MTH 232 ELEMENTARY DIFFERENTIAL

EQUATION
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Course Code MTH 232

Course Title Elementary Differential Equation

Course Developer Ajibola Saheed Oluwatoyin

National Open University of Nigeria
14/16 Ahmadu Bello Way
Victoria Island, Lagos

Course Writer Ajibola Saheed Oluwatoyin

National Open University of Nigeria
14/16 Ahmadu Bello Way
Victoria Island, Lagos

Course Editor Ajibola Saheed Oluwatoyin

National Open University of Nigeria
Victoria Island, Lagos.

Programme Leaders Proffessor Adebanjo

National Open University of Nigeria
Victoria Island, Lagos

NATIONAL OPEN UNIVERSITY OF NIGERIA

ii
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

National Open University of Nigeria

Headquarters
14/16 Ahmadu Bello Way
Victoria Island
Lagos

Abuja Office
No. 5 Dar es Salaam Street
Off Aminu Kano Crescent
Wuse II, Abuja
Nigeria

e-mail: [email protected]
URL: www.nou.edu.ng

Published by:
National Open University of Nigeria 2008

First Printed 2008

ISBN: 978-058-424-2

All Rights Reserved

iii
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

CONTENTS PAGES

Module 1…………………………………………………………… 1

Unit 1 Introduction to The Nature of differential Equations... 1-32

Unit 2 Method of solving Equations of first order and first
Degree……………………………………………….. 33-70

Module 2…………………………………………………………………. 71

Unit 1 Linear Differential Equations………………………... 71-102

Unit 2 Differential Equations of First
Order But not First Degree…………………………... 103-129

Module 3…………………………………………………………… 130

Unit 1 Families of Curve Orthogonal and Oblique trajectories

application to Mechanics and Electricity…………….. 130-143

Modeule 4………………………………………………………….. 144

Unit 1 Higher Order Linear Differential Equation …………. 144-170

Unit 2 Method of Undetermined Coefficients………………. 171-185
Unit 3 Method of Variation of Parameters………………….. 186-209

iv
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

MODULE 1

Unit 1 Introduction To The Nature of differential Equations

Unit 2 Equation of first order and first order and first Degree

UNIT 1 INTRODUCTION TO THE NATURE OF

DIFFERENTIAL EQUATION

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Basic Concept
3.2 Solution of a Different Equation
3.3 Family of curves and differential equations
3.4 Differential equation Arising from physical situation9
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

The subject of differential equation constitutes a part of mathematics that plays an

important role in understanding physical sciences. In fact, it is the source of most
of the ideas and theories which constitute higher analysis. In physics, engineering,
chemistry and many other disciplines it has become necessary to build a
mathematical model to represent certain problems. These mathematical models
often involve the search for an unknown function that satisfies an equation in
which derivatives of the unknown function play an important role. Such equations
are called differential equations. The primary purpose of differential equations is
to serve as a tool for studying change in the physical world.

dy
You may recall that if y = f(x) is a given function then its derivation can be
dx
interpreted as the rate of change of y respect to x . sir Isaac Newton observed that
certain important laws of natural sciences can be phrased in terms of equations
involving rates of change. The most famous example of such a natural law is
Newton’s second law of motion. Newton was able to model the motion of a
particle by an equation involving an unknown function and one or more of its
derivatives.

1
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

As early as the 1690s, scientists such as Isaac Newton, Gottfried Leibniz, Jacques
Bernoulli, Jean Bernoulli and Christian Huygens were engaged in solving
differential equations. Many of the methods which they developed are in use till
today. In the eighteenth century the mathematicians Leonhard Euler, Daniel
Bernoulli, Joseph Lagrange and others contributed generously to the development
of the subject. The pioneering work that led to the development of ordinary
differential equations as a branch of modern mathematics is due to Cauchy,
Riemann, Picard, Poincare, Lyapunov, Birkhoff and others.

Not only are differential equations applied y physicists and engineers, but they are
being used more and more in certain biological problems such as the study of
animal populations and the study of epidemics. Differential equations have also
proved useful in economics and other social sciences. Besides its uses, the theory
of differential equations involving the interplay of functions and their derivatives,
is interesting in itself.

2.0 OBJECTIVES

In this unit, we introduce the basic concepts and definitions related to differential
equations. We also express some of the problems of physical and engineering
interest in terms of differential equations in this unit. We shall give the methods
of solving differential equations of various types in Units 2 and 3. The physical
problems formulated in this unit will be solved in units 2 and 3. The physical
problems formulated in unit will be solved in unit 3 after we have learnt the
various methods of solving the first order equations.

• Distinguish between the order and degree of a differential equation;

• Define the solution of an ordinary differential equation;
• Identify an initial value problem;
• State and use the conditions for existence and uniqueness of first order
ordinary different equations;
• Derive differential equations for some physical problems.

3.0 MAIN CONTENT

3.1 Basic Concepts

In this section we shall define and explain the basic concepts in the theory of
differential equations and illustrate them through examples.

2
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

In unit 1 of MTH 112 Differentialcalculus, you have learnt that if a relation y =

y(x) involving two variables x and y exist then we call x the independent
variable and y the dependent variable.

Further, suppose we are given a relation of the type f(x,t 1, t2, ……., tn) = 0
involving (n + 1) variables (x and t1, t2, ………., tn,); where the value of x depends
on the values of the variables t1, t2, ..…., tn are called independent variables and x
is called the dependent variable and y is dependent variable. Similarly, if z = x2 +
y2 +2xy, the x and y are independent variables and z is a dependent variable.

Any equation which gives the relation between the independent and dependent
variables and the derivatives of dependent variables is called a differential
equation .

In general, we have the following definition.

Definition: An equation involving one (or more) dependent variable derivatives

with respect to one or more independent variables in called a differential equation.

dy
For example, = cosx …(1)
dx
dy a
Y=x +
dx d y/dx

….(2)

∂z ∂z
y2 + xy = nzx
∂x ∂y

….(3)

are all differential equations.

∂z ∂z
In Eqn. (3), and are partial derivatives of z w.r.t.x and y respectively. The
∂x ∂y
partial derivatives of a function of two variables z = f(x, y) w.r.t to one of the
independent variables, can be defined as

∂z ∂f lim f(x + Δx, y) − f(x, y)

= = fx(x, y) =
∂x ∂x Δy → 0 Δx

3
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

∂z
when the limit exist and is independent of the path of approach. is the first
∂x
order partial derivatives of z w.r.t.x and is obtained by differentiating z w.r.t.x
treating y as a constant. It is read as ‘del z by del x’. similarly, first order partial
∂z ∂f
derivative of z w.r.t. y is denoted by (or or fy(x, y), so that
∂y ∂y
∂z ∂f lim f(x, y + Δy) − f(x, y)
= = fy(x, y) =
∂x ∂x Δy → 0 Δy

Note that equations of the type

d dy
(xy) = y + x
dx dx

are not differential equations. In this equation, if you expand the left hand side
then you will find that the left hand side is the same as the right hand side. Such
equations are called identities. Moreover, a differential equation may have more
than one dependent variable. For instance,

d 2 x dy
+ +=y
dt 2 dt

is a differential equation with dependent variable x and y and the independent

variable t.

Differential equations are classified into various types. The most obvious
classification of differential equations is based on the nature of the dependent
variable and its derivatives (or derivatives) in the equation. Accordingly, we
divide differential equations into three classes: ordinary, partial and total. The
following definitions give these three types of equations.

Definition: A differential equation involving only ordinary derivatives (that is,

derivatives with respect to a single independent variable) is called an ordinary
differential equation (abbreviated as ODE).

Equations

d2 y
2
+ y = x2,
dx

4
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

2
 dy  2
  = [sin (xy) + 2] , and
 dx 
2
dy  dy 
y=x + r 1+  
dx  dx 

are all ordinary differentia equations.

The typical form of such equations is

 dy(x) d 2 y(x) d n y(x) 

g x, y(x), , ,......, =0 …(4)
 dx dx 2 dx n 

whenever we talk of Eqn. (4) we assume that g is known real valued function and
the unknown to be determined is y. secondly, in an ordinary differential equation,
y and its derivatives are evaluated at x.

It may be noted that the equation

 dy 
  = (y)x+1
 dx  x
dy
is not a differential equation. This is because y is evaluated at (x + 1) whereas
dx
is evaluated at x.

Similarly, the equation

dy(x) x xs
= ∫ e y(s)ds
dx 0

is not a differential equation since the unknown y is appearing inside an integral.

Also, in this case the values of y on the right hand side of the equation depends on
the interval 0 to x, whereas, in a differential equation, the unknown y has to be
evaluated only at x.

Let us now define partial equation.

Definition: Differential equation containing partial derivatives of one ( or more)

dependent variable with respect to two or more independent variable is called a
partial differential equation. (abbreviated AS PDE)

5
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

The examples of differential equations are

∂ 2v ∂ 2v ∂ 2v
+ 2 + 2 =0
∂x2 ∂y ∂z

∂z ∂z
x + y - z = 0,
∂x ∂y
∂ 3u ∂ u ∂ 2u
+ + xtu = 0.
∂ x3 ∂x ∂t2

You may also note that Eqns. (1`) and (2) given earlier are ordinary differential
equations, whereas, Eqn. (3) is a partial differential equation.

And now an exercise for you.

Besides ordinary and partial differential equations, namely, total differential

equations. Before giving you the definition of total differential equations, we
ascribe a meaning to the symbols dx and dy which permit us to manipulate the
dy
derivative as a quotient of two function y = f(x), we define, the differential of
dx
y, by
Dy = f’ (x) dx

If u = f(x,y) be a function of two independent variables x and y, then we know that

∂u lim f(xΔx,+ y) f(x,
− y)
=
∂x Δx → 0 Δx

and

∂u lim f(x, y + Δy) − f(x, y)

=
∂y Δy → 0 Δy

Let ∆ u be the change in u when both x and y change by the amounts ∆ x and ∆ y
lim
respectively, so that ∆ u = du. Here du is called the total differential.
∆ x, ∆ y → 0

The total differential du of a function u (x, y) is defined as

6
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

∂u ∂u
du = dx + dy …(5)
∂x ∂y

or
du = uxdx + uydy

For instance,
If u = x2y – 3y
then

Du = 2xy dx + (x2 – 3) dy

Now consider the relation u (x, y, z) = c where x, y, z are variables and c is a

constant.
Then
Du = 0

∂u ∂u ∂u
⇒ dx + dy + dz = 0
∂x ∂y ∂z

∂u ∂u ∂u
Here, , , are known functions of x, y and z, and therefore the above
∂x ∂y ∂z
equation can be put in the form

P dx + Q dy + R dz = 0

Which is called the total differential equation in three variables. In this equation
any one of the variables x, y, z can be treated as independent and the remaining
two are then the dependent variables.

Similarly, if u = u(x, y, z, t) then corresponding total differential equation will be

of the form

P dx + Q dy + R dz + T dt = 0.

Remember that a total differential equation always involves three or more

variables.

We now give the following definition.

7
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Definition: A total differential equation contains two or more dependent variables

together with their derivatives with respect to a single independent variable which
may, or may not, exist explicitly into the equation.

For example, equations

Yz (1 + 4 xz) dx – xz (1 + 2 xz) dy – xydz = 0,

and

xdx + ydy = zdz zdx − xdz

+ + 2ax2dx + 3by2dy + 3cz2dz = 0.
x +y +z
2 2 2
x 2 − z2

are total differential equations.

We shall be dealing with only ordinary differential equations in Modules 1ansd 2

and devote Modules 3 and 4 study total and partial differential equations.

We next consider the concepts of order and degree of a differential equation on the
basis of which differential equations can be further classified.

We all know that the nth derivative of a dependent variable with respect to one or
more independent variables is called a derivative of order n, or simply an nth order
derivative.
d2 y ∂ 2z ∂ 2z d3z ∂ 2z
For example, , , are second order derivatives and , are
dx 2 ∂ x 2 ∂ x ∂ y dx 3 ∂ x ∂ y
2

third order derivatives.

Definition: The order of a differential equation is the order of the highest order
derivative appearing in the equation. For instance, the equation

d2 y
+ y = x2 is of second order …(6)
dx 2

d2 y
(because the highest order derivative is , which is of second order), whereas
dx 2
2
 dy 
(x + y)   - 1, is of first order …(7)
 dx 

8
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy
(highest order derivative is ).
dx

Similarly, equation
2 3
 d3 y  d 2 y dy  dy 
 dx 3  + 2 - + x2   = 0 is of third order …(8)
  dx 2
dx  dx 

∂ 2z ∂ 2z ∂z
whereas, 2 + 2 + = 0 is of second order. …(9)
∂x ∂y ∂x

Note that the order of a differential equation is a positive integer.

