You are on page 1of 125

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269571767

Ordinary Differential Equations ‫اﻟﻤﻌﺎدﻻت اﻟﺘﻔﺎﺿﻠﻴﺔ اﻻﻋﺘﻴﺎدﻳﺔ‬

Book · April 2011

CITATIONS READS

0 3,536

1 author:

Arbab Ibrahim Arbab


University of Khartoum
196 PUBLICATIONS   1,016 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Electromagnetism Pedagogy View project

Title of the project Mechanical Wave simulation by fractal algorithm View project

All content following this page was uploaded by Arbab Ibrahim Arbab on 07 June 2017.

The user has requested enhancement of the downloaded file.


Ordinary Differential Equations
———————————————————————

Arbab Ibrahim Arbab

(Ph.D.)
Acknowledgments

This book is dedicated to many people who have been good to me. It is also
dedicated to the memory of my father, who spent his life in teaching people.

Last but not least, I dedicate this book to the memory of Prof. A. Salam, who

remains my ideal of what a genuine scientist and humanist should be.

I wish to thank Professor Al Tahir Al Agib, Gamal Al Abwabi and Badawi


Mohamed for their useful comments in reading the manuscript. I would also like
to thank Ms Sadia El Sir and Mr. Adil Adam for their assistance in preparing the

glossary. My thanks are also due to my students and colleagues who encourage
me for writing this book.

I will be very grateful to the readers of this book for their anticipated feedback

that will be considered in future editions.

1
Preface

This book has been prepared to serve as a textbook for students in science and
engineering taking a first course in ordinary differential equations. The course is

intended to serve as an introduction to the theory of ordinary differential equa-

tions. The solution of differential equations with variable coefficients tackled via

the method of series solution is given in chapter 5.


The famous differential equations like, Bessel, Hermite, Legendre, Laguerre,
Hypergeometric equations, are either solved or given as an exercise. Some phys-

ical applications to differential equations are given in chapter 6. These examples


elucidates the importance of the differential equations in physics and related sci-

ences.

An appendix containing standard integrals is put at the end of the book. To help

the students who had the subject in Arabic, we present a glossary for all words

appearing in the book.

April 2001, Khartoum Arbab, A.I

2
‫ﺑﺴﻢ اﻟﻠّﻪ اﻟﺮﺣﻤﻦ اﻟﺮﺣﯿﻢ‬

‫ﻤﻘﺩﻤـﺔ‬






















، ‫ﺍﻟﻤﺅﻟﻑ‬

‫ﺃﺭﺒﺎﺏ ﺇﺒﺭﺍﻫﻴﻡ ﺃﺭﺒﺎﺏ‬


Contents

1 Differential Equations 1
1.1 The origin of differential equations . . . . . . . . . . . . . . . . . 1
1.1.1 The order of a D.E. . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 The degree of a D.E . . . . . . . . . . . . . . . . . . . . . 2
1.1.3 The solution of differential equations . . . . . . . . . . . . 5
1.2 The general and particular solutions of the D.Es. . . . . . . . . . . 6
1.2.1 First order differential equations . . . . . . . . . . . . . . . 11
1.2.2 Homogeneous Equations . . . . . . . . . . . . . . . . . . . 15
1.2.3 Equations reducible to homogeneous equation . . . . . . . 19

2 Exact Differential Equations 21


2.0.4 Exact differential equation . . . . . . . . . . . . . . . . . . 21
2.1 Solution of an Exact D.E. . . . . . . . . . . . . . . . . . . . 24

3 First Order Linear Differential Equations 29


3.0.1 Linear Differential Equations . . . . . . . . . . . . . . . . . 29
3.0.2 Bernoulli’s Equation . . . . . . . . . . . . . . . . . . . . . 30
3.0.3 First order D.Es. with higher degrees . . . . . . . . . . . . 34
3.0.4 Clairaut’s Equation . . . . . . . . . . . . . . . . . . . . . 37
3.0.5 Ricatti’s Equation . . . . . . . . . . . . . . . . . . . . . . 39
3.0.6 Lagrange’s Equation . . . . . . . . . . . . . . . . . . . . . 40

i
4 Higher Order Linear Differential Equations 43
4.1 Homogeneous D.Es. with constant coefficients . . . . . . . . . . . 45
4.1.1 The complementary solution of a D.E. . . . . . . . . . . . 47
4.2 Non homogeneous linear D.Es. . . . . . . . . . . . . . . . . . . . 50
4.2.1 Operator Method . . . . . . . . . . . . . . . . . . . . . . 50
4.2.2 The properties of L(D) . . . . . . . . . . . . . . . . . . . 54
4.3 Solution of D.Es. by variation of parameters method . . . . . . . . 59
4.4 Solution of D.Es. by undetermined coefficients method . . . . . . . 62
4.5 Euler’s Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
d2 y
4.5.1 Second order D.E. of the form dx2
= f (y) . . . . . . . . . 68
d2 y dy
4.5.2 Second order D.E. of the form dx2
= f (y, dx ) . . . . . . . . 68

5 Series Solution 74
5.1 Series solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.2 Singular points, the method of Frobenius . . . . . . . . . . . . . . 82
5.3 Bessel’s Equation . . . . . . . . . . . . . . . . . . . . . . . . . . 89

6 Applications of Differential Equations 96


6.1 Motion of a particle in a resistive medium . . . . . . . . . . . . . . 96
6.2 Schrödinger Equation . . . . . . . . . . . . . . . . . . . . . . . . 98
6.3 Electric circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.4 Table of Integration . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.5 Answers to exercises . . . . . . . . . . . . . . . . . . . . . . . . . 109
6.6 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

1
Chapter 1

Differential Equations

1.1 The origin of differential equations


A differential equation (D.E) is an equation that involves derivatives.

dy d2 y dy d2 y 2 dy
(i) = x + 5, (ii) + 3 + 2y = 0, (iii) ( ) + + y = x3 ,
dx dx2 dx dx2 dx
∂z ∂z ∂ 2z ∂ 2z
(iv) =z+ , (v) + = x2 + y.
∂x ∂y ∂x2 ∂y 2
If the equation contains a single independent variable, as in (i), (ii) and (iii) the
derivatives are ordinary derivatives and the equation is called an ordinary differ-
ential equation. If the equation contains two or more independent variables, as
in (iv) and (v), the derivatives are partial derivatives and the equation is called
partial differential equation.

1.1.1 The order of a D.E.


Is the order of the highest derivative that appears in the equation. Thus (i) and
(iv) are first order while (ii), (iii) and (v) are second order.

1
1.1.2 The degree of a D.E
Is the degree of the highest ordered derivative that appears in the equation. Thus
(i), (ii) (iv), and (v) are first degree while (iii) is second degree. The origin of a
D.E could be
(a) geometric, (b) physical, (c) primitive. A primitive is a relation between the
variables which involves n arbitrary constants, e.g. y = Ax2 + Bx + C. In general
a primitive involving n arbitrary constants will give rise to a D.E., of order n, free
of arbitrary constants.
A geometrical origin:
dy
A slope of a curve is dx and the equation of the straight line is y = mx + c, where
dy dy
m= dx
, c the intercept. Hence, one can write y = dx
x − 2, as an equation of a
straight line. This equation represents a first order D.E.
Physical origin
A particle of mass m moves vertically downwards due to gravity. If the air resistance
is −kv, k= constant. Find its equation of motion. From Newton’s second law
of motion we have
d2 y dy
ma = mg − kv, a = 2
, v= ,
dt dt
where g is the acceleration due to gravity. This can be written as

d2 y k dy
2
=g− .
dt m dt
This is a second order D.E. of degree 1.
Primitive
(i) Given
y = Ax2 + Bx + C,

where A, B, C are arbitrary constants. Set up the D.E. for this primitive.

dy d2 y d3 y
= 2Ax + B, = 2A, = 0.
dx dx2 dx3

2
Thus
d3 y
= 0,
dx3
is free of any arbitrary constants.
(ii) Find the D.E. associated with

x2 y + y 2 x3 = 1.

Differentiate both sides with respect to (w.r.t) x, we get


dy dy
2xy + x2 + 2yx3 + 3x2 y 2 = 0,
dx dx
which can be written as
dy −3x2 y 2 − 2xy
= ,
dx x2 + 2yx3
which is a D.E. of first order.
(iii) Find the D.E. associated with the primitive

y = A cos ax + B sin ax,

where a, A, B are arbitrary constants.


dy d2 y
= −Aa sin ax + Ba cos ax, 2
= −Aa2 cos ax − Ba2 sin ax,
dx dx
d2 y
2
= −a2 (A cos ax + B sin ax) = −a2 y.
dx
Hence,
d2 y
+ a2 y = 0,
dx2
which is a D.E. of second order.
(iv) Obtain the D.E. associated with the primitive

y = Ae2x + Bex + C,

where A, B, C are arbitrary constants.


dy d2 y d3 y
= 2Ae2x + Bex , = 4Ae 2x
+ Be x
, = 8Ae2x + Bex .
dx dx2 dx3

3
To eliminate the constant, we write

dy d3 y d2 y dy
− 3 = 4Ae2x , − = 2Ae2x ,
dx dx dx2 dx
and hence,
dy d3 y d2 y dy
− 3 = 4Ae2x = 2( 2 − ),
dx dx dx dx
or
d3 y d2 y dy
− 3 + 2 = 0,
dx3 dx2 dx
which is the D.E. that corresponds to the primitive above and is free of any arbitrary
constant.
(v) Find the D.E. associated with the primitive y = Cx2 + C 2 .

dy 1 dy
= 2Cx, C = ,
dx 2x dx
and hence,
1 dy 2 1 dy 2
y= x +( ),
2x dx 2x dx
or
dy 2 dy
( ) + 2x3 − 4x2 y = 0,
dx dx
which is first order and second degree.
(vi) Find the D.E. associated with the family of circles of radius r centered on
the x-axis. This can be written as

y 2 + (x − C)2 = r2 ,

where C is an arbitrary constant. Differentiating both sides, we obtain


dy dy
2y + 2(x − C) = 0, or (x − C) = −y .
dx dx
Therefore,
dy 2
y2 + y2( ) = r2 ,
dx

4
is our required D.E.
Exercise 1.1
(1) Find the order and the degree of the following D.Es.
dy 2
(i) dxq
+ 3xy = 0 , (ii) 3 ddt2z + 4 dz
dt
= z + t.
dy d2 y dy 2 d y 2
(iii) dx
+ y = dx
, (iv) ( dx
2 ) y + dx 2 + x = 0.

(2) Find the D.Es. associated with the following primitives


(i) y = Ax , (ii) y = sin(x + A) , (iii) y = Ax + B,
(iv) y = Ax2 + Bx , (v) y = Aex + Bx,
where A, B are arbitrary constants.
(3) Find the D.E. of all straight lines at a unit distance from the origin.
(4) What are the D.Es. for the following primitives

(i) y = c1 e2x + c2 e−3x , (ii) y = cx + 2x2 , (iii) y = c1 sin x + c2 cos x.

1.1.3 The solution of differential equations


This is that of finding a relation between the variables involving n independent
arbitrary constants, together with the derivatives obtained from it, satisfying the
differential equation.
Example.
d3 y dy 2
dx3
= 0, corresponds to a primitive y = Ax2 + Bx + C and y 2 ( dx ) + y 2 = r2 ,
that corresponds to the primitive y 2 + (x − C)2 = r2 .
Existence Theorem
dy
A D.E. of the form y 0 ≡ dx
= g(x, y) for which
(a) g(x, y) is continuous and single valued over a region R of point (x, y)
∂g
(b) ∂x
exists and continuous at all points in R, admits many solutions f (x, y, C) =
0, where C is an arbitrary constant, such that through each point of R there passes
one and only one curve of the family f (x, y, C) = 0.

5
1.2 The general and particular solutions of the
D.Es.
The primitive is some times called the general solution of a D.E and the particular
solution is obtained from the general solution by assigning definite values to the
arbitrary constants appearing in the primitive. These are some times referred to
as initial conditions.
Example.
d3 y
= 0,
dx3
has the general solution(primitive) y = Ax2 + Bx + C. If we assign certain values
for the constants A, B, C we obtain the particular solutions.
Example. Set A = 1, B = −1, C = 3 then y = x2 − x + 3 is the corresponding
particular solution. If we set A = 0, B = 1, C = −5, we obtain the particular
solution, y = x − 5, and so forth. If the particular solution can not be obtained
from the general solution by certain choice of the arbitrary constant, the solutions
are called singular solutions
Examples.
(i) Show that
y = 2x + Cex (1.1)

is the general solution of the D.E.


dy
− y = 2(1 − x),
dx
dy
and find the particular solution satisfied by x = 0, y = 3. Substituting dx
= 2+Cex
in the above D.E. one obtains

(2 + Cex ) − (2x + Cex ) = 2 − 2x = 2(1 − x) = R.H.S.

Therefore, y = 2x+Cex is the general solution. The particular solution is obtained


by finding the value of C. The point x = 0, y = 3 in equation (1.1) gives 3 = 0+C,

6
or C = 3. Hence, y = 2x + 3ex is the required particular solution.
(ii) Show that
y = C1 ex + C2 e2x + x, (1.2)

is the general solution (primitive) of the D.E.

d2 y dy
2
− 3 + 2y = 2x − 3,
dx dx
and find the equation of the curve through the points (0,0) and (1,0).

dy d2 y
= C1 ex + 2C2 e2x + 1, 2 = C1 ex + 4C2 e2x ,
dx dx
hence,

d2 y dy
2
− 3 + 2y = C1 ex + 4C2 e2x − 3(C1 ex + 2C2 e2x + 1) + 2(C1 ex + C2 e2x + x)
dx dx
= 2x − 3 = R.H.S.

The points (0,0) and (1,0) give, in equation (1.2), the two equations:

0 = C1 + C2 and 0 = C1 e + C2 e2 + 1,

1
which can be solved to yield C1 = −C2 = e2 −e
, thus the equation of the curve
through the given points is

ex e2x
y= − +x
e2 − e e2 − e
which is a particular solution.
(iii) Show that
(y − C)2 = Cx, (1.3)

is the general solution of the D.E.


dy 2 dy
4x( ) + 2x − y = 0,
dx dx

7
and find the equation of the curves through the point (1,2). Differentiating (1.3)
yields
dy
2(y − C) = C,
dx
hence, the D.E. becomes
C C 4xC 2 + 4xC(y − C) − 4y(y − C)2
4x( )2 + 2x( )−y = =0
2(y − C) 2(y − C) 4(y − C)2
=R.H.S.
The point x = 1, y = 2 gives, in equation (1.3), (2 − C)2 = C, or 4 − 4C + C 2 =
C, C 2 − 5C + 4 = 0, which can be written as (C − 1)(C − 4) = 0, C = 1, C = 4.
This gives two particular solutions, viz.,

(y − 1)2 = x and (y − 4)2 = 4x.

(iv) If the general solution of the D.E.,


dy y
= is y = Cx, (1.4)
dx x
where C is an arbitrary constant. Find the equation of the curve through the
points (a) (1,2), (b) (0,0). The point (1,2) gives, in equation (1.4), C = 2 so that
the equation of the curve is y = 2x.
The point (0,0) gives, in equation (1.4), 0=0, and C is not determined!, and all
of the curves that pass through the point (0,0). This is because
y
g(x, y) =
x
is not continuous at the point (origin) and hence, the existence theorem assures
that one and only one curve of the family

y = Cx

passes through each point except the origin.


(v) If
xy = C(x − 1)(y − 1), (1.5)

8
obtain the D.E.
dy
x(x − 1) + y(y − 1) = 0,
dx
we see that the solution y = 0 can be obtained from the general solution equation
(1.5) by putting C = 0, but y = 1 can not be obtained from the general solution
of equation (1.5) by any value of C and hence, this solution is a singular solution.
(vi) The primitive
y = Cx + 2C 2 , (1.6)
and
dy
= C,
dx
give
dy dy
y= x + 2( )2 . (1.6a)
dx dx
The equation of the parabola y = − 18 x2 satisfies equation (1.6a). We see that
the primitive is represented by a family of straight lines (1.6) and it is clear that
equation of a parabola can’t be obtained from this by any value of C, such a
solution is called singular solution.
(vii) Show that the two general solutions

y = C1 cos x + C2 sin x , (1.7)

and
y = A sin(x + B) , (1.8)
where A, B, C1 , C2 are arbitrary constants, satisfies the D.E.
d2 y
+ y = 0,
dx2
and show that how they are made equivalent.
From equation (1.7) we have,
dy d2 y
= −C1 sin x + C2 cos x, 2 = −C1 cos x − C2 sin x = −(C1 cos x + C2 sin x),
dx dx

9
or
d2 y
+y =0 .
dx2
Now consider equation (1.8):
dy d2 y
= A cos(x + B), = −A sin(x + B) = −y,
dx dx2
and
d2 y
+y =0 .
dx2
Now equation (1.8) can be written as

y = A sin(x + B) = A sin B cos x + A cos B sin x,

and by setting
C1 = A sin B, C2 = A cos B,
the above equation becomes

y = C1 cos x + C2 sin x,

which is the same as (1.7) above.


(viii) Show that
y2
ln x2 + ln = A + x,
x2
can be written as
y 2 = Bex ,
where B is an arbitrary constant.
y2 2y
2
ln x2 + ln = ln(x ) = ln y 2 = A + x,
x2 x2
or y 2 = eA+x = eA .ex = Bex , where B = eA =constant.
Exercise 1.2
Which of the following solutions are general solutions and which are particular
solutions

10
dy
(i) y = 2x2 , for the D.E. x dx = 2y.

dy
(ii) x2 + y 2 = C, for the D.E. y dx + x = 0.

dy dy 4
(iii) y = Cx + C 4 , for the D.E. y = x dx + ( dx ).

dy
(iv) (1 − x)y 2 = x3 , for the D.E. 2x3 dx = y(y 2 + 3x2 ).

d2 y
(v) y = C1 ex + C2 e−x , for the D.E. dx2
− y = 0.

d2 y dy
(vi) y = C1 ex + C2 e2x , for the D.E. dx2
− 3 dx + 2y = 0.

1.2.1 First order differential equations


A D.E. of first order and first degree can be written in the form

M (x, y)dx + N (x, y)dy = 0. (1.9)

Example.

dy 2x − y
= ,
dx y + 2xy + 3
can be written as
(2x − y)dx + (y + 2xy + 3)dy = 0,

where
M (x, y) = 2x − y , N (x, y) = y + 2xy + 3 .

If M dx + N dy is a complete (exact) differential of a function µ(x, y), i.e., M dx +


N dy = dµ, then µ = C is the primitive (general solution) of the D.E.
Example.
3x2 y 2 dx + 2x3 ydy = 0,

11
is an exact D.E., since

3x2 y 2 dx + 2x3 ydy = d(x3 y 2 ).

Its primitive is
µ = x3 y 2 = C or y 2 = Cx−3 .
If equation (1.9) is not exact but

ξ(x, y)(M dx + N dy) = dµ,

where ξ(x, y) is known as an integrating factor (I.F.), then µ = C is the primitive


of the D.E.
Example.
3ydx + +2xdy = 0,
is not an exact differential equation, but when multiplied by ξ = x2 y becomes
exact. Thus

x2 y(3ydx + 2xdy) = 3x2 y 2 dx + 2x3 ydy = d(x3 y 2 ),

is exact. Its primitive is µ = x3 y 2 = C, or y 2 = Cx−3 . If equation (1.9) is


not exact and no integrating factor can be found readily, it may be possible by a
change of variables to obtain an equation for which the integrating factor can be
found.
Example.
Solve
(x + y + 1)dx + (2x + 2y + 3)dy = 0.
Now let x + y = t and dx + dy = dt, hence,
t+1
(t + 1)(dt − dy) + (2t + 3)dy = 0 , or dy + dt = 0,
t+2
which can be written as
1
dy + dt − dt = 0,
t+2

12
by integrating this we obtain,

y + t − ln(t + 2) = C , or 2y + x − ln(x + y + 2) = C,

which is the primitive.


Exercise. Try the solution by putting x + y + 1 = t.
A D.E. for which the integrating factor can be found readily has the form

f1 (x)g2 (y)dx + f2 (x)g1 (y)dy = 0,

which when multiplied by (integrating factor)


1
,
f2 (x)g2 (y)

reduces to
f1 (x) g1 (y)
dx + dy = 0,
f2 (x) g2 (y)
and finally
Z Z
f1 (x) g1 (y)
dx + dy = C,
f2 (x) g2 (y)
is the primitive. Such equation is called variables separable.
Examples.
(i) Find the primitive for

(3x2 y − yx)dx + (2x3 y 2 + x3 y 4 )dy = 0.

y(3x2 − x)dx + x3 (2y 2 + y 4 )dy = 0,

in this case

f1 (x) = 3x2 − x, f2 (x) = x3 , g2 (y) = y, g1 (y) = 2y 2 + y 4 .

Hence,
Z Z
3x2 − x 2y 2 + y 4
dx + dy = C,
x3 y

13
and Z Z
3 1
( − 2 )dx + (2y + y 3 )dy = C,
x x
y4
3 ln x + x−1 + y 2 + = C,
4
is the required primitive.
(ii) Find the general solution of the D.E.

