeterminate. With k=0 and using (5), We can express C2,
C4, C6…… in terms of C1 if we assume C1 to be finite. Thus,

Cm ={(m-4)/(m-1)} Cm-2

So that C2=-2C0, C4=0 and hence C6=C8=….=0

And C3=(-C1)/ (2) =-C1/2 C5=C3/ (4) =-C1/8, C7=(3C5)/ (6) =-C1/16.

And so on. Putting k=0 and these values in (2) i.e.

y=x k (C0+C1x+C2x2+C3x3+ C4x4+C5x5+……)

y= x0(C0+C2x2+ C4x4+……) + x0(C1x+C3x3+ C5x5+……), we get

y=C0(1-2x2) +C1(x- x3/2- x5/8+ x7/16-…….) …(7)

which is the required series solution C0 and C1 being two arbitrary constants.

You might also like