Also, if the order of a differential equation is ‘n’ then it is not necessary that the
equation contains some or all lower order derivatives or independent variables
d4 y
explicitly. For instance, equation = 0, is a fourth order differential equation.
dx 4

Definition: The degree of a differential equation is the highest exponent of the

highest order derivative appearing in it after the equation has been expressed in the
form free from radicals and any fractional power of the derivatives or negative
power. For example Equations. (6) and (9) are of first degree and Equations. (7)
and (8) are of second degree.

Equation

3
dy  dy 
y–x =r 1+   ….(10)
dx  dx 

is of degree three for, in order to make the equation free from radicals, we need to
square both the sides, so that

 dy 
2
  dy  3 
 y-x  = r2 1 +   
dx    dx  

dy
since the highest exponent of the highest derivative, that is, is three, thus by
dx
definition the degree of Equation. (10) is three.

9
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Similarly, Equation. (2), that is,

dy a
y–x + is of degree two.
dx dy / dx

dy dy
This is because we multiplied through by to remove negative power of
dx dx
and get
2
dy  dy 
Y =x  dx  + a.
dx

You may now try the following exercise.

We now classify the differential equations depending upon the degree of

dependent variables and its derivatives into two classes, namely, linear and non-
linear.

Definition: When, in an ordinary or partial differential equation, the dependent

variables and its derivatives occur to the degree only, and not as higher powers or
products, we call the equation linear.

The coefficients of a linear equation are therefore either constants or functions of

the independent variable or variables. If an ordinary differential equation is not
linear, we call it non-linear.

For example, the equation

d2 y
2
+ y = x2 , is an ordinary linear differential equation..
dx
dy
However (x + y)2 = 1 is an ordinary non-linear equation, because of the
dx
dy dy
presence of terms like y2 and 2xy .
dx dx

Similar, equation
2
∂ 2z ∂ 2z  ∂ 2z 
+ - =0, is a non-linear partial differential equation.
∂x2 ∂ y 2  ∂ x ∂ y 

10
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Further, if a partial differential equation is not linear, it can be quasi-linear, semi-

linear or non-liner. We will discuss conditions for these classifications in the
later part of this course.

You may now try this exercise.

Normally when we encounter an equation, our natural curiosity is to enquire about

its solution. But, then it is natural for you to ask as to what exactly is the meaning
of a solution of a differential equation. In the next section you will find an answer
to this question. There, we also answer many more questions like

i) Under what conditions does the solution of a given ordinary differential

equation exists?
ii) If the solution exists, then is it a unique solution?

3.2 Solution of a Differential Equation

You have seen that the general ordinary differential equation of the nth order as
given by Equation (4) is

 dy d 2 y dn y 
g  x, y, , 2 ,...., n  = 0
 dx dx dx 

dy d2 y dn y
using the prime notation for derivatives (y’ = , y” = ,…. Y =
n
) we
dx dx 2 dx n
can rewrite Equation . (4) in the form

y(n) = f(x, y, y’, y”, ….. y((n- 1)) …(11)

Let us assume that we can solve Eqn. (11) for y(n), that is, Eqn. (11) can be written
in the form

Y(n) = f(x, y, y’, y”, ……., y(n-1)) ….(12)

It is normally a simple task to verify that a given function = ø (x) satisfies an

equation like (11) or (12). All that is necessary is to compute the derivatives of y
and to show that y = ø (x) and its derivatives, when substituted in the equation,
reduce it to an identity in x. if such a function y exists, we call it a solution of the
Eqn. (11) or (12).

However, usually we assume that

11
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

i) y = ø (x) is defined on some interval [a, b];

ii) y is n times differentiable on [a, b];
iii) We assume that y has a right derivative at point a and a left derivative at b;
iv) y = ø (x) can be real valued function or complex valued function (range is
a subset of C) of x.

We now give the definition of the solution of an ordinary differential equation.

Definition: A real or complex valued function y = ø (x) defined on an interval I is

called a solution or an integral of the differential equation g(x, y, y’,…., y(n)) = 0 if
ø (x) is n time differentiable and if x, ø (x), ø ’(x), ….., ø (n) (x) satisfy this
equation for all x in I.

For example, the first-order differential equation

dy
= 2x – 4x
dx

Note: I could represent any interval ]a,b[, [a, b] ]0, ∞ [,] − ∞ , ∞ ] and so on

has the solution y = 2x + 1 in the interval I = { x : −∞ < x < ∞ } .

This can be checked by computing y’ = 2 and 2y - 4x = 2(1 + 2x) – 4x

In the same way you can check that y = 1 + 2x + ce2x, in the interval −∞ < x < ∞ ,
is also a solution of this equation for any constant c.

In the above definition you might have noticed that a solution of (11) is real
valued or complex valued. In case y is real value it is called a real solution. If y is
complex valued, it is called complex solution. We are usually interested in real
solution of Eqn. (11). To help you clarify what we have just said let us take some
more examples.
Example 1: Show that for any constant c, the function y(x) = cex, x∈ R is a
solution of

dy
= y, x∈ R …(13)
dx

Solution: Here I is R itself. For any x ∈ R, we know that

dy d dy
= (cex) = c (ex) = cex = y
dx dx dx

which shows that y satisfies equation(13).

12
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Example 2: show that for real constants a and b the functions y (x) = a cos 2x and
z(x) = b sin 2x are solutions of the equation bolow;

d2 y
+ 4y = 0, x ∈ R …(14)
dx 2

solution: We will first show that z (x), x ∈ R is a solution of Equation. (14).

d d
Now [z(x)] = dx
(b sin 2x) = 2bcos 2x.
dx
d2 d
∴ 2 [z(x)] = (2bcos 2x) = - 4b sin 2x = - 4z(x).
dx dx

thus,

d2y
+ 4z(x) = 0, x R. ∈

dx 2
That is, z satisfies Equation (14).

By now you must have understood the meaning of z satisfiying Equation (14)._. It
means that Equation (14). holds when y is replaced by z. similarly, you can check
that y(x) = a cos 2x is also a solution of Equation (14).
. You may observe here that the sum y(x) + z(x) that is, a cos 2x + b sin 2x is
again a solution of. Equation (14).

Let us consider another example.

Example 3: Shown that y(x) = eix, x ∈

R is a solution of

d 2y
dx 2
+ y = 0, x ∈

Solution: We have,

dy d
= dx
(eix)\ = i eix
dx

and

13
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

d2y d
dx2
= dx
(ieix) = i2 eix = -eix = - y(x)

d2y
thus, dx2
+y=0

In the examples taken so far, you have seen that the solution(s)differential
equation exist. In Example (1) and (2) the solutions were real valued whereas, the
solution in Example (3) was a complex valued function. But, there are equations
for which real solution does not exist. Suppose that we are looking for real roots
of the equation x2 + 1 = 0. We know that it does not exist. Likewise, the equation

dy
+ y2 + 1 = 0
dx

does not admit a real solution.

 dy 
Similarly, the equation sin   = 2 does not admit a real solution, because real
 dx 
value of the sin of a real function lies between – 1 and + 1.

You may now try the following exercises.

In the above discussion you must note that a differential equation may have more
than one solution. It may even have infinitely many solutions. For instance, each
4
of the functions y = sin x, y = sin x + 3, y = sin x - is a solution of the
5
differential equation y’ = cosx. but from your knowledge of calculus you also
know that any solution of the equation is of the form.

Y = sin x + c …(15)

Where c is a constant. If we regard c as arbitrary then relation (15) represents the

totality of all solutions of the equation. Thus, we can represents even the infinitely
many solutions by a simple formula involving arbitrary constants. Accordingly,
we classify various types of solutions of an ordinary differential equation as
follows.

Definition: The solution of the nth order differential equation with arbitrary
‘n’nconstants is called its general solution.

14
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Definition: Any solution which is obtained from the general solution by giving
particular values to the arbitrary constants is called a particular solution.

For example, y = a cos 2x + b sin 2x, involving two arbitrary constants a and b, is
d2y
the general solution of the second order equation dx2
+ 4y = 0 (ref. Example 2)
whereas, y = 2 sin 2x + cos 2x is its particular solution (taking a = 1 and b = 2).

In some cases there may be further solutions of given equation which cannot be
obtained by assigning a definite value to the constant in the general solution. Such
a solution is called a singular solution of the equation. For example, the equation

y'2- xy’ + y = 0 …(16)

2 x2
has the general solution y = cx – c . A futher solution of Eqn. (16) is y = .
4
Since this solution cannot be obtained by assigning a definite value to c in the
general solution, it is a singular solution of Eqn. (16).

Thus, we have seen the various types of solution of an ordinary differential

equation. We have also seen that a solution of a differential equation may or may
not exist. Even if a solution exists, it may or may not be unique.

We now try to find the conditions under which the solution of a given ordinary
differential equation exists and is unique. Here, we shall confine our attention to
the first order ordinary differential equations only. Let us consider the general
first order equation.

dy
dx
= f(x, y) …(17)

In Eqn. (17) we assume that f is known to us. You may be surprised to know that,
though this equation looks simple, it is very difficult to get its explicit solution.
For clearity, let us look at the following examples.

Example 4: Does the solution y(x) of an ordinary differential equation

dy
dx
= f(x), where f(x) = 0 for x < 0
= 1 for x ≥ 0

exist ∀ x ∈

R?

15
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: The function defined by

 cforx < 0
Y(x) = 
 x + c for x ≥ 0

Satisfies this equation. at the same time this function has no derivative at x = 0,
because of the discontinuity of y(x) at x = 0.

Thus, this differential equation has no valid solution for x = 0.

However, y(x) defined above is the solution of the given differential equation at all
points other than x = 0.

Let us look at another example.

dy
Example 5: Does the equation dx
= - e-y x have a unique solution?

Solution: Rewrite the above equation in the form

dy
dx
(ey) = - x

Integrating, we get the solution of given equation as

x2
ey =- + A,
2
 x2 
or y = In  − + A
 2 

where A is an arbitrary constant.

You know that in x is defined for positive values of x only. So, the solution of the
 x2 
given differential equation will exists as long as  − + A  > 0. Clearly A > 0.
 2 
Also, for different values of A we get different solutions. Moreover these
solutions have different intervals of existence. Thus, the solution of a given
differential equation is not unique.

16
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

As regards the non-unique solutions, it is obvious that the cause for the non-
uniqueness is the arbitrariness of A, (but for A > 0). Thus, we would like to
impose some condition on the solution which might determine A. one such
condition is to specify the value of y at some point x0 where x0 is in the interval of
existence of y. such a condition is called initial condition and the problem of
solving a differential equation together with the initial conditions is called the
initial value problem (IVP). In other words, initial value problem is the problem
in which we look for the solution of a given differential equation which satisfies
certain conditions at a single of the independent variable. Thus, the first order
initial value problem s

dy 
= f (x, y), 
dx  …(18)
y(x 0 ) = y0 

From Example 4 and 5 mainly two questions arise:

1) Under what conditions does an initial value problem of the form (18) have
at least one solution?

2) Under what conditions does that problem have a unique solution, that is,
only one solution?

The above questions are answered by a theorem, known as Existence Uniqueness

Theorem 1. We shall now state this theorem for the first order differential
equation.

Theorem 1: (Existence – Uniqueness):

If f(x, y) is continuous at all points (x, y) in some rectangle (see Fig. 1.
Y
yo + b
R
yo

yo –b

xo-a xo xo+a x

Fig. 1: Rectangle R

17
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

R: |x – x0| < a, |y – y0| < b and bounded in R, say

|f(x, y)| ≤ k (x, y) in R.
∀
…(19)

then the IVP (18) has at least one solution y(x) defined for all x the interval
|x – x0| < h,

∂f
further, if is continuous for all (x, y) in R and bounded say,
∂y

∂f ≤
M, (x, y) in R
∀
…(20)
∂y
then the solution y(x) is the unique solution for all x in that interval |x – x0| < h,

Note: A function f(x, y) is aid to be bounded when (x, y) varies in a region in the
xy-plane and if there is a number k such that |f| k when (x, y) is in that region.
≤

For example.

F = x2 + y2 is bounded, with K = 2 if |x| < 1 and |y| < 1.

We shall not be proving this theorem. The proof of this theorem requires
familiarity with many other concepts which are beyond the scope of this course.
However, we which, we give some remarks which may be helpful for a good
understanding of the theorem.

Remark: Since y’ = f(, y), the condition (19) implies that |y’| ≤ k, that is, the slope
of any solution curve y(x) in R is at least – k and atmost k. Hence a solution
curve which passes through the point (x0, y) must lie in the shaded region in Fig. 2
bounded by the lines 11, and 12 whose slopes are – k and k, respectively
Y Y
11
yo+b 11 R
yo+b
yo yo R
yo-b
12
yo-b 12 h h
h=a h=a a a
x0 x x0 x
(a) (b)
Fig. 2

Now two different cases may arise, depending on the form of R.