(x − 1)2 ydx + x2 (y + 1)dy = 0.

Separating the variables (x and y) one can write the above equation as
(x − 1)2 y+1
2
dx + dy = 0.
x y
Integrate both sides to obtain
Z Z
2 1 1
(1 − + 2 )dx + (1 + )dy = C,
x x y
or
1
x − 2 ln x − + y + ln y = C,
x
is the primitive.
If equation (1.9) admits a solution

f (x, y, C) = 0,

where C is an arbitrary constant, there exist many integrating factors, ξ(x, y) such
that
ξ(x, y)(M (x, y)dx + N (x, y)dy) = 0,
is exact.
Example.
xdy − ydx = 0,
can be put as
dy dx
− = 0,
y x

14
and integrate to obtain
y y
ln y − ln x = ln C, ln( ) = ln C, = C,
x x
and
y = Cx,

is the primitive.
1
(b) If we multiply by x2
, we get

xdy − ydx y
2
= d( ) = 0 i.e., y = C1 x.
x x
1
Similarly, if we multiply the equation by y2
,
we get
xdy − ydx x
2
= d(− ) = 0,
y y
or
−x y 1
= C2 , =− = C = C1 .
y x C2
Thus, the form of the primitive is not unique.

1.2.2 Homogeneous Equations


A function f (x, y) is called homogeneous function of degree n if

f (λx, λy) = λn f (x, y).

Examples.
(i)

f (x, y) = x3 − xy 2 , f (λx, λy) = (λx)3 − (λx)(λy)2 = λ3 (x3 − xy 2 ) = λ3 f (x, y),

hence, the function is homogeneous of degree 3.


(ii) The function
f (x, y) = x2 + y,

15
is not homogeneous, since

f (λx, λy) = (λx)2 + (λy) = λ(λx2 + y) 6= λn f (x, y),

for some number n.


(iii)
1
f (x, y) = ,
x3
+ 2x2 y + 5y 3
1
f (λx, λy) = ,
(λx)3 + 2(λx)2 (λy) + 5(λy)3
1
f (λx, λy) = 3 3 = λ−3 f (x, y),
λ (x + 2x2 y + 5y 3 )
hence, the equation is homogeneous of degree -3. A D.E.

M dx + N dy = 0,

is called homogeneous if both M and N are homogeneous and of the same degree.
A homogeneous D.E. can be made integrable by the transformation y = vx
Examples.
(i) Find the general solution of the D.E.

(x3 + y 3 )dx − 3xy 2 dy = 0.

It is clear this equation is a homogeneous equation of degree 3 (check this). There-


fore one can use the transformation y = vx, so that dy = vdx + xdv, and hence,
our equation becomes

(x3 + v 3 x3 )dx − 3x(vx)2 (vdx + xdv) = x3 (1 − 2v 3 )dx − 3v 2 x4 dv = 0,

or
dx 3v 2 dv
− dv = 0.
x 1 − 2v 3
Integrating this we obtain

ln x + ln (1 − 2v 3 )1/2 = ln(x(1 − 2v 3 )1/2 ) = ln C,

16
or
y3
x2 (1 − 2
) = C 2 , x3 − 2y 3 = C 2 x.
x3
(ii) Find the general solution of the D.E.

(x2 − xy)dx + x2 )dy = 0.

Clearly, the equation is homogeneous of degree 2 (both M and N are homogeneous


of degree 2), therefore one uses the substitution y = vx so that dy = vdx + xdv,
and hence
[x2 − x(vx)]dx + 2x2 [vdx + xdv] = 0,
or Z Z
2 3 dx dv
x (1 − v + 2v]dx + 2x dv = 0, = −2 .
x 1+v
This yields
ln x + 2 ln(1 + v) = ln C, ln x(1 + v)2 = ln C.
Thus, the general solution is

y= Cx − x.
Suppose
dy y
= f( )
dx x
Such a D.E. is called homogeneous equation. In such a case, we make the following
transformation xy = v, so that
dv
x + v = f (v).
dx
Example.
Find the general solution of the following D.E. below.
dy y y
= 2 + ( )2 .
dx x x
y
Let x
= v, so that this equation becomes
dv
x + v = 2v + v 2 ,
dx

17
or
dv
x = v + v2,
dx
which is a separable equation, i.e.,
Z Z Z
dv dv dx
2
= = ,
v+v v(1 + v) x
or
Z Z
1 1 dx Z dv dv Z
dx
( − )dv = , ( − )= , ln v−ln(1+v) = ln x+ln C.
v 1+v x v 1+v x
Thus,
y
v x y Cx2
= Cx, or = = Cx, y = .
1+v 1 + xy x+y 1 − Cx
Exercise 1.3
(1) Find the general solution (primitive) of the following D.Es.

(i) xydx + (1 + x2 )dy = 0 , (ii) (y 2 − x2 )dx + xydy = 0 ,

(iii) (x3 − 3y 3 )dx + 3xy 2 dy = 0 , (iv) (xy + y 2 )dx + (x2 − yx)dy = 0.

(2) Find the solution of the following D.Es. by suitable substitution

(i) (y − xy 2 )dx − x(1 + xy)dy = 0 , (ii) dy − (y − 4x)2 dx = 0.


(3) Find the particular solution of the following D.Es.

(i) (y 2 + xy)dx − x2 dy = 0, when x = 1, y = 1,

(ii) xdy + 2ydx = 0, when x = 2, y = 1,

(iii) (y 2 + x2 )dx + xydy = 0, when x = 1, y = −1.


(4) Solve the following D.E.

(x2 + y 2 )dx − xydy = 0.

18
(5) Solve the following D.E.

dy
2xy 3 + 3x2 y 2
= 0 ; y(1) = 1.
dx
(6) Find all functions f (x) such that the D.E.
dy
y 2 sin x + yf (x) = 0, is exact.
dx
Solve the D.E. for these functions f (x).
(7) The D.E.
dy
+ x2 + y = 0
f (x)
dx
is known to have an integrating factor ξ = x. Find all possible functions f (x).

1.2.3 Equations reducible to homogeneous equation


Consider an equation of the type
dy a1 x + b1 y + c1
= ,
dx a2 x + b2 y + c2
where a1 , a2 , b1 , b2 , c1 , c2 are constants. When c1 = 0 = c2 or a1 b2 − a2 b1 = 0
the solution of the equation has been investigated previously. However, when
a1 b2 − a2 b1 6= 0 one employs a different technique.
Example.
Solve the following D.E.
dy 2x + 3y − 4
= .
dx 4x + y − 3
Let 2X + 3Y = 2x + 3y − 4 , 4X + Y = 4x + y − 3. Solving for X, Y gives
1
X =x− , Y =y−1 . (1.10)
2
dy dY
Since dx
= dX
, the D.E. becomes

dY 2X + 3Y
= .
dX 4X + Y

19
This is a homogeneous equation in X and Y . Its solution can easily be found by
putting Y = V X so that dY = V dX + XdV . Thus,
Z Z
dV 2X + 3V x 2 + 3V dX 4+V
V +X = = , or =− 2
dV.
dX 4X + V X 4+V X V +V −2
Thus,
2 5
ln X = A + ln(2 + V ) − ln(1 − V ).
3 3
Y y−1
Now V = X
= x− 12
so that

1 2 2x + y − 2 5 x − y + 12
ln(x − ) = A + ln − ln .
2 3 x − 12 3 x − 12

Exercise 1.4
Solve the following first order D.Es.
(i) (x + 2y − 3) dx − (2x + y − 3) dy = 0 , (ii) (x − 8y + 7) dy − (x − y) dx = 0 .
(iii) (3x+2y−3)dx−(2x+3y−2)dy = 0 , (iv) (y+x+1)dx−(2y+2x+1)dy = 0 .
(v) (x − 2y + 3)dx − (x − 2y + 5)dy = 0 , (vi) (x + y)dx − (x + y − 1)dy = 0 .

20
Chapter 2

Exact Differential Equations

2.0.4 Exact differential equation


The necessary and sufficient condition for a D.E. of the form

M (x, y)dx + N (x, y)dy = 0, (2.1)

to be exact is that
∂M ∂N
= .
∂y ∂x
( ∂M
∂y
is the derivative of M w.r.t y keeping x constant and ∂N
∂x
is the derivative of
N w.r.t x keeping y constant).
Examples.
(i)
(x2 − y)dx + (y 2 − x)dy = 0,

is exact since
∂M ∂N
= −1, = −1.
∂y ∂x
Therefore,
∂M ∂N
= .
∂y ∂x
(ii)
(y 2 − x)dx + (x2 − y)dy = 0,

21
is not exact since
∂M ∂N
= 2y, = 2x,
∂y ∂x
and
∂M ∂N
6= .
∂y ∂x
If equation (2.1) is exact then it is possible to find µ = µ(x, y) such that

M dx + N dy = dµ, (2.2)

but
∂µ ∂µ
dµ = dx + dy, (2.3)
∂x ∂y
(this is the definition of the total derivative of a function of two variables). Com-
paring (2.2) and (2.3) one obtains
∂µ ∂µ
M= and N = ,
∂x ∂y
This gives Z x
∂µ
µ= M dx + φ(y), and N = ,
∂y
or Z x
∂µ
µ= N dy + φ(x), and M =.
∂x
From these two equations one can find the form of µ(x, y).
R R
(Here x ( y ) means that in the integration y(x) is treated as a constant). If the
equation is not exact then an integrating factor ξ must be sought.
(a) if
∂M ∂N
∂y
− ∂x
= f (x),
N
R
f (x)dx
a function of x only, then the integrating factor ξ(x, y) = e .
Example.
Solve the following D.E.

(x2 + y 2 + x)dx + xydy = 0,

22
∂M ∂N
= 2y , = y.
∂y ∂x
The equation is not exact and we must look for an integrating factor. Now
∂M ∂N
∂y
− ∂x 2y − y 1
= = = f (x),
N xy x
hence, R 1
dx
ξ=e x = eln x = x.

Hence,

x[(x2 + y 2 + x)dx + xydy] = (x3 + xy 2 + x2 )dx + x2 ydy = 0,

is exact.
(b) If
∂M ∂N
∂y
− ∂x
= −g(y),
M
R
g(y)dy
a function of y only, then the integrating factor ξ(x, y) = e .
Example.
Solve
(2xy 4 ey + 2x3 y 2 + y)dx + (x2 y 4 ey − x4 y − 3x)dy = 0,
∂M ∂N
= 8xy 3 ey + 2xy 4 ey + 4x3 y + 1, = 2xy 4 ey − 4x3 y − 3,
∂y ∂x
the equation is not exact and we seek an integrating factor (ξ). Note that
∂M ∂N
∂y
− ∂x 4
= = −g(y),
M y
R R
g(y)dy − y4 dy −4
hence, the integrating factor ξ(x, y) = e =e = e−4 ln y = eln y =
−4
y . Therefore,

y −4 [(2xy 4 ey + 2x3 y 2 + y)dx + (x2 y 4 ey − x4 y − 3x)dy]

= (2xey + 2x3 y −2 + y −3 )dx + (x2 ey − x4 y −3 − 3xy −4 )dy = 0,

23
is exact.
(c) If the equation is homogeneous then the integrating factor is
1
ξ= .
Mx + Ny
Example.
Solve
(x4 + y 4 )dx − xy 3 dy = 0,

the equation is homogeneous of degree 4 (check this) and the integrating factor
1 1 1
ξ= = 4 4 3
= 5.
Mx + Ny (x + y )x + (−xy )y x

Hence,
1
ξ= ,
x5
so that
1 4 4 3 1 y4 y3
[x + y )dx − xy dy] = ( + )dx − dy = 0,
x5 x x5 x4
is exact.

2.1 Solution of an Exact D.E.


Examples.
(i) Find the primitive of the D.E.

(2x3 + 3y)dx + (3x + y − 1)dx = 0.

Now
∂M ∂N
= 3, = 3,
∂y ∂x
and therefore, the equation is exact. Set
Z x
∂µ
µ= M dx + φ(y), and N= .
∂y

24
Therefore,
Z x
x4
µ= (2x3 + 3y)dx + φ(y) = + 3yx + φ(y) ,
2
and
∂µ dφ
3x + y − 1 = = 3x + ,
∂y dy
or

= y − 1.
dy
y2
Integrating this, we obtain φ = 2
− y and the primitive µ = C implies 21 x4 +
3yx + 12 y 2 − y = C.
(ii) Find the primitive of

(4x3 y 3 − 2xy)dx + (3x4 y 2 − x2 )dy = 0.


∂M ∂N
= 12x3 y 2 − 2x, = 12x3 y 2 − 2x,
∂y ∂x
and therefore, the equation is exact.
Set Z x
∂µ
µ= M dx + φ(y) and N = ,
∂y
hence, Z x
µ= (4x3 y 3 − 2xy)dx + φ(y) = x4 y 3 − x2 y + φ(y),
and
∂µ dφ dφ
3x4 y 2 − x2 = = 3x4 y 2 − x2 + , = 0, φ = const.,
∂y dy dy
hence,
µ = x4 y 3 − x2 y + const.
and the primitive is
x4 y 3 − x2 y = C.

Find the constant a such that the following D.E. is exact,


dy
x + ye2xy + axe2xy = 0.
dx

25
We have here, M = x + ye2xy , N = axe2xy , so that
∂M ∂N
= 2yxe2xy + e2xy , = 2ayxe2xy + ae2xy .
∂y ∂x
∂M ∂N
The exact equation requires ∂y
= ∂x
and therefore a = 1.
Find the general solution of the D.E.
1 2 dy
y + 2yex + (y + ex ) = 0.
2 dx
We see that M = 12 y 2 + 2yex , N = y + ex so that ∂M
∂y
= y + 2ex , ∂N
∂x
= ex .
∂M
− ∂N y+ex
However, ∂y
N
∂x
= y+ex
= 1. Hence, the integrating factor ξ = ex . Thus, the
exact D.E. becomes
1 dy 1 dy
ex [ y 2 + 2yex + (y + ex ) ] = y 2 ex + 2ye2x + (yex + e2x ) = 0.
2 dx 2 dx
Therefore, one can write
Z x
∂µ
µ= M dx + φ(y) , N= .
∂y
Therefore,
Z x
1 1
µ= ( y 2 ex + 2ye2x )dx + φ(y) = y 2 ex + ye2x + φ ,
2 2

yex + e2x = yex + e2x + ,
dy
This gives φ0 = 0, or φ = const. hence, µ = 12 y 2 ex + 2ye2x = C.
Find the general solution of the following D.E.
dy
(3xy + y 2 ) + (3xy + x2 ) = 0.
dx
It is evident that the equation is homogeneous, so that we seek the solution with
1
an integrating factor ξ = M x+N y
. Therefore,

1 1 1
ξ= = = .
(3xy + y 2 )x 2
+ (3xy + x )y 2 2
4xy + 4x y 4xy(x + y)

26
Thus the exact D.E. will be
3x + y 3y + x dy
+ = 0.
4x(x + y) 4y(x + y) dx
This gives µ = xy(x + y)2 = C. This D.E. can also be solved by putting y = vx
as is done before.
Find the general solution of the following D.E.

(x + 3y 2 )dx + 2xydy = 0.
∂M ∂N
= 6y , = 2y.
∂y ∂x
It clear that the equation is not exact and we must look for an integrating factor.
We see that,
∂M ∂N
∂y
− ∂x 6y − 2y 2
= = = f (x).
N 2xy x
R 2
dx 2
Hence, the integrating factor ξ = e x = eln x = x2 . Thus the exact D.E. is

(x3 + 3y 2 x2 )dx + 2x3 ydy = 0.

Therefore,
Z x Z x
1 ∂µ dφ
µ= M dx = (x3 +3y 2 x2 )dx+φ(y) = x4 +y 2 x3 +φ(y) , 2yx3 = 2x3 y+ ,
4 ∂y dy
hence, φ0 = 0, φ = const.. Thus,
1
µ = x4 + y 2 x3 = C.
4
Exercise 2.1
Find the primitive of the following D.Es.
(1) (i) (x2 − y)dx − xdy = 0, (ii) (x4 + y 4 )dx − xy 3 dy = 0,
(iii) y(x − 2y)dx − x2 dy, (iv) (2y − x3 )dx + xdy.
(2) The D.E.
dy
ex sec y − tan y + =0
dx

27
has an integrating factor of the form e−ax cos y for some constant a. Find a, and
then solve the D.E.
(3) Determine the constant a so that the following D.E. is exact and solve the
resulting equation
1 1 (ax + 1) dy
2
+ 2+ = 0.
x y y3 dx
(4) Find the general solution of the following D.Es.

(y 2 − 3x2 )ydy − (x2 − 3y 2 )xdx = 0.

(5) Prove that

(3x2 y 2 + cos xy − xy sin xy) dx + (2x3 y − x2 sin xy) dy = 0,

is an exact D.E., and obtain its general solution.

28
Chapter 3

First Order Linear Differential


Equations

3.0.1 Linear Differential Equations


The equation
dy
+ P (x)y = Q(x) , (3.1)
dx
whose left hand side (LHS) is linear in both the dependent and its derivatives, is
called a linear differential equation.
Example.

dy
+ 3xy = sin x,
dx
is a linear D.E., but
dy
+ 3xy 2 = sin x,
dx
R
P (x)dx
is not. Multiply (3.1) by the integrating factor e , to get

dy R P (x)dx R d R R
e + yp(x)e P (x)dx = (ye P (x)dx ) = Q(x)e P (x)dx .
dx dx
Then R Z R
P (x)dx P (x)dx
ye = Q(x)e dx,

29
so that R Z R
− P (x)dx P (x)dx
y=e [ Q(x)e dx + C],

is the primitive (general solution). Equivalently, we substitute y = uv in equation


(3.1) to obtain
du dv
v + u + P (x)uv = Q.
dx dx
Or
dv du
u( + P v) + v = Q.
dx dx
Now choose
dv
+ Pv = 0
dx
which gives R
v = e− P dx
.

The above equation yields v du


dx
= Q, which gives
Z
Q
u= dx + C.
v
Hence,
R Z R Z
− P dx Q − P dx Q
y = uv = e [ dx + C] = e [ R dx + C],
v e P dx

Finally, one writes


R Z R
y = e− P (x)dx
[ Q(x)e P (x)dx
dx + C]

to obtain the primitive, which is the same as equation (3.1).

3.0.2 Bernoulli’s Equation


This is an equation of the form
dy
+ P (x)y = Q(x)y n , (3.2)
dx

30
which can be reduced to
dz
+ (1 − n)P (x)z = (1 − n)Q(x), n 6= 1
dx
by the transformation z = y 1−n . Dividing equation (3.2) by y n we obtain
dy dy dy dz
y −n + P (x)y 1−n = Q(x), = ,
dx dx dz dx
so that
dy 1 dz dy 1 dz
= y n , y −n =
dx 1 − n dx dx 1 − n dx
and hence, equation (3.2) becomes
dz
+ (1 − n)P (x)z = (1 − n)Q(x).
dx
Which is a linear D.E. that can easily be solved.
Examples.
(i) Solve
dy
+ 2xy = 4x,
dx
Z Z
P (x)dx = 2xdx = x2 ,
and hence, R Z R
y = e− P (x)dx
( Q(x)e P (x)dx
dx + C),
yields Z
−x2 2 2
y=e [ 4xex dx + C] = 2 + Ce−x , (Q = 4x)
is the required primitive.
(ii) Solve the following D.E.
dy
x = y + x3 + 3x2 − 2x.
dx
This can be put in the form
dy 1
− y = x2 + 3x − 2,
dx x

31
Z Z
1 1
P (x)dx = − dx = − ln x = ln ,
x x
R 1 1
P (x)dx
e = e− ln x = eln x = ,
x
and hence,
R Z R Z
1
y = e− P (x)dx
[ Q(x)e P (x)dx
dx + C] = x[ (x2 + 3x − 2) dx + C]
x
1
y = x3 + 3x2 − 2x ln x + Cx.
2
(iii) Solve the following D.E.
dy
(x − 2) = y + 2(x − 2)3 .
dx
This can be put in the form
dy 1
− y = 2(x − 2)2 ,
dx x − 2
Z Z
−1
P (x)dx = dx = − ln(x − 2) = ln(x − 2)−1 ,
x−2
and R
P (x)dx −1 1
e = eln(x−2) = ,
x−2
hence, Z
1
y = (x − 2)[ 2(x − 2)2 dx + C],
x−2
or
y = (x − 2)3 + C(x − 2),

is the required primitive.