18
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

b
i) We may have ≥ a. Therefore, h = a, which asserts that the solution exists
k
for all x between x0 – a and x0 + a (see Fig. 2 (a)).
b b
ii) We may have k
< a. Where, h = k
, and we concluded that the solution
b b
exists for all x between x0 - k
and x0 + k
. In this case, for larger or
smaller values of x; the solution curve may leave the rectangle R (see fig. 2
(b)). Since we have not assumed anything about f outside R, nothing can
be concluded about the solution for those corresponding value of x.

The condition stated in Theorem 1 are sufficient but not necessary and can be
relaxed. For example, by the mean value theorem of differential calculus, we have
(ref. Theorem 1).

∂f
f(x, y2) – f(x, y1) = (y2 – y1)
∂y

where (x, y1) and (x, y2) are assumed to be in R. From condition (20) then it
follows that

|f(x, y2) – f(x, y1) | ≤ M| (y2 –y1) |

condition (20) may be replaced by the condition (21) which is known as a

Lipschitx condition, named after the German mathematician, Rudolf Lipschitz
(1831 – 1903).

Thus, we can say that for the existence of the solution of the IVP (18), we must
have

i) f continuous in T.
ii) f bounded in T.

Further the solution is unique if in addition to (i) and (ii), we have

∂f
iii) ∂y continuous in T.
∂f
iv) ∂y bounded in T (or, Lipschitx condition)

19
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

However, if the above conditions do not hold, then the IVP (18) may still have
either (a) no solution (b) more than one solution (c) a unique solution.
This is because theorem provides only sufficient conditions and not necessary.
For instance, consider

dy
= 3y2/3, y (0) = 0.
dx

∂f 2 -1/3 ∂f ∂f
Here, f(x, y) = 3y2/3, ∂y = 3. y for y ≠ 0. ∂y does not exists at y = 0. so ∂y
3
is not bounded but, the solutions y = x3 and y = 0 exist.

Let us examine conditions (i) – (iv) for a few differential equations through
examples.

dy
Example 6: Examine dx
= y ,with y (0) = 1 for existence and uniqueness of the
solution.

Solution: Here f(x, y) =y, fy(x, y) = 1. Also x0 = 0 and y0 = 1.

In this case consider a rectangle T defined by

T: |x – 0 |<a, |y – 1)<b
Where a and b are positive numbers.

In any rectangle T (containing the point (0, 1)) the function f(x, y) is continuous
and bounded. Hence the solution exists. Further fy(x, y) is also continuous and
bounded in any such rectangle T. Therefore, the solution is unique.

You may verify that y = ex is a solution of the given equation satisfying the initial
condition y (0) = 1. Hence, it is the unique solution.

However, if the initial condition is changed to y(0) = 0 then rectangle T will be of

the form

T: |x – 0| < a, |y – 0| <b
And in that case y = 0 will be the unique solution for all x and y in any rectangle
T containing (0, 0).

20
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy
Example 7: Examine dx
= | y | when y (0) = 0, for existence and uniqueness of
solutions.

Solution: Here f(x, y) = | y | , x0 = 0 and y0 = 0. In this case consider the region

T with |x| < a, |y| < b, a and b positive numbers. Function f(x,y) is continuous and
bounded in any rectangle T, containing the point (0,0).

Hence solution exists. In order to test the uniqueness of the solution, consider the
Lipschitz condition.

| f(x, y 2 ) − f(x, y1 ) | | | y 2 | − | y1 | |
=
| y 2 − y1 | | y 2 − y1

for any region containing the line y = 0, Lipschitz condition is violated. Because
for y1 = 0 and y2 > 0, we have

| f(x, y 2 ) − f(x, y1 ) | y2 1
= = , ( y 2 > 0)
| y 2 − y1 | y2 y2

and this can be made as large as we please by choosing y2 sufficiently small,

whereas condition (21) requires that the quotient on the left-hand side of (21) does
not exceed a fixed constant M.

therefore, the solution is not unique.

Further, it can be checked that the given problem has the following solutins

i) y=0 ∀x
 1 2
 4 x for x ≥ 0
ii) y=  1
 − x 2for x ≤ 0
 4

dy  y(1 − 2 x) for x > 0

Example 8: Examine = f(x, y) =  with y(1) = 1
 y(2 x − 1) for x < 0
dx

for existence and uniqueness of solution.

Solution: Here x0 = 1 and y0 = 1. Rectangle T can be any rectangle containing

point (1, 1). You may note that the function is not defined at x = 0. It is

21
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

discontinuous at x = 0. Thus, at x = 0 the solution does not exist. At all other

points the function

 y(1 − 2 x) for x > 0

f(x, y) = 
 y(2 x − 1) for x < 0
is continuous and bounded in T with (x) = 1. Hence, the solution, exists and is
unique for all x other than x = 0. further, you may verify that
2
y = x x − x for x > = 0
2
and y = e x − x for x < 0

is the unique solution of the given problem for all x other than x = 0

you may now try the following exercise.

From the definitions given in page 14, you may have realized that the general
solution of a first order differential equation normally contains one arbitrary
constant which is called a parameter. When this paratmeter is various values, we
obtain a one parameter family of curves. Each of these curves is a particular
solution or integral curve of the given differential equation, and all of them
together constitute its general solution. On the other hand, we expect that the
curve of any one-parameter family are integral curves of some first order
differential equation. In general we pose a problem: given an n-parameter family
of curves, can thus say that differential equations arise from a family of curves. In
the next section we shall take up this.

3.3 Family of Curves and Differential Equations

Let us consider a family of straight lines

Y = mx + c …(22)

Which is a two-parameter family of curves, parameters being m and c.

It is clear from Eqn. (22) that y can be treated as a function of x, x ∈ R .

Differentiating Eqn. (22) w.r.t.x, we have

y' = m …(23)

Again, differentiating (23) we get

y" = 0 …(24)

22
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Equation (23) and (24) are differential equations of order one and two
respectively. The way in which we have arrived at Eqn. (23) or (24) is clear. We
have actually tried to eliminate the parameters, or constants, m and c and the result
is Eqn. (23) or (24).

In general, we represent one-parameter family of curves by an equation

F(x, y, a) = 0 …(25)
where a is a constant.

In Eqn. (25), let us regard y as a function of x and differentiate if w.r.t.x. Suppose

we get
G(x, y, y’, a) = 0 …(26)

In case, we are able to eliminate the constant a between Eqn. (25) and (26), then
we have a relation connecting x, y and y', say

h (x, y, y',) = 0 …(27)

Equation (27) is an ODE of order one. In particular, if Eqn. (25) has the form

ψ (x, y) = a ...(28)
then the elimination of the constant a from Eqn. (28) leads us to the differential
equation
∂ψ ∂ψ
+ y' = 0 …(29)
∂x ∂y
for example, x2 + y2 = a2 …(30)
is the equation of the family of all concentric circles with centre at the origin (fig.
3)

Fig. 3

23
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

For different values of a , we get different circles of the family. Differentiating

Eqn. (30) with respect to x, we get
dy
2x + 2y dx
= 0.
dy
Or x+y dx
= 0, as the differential equation of the given family of circles.

Continuing with equation y = mx + c, if we regard only c as an arbitrary constant

to be eliminated, then y' = m, represents the required differential equation.
Geometrically, for a fixed m, y' = m represents a family or straight lines(in the
plane) whose slope is m (see fig. 4).

=1 =1 =1
M M M =1
M
=1
M
45o 45o 45o 45o 45o
O

Fig. 4

On the other hand, if we assume that in equation y = mx + c both m and c are

constants to be eliminated, then equation y" = 0 represents the required differential
equation. Geometrically, it is the family of the straight lines in the plane (see
Fig. 5)
Y
C =0
= 1,
2
=-

M
,C
=1
M

45o 45o
(-2, 0) O (1, 0) x
M
=1
,C
=1

Fig. 5

24
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

You may now try the following exercise.

In the introduction of this unit we mentioned that there are many problems of
physical and engineering interest which give rise to differential equations. In other
words, we can say that some problems have representations through the use of
differential equations. In the next section we shall take up such problems.

3.4 Differential Equations Arising From Physical Situations

In this section we shall show that differential equations arise not only out of
consideration of families of geometric curves, but an attempt to describe physical
problems, in mathematical terms, also result in differential equations.

The initial-value problem

dy 
= ky 
dx  …(31)
y(t 0 ) = t 0 
where k is a constant of proportionality, occurs in many physical theories
involving either grow or decay. For example, in biology it is often observed that
the rate at which certain bacteria grow is proportional to the number of bacterial
present at any time. In physics an IVP such as Eqn. (31) provides a model for
approximating the remaining amount of a substance that is disintegrating, or
decaying, through radioactivity. The differential Eqn. (31) could also determine
the temperature of a cooling body. In chemistry, the amount of a substance
remaining during certain reactions is also described by Eqn. (31).

Let us now see the formulation of some of these problems.

I: Population Model

Let N(t) = denote the number or amount of a certain species at time t. then the
d
growth of N(t) is given by its derivative N(t). Thus, if N(t) is growing at a
dt
d
constant rate, dt
N(t) = β , a constant. It is sometimes more appropriate to
consider the relative rate of growth defined by

actual rate of growth N' (t) dN(t) /dt

relative rate of growth = = =
size of N(t) N(t) N(t)

25
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

The relative rate of growth indicates the percentage increase in N(t) or decrease in
N(t). For example, an increase of 100 individuals for a species with a population
size of 500 would probably have a significant impact being an increase of 20
percent. On the other hand, if the population were 1,000,000 then the addition of
100 would hardly be noticed, being an increase of 0.01 percent. If we assume that
the rate of change of N at time t is proportional to population N(t), present at the
time t then,

d
N(t) ∝ N(t)
dt
which is written as

d
dt
N(t) = k N(t), …(32)

where k is a constant

if N increases with t, then k > 0 in Eqn. (32)

If N decreases with t, k ≤ 0 in Eqn. (32).

Normally, we have the knowledge of the population, say N0, at some initial time t0.
so along with Eqn. (32) we have

N(t0) = N0. …(33)

Thus, the population N(t) at time t can be found by solving Eqn. (32) with
condition (33). We shall reconsider this problem with some modifications later in
unit 3.

II: Newton’s Law of Cooling

Here we deal with the temperature variations of a hot object kept in a surrounding
which is kept at a constant temperature, say T0. Under certain conditions, a good
approximation to the temperature of an object can be obtained by using Newton’s
law of cooling. Let the temperature of the object be T. If T ≥ T0, we know that the
object radiates heat to the surrounding resulting in the reduction of its (object’s)
temperature. Newton’s law of cooling states that the rate at which the temperature
T(t) changes in a cooling body is proportional to the difference between the
temperature of the body and the constant temperature T0 of the surrounding
medium. That is

26
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

d
dt
T(t) ∝

T(t) – T0
d
or dt
T(t) = k (T(t) – T0), …(34)

where k is a constant of proportionality.

Constant k < 0, because the temperature of the body is reducing (we have assume
that T(t) T0). We observe that that Eqn. (34) is a differential equation of order
≥

one.

III: Radioactive Decay

Many substances are radioactive. This means that the atoms of such a substance
break down into atoms of other substances. In Physic, is has been noticed that the
radioactive material, at time t, decays at rate proportional to its amount y(t). in
other words,

d
dt
y(t) = ky(t) …(35)

where k < 0, is a constant. If the mass of the substance at some initial time, say t =
0, is A, then y(t) also satisfies the initial condition

y(0) = A.

Thus, the physical problem of radioactive decay is modeled by the IVP.

d
dt
y(t) = ky(t), Y(0) = A …(36)

where k is a constant.

Remark: I, II and III above indicate situations where differential equations occur
naturally. In unit 3 we shall give the methods of solving these equations.

You may now try the following exercise.

4.0 CONCLUSION

We now conclude this unit by giving a summary of what we have covered in it.

27
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

5.0 SUMMARY

In this unit, we have covered the following points:

1) a) An equation involving one (or more) dependent variables and

its derivative w.r. to one or more independent variables is called a
differential equations.

b) A differential equation involving only ordinary derivatives is called

an ODE.

c) A differential equation involving partial derivatives is called a

partial differential equation (PDE).

d) The order of a differential equation is the order of the highest order

derivative appearing in the equation.

e) The degree of a differential equation is the highest exponent of the

highest order derivative appearing in it after the equation has been
expressed in the form free from radicals and fractions of the
derivatives.

f) In a differential equation, when the dependent variable and its

derivatives occur in the first degree only, and not as higher powers
or products, the equation is said to be linear.

g) If an ordinary differential equation is not linear, it is said to be non-

linear.

2) a) A real or complex value function = φ (x) defined on an interval I is

called a solution of equation
g(x, y, y', y", ….., y(n)) = 0
if φ (x) is differentiable n times and if φ (x), φ ' (x), φ " (x), ….,(x)
satisfy the above equation for all x in I.
b) The solution of the nth order differential equation which contains n
arbitrary constants is called its general solution.

c) Any solution which is obtained from the general solution by giving

particular values to the arbitrary constants is called a particular
solution of the differential equation.

28
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

d) A solution of a differential equation, which cannot be obtained by

assigning definite values to the arbitrary constants in the general
solution is called its singular solution.

3) a) Conditions on the value of the dependent variable, and its

derivatives, at a single value of the independent variable in the
interval of existence of the solution are called the initial conditions.

b) The problem of solving a differential equation together with the

initial conditions is called the initial value problem.

4) The sufficient conditions for the existence of solution of the first order
equation

dy
= f(x, y), with y(x0) = y0,
dx
in a region T defined by |x – x0| < a and |y – y0| < b are

i) f is continuous in T
and
ii) f is bounded in T.

further if the solution exists, then it is unique if, in addition to (i) and (ii),
we have
∂f
(iii) is continuous in T.
∂y
∂f
iv) is bounded in T (or, Lipschitz condition is satisfied).
∂y

5) The general solution of a first order (nth order) differential equation

represents one-parameter (nm-parameter) family of curves.

6) Many physical situation such as population model, Newton’s law of

cooling, radioactive decay, can be represented by first order differential
equations.

6.0 TUTOR MARKED ASSIGNMENT

1. Which of the following are differential equations? Which of the

differential equations are ordinary and which are partial?

29
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

3
 d2 y  dy
a)  dx 2  + x dx + y = 5x + 2
3

 

x
dy
b) = ∫ sin[xy(s)]ds
dx x

∂ 2u ∂ 2u ∂ 2u
c) + + =0
∂ x 2 ∂ y2 ∂ z 2

dy(x)
d) = 5x y (x + 1)
dx
x
d2 y dy
e) + + y = ∫ sin[xy(s)]ds
dx 2 dx x

∂ 2w 2 ∂ w
2

f) = a
∂t2 ∂x2

2. Find the order and degree of the following differential equations.

2/3
 d2 y  dy
a)  dx 2  = 1 + 2 dx
 

3
d 2 y  dy 
b) +   + y2 = x
dx 2  dx 

 d2 y 
c) sin  2  + x2y2 = 0
 dx 

dy
d) + y3 = 0
dx
3/ 2
  dy  2  d2 y
e) 1 +    =r
  dx   dx 2

2
∂ 4z  ∂ 2z 
f) +  =x
∂x4  ∂x ∂y 

30
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

g) x2 (dx)2 + 2xy dx dy + y2(dy)2 – z2(dz)2 = 0

3. Classify the following differential equations into linear and non-

linear.

∂z ∂z
a) x +y -z=0
∂x ∂y

2
∂ 2z ∂ z  ∂ 2z 
2

b) 2 -   =0
∂x2 ∂y  ∂x ∂y 

dy
c) = (x + y)2
dx

d2 y
2 dy
d) (1 – x ) - 2x + n(n + 1) y = 0
dx 2
dx

d2 y µ
e) (x2 + y2)3/2 + x=0
dx 2

 x2 
 − 2  , and 2cos y = - x are solutions of the
-1 2
4. Verify that y = cos
 
dy
equation sin y dx
= x. Can you state the interval on which y is defined?

1 dy y2
5. Verify that y = (In y + c) is a solution of the equation = for
x dx 1 − xy
every value of the constant c.

6. Verify that y = e2x and y = e3x are both solutions of the second order
equation y” – 5y’ + 6y = 0
Can you find any other solution?

 4 x3y
, when x and y are not both zero
= f(x, y) =  (x 4 + y 2 )
dy
7. Examine dx
 0, when x = y = 0
With y (0) = 0

31
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

For existence and uniqueness of the solution.

8. Assuming y to be a function of x, determine the differential equations by

Eliminating the arbitrary constant (or constants) indicated in the following
problems.

a) xy = c (arbitrary constant is c)
b) y = cos (ax) (arbitrary constant is a).
c) y = A cos (ax) (arbitrary constants are A and a).

9. In the following problems derive the differential equation describing the

given physical situations.

a) A culture initially has P0 number of bacteria. Growth of the bacteria

is proportional to the number of bacteria present. What is the
number p of bacteria at any time t.
b) A quantity of a radioactive substance originally weighing x0 gms
decomposes at a rate proportional to the amount present and half the
original quantity is left after 2 years. Find the amount x of the
substance remaining after t years.

7.0 REFERENCES/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

Indira Gandhi National Open University School Of Sciences Mth-07

32
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

UNIT 2 METHODS OF SOLVING EQUATION OF FIRST

ORDER AND FIRST DEGREE

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Separation of Variable
3.2 Homogenous Equation
3.3 Exact Equation
3.4 Integrating Factor
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 Reference/Further Readings

1.0 INTRODUCTION

In unit 1, we introduced the basic concepts and definitions involved in the study of
differential equations. We discussed various types of solutions of an ordinary
differential equation. We also stated the conditions for the existence and
uniqueness of the solution of the first order ordinary differential equation.
However, we do not seem to have paid any attention to the methods of finding
these solutions. Accordingly, in this unit we shall confine our attention to the
same.

In general, it may not be feasible to solve even the apparently simple equation
dy dy
= f(x, y) or g(x, y, ) = 0 where f and g are arbitrary functions. This is
dx dx
because no systematic; procedure exists for obtaining its solution for arbitrary
forms of f and g. However, there are certain standard types of first order equations
for which methods of solution are available. In this unit we shall discuss a few of
them with special reference to their applications.

2.0 OBJECTIVES

After studying this unit, you should be able to

• Define separable equations and solve them;

33
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

• Define homogeneous equations and solve them;

• Obtain the solution of equations which are reducible to homogenous equations;
• identify exact equations;
• Obtain an integrating factor which may reduce a given differential equation
into an exact one and eventually provide its solution.

3.0 MAIN CONTENT

3.1 Separation of Variables

You know that the problem of finding the tangent to a given curve at a point was
solved by Leibniz. The search for the solution to the inverse problem of tangents,
that is, given the equation of the tangent to a curve at any point to find the
equation of the curve led Leibniz to many important developments. A particular
mention may be made of the method of separation of variables which was
discovered by Leibniz in 1691 by providing that a differential equation of the form
dy
= X (x) Y (y)
dx
is integrable quadratures. However, it is John Bernoulli (1694) who is credited
with the introduction of the terminology and the process of separation of variables.

In short, it is a method for solving a class of differential equations that arises quite
frequently and is defined as follows:

Note: The process of finding the areas of plane regions if called quadrature

Definition: An equation of the form

dy
= f(x, y) …(1)
dx
is called a separable equation or equation in variable separable form if f(x, y)
can be put in the form

f(x, y) = X(x) Y(y), …(2)

where X and Y are given functions of x and y respectively.

In other words, Eqn. (1) is a separable equation if f is a product of two functions,

one of which is a function of x and the other is a function of y. Here X (x) and
Y(y) are real value functions of x and y respectively.

34
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy
For instance, equation = ex+y is a separable equation, since ex+y = ex.ey (here
dx
dy
X(x) = ex and Y(y) = ey). The equation = x2(y2 + y3) is also a separable
dx
dy
equation. But the equation = exy is not a separable equation, because it is not
dx
possible to express exy as a product of two functions in which one is a function of x
dy
only and the other is a function of only. Similarly, equation = x + y is not a
dx
separable equation.

In order to solve Eqn. (1), when it is in variable separable form, we write it as

dy
a(y) + b (x) = 0 …(3)
dx
where a(y) and b(x) are each functions of only one variable

let us assume that there exist functions A and B such that A’(y) = a(y) and B’(x) =
b(x). With this hypothesis, Eqn. (3) can be rewritten as
d
A (y(x)) + B’(x) = 0 …(43)
dx

d dy dy
[by chain rule A (y(x)) = A’ (y(x)) = a (y(x)) ]
dx dx dx

Integrating Eqn. (4) with respect to x, we get

A (y(x)) + B(x) = c …(5)

Where c is a constant.

Thus, any solution y of (3) is implicitly given by (5).

We now take up a few examples to illustrate this method.

dy
Example 1: Solve = ex-y
dx

Solution: This equation may be written as

dy
= ex e-y
dx

35
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy
or ey = ex
dx
ord
(ey) = ex
dx
which, on integration, gives ey = ex+ c, where c is a constant.

In case ex + c ≥ 0, then y(x) = In (ex + c).

Example 2: Solve the equation

(1 + y2) dx + (1 + x2) dy = 0 with y (0) = - 1.

Solution: The given equation can be rewritten as

dx dy
+ = 0.
1+ x 2
1 + x2

Integrating, we get

Tan-1x + tan-1y = c.

The initial condition that y = -1 when x = 0 permits us to determine the value of c

that must be used to obtain the particular solution desired here. Since tan -10 = 0
π π
and tan-1(- 1) = - , c = 0 - . Thus, the solution of the initial value problem is
4 4
π
tan-1x + tan-1y = - .
4

Let us look at another example.

dy
Example 3: Solve = ky – my2, where k > 0 and m > 0 are real constants.
dx

Solution: Let us write the given equation as

1 dy
= 1.
ky − my dx
2

dy
Now we will try do decompose the coefficient of into partial fractions.
dx

36
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1 A B
Let = + …(6)
y(k − my) y k − my
Where A and B are constants to be determined.

From (6), we get

1 = A (k – my) + By
Bk
which gives 1 = Ak and 1 =
m

1 m
or A = and B =
k k

1 1 m 1
Hence = +
y(k − my) ky k (k − my)
Thus the given differential equation can be rewritten as
 1 m 1  dy
 + .  = 1, …(7)
 ky k k − my  dx
for y ≠ 0 and k-my ≠ 0.

In the integration of Eqn. (7) the sign of y and k –my play a important role. We
now discuss the following possible cases:
k
Case 1: y > 0 and k-my > 0 (0 < y < ).
m

For the case under consideration, Eqn. (7) can be exdpressed as

d  1  1 
  Iny  − In(k − my)  = 1
dx   k  k 
which on integrating, yields
1 1
INy - In (k-my) = x + c,
k k
where c is a constant of integration. The above equation can be further express as
In (y)1/k – In (k – my)1/k = x + c
1/ k
 y 
or   = exp(x + c)
 k − my 

Case II: y < 0

When y < 0 then k – my > 0 because m > 0,. In this case, Eqn. (7) on integration
can be written as

37
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1 1
- In (- y) ± In (k – my) = x + c
k k

1 x +c
or ± 1/ k = e
(− y) (k − my)
1/ k

k
Cases III: y> 0 and k – my < 0 (y > ).
m
In this case Eqn. (7) after integration gives
1 1
In(y) - In( - k + my) = x + c
k k
1/ k
 y 
or   = ex +c
 − k + my 
You may now try the following exercises.

Many differential equations that are not separable can be reduced to the separable
form by a suitable substitution. In the next section we shall study one classof such
equations.

3.2 Homogeneous Equations

In this section we shall study equations like

dy x 2 + xy + y 2
=
dx 3x 2 + y 2

This is an example of homogenous differential equations. In 16921 Leibniz made

known to the world the method of solving homogeneous equation differential
equations of the first order.

Before we discuss the method of solving a homogeneous equation, we define

homogeneuous functions of two variables x and y.

Definition: A real-value function h(x, y) of two variables x and y is called a

homogeneous function of degree n, where n is a real number, if we have

h( λ x, λ y ) = λ nh(x, y)

for ∀ x, y and any constant λ > 0.

38
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

For example, h(x,y) = x3 + 2x2y + 3xy2 + 4y3 is homogeneous of degree three

because h( λ x, λ y ) = λ 3h(x, y)

 y
 
 x  + (In|x| - In|y|) xy is homogeneous function of degree 2
Also, h(x,y) = x2 cos
x2
and 2 is homogeneous of degree 0.
x + 2xy + y 2

But, the function h(x, y) = x2 + 2xy + 4 is not homogeneous because

h( λ x, λ y ) ≠ λ n (x2 + xy + 4) for any value of n.

if h(x, y) is a homogeneous function of degree n, that is, h( λ x, λ y) = λ nh(x,y),

1 1
then a useful relation is obtained by letting λ = . This gives n h(x,y) = h
x x
 y  y  y
 1,  = ∅   (say) or , h(x,y) = x ∅   .
n

 x  x  x

We shall be particularly interested in the case where h(x,y) is homogeneous of

degree 0 that is, if h( λ x, λ y) = λ 0h(x,y) = h(x,y). We now give the following
definition.

Definition: A differential equation

y' = f(x,y) …(8)

is called a homogenous differential equation when f is a homogeneous function

of degree 0.

For instance, the following equations are homogeneous differential equations:

dy 2y
i) = ,
dx x
dy 2x + 3y 2 + 3(y / x)
ii) = =
dx 4x 4
dy x + x y + y 1 + (y / x) + (y / x)3
3 2 3

iii) = =
dx 3x 2 y + y3 3(y / x) + (y / x)3

from the above equations, you may have noticed that if an equation can be put in
the form

39
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy f1 (x, y)
= f(x, y) = ,
dx f 2 (x, y)
where f1 and f2 are homogeneous expressions of the same degree in x and y, then f
is a homogeneous function of degree 0.