(iv) Solve
dy
− y = xy 5 .
dx
This is a Bernoulli type equation and we therefore use the transformation

z = y 1−n = y 1−5 = y −4 ,

32
and the equation becomes
dz
+ (1 − n)P z = (1 − n)Q,
dx
dz
− 4z = −4x , (Q = x, n = 5).
dx
This has the usual solution
R Z R Z Z
− P (x)dx P (x)dx
z=e [ Q(x)e dx + C], P (x)dx = − 4dx = −4x,

and R
e− P (x)dx
= e4x ,
and Z
z = e4x [ −4xe−4x dx + C].
Writing the y dependence, we find
Z
1
y −4 = e4x [ −4xe−4x dx + C] = x + + Ce4x .
4
(v) Solve the following D.E.,
dy
+ (tan x)y = y 3 sec4 x.
dx
Putting z = y −2 we find
dz
− 2(tan x)z = −2 sec4 x.
dx
R R
exp[ P (x)dx] = exp[ (−2 tan x)dx] = cos2 x, so that
cos2 x
z cos2 x = C − 2 tan x, or y 2 = .
C − 2 tan x
Exercise 3.1
Find the general solution of the following D.Es.

dy dy 1
(i) + 2xy + xy 4 = 0, (ii) + y = sin x,
dx dx x
dy dy
(iii) = y tan x − sec x, (iv) + y tan x = y 2 sec3 x.
dx dx

33
3.0.3 First order D.Es. with higher degrees
A D.E. of first order has the form
dy
f (x, y, ) = 0.
dx
dy
Upon writing, p = dx
this becomes f (x, y, p) = 0. If the degree of p is greater
than one, the equation is first order but higher degree. The general form is

pn + P1 (x, y)pn−1 + P2 (x, y)pn−2 + ... + Pn (x, y)y = 0,

where P1 (x, ), P2 (x, y), ...., Pn (x, y) are functions of x and y.


We discuss the following situations:
(1) Equation solvable for p:
In this case we will have

(p − F1 )(p − F2 )(p − F3 ).....(p − Fn ) = 0,

where all F’s are functions of x and y, and the solution is

p = F1 (x, y), p = F2 (x, y), ...

which can be solved to obtain

f1 (x, y, C1 ) = 0, f2 (x, y, C2 ) = 0, ...

and the primitive is the product

f1 (x, y, C1 ).f2 (x, y, C2 ).... = 0.

Example.
Solve
p2 − 3xp + 2x2 = 0.
This equation can be written as

(p − x)(p − 2x) = 0,

34
and the solutions are p = x, p = 2x, i.e.,
Z
dy x2 1
p = x, = x, y = xdx = + C1 , y − x2 − C1 = f1 (x, y, C1 ) = 0 ,
dx 2 2
and
Z
dy
p = 2x, = 2x, y = 2 xdx = x2 + C2 , y − x2 − C2 = f2 (x, y, C2 ) = 0 ,
dx
where C1 , C2 are constants. Hence, the general solution is

x2
(y − − C1 )(y − x2 + C2 ) = 0.
2
(2) Equation solvable for x:
i.e., x = f (y, p), differentiate both sides w.r.t. y we obtain

dx ∂f ∂f dp dp
= + ≡ F (y, p, )
dy ∂y ∂p dy dy
and we solve for
1 dp
= F (y, p, ),
p dy
to obtain
φ(y, p, C) = 0

Example.

p3 − 2xyp + 4y 2 = 0.

This can be written as


p2 4y
2x = + ,
y p
and differentiate w.r.t. y to get
dp dp
2 2yp dy − p2 4p − 4y dy
= + ,
p y2 p2

35
which can be written as
dp
(p − 2y )(2y 2 − p3 ) = 0,
dy
that has the solution
dp
p − 2y = 0, p2 = Ky and 2y 2 = p3 .
dy
Substituting the first solution in the main D.E. equation, one obtains
4y
2x = K + √ .
Ky
This can be written as
y
(2x − K)2 = 16
K
or
2y = C(C − x)2 , where 2C = K.

The second solution (2y 2 = p3 ) when substituted in the original equation gives
3 dx
x
= dy
y
, which gives y = Cx3 , where C is a constant.
(3) Equation solvable for y:
i.e., y = f (x, p), differentiate both sides w.r.t. x we obtain

dy ∂f ∂f dp dp
= + ≡ F (y, p, ),
dx ∂x ∂p dx dx
and we solve for
dp
p = F (y, p, ),
dx
to obtain
φ(x, p, C) = 0.

Example.
Solve the following D.E.
y = 2px + p4 x2 .

36
Here we have y = f (x, p) and we therefore differentiate both sides w.r.t. x to
obtain
dy dp dp
= p = 2p + 2x + 4p3 x2 + 2xp4
dx dx dx
which when arranged gives
dp
(p + 2x )(1 + 2p3 x) = 0,
dx
and has the solution
dp dp dx
p + 2x = 0, −2 = .
dx p x
Integrate this to get
C
x= .
p2
To find y we substitute x in our original D.E. by writing the equation as

y − p4 x2 = 2px or (y − p4 x2 )2 = 4p2 x2 ,

hence,
(y − C 2 )2 = 4Cx

as a required primitive to our D.E.

3.0.4 Clairaut’s Equation


Is a D.E. of the form
y = px + f (p).

Its primitive is
y = Cx + f (C).

Examples.
(i) Solve the following D.E.
q
y = px + 4 + p2 .

37
Here q √
f (p) = 4 + p2 , so that f (C) = 4 + C 2,
and hence,

y = Cx + f (C) = Cx + 4 + C 2.
(ii) Solve the following D.E.

(y − px)2 = (1 + p2 ).

We write this as q
y − px = 1 + p2 .
For q q
y − px = 1 + p2 , or y = px + 1 + p2 ,
we have the solution

y = Cx + f (C) = Cx + 1 + C 2,

or
(y − Cx)2 = (1 + C 2 ).
(iii) Solve the following D.E.

y = 3px + 6y 2 p2 .

This can be put in the Clairaut’s equation form by multiplying by y 2 , we then


obtain
y 3 = 3py 2 x + 6y 4 p2 .
Now let y 3 = v, so that
dv 2 dv 2
v=x + ( ),
dx 3 dx
dv
we can set P = dx
to obtain v = P x + 23 P 2 , which has the solution
2
v = Kx + f (K) = Kx + K 2 , K = constant.
3

38
Or
y 3 = 3Cx + 6C 2 ,

where we have put K = 3C.

Exercise 3.2
dy
Obtain the primitive of the following D.E. (p = dx
)

(i) y = px − 2p2 , (ii) xp2 − 2yp + 4x = 0, (iii) x4 p2 − xp − y = 0,

(iv) y 2 p2 + 3px − y = 0, (v) p2 − xp + y = 0,


(vi) y = 2px + y 2 p3 , (vii) xp = y + x2 + y 2 .

3.0.5 Ricatti’s Equation


This has the form
y 0 + a(x)y 2 + b(x)y + c(x) = 0. (3.3)

This equation cannot in general be solved by integration. However, it has the


property that, given one particular solution y1 , the general solution can be found.
If we set y = y1 + z, then

y 0 + a(x)y 2 + b(x)y + c(x) = y10 + z 0 + a(x)(y12 + 2y1 z + z 2 ) + b(x)(y1 + z) + c(x),

y 0 +a(x)y 2 +b(x)y+c(x) = y10 +a(x)y12 +b(x)y1 +c(x)+[z 0 +a(x)(2y1 z+z 2 )+b(x)z],

y 0 + a(x)y 2 + b(x)y + c(x) = z 0 + z[2a(x)y1 + b(x)] + az 2 = 0. (3.4)

This is Bernoulli equation with n = 2.


Example:
Find the general solution of Ricatti equation, given that y1 = − x1 is a solution,
3
y 0 + 2y 2 − = 0.
x2

39
From equation (3.4) one observes that a = 2, b = 0 and c = − x32 . This can be
written as
4
z 0 + (2ay1 + b)z + az 2 = z 0 − z + 2z 2 = 0.
x
This is Bernoulli equation with n = 2 whose solution is given in sect. 3.0.1

3.0.6 Lagrange’s Equation


This has the form
y = xf (y 0 ) + g(y 0 ) , y 0 = p.

This is a generalization of Clairaut’s equation, which can be obtained for

f (y 0 ) = y 0 , g(y 0 ) = f (y 0 ).

We introduced the function p, as before, and consider x as a function of p, i.e.,


x = x(p). It then follows that

dx dx dy 1 dy 1 d(xf (p) + g(p)) 1 dx


= = = = [xf 0 (p) + f (p) + g 0 (p)],
dp dy dp p dp p dp p dp
or
dx f 0 (p) g 0 (p)
=x + .
dp p − f (p) p − f (p)
This is a linear equation for x = x(p), where x is considered as an independent
variable and p the dependent variable. The solution can then be given in the
parametric form
x = x(p)

y = x(p)f (p) + g(p).

Example:
Solve the following D.E.

y = 2xp + 3x + p2 + 6p.

40
This equation can be written as

y = x(2p + 3) + p2 + 6p.

Comparison with Lagrange equation yields

f (p) = 2p + 3 , f 0 (p) = 2 and g(p) = p2 + 6p , g 0 (p) = 2p + 6.

Thus,
dx f 0 (p) g 0 (p) 2 2p + 6 2
=x + = −x − = −x − 2,
dp p − f (p) p − f (p) p+3 p+3 p+3
or
dx 2
=− − 2,
dp p+3
which is a first order linear D.E.( where x is the dependent variable and p is the
independent variable), i.e.,

dx 2
+ x = −2,
dp p + 3

whose solution is (see equation (3.1))


2
x = − (p + 3) + C(p + 3)−2 ,
3
we have thus expressed x in terms of p (x = x(p))
Exercise 3.3
(1) For each of the following D.E., find a solution satisfying the given initial
dy
conditions (y 0 = dx
)

(a) y 0 = sin x , y(0) = 1, (b) y 0 = y , y(0) = 1, (c) y 00 = ex , y(1) = 1, y 0 (1) = 0

(2) Find all solutions of the following equations by separation of variables:


(i) y 0 = ex+y , (ii) y 0 = 3y, (iii) y 0 = (y − 1)(y − 2), (iv) y 0 = x3 y −2 ,

41
(v) sin x cos y dx + tan y cos x dy = 0, (vi) y 0 = xy 2 + y 2 + xy + y.
(3) Solve the following equation after verifying that they are homogeneous:
x−y
(i) y 0 = , ; (ii) xy 0 − y = xey/x , (iii) (3x2 y + y 3 )dx + (x3 + 3xy 2 )dy = 0,
x+y

(iv) y 0 = xy sin y−x


x
.
(4) Show that the Ricatti’s equation

y 0 + a(x)y 2 + b(x)y + c(x) = 0


1
can be transformed into a separable equation by the transformation y = ( ac ) 2 u, if
3
ac > 0 and if (a0 c − ac0 − 2abc)(ac)− 2 = const.
Study Ricatti equation
y 0 + ay 2 = bxm

where a, b are constant, as follows:


(a) Apply the transformation u = xy, t = xm+2 .
(b) Solve the D.E. tu0 − 12 u + βu2 = γt, β, γ=const.
t 1+α
(c) Transform the equation tu0 + αu + βu2 = γt by u = a+v
,v = γ
, and
u = a + vt , a = − αβ
(d) Conclude that y 0 + ay 2 = bxm can be solved in a closed form for m =
4n
− 2n+1 , where n is an integer.
(e) Solve the equations y 0 + y 2 = x−4 , xy 0 + 3y + y 2 = x2 .

42
Chapter 4

Higher Order Linear Differential


Equations

A linear D.E. of order n has the form


dn y dn−1 y dn−2 y
+ a1 + a 2 + .... + an y = Q(x)
dxn dxn−1 dxn−2
where a1 , a2 , a3 , ... are functions of x or constants, and all terms of first degree
(linear). When Q = 0 the D.E. is called homogeneous, and when Q 6= 0 it is
called non-homogeneous.
Example.

d2 y dy
2
− 2x + y = 3x + 1,
dx dx
is a non homogeneous linear D.E.(second order) and

d2 y dy
+ + x2 y = 0,
dx2 dx
is a homogeneous linear D.E. (second order). If y1 , y2 , y3 are solutions to the D.E.
then c1 y1 + c2 y2 + c3 y3 , where c1 , c2 , c3 are arbitrary constants, is also a solution.
The set of solutions y1 , y2 , y3 are called linearly independent if

c1 y1 + c2 y2 + c3 y3 = 0,

43
implies that c1 = c2 = c3 = 0, as the only solution.
Example.
If
y1 = ex , y2 = e−x ,

show that the two solutions are linearly independent.

c1 y1 + c2 y2 = 0,

and by differentiating, we get


dy1 dy2
c1 + c2 = 0.
dx dx
Then,
c1 ex + c2 e−x = 0,

and
c1 ex − c2 e−x = 0.

Solving the two equations, we get c1 = c2 = 0. Hence, the two solutions are
linearly independent.
A necessary and sufficient condition for a set of solutions be linearly indepen-
dent is that ¯ ¯
¯ y1 y2 y3 . . . ¯¯
¯
¯ 0
¯ y1 y20 y30 . . . ¯¯
W = ¯¯ y 00 y200 y300
¯
. . . ¯¯ 6= 0
¯ 1
¯ .. .. .. .. ¯¯
¯ . . . .
(W is called the Wronskian), where

dy d2 y
y0 = , y 00 = , ...
dx dx2
Example.
Show that the two solutions
e−x , e2x ,

44
are linearly independent.
¯ ¯ ¯ ¯
¯y y2 ¯¯ ¯¯ e−x e2x ¯¯
W = ¯¯ 10 = = 3ex 6= 0,
y 1 y20 ¯ ¯ −e−x 2e2x ¯
hence, the two solutions are linearly independent.
Exercise.
Show that y = eax is a solution of the D.E.
d2 y dy
− − 2y = 0,
dx2 dx
for some values of a and find these values.
Exercise 4.1
(1) Compute the Wronskian of the following pairs of functions, and show which
of the pairs are linearly independent.

sin ax, cos bx; sin2 x, 1−cos 2x; x, x ln x; eax , xeax ; eax , ebx ; eax sin bx, eax cos bx

where a, b are constants.


(2) Show that if y1 and y2 are solution of the D.E
y 00 + p(x)y 0 + q(x)y = 0 then their Wronskian (W ) satisfies the D.E.
W 0 + p(x)W = 0, where W 0 = dW dx
.

4.1 Homogeneous D.Es. with constant coeffi-


cients
This can be written in the form
dn y dn−1 y dn−2 y
+ a1 + a 2 + .... + an y = 0, (4.1)
dxn dxn−1 dxn−2
d
where a1 , a2 , a3 , ..., an are constant coefficients. Define the operator D as D = dx
,
then the above equation becomes

Dn y + a1 Dn−1 y + a2 Dn−2 y + ... + an y = 0,

45
which can be written as L(D)y = 0, where

L(D) = Dn + a1 Dn−1 + a2 Dn−2 + ... + an ,

which is function of D.
The equation
L(D)y = 0,

which has the general solution

(D − m1 )(D − m2 )(D − m3 ).....(D − mn )y = 0,

is called the characteristic equation and the roots

m1 , m2 , m3 , ...

are called the characteristic roots. The primitive y obtained from this equation is
called the complementary solution.
Alternatively, one can assume the solution to have the form y = emx and substitute
this in equation (4.1) to obtain an equation for m. Solution of m gives the roots
of the equation. The general solution is the sum of these solutions (emx ).
The operator D has the following properties:
(i) D[cy] = cD[y]
(ii) D[y1 + y2 ] = D[y1 ] + D[y2 ]
such an operator which satisfies the above properties is called a linear operator.
All other operators are nonlinear.
Example.
Write the D.E. in the operator form and find the characteristic roots

d3 y d2 y dy
3
− 2
− 4 + 4y = 0,
dx dx dx
becomes
(D3 − D2 − 4D + 4)y = 0,

46
that can be put in the form

(D − 1)(D − 2)(D + 2)y = 0,

which has the characteristic roots: m1 = 1, m2 = 2, m3 = −2.

4.1.1 The complementary solution of a D.E.


(a) If the roots are real and different, i.e., m1 6= m2 6= m3 .... then the comple-
mentary solution is

y = c1 em1 x + c2 em2 x + c3 em3 x + ...

(b) If the roots are the same, i.e., m1 = m2 6= m3 .... the complementary solution
is
y = c1 em1 x + c2 xem1 x + c3 em3 x + ...

(c) If the roots are complex, i.e., m1 , m2 = a ± ib, where i = −1, the comple-
mentary solution is
y = eax (c1 sin bx + c2 cos bx).

Examples.
Find the complementary solution of the following D.Es.
(i)
d3 y d2 y dy
3
− 2 2
− 4 + 8y = 0.
dx dx dx
This can be written as

(D3 − 2D2 − 4D + 8)y = (D − 2)(D − 2)(D + 2)y = 0,

hence, m1 = 2, m2 = 2, m3 = −2, thus two roots are equal and, therefore, the
complementary solution is

y = c1 e2x + c2 xe2x + c3 e−2x .

47
(ii)
d2 y dy
2
− 5 + 4y = 0.
dx dx
This can be written as

(D2 − 5D + 4)y = (D − 4)(D − 1)y = 0,

hence, the roots are m1 = 4, m2 = 1 (the roots are different) and the complemen-
tary solution is
y = c1 e4x + c2 ex .

(iii)
d3 y d2 y dy
− − 12 = 0,
dx3 dx2 dx
which can be written as

(D3 − D2 − 12D)y = D(D2 − D − 12)y = D(D − 4)((D − 3)y = 0,

hence, the roots are m1 = 0, m2 = 4, m3 = 3, and the complementary solution

y = c1 + c2 e4x + c3 e3x .

(iv)
d2 y dy
2
− 2 + 10y = 0,
dx dx
can be put in the form
(D2 − 2D + 10)y = 0,

the roots are m1 , m2 = 1 ± 3i, and hence, the complementary solution

y = ex (c1 sin 3x + c2 cos 3x).

(v)
d2 y dy
( 2
− 2 + 5)2 y = 0,
dx dx

48
this, in operator form, becomes

(D2 − 2D + 5)2 y = (D2 − 2D + 5)(D2 − 2D + 5)y = 0.

Hence, the roots are m1 , m2 = 1 ± 2i and m3 , m4 = 1 ± 2i, which are the same
(repeated), therefore the complementary solution is

y = ex (c1 sin 3x + c2 cos 3x) + xex (c3 sin 3x + c4 cos 3x).

Exercise 4.2
Find the complementary solutions of the following D.Es.
d3 y 2
d y dy
(1) (i) dx3
− 2 dx2 − dx
+ y = 0,

d2 y dy
(ii) dx2
− 4 dx + 4y = 0,

d2 y dy
(iii) dx2
+ dx
− 2y = 0,

d3 y 2
d y dy
(iv) dx3
− 2 dx2 − 4 dx + 8y = 0,

d2 y dy
(v) dx2
+ 2 dx + 2y = 0.
(2) Let D[y](x) = y 00 − 3xy 0 + 3y, compute D[ex ] , D[x2 ] , D[x] , D[x2 + 3x]
Let
D[y](x) = y 00 + p(x)y 0 + q(x)y
and suppose that D(x2 ) = x + 1 and D(x) = x + 2, show that: y(x) = x − 2x2
is a solution of the above D.E.
√ 1
(3) Show that y1 = x and y2 = x
are solutions of the D.E.

2x2 y 00 + 3xy 0 − y = 0.

(4) Show that the operator D defined by


Z b
D[y] = s2 y(s)ds
a

is linear, that is D[cy] = cD[y] and D[y1 + y2 ] = D[y1 ] + D[y2 ].

49
4.2 Non homogeneous linear D.Es.
4.2.1 Operator Method
This has the form L(D)y = Q(x), which can be written as
1
y= Q(x)
L(D)
or
1
y= Q(x),
(D − m1 )(D − m2 )(D − m3 ).....(D − mn )
or
1 1 1 1
y= ... Q(x). (4.2)
D − m1 D − m2 D − m3 D − mn
To find the solution y, we employ the following method:
Set
Q du
u= , or (D − mn )u = Q, − mn u = Q.
D − mn dx
One can then solve this linear first order D.E. to find u (the solution is u =
R
emn x Qe−mn x dx, see equation (3.1)) and then write
u
w=
D − mn−1
or
dw
− mn−1 w = u
dx
this latter equation can be solved to obtain w. We continue like this until one
reaches the final term in equation (4.2), and finally we solve for y. The general
solution is the sum of the complementary solution (hereafter denoted by yc ) that
obtained from L(D)y = 0 and the above solution, which is called the particular
solution (hereafter denoted by yp ) (for the non-homogeneous part), thus y =
yc + yp .
Examples.
(i) Find the general solution of the following D.E.

(D2 − 3D + 2)y = ex .

50
The complementary solution (yc ) is obtained from the homogeneous part, i.e.,
L(D)y = 0 and the particular solution is due to the non-homogeneous part. Thus

(D − 2)(D − 1)y = 0, gives yc = c1 ex + c2 e2x .

To find the particular solution (yp ) we proceeds as follows:


Write
1 1
yp = ex ,
D−1D−2
and set
ex
u= ,
D−2
hence,
u
yp = .
D−1
Thus one obtains
du
− 2u = ex ,
dx
which has the solution
Z
u = e2x ex e−2x dx = −ex .

Now
u −ex
yp = = ,
D−1 D−1
or
−ex dyp
yp = , Dyp − yp = −ex , − yp = −ex ,
D−1 dx
which has the solution
Z
yp = e x (−ex e−x )dx = −xex .