1
Further, if in Eqn. (8) we let λ = , the we have
x

 y  y
y' = f  1,  = F   …(9)
 x  x

y
This suggests making the substitution v = to solve this equation. since we seek
x
y as a function of x this substitution means

y(x)
V(x) = or y(x) = xv(x)
x
dy dv
and =v+x
dx dx
with this substitution Eqn, (9) reduces to
dv
v+x = f(v)
dx
dv F(v) − v
or = …(10)
dx x

which shows that Eqn. (10) is a separable equation in v and x. if we can solve
Eqn. (10) for v in term of x, using the technique of Sec. 2.2 then the solution of
Eqn. (9) is y = vx and hence we can solve equations of the type (9).

We now illustrate this method with the help of the following examples.

dy 2y 2 + 3xy
Example 4: Solve =
dx x2

Solution: You can easily check that the given equation is homogeneous of degree
0. it can be rewritten as
2
dy  y  y
=2   +3   …(11)
dx  x  x

40
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

y
By making the substitution, v = , Eqn. (11) reduces to
x
dv
x + v = 2v2 + 3v
dx
dv
or x = 2v2 + 2v = 2v(v + 1)
dx
dv 2dx
or =
v(v + 1) x
which is in variable separable form.

1
Resolving into partial fractions, we have
v(v + 1)
1 1  2
 −  dv = dx
 v v + 1 x
which on integration, gives
In |v| - In|v+1| = Inx2 + In |c| …(12)

Where c is an arbitrary constant.

From Eqn. (12), we have

v
= cx2
v+1

y
replacing v by , we get
x

y/x Y
cx2 = =
(y / x) + 1 X + Y
cx 3
or y = ,
1 − cx 2
which is the general solution of the given eqution.

dy y3 y
Example 5: Solve = + , x > 0.
dx x 3 x

Solution: With the substitution y = vx, we have

41
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dv
V+x = v3 + v
dx
1 dv 1
or 3 = …(13)
v dx x

Integration of Eqn. (13) yields

1
- = Inx + In |c|,
2v 2

y
where c is a real constant. On replacing v = , the general solution of the given
x
equation can be expressed as

x2 x2 1
y2 = - 2[Inx + In | c |] or
y2 = - 2 In(x | c |)

Let us consider another example.

Example 6: Solve (x2 + y2) dx – 2xy dy = 0.

Solution: The given equation can be written as

dy x 2 + y2
= …(14)
dx 2xy

Putting y = vx, in Eqn. (14), we get

dv 1 + v 2
V+x = ,
dx 2v
dv 1 + v 2
or x = -v
dx 2v
1 + v 2 − 2v 2 1 − v 2
= =
2v 2v
2v dv 1
or =
1 − v 2 dx x
Integrating, we get,

In |x| |(1 – v2)| = In |c|, where c is a constant of ingration or

X (1- v2) = c

42
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

On substituting for v, we can write the solution of Eqn. (14) in the form;
x2 – y2 = cx.

How about trying some exercise now?

Sometimes it may happen that a given equation is not homogeneous but cann be
reduced to a homogeneous form by considering a transformation of the variables.
We now consider such equations.

Equations reducible to homogeneous form

Consider the wquation of the form

dy ax + by + c
= …(15)
dx a 'x + b'y + c'
where a, b, c, a', b' and c' are all constants.

Eqn. (15) can be reduced to a homogeneous form by using the substitution

X = x' + h and y = y' + k,

Where h and k are constants to be so chosen as to make the given equation

homogeneous. In terms of these new variables, Eqn. (15) becomes

dy dy ' ax '+ by '+ (ah + bk + c

= = , …(16)
dx dx ' a ' x '+ b ' y '+ (a 'h + b 'k + c ')

which will be homogeneous provided h and k are so chosen that

ah + bk + c = 0
a'h + b'k + c' = 0 …(17)

Consequently Eqn. (16) reduces to

dy ' ax '+ by '

= …(18)
dx ' a 'x '+ b ' y '

which can be solved by means of the substitution y' = vx'.

If the solution of the Eqn. (18) is of the form

g(x', y') = 0,

then the solution of Eqn. (15) is

43
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

g(x – h, y – k) = 0,

where h and k are obtained by solving the simultaneous Eqns. (17)

Solving Eqns. (17) for h and k, we get,

bc '− bc ' a 'c − ac '

h= ,k=
ab '− a 'b ab '− a 'b

which are defined except when

a b
ab' – a'b = 0 that is, when = .
a ' b'
a b
If = , then h and k have either infinite values or are indeterminate. But then
a ' b'
a b
the question is what happens if = ?
a ' b'

a b 1
In such cases, we let = = (say)
a ' b' m

Then Eqn. (15) can be written as

dy ax + by + c
= …(19)
dx m(ax + ny) + c '

On putting ax to by = v, Eqn (19) reduces to

1  dv  v+ c
 − a =
b  dx  mv + c '
so that the variables are separated and hence the equation can be solved by the
method given in the Sec 2.2.

We now take up some examples to illustrate the above discussion.

dy y− x+1
Example 7: Solve = …(20)
dx y+ x+ 5

Solution: Comparing the given equation with Eqn. (15), we have

a = 1, b = -1, a' = 1, b' = 1.

44
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

a b a b
∴ = 1, = − |1and ≠
a' b' a ' b'
Putting x = x' + h and y = y' + k in Eqn. (20), we get

dy ' y '− x '+ k − h + 1

= …(21)
dx ' y 'x '+ k + h + 5
we choose h and k such that

k- h + 1 = 0
k+h+5=0 …(22)

on solving Eqn. (22), we get h = - 2 and k = -3 . with these values of h and k, Eqn.
(21) reduces to

dy ' y '− x '

= …(23)
dx ' y '+ x '

which is a homogeneous equation.

on putting y' = vx' in Eqn. (23) and simplifying the resulting equation, we get

1 + v dv 1
− =
1 + v 2 dx ' x '
 1 v  dv 1
or  + 2 
=− …(24)
 1 + v 1 + v  dx '
2
x'

Integration of Eqn. (24) yields

1
Tan-1v + In (1+ v2 ) = - Inx' + c, where c is a constant.
2
1
or, In (1 + v2) x'2 + tan-1v = c.
2

y'
Replacing v by , we have
x'
1 y'
In (x'2 + y'2) + tan-1 = c.
2 x'

substituting x' = x + 2 and y' = y + 3, solution of Eqn. (20) is given by

45
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1  y+ 3
In  (x + 2) + (y + 3)  + tan-1 
2 2
 = c.
2  x+ 2

Example 8: Solve the differential equation (4x + 6y + 5) dy = (3y + 2x + 5) dx.

Solution: The given equation can be written as

dy 3y + 2x + 5
=
dx 4x + 6y + 5

(2x + 3y) + 5
= …
2(2x + 3y) + 5
(250
In this case a = 2, b= 3, a' = 4, b' = 6. Thus,

a b
= . Therefore, we put 2x + 3y = v, and Eqn. (25) reduces to,
a ' b'

1  dv  v+ 5  dy dv 
 − 2 =  here 2 + 3 = 
3  dx  2v + 5  dx dx 
dv 3(v + 5) 3v + 15 + 4v + 10 7v + 25
or = +2 = =
dx 2v + 5 2v + 5 2v + 5

Now variables are separated and we get

2v + 5 dv
=1
7v + 25 dx
2 15  dv
or  −  = 1.
 7 7(7v + 25)  dx

Integrating, we get

2 15  25 
v- In  v +  = x + c, where c is a constant of integration, substituting
7 49  7 
v = 2x+ 3y, we get
2 15  25 
(2x + 3y) - In  2x + 3x +  = x + c,
7 49  7 
25
or, 14(2x + 3y) – 15 In (2x + 3y + ) = 49 (x + c)
7

46
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

or, 42y – 21x – 15 In (14x + 21y + 25) = 49c – 15 In 7 = c 1, say, which is the
required solution.

You may now try the following exercise. In each of the equations in this exercise
a b
you should first see whether = and then decide on the method.
a ' b'

In Unit 1, we defined the total differential of a given function. In the next section
we shall make use of this to define and solve exact differential equations.

3.3 Exact Equations

Let us start with a family of curves h(x, y) = c. Then its differential equation can
be written in terms of its total differential as

dh = 0, or
∂h ∂h
dx + dy = 0.
∂x ∂y

For example, the family x2y3 = c has 2xy3dx + 3x2y3dy = 0 as its differential
equation. suppose we now consider the reverse situation and begin with the
differential equation

a(x, y)dy + b(x,y)dx = 0

If there exists a function h(x,y) such that

∂h ∂h
= b(x, y) and = a(x, y),
∂x ∂y
then Eqn. (26) can be written in the form

∂h ∂h
dx + dy = 0 or dh = 0
∂x ∂y
that is, h(x, y) = constant represents a solution of Eqn. (26).

In this case we call the expression a(x, y)dy + b(x, y)dx an exact differential and
(26) is called an exact differential equation. For instance, equation x2y3 dx + x
1 3 3
+3y+2 dy = 0 is exact, since we have d  x y  = x2y3 dx + x+3y2dy.
3 
Thus, an exact differential equation is formed by equating an exact differential to
zero.

47
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

It is sometimes possible to determine exactness and find the function h by merer

inspection. Consider, for example, the equations.

3x2y4dx + 4x3y3 dy = 0
and xexy dy + (yexy – 2x) dx = 0.

These two equations can be alternatively written as d(x 3y4) = 0 and d(exy – x2) = 0,
respectively. Thus, the general solution of these equations are give by x3y4 = c and
exy = x2 + c, where c is constant.

However except for some cases, this technique of “solution by insight” is clearly
impractical. Consequently we seek an answer to the following question: when
does a function h(x, y) exist such that Eqn. (26) is exact? An answer to this
question is given by the following theorem.

∂a ∂b
Theorem 1: If the functions a(x, y), b(x, y), ax and by = are continuous
∂x ∂y
functions of x and y, then Eqn. (26), namely,

a(x, y) dy + b(x, y) dx = 0 is exact if and only if

∂ ∂
b(x, y) = a(x, y) …(27)
∂y ∂x

Indeed condition (27) is a necessary and sufficient condition for a function h(x, y)
to be such that
∂ ∂
h(x, y) = b(x, y) and h(x, y) = a(x, y) …(28)
∂y ∂y

you may note here that if relation (28) is satisfied, then

∂ ∂
d[h(x, y (x))] = h(x,y (x)) dx + h(x,y(x)) dy
∂x ∂y
= b(x, y(x))dx + a(x, y(x)) dy
and hence Eqn. (26) can be rewritten as
d[h(x, y(x))] = 0
or that the solution of Eqn. (26) is given by
h (x, y) = c,
where c is a constant.

We now give the proof of Theorem 1.

Proof: The condition is necessary

48
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Let the equation

dy
A(x, y) + b(x, y) = 0
dx
Or, a(x, y) dy + b(x, y) dx = 0
Be exact.

Then there exists a function h(x, y) such that

Dh = b(x, y) dx + a(x, y) dy

∂h ∂h
But dh = dx + dy.
∂x ∂y
∂h ∂h
∴ dx + dy = b(x, y) dx + a(x, y) dy
∂x ∂y
Thus, necessarily,
∂h ∂h
= b(x, y) and = a(x, y) …(29)
∂x ∂y
Since a(x, y) and b(x, y) have continuous first order partial derivatives, h possess
∂ 2h ∂ 2h
continuous second order partial derivatives namely, and . Refer Unit
∂ y∂ x ∂ y∂ y
6, Block 2 of MTE-07 for second and higher order partial derivatives.
Now,

∂  ∂ h  ∂ 2h ∂b
 = = (x, y) …(30)
∂ y  ∂ x  ∂ y∂ x ∂ y
∂  ∂h  ∂ 2h ∂a
and   = = (x, y) …(31)
∂y  ∂y  ∂ x∂ y ∂ x
∂b ∂a
Since and are continuous,
∂y ∂x
∂ 2h ∂ 2h
= (ref. Unit 6, Block 2 of MTE-07).
∂ y∂ x ∂ x∂ y
There, from Eqn. (30) and (31), we get
∂a ∂b
(x, y) = (x, y)
∂x ∂y
The condition is sufficient: Now suppose that
∂a ∂b
(x, y) = (x, y)
∂x ∂y
and we shall show that a(x, y) dy + b(x, y) dx is an exact differential

49
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

∂V
Let ∫ b(x, y) = dx = V, then = b(x, y)
∂x
and
∂ 2V ∂b
= (x, y)
∂ y∂ x ∂y
∂a
= (x, y) (using given condition)
∂x
∂a ∂  ∂V 
∴ (x, y) =   …(32)
∂x ∂y  ∂x 

Integration of Eqn. (32) with respect to x, holding y fixed, yields

∂V
a(x, y) = + ø (y)
∂y

where ø, a function of y only, is therefore a constant of integration, when y is held

fixed.