Thus the particular solution is yp = −xex and hence, the general solution is

y = yc + yp = c1 ex + c2 e2x − xex .

51
(ii) Find the general solution of the following D.E.

(D3 + 3D2 − 4)y = xe−2x .

This can be written as

(D − 1)(D + 2)(D + 2)y = xe−2x .

The complementary solution

yc = c1 ex + c2 e−2x + c3 xe−2x .

To find the particular solution, we write


1 1 xe−2x
yp = ,
D−1D+2D+2
and set
xe−2x
u= ,
D+2
to get
du
+ 2u = xe−2x ,
dx
which has the solution
Z
1
u = e−2x xe−2x e2x dx = x2 e−2x .
2
Now
1 1 1 1 1
yp = u= ( x2 e−2x ),
D−1D+2 D−1D+2 2
set
1 1 2 −2x
w= xe ,
D+22
so that
1
(D + 2)w = x2 e−2x ,
2
which can be solved to get
Z
1 2 −2x 2x 1
w = e−2x x e e dx = x3 e−2x ,
2 6

52
and finally, we get
1 3 −2x
1 6
xe
yp = w, so that yp = .
D−1 D−1
or
1
Dyp − yp = x3 e−2x ,
6
which has the solution
Z
1 3 −2x −x 1 2 2
yp = ex x e e dx = − (x3 + x2 + x + )e−2x .
6 18 3 9
Hence,
1 3 2 2
yp = − (x + x2 + x + )e−2x .
18 3 9
and the general solution is given by
1 3 2 2
y = yc + yp = c1 ex + c2 e2x − (x + x2 + x + )e−2x .
18 3 9
(iii) Find the general solution of the following D.E.

(D2 − 3D + 2)y = e5x .

This can be written as


(D − 1)(D − 2)y = e5x .

The complementary solution is

y = c1 ex + c2 e2x .

To find the particular solution, we write


1 1
y= e5x ,
D−1D−2
and set
e5x
u= ,
D−2

53
so that
du
− 2u = e5x ,
dx
which has the solution
Z
1
u = e2x e5x e−2x dx = e5x ,
3
and hence,
1 1 e5x 1
y= u= ⇒ Dy − y = e5x ,
D−1 3D−1 3
which yields Z
1 5x −x 1
yp = ex e e dx = e5x .
3 12
Thus, the general solution is
1 5x
y = c1 ex + c2 e2x + e .
12
Exercise 4.3
(i) Show that the particular solution of the D.E.

(D − a)(D − b)y = Q,

is given by Z Z
y = eax e(b−a)x [ebx Qe−bx dx]dx.

(ii) Find the general solution of the D.E. (D2 + 5D + 4)y = 3 − 2x.
(iii) Find the general solution of the D.E. (D2 − 1)y = 4xex .
(iv) Find the general solution of the D.E. (D3 − 4D)y = x.
(v) Find the general solution of the D.E. (D3 − 5D2 + 8D − 4)y = e2x .

4.2.2 The properties of L(D)


If L(D) is a differential operator with constant coefficients, one can write
n
X
L(D) = ar Dr
r=0

54
If y1 , y2 are solutions then

L(D)(c1 y1 + c2 y2 ) = c1 L(D)y1 + c2 L(D)y2 .

This follows since


n
X
L(D)(c1 y1 + c2 y2 ) = ar Dr (c1 y1 + c2 y2 )
r=0

n
X n
X
= c1 ar Dr y1 + c2 ar Dr y2 = c1 L(D)y1 + c2 L(D)y2 .
r=0 r=0
Theorem I:

L(D)eax = L(a)eax , (a = const.).


Theorem II:

L(D)[eax f (x)] = eax L(D + a)f (x).


Theorem III:

L(D2 ) sin ax = L(−a2 ) sin ax, L(D2 ) cos ax = L(−a2 ) cos ax.
We define the inverse operator L−1 (D) or 1/L(D) by means of
(1)
L(D)[L−1 (D)]f (x) = f (x).
Thus if yp = L−1 (D)f (x), then L(D)yp = f (x) and yp is a particular integral. If
f (x) is a polynomial one can expand L(D) in binomial expansion. Thus,
1 1 1 1 D D D
=− [ D ] = − [1 + + ( )2 + ( )3 + ...].
D−a a 1− a
a a a a

Example:
Solve the D.E.
(D − 1)(D − 3)y = x2 + 2x3 .

55
In this case we have the particular solution

x2 + 2x3 1 1 1
y= = ( − )(x2 + 2x3 ).
(D − 1)(D − 3) 2 D−3 D−1

Hence,
1 1 D D D
y = [− (1 + + ( )2 + ( )3 + ... + (1 + D + D2 + D3 + ...)][x2 + 2x3 ]
2 3 3 3 3
1 4 13 40 186 20 2
= ( + D + D2 + D3 + ...)[x2 + 2x3 ] = + x + 3x2 + x3 .
3 9 27 81 27 3 3
3
We stopped at D since all higher derivatives give zero contribution.
(2)
1 sin ax
2
sin ax = .
L(D ) L(−a2 )
Solve
(D − 2)(D − 3)y = sin 3x.

The particular solution is


1 1 −1
yp = sin 3x = sin 3x = sin 3x,
(D2 − 5D + 6) 2
−3 − 5D + 6 3 + 5D

3 − 5D 3 − 5D
=− 2
sin 3x = − sin 3x,
9 − 25D 9 − 25.(−32 )
1 1
= (3 − 5D) sin 3x = (3 sin 3x − 15 cos 3x).
234 234
(3)
1 ax 1 ax
e = e , L(a) 6= 0.
L(D) L(a)
Solve the following D.E.

(D2 − 5D + 6)y = e5x .


1 1 5x
yp = e5x = e
D2 − 5D + 6 L(5)

56
1 5x 1 5x
yp = e = e ,
52 − 5.5 + 6 6
(where L(D) = D2 − 5D + 6 and L(5) = 52 − 5.5 + 6 = 6).
(4)
1 1 1 xm eax
(i) [eax f (x)] = eax f (x), (ii) eax
= .
L(D) L(D + a) (D − a)m L(D) m!L(a)

(ii) Solve the D.E.


(D2 − 5D + 6)y = e4x x2 .
1 1 1
yp = e4x x2 = e4x x2 = e4x 2 x2 ,
D2 − 5D + 6 2
(D + 4) − 5(D + 4) + 6 (D + 3D + 2)
1 1 1 4x 1 2 1 2
yp = e4x 1 x 2
= e [1 − (D + 3D) + (D + 3D)2 + ..]x2 ,
2 [1 + 2 (D2 + 3D)] 2 2 4
1 3 7
yp = e4x (1 − D + D2 + ..)x2 .
2 2 4
Hence,
1 7
yp = e4x (x2 − 3x + ).
2 2
(iii) Solve the D.E.
(D − 1)(D − 3)y = e3x .
1 1 e3x
yp = e3x = [ ]
(D − 1)(D − 3) D−1 D−3
1 1
yp = e3x [ .1]
D−1 D+3−3
1 3x 1 1
yp = e [ .1] = xe3x ,
3−1 D 2
1 1
where we have used D−3
e3x = D−3
[e3x .1] (we have put f (x) = 1, in (4) above)
1
and the fact that D
.1 = x. We use this trick when L(a) = 0.
More generally, we can evaluate
1 1
f (x) = eλx (e−λx f (x))
D−λ D−λ

57
Z
1 1
eλx (e−λx f (x)) = eλx (e−λx f (x)) = eλx e−λx f (x)dx.
D−λ+λ D
Hence, Z
1 λx
f (x) = e e−λx f (x)dx. (4.3)
D−λ
(iv) Solve the D.E.

(D2 − 3D + 2)y = (D − 2)(D − 1)y = e3x .

The particular solution is


1 3x 1 3x
yp = e = e , where, L(3) = 32 − 3.3 + 2.
L(3) 2
(v) Solve the D.E.
(D − 2)(D − 3)y = e4x sin 3x.
e4x sin 3x sin 3x
yc = c1 e2x + c2 e3x , yp = = e4x
(D − 2)(D − 3) (D + 2)(D + 1)
sin 3x sin 3x sin 3x
yp = e4x 2 = e4x 2
= e4x
(D + 3D + 2) (−3 + 3D + 2) (3D − 7)
3D + 7 3D + 7 1
yp = e4x sin 3x = e4x
sin 3x = (9 cos 3x+7 sin 3x)e4x .
(9D2 − 49) 9(−32 ) − 49] −130
(vi) Solve the D.E.
y 00 + 3y 0 − 4y = sin 2x.
The particular solution is
1 1 1 1
yp = sin 2x = ( − ) sin 2x,
(D − 1)(D + 4) 5 D−1 D+4
1 1 1 1 1 x Z −x 1 −4x Z 4x
yp = sin 2x− sin 2x = e e sin 2x dx− e e sin 2x dx.
5D−1 5D+4 5 5
Using equation (4.3).
(5)
1 x cos ax 1 x sin ax
(i) sin ax = − , cos ax = ,
D2 +a 2 2a D2 +a 2 2a

58
1 sin bx 1 cos bx
(ii) sin bx = 2 , cos bx = 2 , a 6= b,
D2
+a 2 a −b 2 2
D +a 2 a − b2
Z
1 1 1 Z Z Z
(iii) f (x) = f (x)dx, 2 f (x) = f (x)dx = [ f (x)dx]dx.
D D D
(vii) Solve the following D.Es.
(a) (D2 + 4)y = sin 2x, (b) (D2 − 2D + 1)y = 2xex .
1
(a) yp = D2 +22
sin 2x = − 41 x cos 2x, (b) yp = 1
D2 −2D+1
2xex = 1
2ex (D+1)2 −2(D+1)+1 x,
R R
yp = 2e D2 x = 2ex D1
x 1
xdx = ex D1 x2 = e x
x2 dx = 31 ex x3 .
One can accordingly solve example (vi) as

(D + 1)(D − 4) (D + 1)(D − 4)
yp = 2 2 2
sin 2x = sin 2x,
(D − 1)(D − 4 ) (−22 − 1)(−22 − 42 )
1 1
yp = (D2 − 3D − 4) sin 2x = − (4 sin 2x + 3 cos 2x),
100 50
Exercise 4.4
(1) Prove Theorem I, II, III stated before.
(2) Find the general solution of the following D.Es.
(i) y 00 + 36y = 2 sin 3x, (ii) y 00 − 4y 0 + 4y = 3 sin 9x ,
(iii) y 00 + 4y 0 + 13y = 2x + 7e3x , (iv) y 00 + 5y 0 + 6y = 2 cosh x.
1 xm eax 1 xm eax
(3) Show that (D−a)m
eax = m!
and (D−a)m L(D)
eax = m!L(a)
.

4.3 Solution of D.Es. by variation of parame-


ters method
The D.E. can be put in the general form

L(D)y = Q(x).

The complementary solution has the form

yc = c1 y1 + c2 y2 + c3 y3 + ...

59
We now allow the constant (c’s) to vary and write instead,

y = L1 (x)y1 + L2 (x)y2 + L3 (x)y3 + ...

as our particular solution. The problem is to find the parameters L1 , L2 , L3 , ... as


dy d2 y d3 y
functions of x from the following equations:(y 0 = dx , y 00 = dx 2,y
000
= dx 3)

y 0 = L01 y1 + L02 y2 + L03 y3 + (L1 y10 + L2 y20 + L3 y30 ),

and set
L01 y1 + L02 y2 + L03 y3 = 0, (4.4)

y 00 = L01 y10 + L02 y20 + L03 y30 + (L1 y100 + L2 y200 + L3 y300 ),

and set
L01 y10 + L02 y20 + L03 y30 = 0, (4.5)

y 000 = L01 y100 + L02 y200 + L03 y300 + (L1 y1000 + L2 y2000 + L3 y3000 ),

and set
L01 y100 + L02 y200 + L03 y300 = Q, (4.6)

so that, if we write

L(D) = D3 + P1 D2 + P2 D + P3 , L(D)y = Q,

and substitute for D3 y, D2 y, Dy, we obtain

L1 (y1000 + P1 y100 + P2 y10 + P3 y1 ) + L2 (y2000 + P1 y200 + P2 y20 + P3 y2 )

+L3 (y3000 + P1 y300 + P2 y30 + P3 y3 ) + Q = 0,

and since y1 , y2 , y3 are solutions of the equation L(D)y = 0, all terms in the
brackets are zero.
Examples.
Find the general solution of the following D.Es. by the method of variation of

60
parameters
(i)
(D2 − 2D)y = ex sin x.

The complementary solution is

yc = c1 + c2 e2x ,

so that we write for the particular solution the form

y = L1 + L2 e2x .

Apply the conditions (4.4)-(4.6) above, we obtain

L01 + L02 e2x = 0, 2L02 e2x = Q = ex sin x.

Hence,
1
L02 = e−x sin x,
2
which can be integrated to yield
1 Z −x 1
L2 = e sin x dx = − e−x (sin x + cos x),
2 4
and hence,
1
L01 = −L02 e2x = − ex sin x,
2
which can be integrated to get
1Z x 1
L1 = − e sin x dx = − ex (sin x − cos x).
2 4
Thus
1
yp = L1 + L2 e2x = − ex sin x,
2
and finally the general solution is
1
y = c1 + c2 e2x − ex sin x.
2

61
(ii)
(D3 + D)y = csc x.

This can be written as


D(D2 + 1)y = csc x,

which has the complementary solution

yc = c1 + c2 cos x + c3 sin x,

so that the particular solution is of the form

y = L1 + L2 cos x + L3 sin x.

To determine L1 , L2 , L3 we apply the conditions (4.4)-(4.6) above. Hence,


(i) L01 + L02 cos x + L03 sin x = 0,
(ii) − L02 sin x + L03 cos x = 0,
(iii) − L02 cos x − L03 sin x = Q = csc x.
Adding (i) and (iii), we get: L01 = csc x and hence, L01 = − ln(csc x + cot x),
solving (ii) and (iii) we obtain: L03 = −1, hence, L3 = −x and L02 = − cot x , L2 =
− ln(sin x). Thus

yp = L1 + L2 cos x + L3 sin x = − ln(csc x + cot x) − ln(sin x) cos x − x sin x,

and finally, the general solution is

y = c1 + c2 cos x + c3 sin x − ln(csc x + cot x) − ln(sin x) cos x − x sin x.

4.4 Solution of D.Es. by undetermined coeffi-


cients method
If L(D)y = x3 , we can write the particular solution as

yp = Ax3 + Bx2 + Cx + E,

62
where A, B, C and E are undetermined coefficients.
If L(D)y = ex + e3x , then the particular solution of the D.E. takes the form

yp = Aex + Be3x .

Special Cases:
(1) If a term of Q is also a term of yc , say u, which has a s-fold root, then we
introduce a term of xs u plus terms arising from it by differentiation
Example.
(D − 2)2 (D + 3)y = e2x + x2 ,
then
yc = c1 e2x + c2 xe2x + c3 e−3x ,
then the term e2x appears in Q and in yc in this case yp has a term with multi-
plicity of 2, hence, yp should have a term x2 e2x and all terms arising from it by
differentiation. Thus,

yp = Ax2 e2x + Bxe2x + Ce2x + Ex2 + F x + G.

(2) If a term of Q is a xr u and u is a term of yc , if u corresponds to s-fold root,


yp must contain a tern xr+s u plus all terms arising from it by differentiation.
Examples.
(1) Solve the following D.E.

(D − 2)3 (D + 3)y = x2 e2x + x2 ,

the root m = 2 is three fold (three equal roots), i.e., s = 3. The complementary
solution is
yc = c1 e2x + c2 xe2x + c3 x2 e2x + c4 e−3x ,
and we see that the term x2 e2x (in Q) is also in yc , hence, yp must contain a term
x2+3 e2x , and all terms arising from it by differentiation. Thus

yp = Ax5 e2x + Be2x x4 + Cx3 e2x + Gx2 + Hx + J.

63
(the terms Ex2 e2x + F xe2x + M e2x are all already in the RHS).
(2) Solve the D.E.
(D2 − 2D + 3)y = x3 + sin x.

This has the roots m1 , m2 = 1 ± 2 i, so that the complementary solution is
√ √
yc = ex (c1 sin 2x + c2 cos 2x).

We write the particular solution as

yp = Ax3 + Bx2 + Cx + E + F sin x + G cos x.

Substitute this in our differential equation to determine the constant A, B, C, E, F, G.


Now

Dyp = 3Ax2 +2Bx+C +F cos x−G sin x, D2 yp = 6Ax+2B−F sin x−G cos x.

Hence,

(6Ax + 2B − F sin x − G cos x) − 2(3Ax2 + 2bx + C + F cos x − G sin x)

+3(Ax3 + Bx2 + Cx + E + F sin x + G cos x) = x3 + sin x.


Equating the coefficient of the different expressions in both sides yields
1 2 2 16 1 1
A = ,B = ,C = − ,E = − ,F = ,G = .
3 3 9 27 4 4
(3) Solve the D.E.
(D2 + 4)y = x2 sin 2x.
The roots for (D2 + 4)y = 0 are m1 , m2 = ±2 i,

yc = c1 sin 2x + c2 cos 2x,

but the term sin 2x in Q is also in yc with multiplicity of 1 (1-fold). Hence, yp


must contain a term x2+1 sin 2x = x3 sin 2x. Therefore

yp = Ax3 sin 2x + Bx2 sin 2x + Cx sin 2x + Ex3 cos 2x + F x2 cos 2x + Gx cos 2x.

64
Note that sin 2x + cos 2x are not included in yp since they are already in yc . Sub-
stitute y = yp in the original D.E above and equate the coefficients of different
1
expressions to get A, B, C, E, F, G to complete the solution (A = 0, B = 16
,C =
1 1
0, E = − 12 , F = 0, G = − 32 ).

Exercise 4.5
(1) Solve the following D.Es by the method of undetermined coefficients
(i) (D2 + 2)y = ex + 2, (ii) (D2 + 2D + 2)y = sin x + x2 ,
(iii) (D2 − 1)y = ex sin 2x, (iv) (D2 − 5D + 6)y = x2 e4x + 9e4x .
(2) Solve the following D.Es by the method of variation of parameters
(i) (D2 + 1)y = cos x, (ii) (D3 + 4D2 + 3D)y = x2 , (iii) (D2 + 4)y = sin 2x,
e3x
(iv) y 00 + y = tan x, (v) (D2 − 6D + 9)y = x2
.

4.5 Euler’s Equation


The homogeneous equation has the form

d2 y dy
x2 2
+ ax + by = 0, (4.7)
dx dx
where a, b are constants. This can be changed to an equation with constant
coefficients upon the transformation x = et so that, t = ln x. Now
dy dy dt 1 dy
= = ,
dx dt dx x dt
hence,
dy dy
= e−t ,
dx dt
we see that
d d
= e−t ,
dx dt
and
d2 y d dy −t d −t dy
2
−t −t d y dy
2
= ( ) = e (e ) = e (e 2
− e−t ),
dx dx dx dt dt dt dt

65
or
d2 y −2t dy
2
−2t d y dy
e−2t 2
− e = e ( 2
− ).
dt dt dt dt
Writing
dy d2 y
ẏ = , ÿ = ,
dt dt2
one yields
dy 1 d2 y 1
= e−t ẏ = ẏ, = (ÿ − ẏ),
dx x dx2 x2
and hence, equation (4.7) becomes

(ÿ − ẏ) + aẏ + by = 0,

or
ÿ + (a − 1)ẏ + by = 0,

which has the solution


y(t) = c1 em1 t + c2 em2 t ,

and finally substitute t = ln x in this equation.


Examples.
(i) Solve the Euler’s equation

d2 y dy
x2 2
+ 5x + 5y = 0.
dx dx
Letting x = et one obtains

ÿ + (a − 1)ẏ + by = 0,

which becomes (a = 5, b = 5)

ÿ + (5 − 1)ẏ + 5y = ÿ + 4ẏ + 5y = 0,

which has the solution (in terms of t the roots are m1 , m2 = −2 ± i)

y = e−2t (c1 sin t + c2 cos t),

66
or
1
y= (c1 sin(ln x) + c2 cos(ln x)), since t = ln x.
x2
(ii) Solve the non homogeneous D.E.
d2 y dy
x2 2
− x + 2y = x ln x.
dx dx
t
Let x = e , the equation becomes
ÿ + (a − 1)ẏ + by = x ln x,
or
ÿ − 2ẏ + 2y = x ln x = tet ,
since a = −1, b = 2
or
ÿ − 2ẏ + 2y = tet
which can be solved by one of the previous methods to give
y(x) = x(c1 cos(ln x) + c2 sin(ln x)) + x ln x.
(iii) Solve the following D.E.
d2 y dy
x2 2
+ 3x + 2y = x2 + 2x.
dx dx
Putting x = et , we have (in terms of t)
ÿ + 2ẏ + 2y = e2t + 2et ,
whose complementary solution yc = e−t (A cos t + B sin t) and the particular solu-
1 2t
tion is yp = 10 e + 25 et . Hence, in terms of x, we have
1 1 2
y(x) = (A cos ln x + B sin ln x) + x2 + x.
x 10 5
Exercise 4.6
(1) Solve the following differential D.Es.
2
d y dy 2
d y
(i) x2 dx2 − 5x dx + 9y = 0, (ii) x2 dx2 − 2y = 0,
2d y dy 2
2d y dy
(iii) x2 dx 2
2 + 3x dx − 3y = 5x , (iv) x dx2 + x dx + y = x(6 − ln x),
2
d y 2
d y dy
(v) x2 dx2 − 2y = ln x , (vi) x2 dx 2
2 + 5x dx + 13y = ln x + x .