Thus,

∂V ∂V
a(x, y) dy + b(x, y) dx = dy + ø(y) dy + dx
∂y ∂x
= d[V(x, y) + ø(y) dy]

which establishes that a(x, y) dy + b(x, y) dx is an exact differential implying

thereby that, a(x, y) dy + b(x, y)dx = 0 is an exact differential equation. This
completes the proof of Theorem 1.

We shall now illustrate this theorem with the help of the following examples.

Example 9: Solve the differential equation

Sin(y) + x cos (y)y' = 0.

Solution: For the case under consideration, a(x, y) = xcos(y) and b(x, y) = sin(y).
also

∂ ∂
a(x, y) = cos(y) = b(x, y)
∂x ∂y
which shows that the given equation is an exact equation.

50
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

∂h ∂h
Therefore, there exists a function h(x, y) = constant such that, = b(x, y) and
∂x ∂y
= a(x, y)

Then we have
∂h
∂x
= sin y …(33)
and
∂h
= x cos y …(34)
∂y

Integrating Eqn. (33) with respect to x, treating y as a constant, we get

H(x, y) = x sin y + ø (y) …(35)

Where ø(y) is a constant of integration. Differentiating Eqn. (35) partially w.r.t.y,

we get

∂
∂y h(x, y) = x cos y + ø'(y) …(36)

from Eqns. (34) and (36), we get

x cos y = x cos y + ø'(y)
which shows that ø'(y) = 0 ⇒ ø(y) = constant = c1. Hence from Eqn. (35), we can
write

h(x, y) = xsin y + c1
so the required solution, h(x, y) = constant, is
x sin y + c1 = c2, where c2 is a constant or,
x sin y = c,

where c = c2 – c1 is a constant.

Example 10: Solve ex sin y + ex cosy y' + 2x = 0.

Solution: Comparing with Eqn. (26), we have a(x, y) = excos y and b(x, y) = exsin
y + 2x. Therefore,

∂
∂x
a(x, y) = excos y
∂ ∂
and ∂y b(x, y) = excos y = ∂x
a(x, y).

51
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Hence the given equation is exact and can be written in the form dh(x, y(x) = 0
where
∂
∂x
h(x, y) = excos y + 2x …(37)
∂
and ∂y h(x, y) = ex cos y …(38)

Integrating Eqn. (37) w.r.t.x, we get

h(x, y) = ex sin y + x2 + ø(y) …(39)

Where ø, a function of y only, is a constant of integration

From Eqns. (38) and (39), we get

∂
∂y h(x, y) = excos y + ø' (y) = ex cos y

So we have ø' (y) = 0 or ø (y) = c1 where c1 is a constant.

Hence from Eqn. (39) we have the required solution as

h(x, y) = ex sin y + x2 = - c1 = c

where c is a constant.

On the basis of Theorem 1 and Example (9) and (10) we can say that various steps
involved in solving an exact differential equation b(x, y) dx + a(x, y) dy = 0 are as
follows:

Step 1: Integrate b(x, y) w.r.t.x, regarding y as a constant.

Step 2: Integrate, with respect to y, those terms in a(x, y) which do not involve x.

Step 3: The sum of the two expressions obtained in steps 1 and 2 equated to a
constant is the required solution.

We now illustrate these various steps with the help of an example.

Example 11: Solve (x2 – 4xy – 2y2) dx + (y2 – 2x2) dy = 0.

Solution: Here a(x, y) = y2 – 4xy – 2x2 and b(x, y) = x2 – 4xy – 2y2

52
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

∂a ∂b
∴ = - 4y – 4x and = - 4x – 4y
∂x ∂y
∂a ∂b
∴ = ; hence it is an exact equation.
∂x ∂y

Step1: Integrating b(x, y) w.r.t.x. regarding y as a constant, we have

x3
∫ (x2 – 4xy – 2y2) dx = - 2xxy – 2xy2.
3

Step 2: We integrate those terms in a(x, y) w.r.t.y, which do not involve x. there
is only one such term namely, y2.

y3
∴ ∫ y2 dy = .
3

Step 3: The required solution is the sum of expressions obtained from Steps 1 and
2 equated to a constant, that is.

x3 y3
- 2x2y – 2xy2 + = c1,
3 3
or x3 – 6x2y – 6xy2 + y3 = c.

where c and c1 are constants.

Note that the test for an exact differential equation and the general procedure for
finding the solution can sometimes be simplified. We can pick out those terms of
a(x, y) dy + b(x, y) dx = 0 that obviously form an exact differential or can take the
form f(u) du. The remaining, expression which is less cumbersome than the
original can then be tested and integrated. This is illustrated by the following
example.

xdy − ydx
Example 12: Solve xdx +ydy + = 0.
x2 + y 2

Solution: Note that the first two terms on the left hand side of the given equation
are exact differentials and hence need not be touched. Dividing the numerator and
denominator of the last term by x2, we get

d ( y / x)
Xdx + ydy + =0
1 + ( y / x) 2

53
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Now each term of the above equation is an exact differential. Integrating, we get

x 2 y2 y
+ + tan-1 = c
2 2 x
as the required solution with c as a constant.

You may now try the following exercises.

In practice the differential equations of the form a(x, y) dy + b(x, y) dx = 0 are

rarely exact, since the condition in Theorem 1 requires a precies balance of the
functions a(x, y) and b(x, y). but they can often be transformed into exact equatins
on multiplication by a suitable function F(x, y) ≠ 0. This function is then called an
integrating factor. The question we, now, must ask is: if

a(x, y( dy + b(x, y) dx = 0

is not exact, then how to find a function

F(x, y) ≠ 0 so that
F(x, y) [a dy + b dx] = 0
Is exact? In the next section we shall give an answer to this question.

3.4 Integrating Factor

We begin with a very simple equation, namely,

y' + y = 0 …(40)

∂
In this case a(x, y) = 1 and b(x, y) = y. Here a(x, y) = 0
∂x
∂
and b(x, y) = 1 and hence the given equation is not exact. Let us multiply Eqn.
∂y
(40) by ex to get
exy' + exy = 0 …(41)

you may now check that Eqn. (41) is an exact equation. Thus Eqn. (40) is not
exact whereas when we multiply EQn. (40) by ex the resulting equation becomes
an exact one.

Here ex is termed as integrating factor for Eqn. (40).

We now give the following definition.

54
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Definition: A factor, which when multiplied with a non-exact differential equation

makes it exact, is known as an integrating factor (abbreviated as I.F.).

The term I.F., to solve a differential equation, was first introduced by Fatio de
Duillier in 1687.

For a given equation, there may not be a unique integrating factor.

Consider, for example, the equation

ydx – xdy = 0 …(42)

1
You can check that Eqn. (42) is not exact, but when multiplied by , it becomes
y2

ydx− xdy
=0
y2
 x
which is exact. This can now be written as d   = 0 and thus has for its solution
y 
x
= c with c being an arbitrary constant.
y

1
Further, when Eqn. (42) is multiplied by , it becomes
xy

dx dy
− = 0,
x y

which is given exact and has its solution as Inx –Iny = c.

you may notice that this solution can be transformed into the earlier solution
1 1
obtained through the I.F. 2 . Also Eqn. (42) when multiplied by reduces to
y x2
y dy  y y
an exact equation 2 dx - = 0 or, - d   = 0 with | - = c as its solution.
x x  x x

55
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1 1
Thus, we have seen that some of the integrating factors for Eqn. (42) are 2 ,
y xy
1
and
x2

Now the question arises: Is this the case only with Eqn. (42) or, in general, does an
equation of the form a(x, y) dy + b(x, y) dx = 0 have infinitely many integrating
factors?

An answer to this question is given in Theorem 2.

Before we give you this theorem, here is an exercise for you.

Theorem 2: The number of integrating factors for the equation

A(x, y) dy + b)x, y) dx = 0
9s infinite

proof: Let g(x, y) be an integrating factor of the given equation. Then, by

definition
 dy 
g(x, y)  a(x, y) + b(x, y) = 0 …(43)
 dx 
is an exact differential equation.

Therefore, there exist a function h(x, y) such that

 dy 
dh = g(x, y)  a(x, y) + b(x, y) = f(h)dh = d [ ∫ f(h) dh] …(44)
 dx 

since the term on the right hand side of Eqn (44) is an exact differential, the term
in the left must also be an exact differential. Therefore, g(x, y).f(h) is an
integrating factors of the given differential equation.

Since f(h) is an arbitrary function of h, hence the number of integrating factors for
equation a(x, y) dy + b(x, y) dx = 0 is infinite.

This fact is, however, of no special assistance in solving the differential equations.

So far, in our discussion we have not paid any attention to the problem of finding
the integrating factors. In general, it is quite difficult to obtain an integrating
factor for a given equation. However, rules for finding the integrating factors do
exist, we shall now take up these rules one by one.

56
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Rules for finding integrating factors.

Rule 1: Integrating factors of obtainable inspection: Sometime integrating

fators of a differential equation can be seen at a glance, as in the case of EQn. (42)
abve. We give below some more examples in this regard.

Example 13: Solve (1 + xy) ydx + (1 – xy)xdy = 0, x>0, y > 0. …(45)

Solution: Rearranging the terms of Eqn. (45), we get

ydx + xdy + xy2dx – x2ydy = 0
⇒ d(xy) + xy2dx – x2ydy = 0 …(46)

1
It is immediately seen that multiplication by makes Eqn. (46) eact and the
x y2
2

equation becomes
d(xy) dx dy
+ − =0
x 2 y2 x y

integrating, we get
1
- + Inx – Iny = Inc,
xy
or x = cy e1/(xy) (where c is a constant).

Example 14: Solve (x4ex – 2my2x) dx + 2mx2y dy = 0.

Solution: We can write the given equation as

x4ex dx + 2m (x+2ydy – xy2dx) = 0

 y
⇒ x4ex dx + 2mx3y d   = 0
 x
Dividing by x4, we get
y  y
dx + 2m d  = 0
x  x
 x  y 
2

⇒ d  e + m   = 0
  x  

1
thus has served the role of an integrating factor in this case.
x4

The required solution is, then, given by

57
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

2
 y
ex + m   = c with c as a constant.
 x

We would like to mention that determination of an integrating factor by inspection

is a skill and can be developed through practice only.

At this stage you may try the following exercises by finding an integrating factor
through inspection.

Rule II: For a homogeneous equation a(x, y)dy + b(x, y) dx -= 0, when bx + ay

1
≠ 0, then is an integrating factor.
bx + ay

Proof: Consider an equation

a(x, y) dy + b(x, y) dx = 0

1  dx dy   dx dy  
Now ady + bdx =  (bx + ay)  +  + (bx − ay) − 
2  x y  x y  
ady+ bdx 1   dx dy  bx − ay  dx dy  
∴ =   +  +  − 
ay+ bx 2  x y  x + ay  x y  

since the given equation is homogeneous, therefore a and b are of the same degree
bx − ay x  x
in x and y and therefore can be written as a function of , say f   .
bx + ay y  y

ady+ bdx 1   dx dy   x   dx dy  
∴ =   +  + f    − 
ay+ bx 2  x y   y  x y  
 

1  x 
 d(In xy) + f(e ) d In  
Inx/y
=
2  y 
1  x 
=  d(In xy) + dF In  
2  y 

1 1  x 
= d Inxy+ F In   …(47)
2 2  y  

58
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

 x  x
where d F   = f(eInx/y) d  In  .
 y  y

1
Since right hand side of Eqn. (47) is an exact differential, it shows that is
ay + bx
an integrating factor for the homogeneous equation a(x, y)dy + b(x, y)dx = 0.

We illustrate this rule by the following example.

Example 15: Solve (x2y – 2xy2) dx – (x3 – dx2+y) dy = 0.

Solution: Here the given equation is homogeneous and
a(x, y) = x3 + 3x2y and b(x, y) = x2y – 2xy2
∴ bx + ay = x(x2y – 2xy2) + y( - x3 + 3x2y) = x2y2 ≠ 0,
1
∴ is an integrating factor.
x y22

1
Multiplying the given differential equation by , we get
x y2
2

 1 2  x 3
 −  dx -  2 −  dy = 0,
 y x y y
dx dy dx x
or + 3 − 2 − 2 dy = 0,
y y x y
 dx x  dy dx
or  − 2 dy  + 3 − 2 = 0,
 y y  y x
 x
or d   + d(3 Iny + 2 Inx) = 0.
 y

Therefore, the solution is

x
+ 2 Iny – 2 Inx = c1 and c are constants.
y

a y dy y
Note: In case bx + ay = 0, then = − and the given equation reduces to =
b x dx x
,
Whose solution is straightaway obtained as x = cy.

You may now try this exercise.

59
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Rule III: when bx – ay ≠ 0 and the different equation a(x, y) dy + b(x, y) dx =

1
0 can be written in the form yf1(x y) dx + xf2(xy) dy = 0 then is an
bx − ay
integrating factor.

Proof: If equation a(x, y) dy + b(x, y)dx = 0 can be written in the form yf 1(xy)dx
+ xf2(xy)dy = 0
Then evidently,

a b
= = λ , say
xf 2 (xy) yf1 (xy)
∴ a = λ xf2(xy) and b = λ yf1(xy)

1  dx dy   dx dy  
Also ady + bdx =  (bx + ay)  +  + (bx − ay) − 
2  x y  x y  
ady+ bdx 1  bx + ay  dx dy   dx dy  
∴ =   + + − 
bx + ay 2  bx − ay  x y   x y  
1 f + f  x 
=  1 2 d(In xy) + d In  
2  f1 − f 2  y 
1  x 
=  f(xy) d(In xy) + d In  
2  y 
f1 + f 2
where f(xy) =
f1 − f 2
1   x  
=  dF(Inxy) + d In   
2    y   
1  x  1 
= d  In  + F(In(xy))
2  y 2 
where dF (In xy) = f(x, y) d(In xy),
which is an exact differential.

1
Hence, is an integrating factor.
bx − ay

We now illustrate this through the following example.

Example 16: Solve y(xy + 2x2y2) dx + x(xy – x2y2)dy = 0,

Solution: Here a = x(xy – x2y2) and b = y(xy + 2x2y2)

60
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

∴ bx – ay = xy[xy + x2y2 – xy + 2x2y2]

= 3x3y3 ≠ 0.

1 is an I.F.
∴
3 x 3 y3

1
Multiplying the given equation by , we get
3 x 3 y3
1 2 2
1
3 2 (xy + 2x y ) dx + (xy – x2y2) dy = 0
3x y 3 x 2 y3
dx 3 dx dy dy
or 2 + + 2
− =0
3x y 3 x 3 xy 3y
 dx dy  2 dx 1 dy
or  2
+ 2
+ − =0
 3 x y 3 xy  3 x 3 y
 1 1 2 1 
or d  − + Inx − Iny  = 0.
 3 xy 3 3 

Therefore, the solution is

1 2 1
− + Inx − Iny = c1 where c1 is a constant.
3 xy 3 3
1
or − + Inx 2 − Iny = 3 c1 = c for c being an arbitrary constant.
xy
 x2  1
or In   = c + .
 y xy

a y dy y
Note: If bx – ay = 0, i.e., = , then given equation will be of the form =−
b x dx x
and have a solution xy = c.

Before we go to the next rule here is an exercise for you.

1 ∂ b ∂a 
Rule IV: When  −  is a function of x alone, say f(x), then e ∫ f(x) dx is an
a  ∂ y ∂ x 
I.F. of the equation ady + bdx = 0.

Proof: Consider the equation e ∫ f(x) dx (ady + bdx) = 0 …(48)

Let c = b e ∫ f(x) dx and d = a e ∫ f(x) dx

61
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Then Eqn. (48) reduces to c dx + d dy = 0

∂ c ∂ b ∫ f(x) dx
Now, =
∂y ∂y e
∂ d ∂ a ∫ f(x) dx
and = e + a e ∫ f(x) dx f(x)
∂x ∂x
∂a 
= e ∫ f(x) dx  + a f(x)
∂ x 
∂a ∂b ∂a  1 ∂ b ∂a 
= e ∫ f(x) dx  + −   because  −  = f(x)
∂ x ∂ y ∂ x  a  ∂ y ∂ x  
∂ b ∫ f(x) dx
=
∂y e
∂c
=
∂y

therefore, the equation c dx + d dy = 0 is exact.

Hence e ∫ f(x) dx is an I.F. of the equation ady + bdx = 0.

We illustrate this rule with the help of the following example.

Example 17: Solve (x2 + y2) dx – 2xydy = 0.

Solution: Here a = 2xy, b = x2 + y2

∂b ∂a ∂b ∂a
∴ = 2y and = - 2y. thus, ≠ .
∂y ∂x ∂y ∂x
1 ∂ b ∂a  1 2
Here  −  = (2y + 2y) = - , which is a function of x alone.
a∂y ∂x − 2 xy x

2 1
∴ I.F. = ∫ − dx
=e
− 2 Inx
=
e x
x2

1
Multiplying the given equation by , we get
x2
1 2 2 2y
2 (x + y ) dx - dy = 0,
x x

y2 2y
i.e., dx 2 dx - dy = 0,
x x

62
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

 y2 
i.e., dx + d  −  =0 …(49)
 x 

integrating Eqn. (49), the required solution is obtained as

y2
x- = c (a constant)
x

You may now try this exercise.

1 ∂ b ∂a 
Rule V: When  −  is a function of y alone, say f(y), then e ∫ f(x) d y is an
b  ∂ y ∂ x 
I.F. of the differential equation
a dy + b dx = 0.

The proof of this rule is similar to the proof of Rule IV above and we leave this as
an exercise for you (see E 13).

We however illustrate the use of Rule V with the help of following example.

Example 18: Solve 93x2y4 + 2xy) dx + (2x3y3 – x2) dy = 0.

Solution: Here a = 2x3y3 – x2 and b = 3x2y+4 + 3xy

∂b ∂a
∴ = 12x2y3 + 2x and = 6x2y3 – 2x
∂y ∂x
1 ∂ b ∂a  1
Here  −  = (12x2y3 + 2x – 6x2y3 + 2x)
b ∂ y ∂ x 3 x y + 2 xy
2 4

2(3 x 2 y3 + 2 x 2
= = , which is a function of y alone.
y(3 x y + 2 x
2 3
y

2 1
Hence, I.F. = e ∫ − y
dy |
|e-2Iny = y-2 = .
y2

Multiplying the given equation by the I.F. = y-2, and on rearranging th terms, we
get

 2x x2 
2 2 3
(3x y dx + 2x ydy) +   dx − dy  = 0
 y y 2 
 x2 
i.e., d(x y ) + d   = 0
3 2

 y

63
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Integrating the above equation, we get

3 2
x2
xy + = c, where c is a constant of integration.
y
i.e., x3y3 + x2 = cy, which is the required solution.

And now an exercise for you.

Rule VI: If the differential equation is of the form

x α yβ (mydx + nxdy) = 0, where α, β , m and n are certain constants, then xkm-I-
α ykn-I- β is an integrating factor, where k can assume any value.

Proof: Multiplying the given equation by I.F., we get

Xkm-1 ykn-1 (mydx + nxdy) = 0,

or km xkm-1 ykn dx + kn xkmykn-1 dy = 0
or (xkm-1ykn) = 0, which is an exact differential.

It may be noted that if the given differential equation is of the form

x α yβ (mydx + nxdy) + x α 1yβ 1 (m1ydx + n1xdy) = 0
then also I.F. can be determined.
k 1 n 1 − 1− β
By Rule VI, x km − 1− α y kn − 1− β will make the first term exact, while x k m − 1− α y
1 1 1

will make the second term exact, where k and k1 can have any value.

Now these two factors will be identical if

kn-1- α = k1m1 – 1- α
1
and kn -1- β = k1n1-1- β
1 .

Values of k and k1 can be found tp satisfy these two algebraic equations. Then
either factor is an integrating factor of the above equation.

We now consider an example to illustrate this rule.

Example 19: Solve (y3 – 2yx2) dx + (2xy2 – x3) dy = 0.

Solution: On rearranging the term of the given equation, we can write

Y2(ydx + 2xdy) – x2(2ydx + xdy) = 0 …(50)
For the first term, = 0, = 2, m = 1 and n = 2 and hence its I.F.is xk-1y2k-1-2
α β

For the second term α1 = 2, β1 = 0, m1 = 2, n1 = 1 and hence for the second term
I.F. is

64
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

x 2 k1 − 1− 2 y k 1 − 1
these two integrating factors will be identical if
k-1 = 2k1 -1-2
and 2k-1-2 = k1 – 1 …(51)

solving the system fo Eqn. (51) for k and k1, we get k = 2 and k1 = 2 and,
therefore, integrating factor for Eqn. (50) is
x2-1 y4-1-2, i.e., xy.

Multiplying Eqn. (50) by xy, we get

xy3 (ydx + 2xdy) – x3y (2ydx + xdy) = 0
⇒ xy4dx + x2y3dy – (2x3y2dx + x4ydy = 0
1 1
⇒ (2xy4dx + 4x2y3dx) – (4x3y2dx + 2x4y4dy) = 0
2 2
1 1
⇒ d[2x2y4] - d[x4y2] = 0 …(52)
2 2

Integrating Eqn. (52), we get the required solution as

x 2 y4 − x 4 y2
= c1 (a constant)
2
or x2y2(y2 – x2) = 2c1 = c ( a constant).

You may now apply your knowledge about these rules and try to solve the
following exercises.

4.0 CONCLUSION

We now end this unit by giving a summary of what we have covered in it.

5.0 SUMMARY
In this unit we have covered the following:
dy
1) An equation dx
= f(x, y) is called a separable equation or an equation with
separable variables if f(x, y) = X(x) Y(y). to solve a separable equation, we
can write it as
dy
A(y) dx
+ b(x) = 0
For some a(y) and b(x). Integrating w.r. to x and equating it to a constant,
we get its solution.
2) a) A real-valued function h(x, y) of two variables x and y is
called a homogeneous function of degree n, if

65
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

n
h( λ x, y) = λ λ
h(x, y), where n is a real number and λ
is
any constant.
b) A differential equation
dy
dx
= f(x, y)
is called a homogeneous differential equation of first order when f is a
homogeneous function of degree zero.
c) A homogeneous differential equation reduces to separable equation by the
substitution y = vx, where v is some function of x.
3) Equations of the form
dy ax + by+ c a b
= where a, b, c, a', b', c' are constants and ≠ can be
dx a' x + b' y + c' a' b'
reduced to homogeneous equations by the substitution x' = x + h, y' = y +
k, where h and k are such that ah + bk + c = 0 and a'h + b'k + c' = 0.
a b 1
In case ≠
a' b'
= , say, then substitution ax + by = v reduces this type of
m
equations to separable equations.
4) An exact differential equation s formed by equating an exact differential to
zero.
5) The differential equation
A(x, y)dy + b(x, y)dx = 0

6.0 TUTOR MARKED ASSIGNMENT

1. Solve the following equations.

a) (1 – x)dy – (1 + y) dx = 0

dy 1 − y2
b) y–x + =0
dx 1 − x2

dy  2 dy 
c) y–x =a y + 
dx  dx 

d) 3ex tan ydx + (1 – ex) sec2 ydy = 0

dy
c) = ex-y + x2 e-y
dx
2. Solve the following equations satisfying initial condition indicted
alongside.

66
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy
a) 2xy = 1 + y2, y(2) = 3∀ x, y > =
dx

dy
b) = - 4xy, y(0) = y0 ∀ y > 0
dx

dy y− x 2
c) xe , y(0) = 0
dx

dy
d) y = g, y (x0) = y0
dx

3. Solve the following equations.

dy y
a) = for x ∈ ]0, ∞ [and for x ∈ ] - ∞ , 0[
dx x

dy 2x + y
b) =
dx 3x + 2y

y y
c) (x sin ) dy – (y sin-1 - x) dx = 0
x x

dy
d) x = y (Iny – Inx + 1)
dx

e) x dy – y dx = x 2 − y 2 dx.

4. Solve the following equations subject to the indicated initial conditions.

dy
a) 2x2 = 3xy + y2, y(1) = -2
dx

b) (x + yey/x) dx – xey/x dy = 0, y(1) = 0

c) (y2 + 3xy) dx = (4x2 + xy) dy, y(1) = 1.

d) y2dx + )x2 + xy + y2) dy = 0, y (0) = 1.

5. Solve the following equations.

67
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy 2y − x − 4
a) =
dx y − 3x + 3

dy
b) (7y – 3x + 3) + (3y – 7x + 7) = 0
dx

c) (2x + y + 1) dx + (4x + 2y – 1) dy = 0

d) (x + y) dx + (3x + 3y – 4) dy = 0

6. Prove that the following equations are exact and solve them.

a) (ycos(x) + 2x ey) + (sin(x) + x2ey + 2)y' = 0

ax + by
b) y' = - (a,b,c,d are given real constants).
bx + cy
c) 96x + y/x) + (Inx + y)y' = 0, x ≥ 1.

7. Determine the values of k for which the equations given below are exact
and find the solution for these values of k.

a) x + kyy' = 0 (k ≠ 0)
b) y + kxy' = 0 (k ≠ 0)
c) (2y e2xy + 2x) + k x e2xyy' = 0

8. In each of the following equations verify that the function F(x, y), indicated
alongside is an I.F. of the equation:

i. 6xy dx + (4y + 9x2) dy = 0; F(x, y) = y2

1
ii. –y2 dx + (x2 + xy) dy = 0; F(x, y) =
x2y
iii. ( - xy sin x + 2y cos x) dx + 2x cos x dy = 0; F(x, y) = xy.