67
d2 y
4.5.1 Second order D.E. of the form dx2 = f (y)
dy
Put p = dx
so that
d2 y dp dy dp
2
= =p .
dx dy dx dy
Thus Z
dp 1 2
p = f (y) , or p = A + f (y)dy,
dy 2
hence we can find y(x).
Example:
Solve
d2 y
= y 2 given that y 0 (y = 0) = 0, y(x = 0) = ∞.
dx2
dy
Writing p = dx
gives
dp 1 2 1 3
p = y2 , p = y + C.
dy 2 3

But p(y = 0) = 0, y(x = 0) = ∞ gives C = 0 and hence x = ∓ 6y −1/2 .

d2 y dy
4.5.2 Second order D.E. of the form dx2 = f (y, dx )
dy dp
The assumption p = dx
gives p dy = f (y, p) which could be integrable.
Example:
Solve
yy 00 = y 02 − y 03 .
dy dp
Put p = dx
so that y 00 = p dy . Hence

dp dp
yp = p2 − p3 , p(y − p + p2 ) = 0,
dy dy
dp
so that either p = 0 which implies that y = const. or y dy = p − p2 , which implies
dy dp dp p dy y
y
= p
+ 1−p
, which upon integration gives y = C 1−p or p = dx
= C+y
. Inte-
grating this one obtains C ln y + y = x − B, C, B = const.

68
Exercise 4.7
(1) Find the general solution of yy 00 + 2y 02 = 0.
(2) Find the general solution of the following D.Es.
d2 y d y 2 dy 4
(i) dx2
+y = cot x, (ii) x dx2 + 6 dx = x ,
d2 y dy d y 2
(iii) dx2 + dx + 4y = ex , (iv) dx2 − y = x + 1,
d2 y dy d2 y dy
(v) dx 2 + 3 dx + 2y = sin x, (vi) dx 2 + dx − 2y = 2x,
dy
and find the particular solution if dx
= 0, y = 1 at the point x = 0,
d2 y
(vii) dx2
+ y = 2 cos x,
dy
and find the particular solution if y = 0, dx = 1 at x = 0.
(3) Solve the Euler’s equation

x2 y 00 − 2nxy 0 + n(n + 1)y = 0.

(4) Find the transformation y = v(x)g(x) such that the equation

y 00 + p(x)y 0 + q(x)y = 0,

would become
v 00 + r(x)v = 0,
and find r(x) in terms of p and q.
(5) Given
y 00 + P (x)y 0 + Q(x)y = R(x).
(a) Prove that if Q − 21 P 0 − 14 P 2 is a constant, the given equation may be solved
R
by putting y = u(x) exp(− 12 P (x)dx). Hence, solve the equations
(i) y 00 − 2y 0 cot x + 2y cot2 x = sin 2x,
(ii) x2 y 00 − 2x2 y 0 + [x2 − n(n + 1)]y = 0.
(b) Letting t(x) be the new independent variable, show that the equation reduces
to
d2 y 0 dy
t02
+ (t00
+ P t ) + Qy = R,
dt2 dt
0 dt 02
where t = dx , etc. Now choose t so that t = kQ, where k is a constant.
Prove that if (t00 + P t0 )/t02 is a constant, then the equation reduces to one of the

69
constant coefficients.
(c) Use the method to solve

y 00 + (2 cos x + tan x)y 0 + y cos2 x = cos4 x.

(6) Let y1 and y2 be solutions of Bessel’s equation

x2 y 00 + xy 0 + (x2 − n2 )y = 0,

with y1 (1) = 1, y10 (1) = 0, y2 (1) = 0 and y20 (1) = 1, compute the Wronskian for
y1 and y2 .
(7) Show that y = xr is a solution of the D.E.

x2 y 00 + αxy 0 + βy = 0,

provided that r2 + (α − 1)r + β = 0 and hence, find the solutions of the following
D.Es.
(i) x2 y 00 + 5xy 0 − 5y = 0, (ii) x2 y 00 − xy 0 − 2y = 0; y(1) = 0, y 0 (1) = 1.
(8) Given the equation

xy 00 − (1 + 3x)y 0 + 3y = 0,

has a solution of the form ecx , for some constant c, find the general solution.
(9) Find the general solutions of the following D.Es.
(i) x2 y 00 + 3xy 0 + y = 0,
(ii) x2 y 00 − xy 0 + y = 0.
(10) Show that if y1 is a solution of

y 00 + p(x)y 0 + q(x)y = 0, (4.8)


R
R e
− p(x)dx
then y2 = y1 u(x)dx, where u = y2 (x) , is also a solution. (Hint use the
1
transformation y2 = y1 v(x).) This method is called the reduction of order, since
the substitution y2 = y1 v(x) reduces the problem of solving the second order

70
equation to that of solving a first order equation.
Examples:
(1) Find the general solution of the following D.E.

(1 − x2 )y 00 + 2xy 0 − 2y = 0, y(0) = 3, y 0 (0) = −4.

Clearly, y1 = x is one solution of the D.E.


R To find the second solution we use the
R e
− p(x)dx
above formula y2 = y1 u(x)dx, u = y12 (x)
. First of all, we write the equation
in the standard form (equation (4.8))
2x 0 2
y 00 + 2
y − y = 0.
1−x 1 − x2
R 2x dx

2x e 1−x2 1−x2
Thus, p(x) = 1−x2
and hence, u = x2
= x2
, and
Z Z
1 − x2
y2 = y1 u dx = x 2
dx = −(1 + x2 ),
x
is the second solution. Therefore, the general solution is y = c1 x−c2 (1+x2 ). One
can determine the coefficients c1 , c2 according to the boundary conditions stated
above.
Thus, 3 = y(0) = −c2 and −4 = y(0 0) = c1 . Hence, y = −4x + 3(1 + x2 ).
(2) Find the solution of the D.E.

x2 y 00 − 7xy 0 + 16y = 0.

Clearly, the substitution xk , k = const., is a solution of the D.E. for k = 4, hence


the second solution is
R R
Z Z Z Z 7 7 Z
4 e− p(x)dx 4 e x dx 4 x 4 dx
y2 = y1 udx = x 2
dx = x 8
dx = x 8
dx = x = x4 ln x,
y1 (x) x x x

hence,
y2 = x4 ln x.

71
and the general solution is y = c1 y1 + c2 y2 = c1 x4 + c2 x4 ln x.
(3) Find the general solution of the D.E.

(x2 + 1)y 00 − 2y = 0.

Clearly, y1 = x2 + 1 is a solution of the D.E. and hence one can obtain the second
solution. Now p(x) = 0 so that
R
Z Z Z
e− p(x)dx 2 e0
y2 = y1 udx = y1 dx = (x + 1) dx
y12 (x) (x2 + 1)2
Z
1 x 1
y2 = (x2 + 1) dx = (x 2
+ 1)[ + tan−1 x],
(x2 + 1)2 2(x2 + 1) 2
or
1 1
y2 = x + (x2 + 1) tan−1 x.
2 2
Exercise 4.8
(1) Verify that y1 = x(x − 1)−2 is a solution of the D.E.

x(x − 1)y 00 + 3xy 0 + y = 0,

and find the second solution.



(2) Using the substitution t = x, show that the general solution of the D.E.
√ √
4xy 00 + 2(1 − x)y 0 − 6y = e−2 x , is
√ √ 1√ √
y = A exp(3 x) + B exp(−2 x) − x exp(−2 x).
5
(3) Find the second solution of the following D.Es. with the given one solutions
and hence write down the general solution:
6 1
(a) y 00 − xy 0 + y = 0, y1 = x; (b) y 00 − x2
y = 0, y1 = x2
.
(c) (1 − x2 )y 00 − 2xy 0 + 2y = 0 , y1 = x; (d) x2 y 00 + 4xy − 10y = 0, y1 = x2 .
0

(4) Given that the D.E.

4x(1 − x)y 00 + 2(3 − 4x)y 0 − y = 0,

72
has a solution of the form xk , determine k and hence solve the equation completely.
(5) Find the complete solution of the D.E.

x2 y 00 − 2xy 0 + 2y = x2 + ln x.
2 /2
(6) Solve the D.E. below, given that y1 = e−x is one solution

y 00 + xy 0 + y = 0,

such that y(0) = 0 and y 0 (0) = 1.


(7) Prove that the equation

y 00 + P (x)y 0 + Q(x)y = 0,

can be transformed into


d2 y dx
2
+ Q( )2 y = 0.
dt dt
Hence, solve the D.E.

(1 − x2 )y 00 − xy 0 + n2 y = 0.

(8) By putting x = sin t solve the following D.E.


q
(1 − x2 )y 00 − xy 0 + 4y = 2x (1 − x2 ).

73
Chapter 5

Series Solution

5.1 Series solution


Consider the homogeneous linear D.E.

d2 y dy
P (x) 2
+ Q(x) + R(x)y = 0, (5.1)
dx dx
with P (x) 6= 0. If P (x), Q(x) and R(x) are polynomial in x, in this case, the
form of the D.E. suggests that we guess a polynomial solution of y. Thus, in
principle we can determine a polynomial solution of y(x) by setting the sums of
the coefficients of like powers of x in (5.1) to zero.
Example.
Find the general solution of the D.E.

y 00 − 2xy 0 − 2y = 0.

We set ∞
X
y = a0 + a1 x + a2 x2 + a3 x3 + .. = an xn (5.2)
n=0

Hence,

X ∞
X
y0 = nan xn−1 , y 00 = n(n − 1)an xn−2
n=0 n=0

74
so that the D.E. becomes

X ∞
X ∞
X
n(n − 1)an xn−2 − 2x nan xn−1 − 2 a n xn = 0
n=0 n=0 n=0

or ∞ ∞ ∞
X X X
n(n − 1)an xn−2 − 2 nan xn − 2 an xn = 0. (5.3)
n=0 n=0 n=0

We can now factorize xn but the first term needs some amendments, i.e.

X ∞
X
n(n − 1)an xn−2 = (n + 2)(n + 1)an+2 xn
n=0 n=−2

where we have replace n by n + 2 in all terms without affecting the summation,


but the limits changes from 0 → ∞ to n = −2 → ∞. However, one can write

X ∞
X
(n + 2)(n + 1)an+2 xn = (n + 2)(n + 1)an+2 xn
n=−2 n=0

since the summation over n = −1 and n = −2 are zero !. Therefore, equation


(5.3) becomes

X ∞
X ∞
X
[ (n + 2)(n + 1)an+2 − 2 nan − 2 an ]xn = 0
n=0 n=0 n=0

which has the solution



X ∞
X ∞
X
(n + 2)(n + 1)an+2 − 2 nan − 2 an = 0
n=0 n=0 n=0

which yields
2
an+2 = an , n = 0, 1, 2, ... .
n+2
Setting n = 0 we obtain a2 = a0 , n = 1 gives a3 = 23 a1 and a4 = 12 a0 , a5 = 52 a3 =
22
a , a = 13 a4 = 12 13 a0 , ....
53 1 6
We see that one can write all the coefficients of the series (5.2) in terms of only
a0 and a1 . Note that

y 0 (x) = a1 + 2a2 x + 3a3 x2 + ....

75
together with the series (5.2) gives y(0) = a0 and y 0 (0) = a1 . Thus a0 and a1
must be arbitrary until specific conditions (e.g. y(0), y 0 (0)) are imposed on y. To
find two solutions of our original D.E., we choose two different set of values of a0
and a1 . The simplest possible choices are (i) a0 = 1, a1 = 0, (ii) a0 = 0, a1 = 1.
(i) a0 = 1, a1 = 0
In this case a3 = 0, a5 = 0, ... i.e., all the odd coefficients are zero since a3 =
2
a ,a
3 1 5
= 25 a3 , a7 = 27 a5 and so on. The even coefficients are determined from the
relations
1 1 1
a2 = a0 =, a4 = a2 , a6 = a4 = a0 ,
2 3 2.3
and so on. By induction one can write for any even coefficients (a2n )
1 1
a2n = = .
1.2.3.4...n n!
x4 x6 P∞ x2n 2
Hence, the first solution is y1 (x) = 1 + x2 + 2!
+ 3!
+ ... = n=0 n! = ex ,
x2 x3 x4
since ex = 1 + x + 2!
+ 3!
+ 4!
+ .. (and here we replace x by x ). 2

(ii) a0 = 0, a1 = 1
In this case all even coefficients are zero, and the odd terms are defined by
2 2 2 22 2 222
a3 = a1 = , a5 = a3 = , a 7 = a5 =
3 3 5 53 7 753
and by induction one can write for any odd coefficient (a2n+1 , since 2n + 1 is
always an odd number)
2n
a2n+1 = .
3.5.7....(2n + 1)
2
Thus y2 (x) = x + 23 x3 + 3.5
2
x5 + .. is the second solution.
Now observe that y1 (x), y2 (x) are polynomials of infinite degree even though the
coefficients P (x) = 1, Q(x) = −2x and R(x) = −2 are polynomials of finite
degree. Such polynomials are called Power Series. Many of the functions f (x)
that arise in applications can be expanded in power series; that is, we can find
coefficients a0 , a1 , a2 , a3 , ... so that

76
P
f (x) = a0 + a1 (x − x0 ) + a2 (x − x0 )2 + a3 (x − x0 )3 + ... = ∞ n
n=0 an (x − x0 ) .
Such functions are said to be analytic at x = x0 , and the series is called Taylor
Series of f (x) about x = x0 .
Exercise. n
( ddxnf )|x=x0
Show that if f (x) admits an expansion, then an = n!
.

Examples.
(1) Find two linearly independent solutions of
3x 0 1
y 00 + 2
y + 2 y = 0. (5.4)
1+x x +1
This can be written as

(1 + x2 )y 00 + 3xy 0 + y = 0.

Setting

X ∞
X ∞
X
y= an xn , so that, y0 = nan xn−1 , y 00 = n(n − 1)an xn−2 .
n=0 n=0 n=0

Hence, equation (5.4) becomes



X ∞
X ∞
X ∞
X
n(n − 1)an xn−2 + x2 n(n − 1)an xn−2 + 3x nan xn−1 + an xn = 0,
n=0 n=0 n=0 n=0

or

X ∞
X ∞
X ∞
X
n−2 n n
n(n − 1)an x + n(n − 1)an x + 3 nan x + an xn = 0. (5.5)
n=0 n=0 n=0 n=0

We see that the term xn can be made a common factor but the first term needs
some amendments, i.e.,

X ∞
X
n−2
n(n − 1)an x = (n + 2)(n + 1)an+2 xn ,
n=0 n=−2

77
where we have changed n by n + 2 every where in the summation. However, the
sum for n = −2, n = −1 gives zero and hence, one can write

X ∞
X
(n + 2)(n + 1)an+2 = (n + 2)(n + 1)an+2 .
n=−2 n=0

Hence, equation (5.5) yields



X ∞
X ∞
X ∞
X
( (n + 2)(n + 1)an+2 + n(n − 1)an + 3 nan + an )xn = 0,
n=0 n=0 n=0 n=0

which implies that the bracket is zero, i.e.,


n+1
an+2 = − an , n = 0, 1, 2, ...
n+2
Such a relation is called a recurrence formula for the coefficients a2 , a3 , ... in terms
of a0 and a1 . To find two linearly independent solutions of equation (5.4), we
choose the two simplest cases:
(i) a0 = 1, a1 = 0
In this case, all odd coefficients are zero since a3 = − 23 a1 = 0, a5 = − 45 a3 = 0,
and so on. The even coefficients are determined from the relations
1 1 3 1.3 5 1.3.5
a2 = − a0 = − , a4 = − a2 = , a6 = − a4 = − ,
2 2 4 2.4 6 2.4.6
and so on, and by induction one can write, for any even coefficient (a2n )

1.3.5...(2n − 1) 1.3.5...(2n − 1)
a2n = (−1)n = (−1)n .
2.4.6...2n 2n n!
Thus, y1 (x) = 1 − 12 x2 + 1.3 4
2.4
x + ... is one solution.
(ii) a0 = 0, a1 = 1
In this case, all even coefficients are zero, and all odd coefficients can be obtained
from the relations:

2 2 4 2.4 6 2.4.6
a3 = − a1 = − , a5 = − a3 = , a7 = − a5 = − ,
3 3 5 3.5 7 3.5.7

78
and so on, and by induction one can write for any odd coefficient the relation
(a2n+1 )
2.4.6...2n 2n n!
a2n+1 = (−1)n = (−1)n .
3.5.7...(2n + 1) 3.5...(2n + 1)
Thus, y2 (x) = x − 23 x3 + 2.4 5
3.5
x + ... is the second solution.
(2) Solve

y 00 + x2 y 0 + 2xy = 0, with y(0) = 1, y 0 (0) = 0. (5.6)

Setting

X ∞
X ∞
X
n 0 n−1 00
y= an x , so that, y = nan x , y = n(n − 1)an xn−2 .
n=0 n=0 n=0


X ∞
X ∞
X
n−2 2 n−1
n(n − 1)an x +x nan x + 2x an xn = 0.
n=0 n=0 n=0
or ∞ ∞ ∞
X X X
n(n − 1)an xn−2 + nan xn+1 + 2 an xn+1 = 0. (5.7)
n=0 n=0 n=0

We see that the term xn+1 can be made a common factor but the first term needs
some amendments, i.e.,

X ∞
X
n(n − 1)an xn−2 = (n + 3)(n + 2)an+3 xn+1
n=0 n=−3

where we have changed n by n + 3 every where in the summation. However, the


sum for n = −3, n = −2 gives zero (but n = −1 gives 2a2 ) and hence, one can
write ∞ ∞
X X
(n + 3)(n + 2)an+3 = 2a2 + (n + 3)(n + 2)an+3 .
n=−2 n=0

Hence, equation (5.7) yields



X ∞
X ∞
X
2a2 + ( (n + 3)(n + 2)an+3 + nan + 2 an )xn+1 = 0,
n=0 n=0 n=0

79
implies that the bracket is zero, i.e.,
1
an+3 = − an , n = 0, 1, 2, ...
n+3
Setting the sums of the coefficients of like powers of x equal to zero gives 2a2 = 0,
a3 = − 13 a0 , a4 = − 14 a1 , a5 = − 15 a2 , a6 = − 61 a3 , a7 = − 17 a4 , a8 = − 18 a5 , a9 =
− 19 a6 , but since a2 = 0, a5 = 0, a8 = 0, a11 = 0, ... and a3 = − 13 a0 , a4 =
− 14 a1 , a6 = − 16 a3 = 16 13 a0 , a7 = 17 41 a1 , ..., and by induction one can write for any
odd coefficients
(−1)n (−1)n (−1)n
a3n = a0 = n a0 = n a0 .
3.6...3n 3 .2.3.4 3 n!
The value of the coefficient a0 and a1 are determined from y(0) = a0 , y 0 (0) = a1 ,
hence, a0 = 1, a1 = 0 (see equation (5.2)). Hence,

1 x6 x9 X (−1)n 3n
y(x) = 1 − x3 + − + ... = n
x .
3 3.6 3.6.9 n=0 3 n!

(3) Solve the D.E.

(x2 − 2x)y 00 + 5(x − 1)y 0 + 3y = 0, with y(1) = 7, y 0 (1) = 3

Since the initial condition are given at x = 1, we express the coefficients of the
D.E. above as polynomials in (x − 1), and we find y(x) as a power series centered
about x = 1. To this end, we observe that

x2 −2x = x(x−2) = [x−1+1][(x−1−1)] = [(x−1)+1][(x−1)−1)] = [(x−1)2 −1].

Hence, one can write the D.E. as

[(x − 1)2 − 1]y 00 + 5(x − 1)y 0 + 3y = 0. (5.8)

Setting

X ∞
X ∞
X
y= an (x−1)n , so that, y 0 = nan (x−1)n−1 , y 00 = n(n−1)an (x−1)n−2 .
n=0 n=0 n=0

80
Hence, equation (5.8) yields

X ∞
X ∞
X
[(x−1)2 −1] n(n−1)an (x−1)n−2 +5(x−1) nan (x−1)n−1 +3 an (x−1)n−1 = 0,
n=0 n=0 n=0

or

X ∞
X ∞
X ∞
X
n(n−1)an (x−1)n − n(n−1)an (x−1)n−2 +5 nan (x−1)n +3 an (x−1)n = 0,
n=0 n=0 n=0 n=0
(5.9)
n
We factorize (x − 1) , but the second term should be changed as

X ∞
X
n(n − 1)an (x − 1)n−2 = (n + 2)(n + 1)an+2 (x − 1)n ,
n=0 n=−2


X
= (n + 2)(n + 1)an+2 (x − 1)n ,
n=0

since the summation for n = −2 and n = −1 give zero contribution. Hence,


equation (5.9) gives
n+3
an+2 = an , n = 0, 1, 2, ... .
n+2
The initial conditions determine a0 , a1 , which give a0 = y(1) = 7, a1 = y 0 (1) = 3
or a0 = 7 and a1 = 3. We see that a2 = 23 a0 , a3 = 43 a1 , a4 = 54 a2 = 5.3
a ,a
4.2 0 5
=
6 64
a
5 3
= a ,a
53 1 6
= 76 a4 = 753
a,
642 0
and by induction one can write for any coefficients

3.5....(2n + 1) 4.6....(2n + 2)
a2n = .7, and a2n+1 = .3
2.4.6...2n 3.5.7.(2n + 1)

Hence, y(x) = 7+3(x−1)+ 32 .7(x−1)2 + 43 .3(x−1)3 +... is the required solution.


Exercise 5.1
Solve the following D.E. with the given initial conditions
(i) y 00 + x2 y 0 = 0, y(0) = 2, y 0 (0) = −1,
(ii) y 00 + y 0 + xy = 0, y(0) = −1, y 0 (0) = 2,
(iii) y 00 + x3 y 0 + 3x2 y = 0, y(0) = 2, y 0 (0) = −1,
(iv) y 00 − x3 y 0 = 0, y(0) = 0, y 0 (0) = −2.

81
5.2 Singular points, the method of Frobenius
The D.E.
P (x)y 00 + Q(x)y 0 + R(x)y = 0

is said to be regular at x = x0 , if P (x0 ) = 0. There are certain special cases where


we can find solutions y(x) of the form

X
y(x) = (x − x0 )r [a0 + a1 (x − x0 ) + ....] = an (x − x0 )n+r ,
n=0

where n is an integer but r need not be an integer.


Theorem:
Now consider
P (x)y 00 + Q(x)y 0 + R(x)y = 0, x ≥ x0 , (5.10)

and assume that x = x0 is a singular point of equation (5.10). For simplicity we


set x0 = 0. It is possible, sometimes, to solve equation (5.10) by the method
xQ(x) x2 R(x)
of Frobenius. This requires that P (x)
and P (x)
to be analytic at x = x0 .
Equivalently,
Q(x) b0
= + b1 + b2 x + b3 x2 + ... (5.10a)
P (x) x
and
R(x) c0 c1
= 2 + + c2 + c3 x + ... (5.10b)
P (x) x x

In this case we say that x = x0 is a regular singular point of equation (5.10).


Then, we can always find at least one solution y(x) of the form

X
r 2 3
y(x) = x [a0 + a1 x + a2 x + a3 x + ...] = an xn+r
n=0

for some r (possibly complex) [xr = xα+iβ = xα xiβ = xα eiβ ln x ]. Now


xQ(x) x2 R(x)
P (x)
= b0 + b1 x + b2 x2 + b3 x3 + ... and P (x)
= c0 + c1 x + c2 x2 + c3 x3 + ...

82
will converge for |x| < ρ. Now substitute these in equation (5.10) to obtain the
equation
r2 + (b0 − 1)r + c0 = 0,
which is called the indicial equation with r1 ≥ r2 if they are real, where r1 and
r2 are its roots. Then the D.E. has two linearly independent solutions y1 and y2
on the interval 0 < x < ρ of the following form1 :
(a) if r1 − r2 is not a positive integer, then

X ∞
X
y1 = xr1 an xn and y2 = xr2 b n xn .
n=0 n=0

(b) if r1 = r2 , then

X ∞
X
r1 n r1
y1 = x an x and y2 = y1 ln x + x bn xn .
n=0 n=0

(c) if r1 − r2 = N is a positive integer, then



X ∞
X
y1 = xr1 an xn and y2 = ay1 ln x + xr2 b n xn .
n=0 n=0

The correct a may turn out to be zero.


Examples.
(1) Find the general solution of the D.E.

x2 y 00 − 2xy 0 + 2y = 0.

We have (see equations (5.10), (5.10a) and (5.10b)) b0 = −2 (the coefficient of x1 ),


1
c0 = 2 (the coefficient of x2
), so that the indicial equation r2 + (b0 − 1)r + c0 = 0,
or r2 − 3r + 2 = 0, i.e., r = 1, 2. The general solution can then be obtained (see
(a)-(c) above).
(2) Find the indicial equation, its roots, and the general solution of the following
D.Es.
1 0 1 + x2 00 2 0 00 3 0 1 + x
(a) y 00 − y+ y = 0, (b) y + y +xy = 0, (c) y + y + 2 y = 0.
2x 2x2 x x x
1
see page 108

83
(a): We see that b0 = − 12 , c0 = 21 , so that the indicial equation r2 − 23 r + 1
2
= 0,
1
yields r1 = 1, r2 = 2
, hence r1 − r2 is not a positive integer. Thus, y1 =
P∞ 1 P∞
x n=0 a n xn , y 2 = x 2
n=0 bn x
n
can be found.
(b): We have b0 = 2, c0 = 0 so that the indicial equation r2 + r = 0 yields
P
r1 = 0, r2 = −1, so r1 − r2 is a positive integer. Therefore, y1 = ∞ n
n=0 an x , y2 =
P∞
x−1 n=0 bn x
n
or logarithmic.
(c): Here we have, b0 = 3, c0 = 1 so that the indicial equation r2 + 2r + 1 = 0
yields r1 = −1, r2 = −1, i.e., r1 = r2 . This is a logarithmic case and y1 =
P P
x−1 ∞ n
n=0 an x , y2 = y1 a ln x + x
−1 ∞ n
n=0 bn x .
(3) Find two linearly independent solutions of the D.E.

2xy 00 + y 0 + xy = 0, 0<x<∞. (5.11)


P∞
Let y = n=0 an xn+r , a0 6= 0. Compute

X ∞
X
y0 = (n + r)an xn+r−1 , y 00 = (n + r)(n + r − 1)an xn+r−2 ,
n=0 n=0

we see that equation (5.11) yields,



X ∞
X
an [2(n + r)(n + r − 1) + (n + r)]xn+r−1 + an xn+r = 0.
n=0 n=0
P∞
Summing for n = 0, 1 separately and n=2 together give,

[2r(r − 1)a0 + ra0 ]xr−1 + [2(1 + r)ra1 + (1 + r)a1 ]xr



X
+ [2(n + r)(n + r − 1)an + (n + r)an + an−2 ]xn+r−1 = 0.
n=2

Setting the coefficients of each power of x to zero gives:


(i) 2r(r − 1)a0 + ra0 = r(2r − 1)a0 = 0,
(ii) 2(r + 1)ra1 + (r + 1)a1 = (r + 1)(2r + 1)a1 = 0,
(iii) 2(n + r)(n + r − 1)an + (n + r)an = (n + r)[2(n + r) − 1]an = −an−2 , n ≥ 2.

84
1
Equation (i) determines r: i.e., r = 0, r = 2
Equation (ii) forces a1 to be zero
Equation (iii) determines an for n ≥ 2.
(1) r = 0. In this case, equation (iii) becomes
1
an = − an−2 , n ≥ 2.
n(2n − 1)
Since a1 = 0, we shall see that all the odd coefficients are zero. The even coeffi-
cients are determined from the relations
1 1 1 1 1
a2 = − 2.3 a0 , a4 = − 4.7 a2 = 2.4.3.7 a0 , a6 = − 6.11 a4 = − 2.4.6.3.7.11 a0 , and so on.
Setting a0 = 1, we see that
1 2 1
y1 (x) = 1 − 2.3 x + 2.4.3.7 x4 + .... is one of the solutions of equation (5.11) on the
interval 0 < x < ∞.
(2) r = 12
In this case, equation (iii) gives
1
an = − an−2 , n = 2, 3, 4, 5, ...
n(2n + 1)
Again, all the odd coefficients are zero. The even coefficients are determined from
the relations
1 1 1 1 1
a2 = − 2.5 a0 , a4 = − 4.9 a2 = a ,a
2.4.5.9 0 6
= − 6.13 a4 = a,
2.4.5.6.9.13 0
and so on.
Setting a0 = 1, we see that
1 1 2 1
y2 (x) = x 2 (1 − 2.5
x + 2.4.5.9
x4 + ...) is the second solution of equation (5.11).
(2) Solve the following D.E.
1
x2 y 00 + xy 0 + (x2 − )y = 0.
4
P∞
Putting y = xr n=0 xn and substituting for y 0 , y 00 , we have

X X∞
1
an [(r + n)(r + n − 1) + (r + n) − ]xr+n + an xr+n+2 = 0.
n=0 4 n=0

The above results holds for all values of x and so we may equate to zero the
coefficients of the various powers of x. We note that the smallest power of x is r

85
and equating to zero the coefficient of xr gives
(i)
1 1
a0 (r + )(r − ) = 0.
2 2
The next smallest power of x is r + 1 and equating to zero the coefficient of xr+1
gives
(ii)
3 1
a1 (r + )(r + ) = 0.
2 2
For all n ≥ 2, equating to zero the coefficient of xn+r gives
(iii)
1 1
an (n + r + )(n + r − ) + an−2 = 0,
2 2
P∞ n+r+2 P∞
where we have used n=0 an x = n=2 an−2 xn+r . If we assume a0 6= 0,
equation (i) gives r = ± 12 . Thus possible values for the index r in the assumed
expansion have been found from equation (i). Consequently, equation (i) is called
the indicial equation.
Let us assume r = 21 , then equation (ii) gives a1 = 0, and equation (iii) becomes
1
an (n + 1)n + an−2 = 0, or an = − an−2 , (n = 2, 3, 4, ..).
n(n + 1)
Equation (iii) is called the recurrence relation. We see from above that a1 = a3 =
a0
a5 = ... = 0, and a2 = − 2.3 = − 3!1 a0 , a4 = − 4.5
1
a2 = a0
5!
, .... Thus with r = 12 , we
obtain the solution
1 1 2 1 4
y(x) = a0 x 2 (1 − x + x − ....).
3! 5!
Now consider the case r = − 12 . Equation (iii) yields
1
an = − an−2 , (n = 2, 3, 4, ...).
n(n − 1)
In this case, one finds
1 1 2 1 4
y(x) = a0 x− 2 (1 − x + x − ....).
2! 4!

86
Cauchy product formula
P P∞
The two series ∞ n
n=0 an (x − x0 ) and
n
n=0 bn (x − x0 ) have the product

X ∞
X ∞ X
X ∞
( an (x − x0 )n )( bn (x − x0 )n ) = ( aj bn−j )(x − x0 )n = a0 bb +
n=0 n=0 n=0 j=0

(a0 b1 +a1 b0 )(x−x0 )+(a0 b2 +a1 b1 +a2 b0 )(x−x0 )2 +...+(a0 bn +a1 bn−1 +...+an b0 )(x−x0 )n .
Therefore,

X
[ an (x − x0 )n ]2 = a20 + 2a0 a1 (x − x0 ) + (a21 + 2a0 a2 )(x − x0 )2 + ...
n=0

In the reduction of order formula (page 71) there appears an expression of the form
1
y2
, so if y1 is represented by a power series, it can be squared by using the above
1
formula. But then it must be inverted or, equivalently, the series representation
1
of an expression of the form P∞ b (x−x0 )n
must be computed. To accomplish
n=0 n P∞
this, let the last expression equal the series n=0 qn (x − x0 )n , where qn are to be
determined. Therefore we have the relation

X ∞
X
bn (x − x0 )n qn (x − x0 )n = 1.
n=0 n=0

Using the Cauchy product formula on the LHS and equate the coefficients of like
powers of (x − x0 ) to obtain

b0 q0 = 1, b0 q1 +b1 q0 = 0, b0 q2 +b1 q1 +b2 q0 = 0, ... b0 qn +b1 qn−1 +...+bn q0 = 0

Now solve for q0 from first equation, substitute its value in the second equation,
and solve for q1 , use q0 and q1 in the third equation to find q2 , etc.
Example:
Solve the D.E.( let y1 = J0 ) is the first solution
1
y 00 + y 0 + y = 0.
x
1 1 1 6
J0 = y1 = 1 − x2 + x4 − x + ...
4 64 2304

87
R
Z − 1
dx Z
e x 1
y2 = y1 dx = J0 dx.
y12 xJ02
From Cauchy product formula one obtains (a0 = 1, a1 = 0, a2 = − 14 , a3 = 0, a4 =
1 1
64
, a5 = 0, a6 = − 2304 )
1 3
J02 = 1 − x2 + x4 + ...
2 32
To find J12 we solve the following equations (b0 = 1, b1 = 0, b2 = − 12 , b3 = 0, b4 =
0
3
32
, ... so that q0 = 1, q1 = 0, q2 = − 12 q0 = 0(q2 = 12 ), b0 q3 + b1 q2 + b2 q1 + b3 q0 =
5
0(q3 = 0), b0 q4 + b1 q3 + b2 q2 + b3 q1 + b4 q0 = 0(q4 = 32
), ... Therefore,
Z
1 1 5 1 5 4
y2 = J0 (1 + x2 + x4 + ...)dx = J0 (ln x + x2 + x + ...)
x 2 32 4 128
We see that the solution for this case has a logarithmic form.
Example
Solve the D.E.
2
y 00 + y 0 + xy = 0.
x
2
The indicial equation is r + r = 0, i.e., r1 = 0, r2 = −1. Since r1 − r2 = 1 (a
positive integer) this may or may not be a logarithmic case. The first solution has
P∞ R dx R
the form y1 = n=0 an xn (r = 0) and y2 = y1 x2 y 2
= y1 x12 (b0 + b1 x + b2 x2 +
1
−1
...)dx = y1 (−b0 x + b1 ln x + b2 x + ...). However, we will get (b1 = 0) from the
P∞
calculations, and hence y2 = x−1 n=0 cn x
n
(From Cauchy formula).
Exercise 5.2
(1) Determine whether the specified values of x is a regular singular point of the
given D.E.
(i) x(x − 2)2 y 00 + xy 0 + y = 0, x = 0.
(ii) x(x − 2)2 y 00 + xy 0 + y = 0, x = 2.
(iii) x(ex − 1)y 00 + ex y 0 + y = 0, x = 0.
(2) Find the series solution of the following D.Es.
(i) (1 − x2 )y 00 − xy 0 + y = 0.
(ii) 2x(1 − x)y 00 + (1 − x)y 0 + y = 0.

88
(iii) 2xy 00 + (1 + x)y 0 − 3y = 0.
(3) Show that the indicial equation of

x2 y 00 + xy 0 + (1 + x)y = 0

has complex roots r = ±i.


Show that the D.E. has two linearly independent solutions of the form

X ∞
X
n
y(x) = sinh(ln x) an x + cosh(ln x) b n xn .
n=0 n=0

(4) The D.E.


xy 00 + (1 − x)y 0 + λy = 0 , λ = const,

is called Laguerre differential equation.


(a) Show that the indicial equation is r = 0.
P
(b) Find a solution y(x) of the form y(x) = ∞ n
n=0 an x .
(c) Show that this solution reduces to a polynomial of degree n if λ = n.
d n
(d) Show that this polynomial can be written as Ln (x) = ex dx n −x
n (x e ).

5.3 Bessel’s Equation


Use the method of Frobenius to solve Bessel’s equation of zero order given by

x2 y 00 + xy 0 + x2 y = 0, x > 0. (5.12)

xQ(x) x2 x2 R(x) x2 .x2


Here we have P (x)
= x2
= 1 and P (x)
= x2
= x2 as both analytic functions
of x at x = 0.
Therefore, x = 0 is a regular singular point of Bessel’s equation above, and we
can find a solution y(x) of the equation of the form

X
y(x) = an xn+r .
n=0

89
P P∞
Now y 0 = ∞ n=0 (n + r)an x
n+r−1
, y 00 = n=0 (n + r)(n + r − 1)an xn+r−2 . The
equation (5.12) yields

X ∞
X ∞
X
an (r + n)(r + n − 1)xn+r + (n + r)an xn+r + an xn+r+2 = 0,
n=0 n=0 n=0
P∞ P∞
and using n=0 an xn+r+2 = n=2 an−2 xn+r , one finds

[2r(r − 1)a0 + ra0 ]xr−1 + [2(1 + r)ra1 + (1 + r)a1 ]xr



X
+ [(2(n + r)(n + r − 1)an + (n + r)an + an−2 ]xn+r−1 = 0,
n=2

where we have summed for n = 0 and n = 1 in the first two square brackets, and
replace n by n − 2 in the last term. Setting the sum of the coefficients of like
powers of x equal to zero gives
(i) [r(r − 1) + r]a0 = r2 a0 = 0,
(ii) [(r + 1)r + (r + 1)]a1 = (r + 1)2 a1 = 0,
(iii) [(r + n)(r + n − 1) + r + n]an = (r + n)2 an = −an−2 , n ≥ 2.
Equation (i) is the indicial equation of the D.E. and it has a double root r = 0.
Equation (ii) forces a1 to be zero, and equation (iii), which can be written as
an = − n12 an−2 , determines a2 , a3 , ... in terms of a0 . Specifically, all the odd
coefficients are zero, and
a2 = − 212 a0 , a4 = − 412 a2 = 1
a,
22 42 0
a6 = − 612 a4 = 1
a,
22 42 62 0
and so on.
Setting a0 = 1, we see that
y1 (x) = 1 − 212 x2 + 22142 x4 + .... is one solution of Bessel equation. This solution is
known as Bessel solution of order zero, and is denoted by J0 (x). A second solution
R
is y2 = J0 (x) xJdx
2 (x) can be found. This solution when properly normalized is
0
known as Neumann function of order zero, and is denoted by Y0 .

Using the substitution z = xy, show that Bessel’s modified equation of order
1
2
, viz.
1
x2 y 00 + xy 0 − (x2 + )y = 0 ,
4

90
has the general solution
1
y = x− 2 (Aex + Be−x ) ,

where A, B are constants. Consider the D.E.


d2 y dy
+ P (x) + Q(x)y = R(x).
dx2 dx
Suppose y = u(x) is a known particular solution of either of the equations
d2 y dy
2
+ P (x) + Q(x)y = R(x),
dx dx
or
d2 y dy
2
+ P (x) + Q(x)y = 0.
dx dx
To find the second solution we write,

y = u(x)v(x).

Then the given equation becomes

(u00 v + 2u0 v + uv 00 ) + P (u0 v + uv 0 ) + uvQ = R.

But
u00 + P u0 + Qu = R or 0.
Hence,
uv 00 + (2u0 + P u)v 0 = 0 or R. (5.13)
This permits us to find v.
Example:
If f1 (x) + f2 (x) + f3 (x) = 0, prove that y = ex is a solution of

f1 (x)y 00 + f2 (x)y 0 + f3 (x)y = 0.

Solve the D.E. x2 y 00 + 2x(1 − x)y 0 + x(x − 2)y = 0. When

y = ex , [f1 (x) + f2 (x) + f3 (x)]ex = 0.

91
Hence, y = ex is solution of f1 y 00 +f2 y 0 +f3 y = 0, (since x2 +2x(1−x)+x(x−2) =
0, y = ex is a solution.) To find the complete solution, we write y = ex v. Using
equation (5.13) the above D.E. yields
2
v 00 + v 0 = 0,
x
or
x2 v 0 = 0, v 0 = A/x2 ,
which can be solved to give

y = ex (B − A/x),

where A, B are constants.


Exercise 5.3
(1) Use the method of Frobenius to verify the solution of the D.E.

d2 y
+ ω 2 x = 0,
dx2
which has the solutions y1 = c1 sin ωx and y2 = c2 cos ωx.
(2) Find the general solution of the D.E.

(1 − x2 )y 00 − 2xy 0 + 2y = 0.

(Try another solution by considering that y = x is one solution.)


(3) Prove that y = x2 is a solution of xy 00 + ( 12 x2 − 1)y 0 − xy = 0, and obtain the
general solution.
1
(4) Show that x− 2 cos x is a solution of
3
x2 y 00 + (x2 − x tan x − )y = 0,
4
and find the general solution.
(5) Given that the D.E.

4x(1 − x)y 00 + 2(3 − 4x)y 0 − y = 0

92
has a solution of the form xk , determine k and hence solve the equation completely.
(6) It is given that y = e2x satisfies

(x + 2)y 00 − (2x + 5)y 0 + 2y = 0.

Find the general solution of

(x + 2)y 00 − (2x + 5)y 0 + 2y = (x + 1)ex .

(7) The real function u satisfies the equation

u00 + (E − x2 )u = 0,

where E is a constant, and is related to the function u = H(x) exp(− 12 x2 ). Prove


that
H 00 − 2xH 0 + (E − 1)H = 0.

By considering the series solution of H(x), prove that if E = 2n + 1, where n is


an integer, there is a bounded solution for u(x). Obtain the bounded form of u(x)
in the case n = 0 and n = 1. The function H(x) is known as Hermite polynomial
function.
(8) Show that the two series
1 1 6
Ai(x) = 1 + x3 + x + ...
6 180
1 4 1 7
Bi(x) = x + x + x + ...
12 504
are the solutions to Airy’s equation

y 00 − xy 0 = 0.

(9) The hypergeometric equation is given by

x(x − 1)y 00 + [c − (a + b + 1)x]y 0 − aby = 0,

93
where, a, b, c are constants.
(a) Show that the equation has regular singular point at x = 0 and that the
indicial equation has roots 0, 1 − c.
(b) Show that if c is not negative or 0, than one solution is given by
a.b a(a + 1) b(b + 1) 2
y1 (x) = 1 + x+ x + ..., |x| < 1
1.1 2! c(c + 1)
This function is denoted by F (a, b, c; x) and its series is termed the hypergeometric
P
series. When a = 1 and b = 1, the series reduces to geometric series n xn .
(c) Show that if c is not an integer, then two linearly independent solutions for 0 <
|x| < 1 are given by y1 = F (a, b, c; x) and y2 = x1−c F (a−c+1, b−c+1, 2−c; x).
(10) Legendre equation is given by

(1 − x2 )y 00 − 2xy 0 + n(n + 1)y = 0.

(i) Show that the equation has regular singular points at x = ±1.
(ii) To study the singular point at x = 1, let t = x − 1 to obtain the equation

d2 y dy
(t2 + 2t) + 2(1 + t) − n(n + 1)y = 0,
dt2 dt
with a singular point at t = 0. Show that the indicial equation has roots 0, 0.
(11) By applying the method of Frobenius to the D.E.

(2x + x3 )y 00 − y 0 − 6xy = 0,

obtain the indicial equation r(2r − 3) = 0. Hence obtain the complete solution of
the D.E. as far as the term in x8 .
(12) Use the substitution z = et to transform the equation

d2 y 1
2
+ (e2t − )y = 0,
dt 9
into an equation
d2 y 1 dy 1
+ + (1 − )y = 0.
dz 2 z dz 9z 2

94
P
By substituting y = ∞ n=0 an z
n+r
in this equation find the general solution of the
original equation as the sum of two series which are independent of each other.
(13) Show that the Hermite equation

y 00 − axy 0 + any = 0, a const.,


1 2 dn 1 2
has the polynomial solution Hn (x) = (−1)n e 2 ax dxn
(e− 2 ax ), if n is an integer.
This polynomial is called Hermite polynomial.
(14) Obtain a series solution in descending power of x of the D.E.

4x3 y 00 + 6x2 y 0 + y = 0.

(a) By putting x = 1/t, write the D.E. in terms of t and hence solve the resulting
D.E. (we see that as t increases x decreases (descends))
(b) Show that the solution may be written in the form
1 1
y = A cos √ + B sin √ ,
x x

where A and B are constants (see page 105 for sin x and cos x expansion).
x4 x6
(15) Show that the series x2 + 3
+ 5
+ ... is absolutely convergent2 .
(16) Solve the following D.E.

xy 00 − 2(x + 1)y 0 + (x + 2)y = (x − 2)e2x .

2
See page 106

95
Chapter 6

Applications of Differential
Equations

6.1 Motion of a particle in a resistive medium


(1) Consider a mass m bounced on a spring and immersed in a liquid (resistive
medium) so that the mass experiences a force proportional to its velocity. Find
the motion of the mass. We set up the equation of motion of this mass. The total
force is
F = ma = −kx − cv, (6.1)
where −kx is the force due to the spring and −cv is the resistance force of the
medium, c and k are the proportionality the constants. x,v and a are the displace-
ment of the mass from its equilibrium position, the velocity and the acceleration,
respectively. Thus,
d2 x dx
m 2
= −kx − c , (6.2)
dt dt
so that
d2 x c dx k
2
+ + x = 0. (6.3)
dt m dt m
This is a linear D.E. of second order. The roots of this equation are
√ √
−c + c2 − 4km −c − c2 − 4km
m1 = , m2 = . (6.4)
2m 2m

96
We discuss the following cases:
(i) c2 − 4km > 0: In this case, m1 and m2 are negative, and the solution of
equation (6.3) has the form,

x(t) = c1 em1 t + c2 em2 t . (6.5)

(ii) c2 − 4km = 0 (equal roots): In this case the solution is represented by

x = c1 e−ct/2m + c2 te−ct/2m . (6.6)

(iii) c2 − 4km < 0: In this case we have complex roots. Thus,

x = e−ct/2m (c1 sin βt + c2 cos βt), (6.7)



4km−c2
where β = 2m
. The first two cases referred to as overdamped and critically
damped motion, respectively. The third case, which is referred to as underdamped
motion, happens quite often in mechanical systems. This is because in all cases x
involves a term that decreases rapidly with time t.
(2) Free vibration:
This is the simplest case of free undamped motion. In this case, equation (6.3)
reduces to
d2 x k
2
+ x = 0. (6.8)
dt m
Or
d2 x
+ ω02 x = 0, (6.9)
dt2
k
where ω02 = m . The general solution of equation (6.9) is

x(t) = A sin ω0 t + B cos ω0 t, (6.10)

where A and B are constants.


(3) Forced free vibration:
We now remove the damping term from our system and consider the case of forced

97
free vibration, where the forcing term is periodic and has the form F = F0 cos ωt.
In this case the D.E. governing the motion of the mass m is given by

d2 x F0
2
+ ω02 x = cos ωt. (6.11)
dt m
In this case the general solution of the D.E. yields
F0
x(t) = c1 cos ω0 t + c2 sin ω0 t + cos ωt. (6.12)
m(ω02 − ω 2 )

Now the very interesting case is when ω = ω0 ; that is, when the frequency ω of
the external force equals the natural frequency of the system. This case is called
the resonance case, and the D.E. of the motion for the mass m is

d2 x F0
2
+ ω02 x = cos ω0 t. (6.13)
dt m
We see that the coefficient of the cosine term (the amplitude) in equation (6.12)
becomes infinitely large as ω → ω0 ! Such a phenomenon was responsible for the
collapse of the Tacoma Bridge (USA) in 1940.

6.2 Schrödinger Equation


This equation describes the motion of micro-particles (e.g. electron, proton, neu-
tron, etc.). It is a second order D.E., it has the form

d2 ψ 2m 2m
+ 2 (E − V )ψ = 0, or ψ 00 + (E − V)ψ = 0 (6.14)
dx2 h̄ h̄2
where E = constant, is the total energy of the particle of mass m, V is the
potential energy of the particle, h = 2πh̄ is known as Planck’s constant. ψ is
called the wave function of the particle, since micro-particles have wave nature.
Thus, the solution ψ depends only on the form of the potential V (x).

98
Examples:
(1) V (x) = constant,
in this case Schrödinger equation reads,

ψ 00 + k 2 ψ = 0,

2m
where k 2 = h̄2
(E − V )=const., thus one has, if E > V ,

ψ = c1 sin(kx) + c2 cos(kx),

and, if E < V ,
ψ = c1 ekx + c2 e−kx .

(2) V (x) = 12 kx2 , then one gets the solution of a simple harmonic motion (quan-
tum mechanically). Such a system represents the vibrational motion of atoms in
a solid crystal. The resulting D.E. can be written as,

ψ 00 + aψ − bx2 ψ = 0, a , b, const..

In this case the solution is obtained by the method of series solution discussed
previously (see exercise (7) page 93).

6.3 Electric circuit


Consider an electric circuit consisting of a capacitor (C), an inductor (L), a resistor
(R) and a power generator (E). The total voltage is the sum of the voltages of
each component. Therefore,
dI Q
E(t) = L + IR + ,
dt C
where L dI
dt
is the voltage drop across the inductor (L is called the inductance),
IR is the voltage drop across the resistor and QC
, is the voltage drop across the
capacitor (C is called capacitance, Q is the charge).

99
dQ
The above equation can be written as (since I = dt
)

d2 Q dQ Q
L 2
+R + = E(t),
dt dt C
which is a non-homogeneous second order D.E. with constant coefficients (L, R, C).
The solution of such an equation is already given in sec.4.2. As discussed previ-
ously, this system has a resemblance with the mechanical system investigated in
sec.6.1. For example, the tuning knob of a radio is used to vary the capacitance
in the tuning circuit. In this manner, the resonant frequency is changed until it
coincides with the frequency with one of the incoming radio signals. In this case
the amplitude of the current produced by this signal will be the maximum one
in relation to other signals. In this way, the tuning circuit picks out the desired
station.
Examples:
(1) Consider a C-L circuit (i.e., R = 0) then our D.E. becomes
Q
LQ̈ + = E0 sin ωt.
C
This is a D.E. and can be written as
Q E0
Q̈ + = sin ωt.
LC L
d
Hence, (D = dt
)
1 E0
(D2 + )Q = sin ωt.
LC L

100
The complementary solution is
1 1
Qc = A sin √ t + B cos √ t,
LC LC
and the particular solution is
E0 E0
L
sin ωt L
sin ωt
Qp = 2 1 = 1 .
D + LC −ω 2 + LC
The general solution is
E0
1 1 sin ωt
Q = A sin √ t + B cos √ t+ L 2 1 .
LC LC −ω + LC
(2) Solve the following simultaneous D.Es.
dx 1 dy 3
+ 18x + + y + 5e2t = 0,
dt 2 dt 2
dx dy
5 + 22x + + 2y + 7e−2t = 0.
dt dt
Here x, y are dependent variables and t is the independent variable. We have to
find x, y in terms of t and possible integration constants. First write the equations
in the forms
1
(D + 18)x + (D + 3)y = −5e2t , (6.15)
2
(5D + 22)x + (D + 2)y = −7e−2t , (6.16)
where D = dtd . Now operate with 2(D + 2) on equation (6.15) and with (D + 3)
on equation (6.16) and subtract to eliminate y and its derivative. Then

[2(D + 2)(D + 18) − (D + 3)(5D + 22)] = −10(D + 2)e2t + 7(D + 3)e−2t ,

or
10 2t 7 −2t
(D − 2)(D + 1)x = e − e . (6.17)
3 3
The complementary solution of the above equation is

xc = Ae2t + Be−t .

101
The particular solution is obtained from
10 2t
3
e − 73 e−2t
xp = .
(D − 2)(D + 1)
40 2t 7
xp = Ae2t + Be−t + te − e−2t . (6.18)
9 12
To find y, we now eliminate Dy from equations (6.15), (6.16) to obtain y directly
in terms of x, Dx and know function of t. To this end, multiply equation (6.15)
by 2 and subtract equation (6.16) to give

(−3D + 14)x + y = −10e2t + 7e−2t ,

yp = (3D − 14)xp − 10e2t + 7e−2t .

This gives
10 320 2t 56 −2t
yp = −8Ae2t − 17Be−t + ( − t)e + e .
3 9 3
This equation finds an application in electric circuit (a transformer circuit).
Exercise 6.1
(1)-(a) In a C-L circuit, find the resonant frequency of the system in terms of C, L;
if the driving voltage is E = E0 sin ωt.
(1)-(b) If the circuit does not contain an inductor and E = 0, find the time at
1
which the charge (Q) reduces to e
of its initial value.
(2) Find the equation of motion of a simple pendulum subjected to viscous drag
= −k`θ̇, where θ is angle between the pendulum string and the vertical, k =const.,
` is the length of the string, [consider a small vibration of the pendulum]
(3) Let a particle of mass m move on the x-axis subject to a force F = −kx+g(v),
where k is a positive constant and g(v) = −bv 2 , b is a constant and v is the velocity.
Find v in terms of x and give a physical interpretation.
(4) Let a particle of mass m move on the x-axis subject to a force F ; let x =
x0 , v = v0 when t = 0. Find x and v in terms of t for each of the following

102
cases (m, g, b, and k are constants):
(a)
F = −mg − bv ,

(b)
F = −bv + k sin t ,

(5) Solve the following simultaneous D.Es.

dx dy
3 + − x + y = 4 cos t,
dt dt

dx dy
+ − 2x = e−t ,
dt dt
given that at t = 0, x = 0, y = 0.

(6) Solve the following simultaneous D.Es.

dx dy
+ 2 − x − 5y = sin 2t + cos 2t,
dt dt

dx dy
+ 3 − 2x − 6y = 9et ,
dt dt
given that at t = 0, x = 0, y = 0.

103
References
Zalman Rubinstein, A course in ordinary and partial differential equations,
Academic Press(U.K.),1969
M. Braun, Differential equations and their applications, Springer-Verlag (U.S.A.),
1975
Frank Ayres, Jr., Differential equations, Schaum’s outline series, McGraw-Hill
Book Company (1952)
Wilfred Kaplan, Ordinary differential equations, Addison-Wesely Publishing
Company (USA), 1958
Frank Chorlton, Ordinary differential and differential equations, D.Van Nos-
trand Company LTD.(UK), 1965
R. A. Agnew, Differential Equations, McGraw-Hill Book Company (1960)
David. A. Sanchez. Differential equations, Addison-Wesely Publishing Com-
pany (USA), 1988

104
6.4 Table of Integration
R xn+1
xn dx = n+1
+ C, n 6= −1
R dx
x
= ln |x| + C
R x x
a dx = lna a + C , a > 0
R ax 1 ax
e dx = a e +C
R 1
cos ax dx = a
sin ax + C
R
sin ax dx = − a1 cos ax + C
R
csc2 x dx = − cot x + C
R
sec2 x dx = tan x + C
R
√ dx
a2 −x2
= sin−1 xa + C
R dx
a2 +x2
= a1 tan−1 xa + C
R dx 1
x2 −a2
= 2a ln x−a
x+a
+C
R dx −1 x

x2 −a2
= cosh a + C
R dx

2
x ±a 2 = ln (x + x2 ± a2 ) +C
R
cosh x dx = sinh x + C
R
sinh x dx = cosh x + C
R dx
1+cos x
= tan x2 + C
R 2
tan x dx = tan x − x + C
R dx
sin x
= ln | tan x2 | + C

Taylor expansion of some common functions


ex = 1 + x + 2!1 x2 + 3!1 x3 + ....
1/(1 − x) = 1 + x + x2 + x3 + x4 + ..., |x| < 1
x2 3 4
ln(1 − x) = −(x + 2
+ x3 + x4 + ...), |x| < 1
1 3 1 5
sin x = x − 3!
x + 5!
x − ....
1 2 1 4
cos x = 1 − 2!
x + 4!
x + ....

105
Convergence of a series
In obtaining the Frobenius series, it is important to know the range of the values
of x for which the series converges. To this end, we summarize the ratio test
for convergence. If each term of the series

u0 + u1 + u2 + ... + un + ...

is positive (i.e., ur > 0 (r = 1, 2, 3, ...)), and if further


un+1
lim = ` < 1,
n→∞ un
then the series converges. When, however, we have a power series of the form

a0 + a1 x + a2 x2 + ... + an xn + ...,

the convergence is studied in the following way. Write un = an xn (n = 0, 1, 2, ...).


Then
un+1 an+1
| |=| x | . | |.
un an
Now impose
an
lim | |→ R as n → ∞.
n→∞ an+1
Then
un+1
lim | |=| x | /R.
n→∞ un
R is called the radius of convergence. The series converges for |x| < R and
diverges for |x| > R. Let us suppose | x |< R. Then,
un+1
lim | |< 1,
n→∞ un
so that by the ratio test the positive series

| u0 | + | u1 | + | u2 | +...+ | un | +...

106
is convergent. This implies that the series u0 + u1 + u2 + ... + un + ... is absolutely
convergent and hence also convergent for all values of x satisfying |x| < R.
Example:
1
Consider y = a0 x 2 (1 − 3!1 x2 + 5!1 x4 + ...). We have to discus the convergence of
(−1)n x2n (−1)n
(1 − 3!1 x2 + 5!1 x4 + ...). Write un = (2n+1)!
= a2n x2n , where a2n = (2n+1)!
. Then

a2n+1 1
lim | |= n→∞
lim = 0.
n→∞ a2n (2n + 2)(2n + 3)

Regular Singular Points


Consider the equation

P (x)y 00 + Q(x)y 0 + R(x)y = 0,


P P P
where P (x) = ∞ n
n=0 an x , Q(x) = ∞ n
n=0 bn x , R(x) = ∞ n
n=0 cn x . This
equation is said to have a regular singular point at x = 0 if, after cancellation of
the highest power of x which is a common factor of the coefficients, the equation
takes the form

(i) (a1 x + a2 x2 + ...)y 00 + (b0 + ....)y 0 + (c0 + ...)y = 0, (a1 6= 0),

(ii) (a2 x2 + a3 x3 + ...)y 00 + (b1 x + ....)y 0 + (c0 + ...)y = 0, (a2 6= 0).

When P (x0 ) 6= 0, we say that x0 is an ordinary point.


Example:
Consider the D.E.
x2 y + xy 0 + (x3 − 1)y = 0.

Here we have P (x) = x2 , i.e. a0 = 0, a1 = 0, a2 = 1, a3 = a4 = ... = 0(a2 =6


0), Q(x) = x, i.e. b0 = 0, b1 = 1, b2 = b3 = ... = 0, R(x) = x3 − 1, i.e. c0 =
−1, c1 = 0, c2 = 0, c3 = 1, c4 = c5 = ... = 0 so that (ii) applies (our D.E. above
can be written as equation (ii)) and we have a regular singular point at x = 0

107
Theorem
Consider a second order D.E. of the form L(y) = 0, at a regular singular point at
x = 0: where,
L = P (x)D2 + Q(x)D + R(x),
L(n2 ) = f (n)xn+h + g(n)xn+h+k ,
f (n) of degree 2, k > 0, f (r) = 0 for r = r1 , r2 ,

X g(r)...g[r + (s − 1)k]
φ(x, r) = xr (1 + (−1)s xks ),
s=1 f (r + k)...f (r + sk)

∂φ
φ1 (x, r) = ,
∂r
∂ψ
ψ(x, r) = (r − r1 )φ(x, r), ψ1 (x, r) = .
∂r
(1) If (r2 − r1 )/k is not an integer or r2 = r1 + N k, g(r1 + νk) = 0, ν, N
integers, 0 ≤ ν ≤ N − 1 then, y1 = φ(x, r1 ), y2 = φ(x, r2 )
(2) r1 = r2 then y1 = φ(x, r1 ), y2 = φ1 (x, r1 )
(3) r2 = r1 + N k, N is a positive integer and g(r1 + νk) 6= 0 for ν = 0, ..., N − 1
then y1 = ψ1 (x, r1 ), y2 = φ(x, r2 ).
The general solution is written as y = c1 y1 + c2 y2 .

108
6.5 Answers to exercises
Exercise 1.1
(1) (i) order 1, degree 1 (ii) order 2, degree 1, (iii) order 2, degree 2 (iv) order 2,
degree 1.
(2) (i) y = y 0 x, (ii) y 00 + y = 0, (iii) y 00 = 0, (iv) x2 y 00 − 2xy 0 + 2y = 0,
(v) y 00 (1 − x) + xy 0 − y = 0
dy dy 2
(3) (x dx − y)2 = 1 + ( dx ).
(4) (i) y 00 + y − 6 = 0, (ii) y = y 0 x − 2x2 , (iii) y 00 + y = 0
Exercise 1.2
(i) particular, (ii) general, (iii) general (iv) particular, (v) general (vi) general
Exercise 1.3
(1) (i) y = C(1 + x2 )−1/2 , (ii) y 2 = Cx−2 + 12 x2 , (iii) ln x = −y 3 x−3 + C,
(iv) xy = Cey/x
2+y−4x
(2) (i) ln xy = xy + C, (ii) ln 2−y+4x = 4x + C
(3) (i) y ln x + x = y, (ii) y = 4x−2 , (iii) y 2 + 12 x2 = 23 x−2
(4) y 2 = 2x2 ln x + Cx2 .
(5) y 3 = x−2 .
(6) f (x) = −2 cos x + C, x4 + 2yx2 = C.
(7) f (x) = 21 x + Cx−1 .
Exercise 1.4
(1) (i) (y − x)3 = C(y + x − 2), (ii) x2 − 2x + 8y 2 − 14y = C, (iii) (y − x + 1)5 (y +
x − 1) = C, (vi) 6y − 3x = ln(3y + 3x + 2) + C, (v) x − y = C − 2 ln (x − 2y + 1),
(vi) 2(x − y) + ln (2y + 2x − 1) = C.
Exercise 2.1
y 4
(1) (i) 13 x3 − yx = C, (ii) ln x − 4x x 2 1 5
4 = C, (iii) ln x − 2y = C, (iv) yx − 5 x = C.

(2) a = 1, x + e−x sin y = C.

109
(3) a = −2, (2x−1)
2y 2
− x1 = C.

(4) y 1 − x2 = C + sin−1 x.
Exercise 3.1
2
(1) (i) y −4 = − 12 + e3x , (ii) y = − cos x + sinx x + Cx , (iii) y = −x sec x + C sec x,
(iv) y(C cos x − sin x) = cos2 x.
Exercise 3.2
2
(i) y = Cx − 2C 2 , (ii) y = C
+ 12 x2 , y = x, (iii) y = C 2 − Cx−1 , y = − 41 x−2 ,
(iv) y 3 = C 2 +3Cx, (v) y = 14 x2 , y = Cx−C 2 , (vii) y 2 = 2Cx+C 3 , y 4 = − 21 x3 ,
(viii) y = 21 Cx2 − 1
2C
.
Exercise 3.3
(1) (a) y = 2 − cos x, (b) y = ex , (c) y = ex − ex + 1.
(2) (i) ex + e−y = C, (ii) y = Ce3x , (iii) y − 2 = Cex (y − 1), (iv) 4y 3 = 3x4 + C,
y
(v) e− cos y cos x = C, (vi) ln y+1 = 12 x2 + x + C.
(3) (i) x2 − 2xy − y 2 = C, (ii) e−y/x + ln x = C, (iii) x3 + xy 3 = C,
(iv) tan[ (y−x)
2x
] = Cx.
Exercise 4.1
(1) all are independent W 6= 0.
Exercise 4.2
(1) (i) y = c1 ex + c2 e2x + c3 e−x , (ii) y = c1 e2x + c2 xe2x , (iii) y = c1 ex + c2 e−2x ,
(iv) y = c1 e2x + c2 xe2x + c3 e−2x , (v) y = e−x (c1 cos x + c2 sin x).
(2) (4 − 3x)ex , 2 − 3x2 , 0, 2 − 3x2 .
Exercise 4.3
11
(ii) y = c1 e−x + c2 e−4x + 8
− 12 x, (iii) y = c1 ex + c2 e−x + ex (x2 − x + 12 ),
(iv) y = c1 + c2 e2x + c3 e −2x
− 18 x2 − 1
16
, (v) y = c1 ex + c2 e2x + c3 xe2x + 12 x2 e2x .
Exercise 4.4
2
(2) (i) y = c1 sin 6x + c2 cos 6x + 27
sin 3x,
3
(ii) y = c1 e2x + c2 xe2x + 7225
(−77 sin 9x + 36 cos 9x), (iii) y = e−2x (c1 sin 3x +

110
2 7 3x 8 1 x
c2 cos 3x) + 13
x + 34
e − 169
, (iv) y = c1 e−2x + c2 e−3x + 12
e + 12 e−x .
Exercise 4.5
√ √
(1) (i) y = c1 sin 2x + c2 cos 2x + 13 e3x + 1, (ii) y = c1 e−x + c2 xe−x − 15 (sin x −
2 cos x) + 21 x2 − x + 21 (iii) y = c1 ex + c2 e−x − 18 ex (cos 2x + 2x sin 2x + 2x cos 2x),
(iv) y = c1 e2x + c2 e3x + 14 e4x (2x2 − 6x + 25), (v) c1 e3x + c2 xe3x − e3x ln x.
(2) (i) y = c1 sin x + c2 cos x + 21 x sin x, (ii) y = c1 + c2 e2x + c3 e−x + 91 (x3 − 4x2 −
26
3
x),
(iii) y = c1 sin 2x+c2 cos 2x− 14 (x cos 2x), (iv) y = c1 cos x+c2 sin x−cos ln(sec x+
tan x).
Exercise 4.6
(1) y = c1 x3 + c2 x3 ln x.
(2) (i) y = c1 x3 + c2 x3 ln x, (ii) y = c1 x2 + c2 x−1 , (iii) y = c1 x + c2 x−3 + x2 ,
(iv) y = c1 sin ln x + c2 cos ln x − 21 x(ln x − 1), (v) y = c1 x2 + c2 x−1 + 12 x2 ln x.,
1 4 1 2
(vi) y = x−2 [c1 cos(3 ln x) + c2 sin(3 ln x)] + 13
ln x − 169
+ 25
x.
Exercise 4.7
(1) y 3 = 3c1 x + c2 .

(2) (i) y = c1 cos x + c2 sin x − sin x ln | csc x + cot x| , (ii) y = c1 sin 5x +
√ 1
√ √
c2 cos 5x − 45 , (iii) y = e− 2 x (c1 sin 215 x + c2 cos 215 x) + 16 ex , (iv) y = c1 ex +
1
c2 e−x − (x + 1), (v) y = c1 e−x + c2 e−2x + 10
(sin x − 3 cos x), (vi) y = 43 ex +
1 −2x
6
e − (x + 12 ), (vii) y = sin x + x sin x.
(3) y = c1 xn + c2 xn+1 .
(5) (i) y = (c1 ex + c2 e−x − cos x) sin x, (ii) y = (c1 xn+1 + c2 x−n )ex ,
(c) y = (c1 + c2 sin x) exp(− sin x) − sin2 x + 4 sin x − 5.
(6) W = y1 y20 − y2 y10 = 1 .
√ √
1 1
(7) (i) y = c1 x + c2 x−5 , (ii) y = √
2 3
x (1+ 3)
− √
2 3
x (1− 3)
.
(3−c)x
(8) y = c1 ecx + c2 ce(3−2c) [x − c
(3−2c)
].
(9) (i) y = c1 x−1 + c2 x−1 ln x, (ii) y = c1 x + c2 x ln x.

111
Exercise 4.8
(1) y2 = (x ln x + 1)(x − 1)−2
(3) (a) y2 = x ln x, (b) y2 = 51 x3 , (c) y2 = 12 x ln 1+x
1−x
− 1, (d) y2 = − 71 x−5 .
1
(4) k = − 12 , y = x− 2 [c1 + c2 sin−1 (2x − 1)],
(5) y = c1 x + c2 x2 + 12 ln x + 34 x2 ln x.
1 2 1 2 R − 1 x2
(6) y = c1 e− 2 x + c1 e− 2 x e 2 dx.
(7) y = c1 cos(n sin−1 x) + c2 sin(n sin−1 x).

(8) y = (c1 − 14 sin−1 x)(1 − 2x2 ) + 2c2 x 1 − x2 .
Exercise 5.1
(i) y = 2−x− 12 x4 + 15 x5 + 16
1 6
x +..., (ii) y = −1+2x−x2 + 13 x3 + 24
7 4 1
x − 120 x5 +...,
(iii) y = 2 − x − 12 x4 + 15 x5 + 7
144
x8 + ... , (iv) y = −2x − 1 5
10
x − 1
144
x9 + ....
Exercise 5.2
(1) all are regular singular points
√ 1 x x2 x3
(2) (i) y = a0 1 − x2 + a1 x, (ii) y = c1 (1 − x) + c2 x 2 (1 − 1.3 − 3.5 − 5.7 − ...
1 3 √ (− 5
) (− 5
)( −3
) ( −5 −3
)( )(− 1
)
(iii) y = c1 (1+3x+x2 + 15 x )+ c2 x[1− 3.1!2 x+ 3.5.2!
2 2
x2 − 2 3.5.7.3!
2 2
x3 +..
Exercise 5.3
(2) y = a0 (1 − x2 − 31 x4 − 15 x6 − ...) + a1 x or y = a0 [1 − 12 x ln( 1+x
1−x
)] + a1 x
R exp(− 41 x2 )
(3) y = x2 [c1 x3
dx + c2 ]
− 12
(4) y = x cos x[c1 + c2 (x tan x + ln cos x)]
1
(5) k = − 12 , y = x− 2 [c1 + c2 sin−1 (2x − 1)],
(6) y = c1 e2x + c2 (2x + 5) − ex
22 .1.3 4 3
(7) n = 0, u = exp(− 12 x2 )[c1 + c2 x(1 + 23 x2 + 5!
x + 2 .1.3.5.
7!
x6 + ...]
2 3
.... n = 1, u = exp(− 12 x2 )[c3 + c4 (1 − 2!2 x2 − 24!.1 x4 − 2 6!.1.3 x6 + ...]
3
(11) y = c1 [1 + 3x2 + x4 + 17 x6 − 77 x + ...] + c2 x 2 [1 + 34 x2 − 16
3 7 1 4
x + 64 1 6
x 9
− 1792 ]
t 2t 2 2t t 2t 2 2t
(12) y = c1 e 3 [1 − 3e8.2
3 .e
+ 8.14.2.4 − ....] + c2 e− 3 [1 − 3e 2.4
3 .e
+ 2.4.410 − ....]
2
(14) (a) 4t ddt2y + 2 dy
dt
+ y = 0,

112
1 1
y = c1 [1 − (2!x)
+ (4!x1 2 ) − (6!x1 3 ) + ...] + c2 x− 2 [1 − 1
(3!x)
+ 1
(5!x2 )
− 1
(7!x3 )
+ ...]
(16) y = ex (c1 + c2 x3 + ex ).
Exercise 6.1
(1) (a) ω = √ 1 , (b) t = RC.
LC
(2) θ̈ + g` θ̇ + m
k
θ = 0.
(3) (v 2 − 12 mkb−2 + kb−1 x)e2bx/m = C.
(4) (a) v = e−bt/m (v0 +gmb−1 )−gmb−1 , x = x0 −mgb−1 t+(1−e−bt/m )b−2 (m2 g+
mbv0 ), (b) v = (b2 + m2 )−1 [[(b2 + m2 )v0 + km]e−bt/m + (kb sin t − m cos t)],
x = (b3 + m2 b)−1 [[−mv0 (b2 + m2 ) − km2 ]e−bt/m − kb(b cos t + m sin t)] + x0 +
(mv0 + k)/b.
(5) x = 21 sin t − t cos t, y = (2t − 12 ) sin t + (1 − t) cos t − e−t
(6) x = 94 e2t + 27 −2t
4
e + 32 sin 2t − 9e−t , y = 94 e2t − 94 e−2t − 12 sin 2t.

113
6.6 Glossary

114
Glossary  
‫ א و‬
bounced ‫ﻤﺘﺄﺭﺠﺢ‬
above ‫ﺃﻋﻼﻩ‬ bracket ‫ﻗﻭﺱ‬
accordingly ‫ﻭﻓﻘﹰﺎ ﻝـ‬ bridge ‫ﻜﺒﺭﻯ‬
across ‫ﻋﺒﺭ‬ by ‫ﺒﻭﺍﺴﻁﺔ‬
admit ‫ﻴﻘﺒل‬ called ‫ ﻴﺴﻤﻰ‬/ ‫ﻴﻌﺭﻑ‬
affect ‫ﻴﺅﺜﺭ‬ case ‫ﺤﺎﻝﺔ‬
already ‫ﻤﻥ ﻗﺒل‬ certain ‫ﻤﻌﻴﻨﺔ‬
alternatively ‫ﺒﻁﺭﻴﻘﺔ ﺃﺨﺭﻯ‬ change ‫ﺘﺤﻭﻴل‬
amendment ‫ﺘﻌﺩﻴل‬ characteristic ‫ﻤﻤﻴﺯ‬
angle ‫ﺯﺍﻭﻴﺔ‬ choice ‫ﺍﺨﺘﻴﺎﺭ‬
any ‫ﺃﻱ‬ choose ‫ﻴﺨﺘﺎﺭ‬
appear ‫ﻴﻅﻬﺭ‬ circuit ‫ﺩﺍﺌﺭﺓ‬
apply ‫ﻴﻁﺒﻕ‬ clear ‫ﻭﺍﻀﺢ‬
arbitrary ‫ﻏﻴﺭ ﻤﺤﺩﺩ‬ coefficient ‫ﻤﻌﺎﻤل‬
arising ‫ﻨﺎﺸﺊ‬ coincide ‫ﻤﺘﻁﺎﺒﻕ‬
arrange ‫ﻴﺭﺘﺏ‬ collapse ‫ﻴﻨﻬﺎﺭ‬
assign ‫ﻴﻀﻊ‬ compare ‫ﻴﻘﺎﺭﻥ‬
associated ‫ﻤﺭﺘﺒﻁ‬ comparison ‫ﻤﻘﺎﺭﻨﺔ‬
assume ‫ ﻴﻔﺭﺽ‬/‫ﻴﺄﺨﺫ‬ complementary ‫ﻤﻜﻤل‬
assure ‫ﻴﻀﻤﻥ‬ complete ‫ﻜﺎﻤل‬
because ‫ﺒﺴﺒﺏ‬ complex ‫ﻤﺭﻜﺏ‬
become ‫ ﻴﺼﻴﺭ‬/ ‫ﻴﺼﺒﺢ‬ compute ‫ﻴﺤﺴﺏ‬
before ‫ﻗﺒل‬ conclude ‫ﻴﺨﺘﻡ‬
below ‫ ﺃﺴﻔل‬/ ‫ﺘﺤﺕ‬
condition ‫ﺸﺭﻁ‬ drag ‫ﺴﺤﺏ‬
confirm ‫ﻴﺅﻜﺩ‬ driving ‫ ﺠﺎﺫﺏ‬/‫ﻓﺎﻋل‬
consider ‫ﻴﺄﺨﺫ‬ due to ‫ ﺒﺴﺒﺏ‬/ ‫ﻨﺘﻴﺠﺔ ﻝـ‬
consist ‫ﻴﺘﻜﻭﻥ‬ eliminate ‫ﻴﺘﺨﻠﺹ‬
constant ‫ﺜﺎﺒﺕ‬ employ ‫ﻴﺴﺘﺨﺩﻡ‬
contain ‫ﻴﺤﻭﻯ‬ equal ‫ﻴﺴﺎﻭﻯ‬
continue ‫ﻴﺴﺘﻤﺭ‬ equation ‫ﻤﻌﺎﺩﻝﺔ‬
continuous ‫ﻤﺴﺘﻤﺭ‬ equivalent ‫ﻤﻜﺎﻓﺊ‬
correct ‫ﺼﺤﻴﺢ‬ evaluate ‫ﻴﺤﺴﺏ‬
correspond ‫ﻴﻘﺎﺒل‬ even number ‫ﻋﺩﺩ ﺯﻭﺠﻲ‬
curve ‫ﻤﻨﺤﻨﻰ‬ evident ‫ﻭﺍﻀﺢ‬
decrease ‫ﻴﻨﻘﺹ‬ exact ‫ﺘﺎﻡ‬
define ‫ﻑ‬‫ﻴﻌﺭ‬ example ‫ﻤﺜﺎل‬
definition ‫ﺘﻌﺭﻴﻑ‬ except ‫ﻋﺩﺍ‬
degree ‫ﺩﺭﺠﺔ‬ exercise ‫ﺘﻤﺭﻴﻥ‬
denote ‫ﻴﺭﻤﺯ‬ exist ‫ﻴﻭﺠﺩ‬
dependence ‫ﺍﻋﺘﻤﺎﺩ‬ existence ‫ﻭﺠﻭﺩ‬
derivative ‫ﻤﺸﺘﻘﺔ‬ experience ‫ﻴﺨﻀﻊ‬
determine ‫ﻴﺤﺩﺩ‬ express ‫ﻴﻌﺒﺭ‬
different ‫ﻤﺨﺘﻠﻑ‬ expression ‫ ﺘﻌﺒﻴﺭ‬/ ‫ﻤﻘﺩﺍﺭ‬
differentiate ‫ﻴﻔﺎﻀل‬ external ‫ﺨﺎﺭﺠﻲ‬
discuss ‫ﻴﻨﺎﻗﺵ‬ factor ‫ﻋﺎﻤل‬
divide ‫ﻴﻘﺴﻡ‬ family ‫ﻤﺠﻭﻋﺔ‬
downward ‫ﺇﻝﻰ ﺃﺴﻔل‬ limit ‫ﻨﻬﺎﻴﺔ‬
linear ‫ﺨﻁﻰ‬ replace ‫ﻴﺒﺩل‬
look for ‫ﻴﺒﺤﺙ‬ represent ‫ﻴﻤﺜل‬
phenomenon ‫ﻅﺎﻫﺭﺓ‬ required ‫ﻤﻁﻠﻭﺏ‬
point ‫ﻨﻘﻁﺔ‬ resemblance ‫ﻤﺸﺎﺒﻬﺔ‬
polynomial ‫ﻜﺜﻴﺭﺓ ﺍﻝﺤﺩﻭﺩ‬ resistance ‫ﻤﻘﺎﻭﻤﺔ‬
possible ‫ﻤﻤﻜﻥ‬ respectively ‫ﻋﻠﻰ ﺍﻝﺘﺘﺎﻝﻰ‬
power ‫ ﻗﺩﺭﺓ‬/‫ ﻗﻭﺓ‬/ ‫ﺃُﺱ‬ responsible ‫ﻤﺴﺌﻭل‬
previously ‫ﺴﺎﺒﻘﹰﺎ‬ roots ‫ﺠﺫﻭﺭ‬
proceed ‫ﻴﺴﺘﻤﺭ‬ said to be ‫ﻴﻘﺎل ﺒﺄﻨﻪ‬
product ‫ﻨﺎﺘﺞ‬ same ‫ﻨﻔﺱ‬
properly ‫ﺒﻁﺭﻴﻘﺔ ﺼﺤﻴﺤﺔ‬ satisfy ‫ﻴﺤﻘﻕ‬
properties ‫ﺨﻭﺍﺹ‬ seek ‫ﻴﺒﺤﺙ‬
proportional ‫ﻴﺘﻨﺎﺴﺏ‬ separable ‫ﻴﻤﻜﻥ ﻋﺯﻝﻪ‬
prove ‫ ﻴﺜﺒﺕ‬/ ‫ﻴﺒﺭﻫﻥ‬ series ‫ﻤﺘﻭﺍﻝﻴﺔ‬/ ‫ﻤﺘﺴﻠﺴﻠﺔ‬
provided that ‫ﺸﺭﻴﻁﺔ ﺃﻥ‬ set ‫ﻴﻀﻊ‬
put ‫ﻴﻀﻊ‬ set up ‫ﻥ‬‫ ﻴﻜﻭ‬/‫ﻴﻌﺩ‬
rapidly ‫ﺴﺭﻴﻌﹰﺎ‬ show ‫ﺢ‬‫ ﻴﻭﻀ‬/ ‫ﻥ‬‫ﻴﺒﻴ‬
readily ‫ﺒﺴﻬﻭﻝﺔ‬ simplest ‫ﺃﺒﺴﻁ‬
reduce to ‫ﻴﺅﻭل ﺇﻝﻰ‬ since ‫ﺒﻡ ﺃﻥ‬
refer to ‫ﻴﺭﺠﻊ ﺇﻝﻰ‬ single ‫ﻤﻨﻔﺭﺩ‬
region ‫ﻤﻨﻁﻘﺔ‬ situation ‫ ﻭﻀﻊ‬/ ‫ﺤﺎﻝﺔ‬
relation ‫ﻋﻼﻗﺔ‬ so forth ‫ﻭﻫﻜﺫﺍ‬
remove ‫ﻴﺒﻌﺩ‬ solution ‫ﺤل‬
solve ‫ﻴﺤل‬ turn out ‫ﻴﺘﻀﺢ‬
some ‫ﺒﻌﺽ‬ type ‫ﻨﻭﻉ‬
special ‫ﺨﺎﺹ‬ undetermined ‫ﻏﻴﺭ ﻤﺤﺩﺩ‬
specific ‫ ﻤﺤﺩﺩ‬/ ‫ﻤﻌﻴﻥ‬ unit ‫ﻭﺤﺩﺓ‬
standard ‫ﻗﻴﺎﺴﻲ‬ until ‫ﺤﺘﻰ‬
stop ‫ﻴﻘﻑ‬ use ‫ﻴﺴﺘﻌﻤل‬
subjected ‫ﺨﺎﻀﻊ‬ usual ‫ﻋﺎﺩﻯ‬
substitute ‫ﺽ‬‫ﻴﻌﻭ‬ value ‫ﻗﻴﻤﺔ‬
such ‫ﻤﺜل‬ variable ‫ﻤﺘﻐﻴﺭ‬
sufficient ‫ﻜﺎﻓﻲ‬ vary ‫ﻴﻐﻴﺭ‬
suggest ‫ﻴﻘﺘﺭﺡ‬ verify ‫ﻴﺘﺤﻘﻕ ﻤﻥ‬
suitable ‫ﻤﻨﺎﺴﺏ‬ vertical ‫ﺭﺃﺴﻲ‬
sum ‫ﻤﺠﻤﻭﻉ‬ vertically ‫ﺭﺃﺴﻴﹰﺎ‬
suppose ‫ﻴﻔﺘﺭﺽ‬ vibration ‫ﺍﻫﺘﺯﺍﺯ‬
system ‫ﻤﻨﻅﻭﻤﺔ‬ viscous ‫ﻝﺯﺝ‬
term ‫ﺤﺩ‬ wave ‫ﻤﻭﺠﺔ‬
therefore ‫ ﺇﺫﹰﺍ‬/ ‫ﺇﺫﻥ‬ which ‫ﺍﻝﺘﻰ‬
though ‫ﺭﻏﻡ‬ while ‫ﺒﻴﻨﻤﺎ‬
thus ‫ﻫﻜﺫﺍ‬ with respect to ‫ﺒﺎﻝﻨﺴﺒﺔ ﻝـ‬
together ‫ﻤﻊ‬ without ‫ﺒﺩﻭﻥ‬
total ‫ﻜﻠﻰ‬ write ‫ﻴﻜﺘﺏ‬
transformation ‫ﺘﺤﻭﻴل‬ yield ‫ ﻴﻌﻁﻰ‬/ ‫ﻴﻨﺘﺞ‬
true ‫ﺼﺤﻴﺢ‬

View publication stats

You might also like