9. Solve the following equations.

a) y(2yx + ex) dx – ex dy = 0
b) ydx – xdy + Inx dx = 0 ∀ x, y > 0.
c) (xy – 2y2) dx – (x2 – 3xy)dy = 0

10. Solve (x4 + y4) dx – xy3dy = 0.

11. Solve y(x2y2 + 2) dx + x(2 – 2x2y2) dy = 0.

68
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

12. Solve (x2 + y2 + x) dx + xy dy = 0.

13. Prove Rule V above.

14. Solve (2xy4ey + 2xy3 + y) dx + (x2y4ey – x2y2 – 3x) dy = 0

15 Solve the following equations.

a) (x2+ y2 + 2x) dx + 2y = 0
b) x2y dx – (x3 + y3) dy = 0
c) (y4 + 2y)dx + (xy3 + 2y4 – 4x) dy = 0 ∀ x, y > 0.
d) (y2 + 2x2y)dy + (2x3 – xy)dy = 0
e) (2x2y – 3y4) dx + (3x3 + 2xy3) dy = 0

16. Solve the following equations.

dy
a) (x + y)2 = a2
dx
b) ydx + dy = 0
x  dy
c) 1 +  − siny  =0
y  dx
dy
d) (3y2 + 2xy) = (2xy + x2) =0 ∀
x > 0, y > 0.
dx
 y  dy
 2 xy + 
+  1 + y  dx
e) Y + y2 =0
dy
f) 2x2y2 + 3x(1 + y2) dx
=0
dy 2y
g) dx
+ = 0, y ≥ 0
x

7.0 REFERENCES/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

69
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Indira Gandhi National Open University School Of Sciences Mth-07

70
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

MODULE 2

Unit 1 Linear Differential Equations

Unit 2 Differential Equations of First order but not of first degree

UNIT 1 LINEAR DIFFERENTIAL EQUATION

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Classification of First Order Differential Equation
3.2 General solution of Linear Non-homogeneous Equation
3.2.1 Method of Undetermine Coefficients
3.2.2. Method pf variation of Parameters
3.3 Properties of the solution of Linear Homogeneous differential
Equation
3.4 Equation reducible to linear Equation
3.5 Applications of Linear Differential Equation
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 Reference/Further Readings

1.0 INTRODUCTION

In unit 2, we have discussed methods of solving some first order first degree
differential equations, namely,

i) differential equations which could be integrated directly i.e., separable and

exact differential equations,

ii) equations which could be reduced to these forms when direct integration is
not possible. These includes homogeneous equations, equations reducible
to homogeneous form and equations that become exact when multiplied by
an I.F.

in this unit, we focus our attention on another very important type of first order
first degree differential equations known as linear equations. These equations are
important because of their wide range of applications, for example, the physical

71
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

situations we gave in Sec. 1.5 of unit 1 are all governed by linear differential
equations. In this unit, we shall solve some of these physical problems.

The problem of integrating a linear differential equation was reduced to quadrature

by Leibniz in 1692. In December, 1695, James Bernoulli proposed a solution of a
non-linear differential equation of the first order, now known as Bernoulli’s
equation.

In 1696, Leibniz pointed out Bernoulli’s equation may be reduced to a linear

differential equation by changing the dependent variable. We shall discuss this
equation in the later part of this unit along with some other equations, which may
not be of first order or first degree but which can be reduced to linear to linear
differential equations.

2.0 OBJECTIVES

After studying this unit, you should be able to

• Identify a linear differential equation;

• Distinguish between homogeneous and non-homogeneous linear differential
equation;
• Obtain the general solution of a linear differential equation;
• Obtain the particular integral of a linear equation by the methods of
undetermined coefficients and variation of parameters;
• Use general properties of the solutions of homogeneous linear equations for
finding their solutions;
• Obtain the solution of Bernoulli’s equation;
• Obtain solution to linear equations modeled for certain physical situations.

3.0 MAIN CONTENT

3.1 Classification of First Order Differential Equations

We begin by giving some definitions in this section. You may recall that in Unit 1
we defined the general form of first order differential equation to be

 dy 
g  x, y  =0
 dx 
and if the equation is of first degree, then it can be expressed as
dy
dx
= f(x, y)

72
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

In the above equation if the function f(x, y) be such that it contains dependent
variable y in the first degree only, then it is called a linear differential equation.
Formally, we have the following definition.

Definition: We say that a differential equation is linear if the dependent variable

and all its derivatives appear only in the first degree and also there is no term
involving the product of the derivatives or any derivative and the dependent
variable.

dy 2y d2 y dy
For example, equation dx
+ x
= x3 and + dx
= x sin x are linear
dx 2
dy
differential equations. However dx
+ x2 = 10 is not linear equation of the presence
dy
of the term y dx
.
The general form of the linear differential equation of the first order is
dy
a(x) dx
= b(x)y + c(x) …(1)
where a(x), b(x) and c(x) are continuous real valued functions in some interval I ⊆
R.
If c(x) is identically zero, then Eqn. (1) reduces to

dy
a(x) dx
= b(x)y …(2)

Eqn. (2) is called a linear homogeneous differential equation.

When c(x) is zero, Eqn. (1) is called non-homogeneous (or inhomogeneous)
linear differential equation

Note: You may note that the word homogeneous as it is used here has a very
different meaning from that used in Sec. 2.3, unit 2.

Any differential equation of order one which is not of type (1) or (2) is called a
non-linear differential equation.

On dividing Eqn. (1) by a(x) for x s.t a(x) ≠ 0, it can be put in the more useful
form
dy
dx
+ P(x) y = Q(x), …(3)
where P a. I Q are functions of x alone or are constants. Consider, for instance, the
dy
equation dx
=y

73
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

It is a linear homogeneous equation. Here a(x) = 1 and b(x) = 1. Similarly,

dy dy
dx
= 0, dx
= exy are also linear homogeneous equation of order one with

a(x) = 1, b(x) = ex and c(x) = x.

dy
Next consider the differential equation dx
= |y|.

You know that |y| = y for 0 and |y| = - y for y < 0. Hence, in order to solve this
≥

equation, we will have to square it and the resulting equation is neither of type (1)
dy
nor of (2). It is a case of non-linear equation. Similarly, = y is a non-linear
dx
dy dy
equation because of the term dx
. Again dx
= cos y is a non-linear equation (as
cos y can be expressed as an infinite series in powers of y).

You may now try this exercise.

You will realize the need for classification of linear differential equation into
homogeneous and non-homogeneous equations when we discuss some properties
involving the solution of linear homogeneous differential equations. But first let
us talk about the general solution of linear non-homogeneous equation of type (1)
or (3).

3.2 General Solution of Linear Non-Homogeneous Equation

Consider Eqn. (3) , viz.,

dy
dx
+ P(x)y = Q(x)

In the discussion that follows, we assume that Eqn. (3) has a solution. You can
see that in general, Eqn (3) is not exact. But we will show that we can always find
an integrating factor μ (x), which makes this equation exact – a useful property of
linear equations.

Let us suppose that Eqn. (3) is written in the differential form

dy + [P(x)y – Q(x)] dx = 0 …(4)

Suppose that μ
(x) is an I.F. of Eqn. (4). Then

74
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

(x) dy +
μ
(x) [P(x)y – Q(x)] dx = 0 μ
…(5)
is an exact differential equation. By Theorem 1 of Unit 2, we know that Eqn. (5)
will be an exact differential if
∂ ∂
( μ
(x) = ( μ
(x)[P(x)y – Q(x)]} …(6)
∂x ∂y

dμ
or = μ
P(x)
dx

This is a separable equation from which we can determine μ

(x). we have
dμ
= P(x)dx
μ
or In | μ
| = ∫ P(x) dx …(7)

∫ P(x)dx
so that (x) = e is an integrating factor for Eqn. (4).
μ

Note that we need not use a constant of integration in relation (7) since Eqn. (5) is
unaffected by a constant multiple. Also, you may note that Eqn. (4) is still an
exact differential equation even when Q9x) = 0. in fact Q(x) plays no part in
determining

∂
μ
(x) since we see from (6), that ∂y
μ
(x) Q(x) = 0. Thus both
∫ P(x)dx ∫ P(x)dx
e dy + e [P(x)y – Q(x)] dx and
∫ P(x)dx ∫ P(x)dx
e dy + e P(x)y dx

are exact differentials.

We, now, write Eqn. (3) in the form

 dy 
e ∫ Pdx  + + y = Q e ∫ Pdx
 dx 

This can also be written as

d ∫ Pdx ∫ Pdx
(y e )=Qe
dx
Integrating the above equation, we get

y e ∫ Pdx ∫ Q e∫ Pdxdx + α , where α

is a constant of integration
or y = e ∫ Pdx
∫ Q e dx + ∫ Pdx
e α ∫ Pdx
…(8)

75
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

For initial value problem, the constant α

in Eqn. (8) can be determined by using
initial conditions. Relation (8) gives the general solution of Eqn. (3) and can be
used as a formula for obtaining the solution of equation of the form (3). As a
matter of advice we may put it that one need not try to learn the formula (8) and
apply it mechanically for solving linear equations. Instead, one should use the
procedure by which (8) is derived: multiply by e and integrate.
∫ Pdx

In case of linear homogeneous equation, the general solution can be obtained by

putting Q = 0 in Eqn. (8) as

y = α e− ∫ Pdx

Note that the first term on the right hand side of Eqn. (8) is due to non-
homogeneous term Q of Eqn. (3). It is termed as the particular integral of the
linear non-homogeneous differential equation, that is, particular integral of Eqn.
(3) is
e − ∫ Pdx ∫ Q e dx. ∫ Pdx

The particular integral does not contain any arbitrary constant.

The solution of linear non-homogeneous equation and its corresponding linear

homogeneous equation are nicely interrelated. We give the first result, in this
direction, in the form of the following theorem:

Theorem 1: In I ⊆ R, if y1 be a solution of linear non-homogeneous differential

Eq. (3), that is,

dy
dx
+ P(x)y = Q(x)
and if z be a solution of corresponding linear homogeneous differential equation

dy
dx
+ +(x)y = 0, …(9)
then the function y = y1 + z is a solution of Eqn. (3) on I.

∴ dy
= dy1 + d z ..(10)
dx dx dx

Since y1 P(x)y1 = Q(x)

dy1
+ P(x)y1 = Q(x) …(11)
dx

76
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Also since z is a solution of (9), therefore

dz
+ P(x)z = 0 …(12)
dx

On combining Eqns. (10) – (12), we get

dy
dx
= [Q(x) – P(x)y1] + [-P(x)z]
= Q(x) – P(x) [y1 + z]
= Q(x) – yP(x) as ( y1 + z = y),
dy
i.e., dx
+ P(x)y = Q(x).

Hence y = y1 + z is a solution of Eqn. (3) and this completes the proof of the
theorem.

From this theorem, it should be clear that any solution of Eqn. (3) must contain
solution of Eqn. (9) (corresponding linear homogeneous equation).

In case, the function Q(x) on the right-hand side of Eqn. (3) is a linear combination
of functions, then we can make use of the following theorem:

Theorem 2: Let yi be a particular solution of

dy
dx
+ P(x)y = Qi(x),

where Qi(x) are continuous functions defined on an interval I for I = 1, 2,…., n.

then the function yp = y1 + y2 +…+ yn, defined on I, is a particular solution of

dy
dx
+ P(x)y = Q(x),

where Q(x) = Q1(x) + Q2(x) +….+ Qn(x), ∀ x ∈ I.

The proof of this theorem is simple and is left as an exercise for you.

We now take up some examples and illustrate the method of finding the solution
of linear non-homogeneous differential equations.
dy
Example 1: Solve x dx
+ y = x3

77
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: The given differential equation can be written as

dy 1
dx
+ y = x2 …(13)
x

it is a linear equation. Comparing it with Eqn (3), we have

1
P= x
. So I.F. = e∫ Pdx = e ∫ (1 / x ) dx = eInx = x

Multiplying Eqn. (13) by x, we get

dy
x dx
+ y = x3
d
i.e., (xy) = x3, which is exact.
dx

Integrating, we get
x4
xy = + c,
4

c being a constant, as the required solution.

dy
Example 2: Solve x dx
- ay = x + 1

Solution: Clearly the given equation is linear and can be written in the form
dy a x+ 1
dx
- y=
x x
1
∴ I.F. = e ∫ ( − a/x) dx = e-aInx = e Inx − a =
xa

1
Multiplying the given equation by , we get
xa
1 dy
- aa + 1 y = x +a + 11 ,
xa x dx x
d  y  x+ 1
i.e.,  =
dx  x a  x a + 1

Integrating the above equation w.r.t.x, we get

y x+ 1
∫
xa = x a + 1 dx + c (c is constant)

78
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

x − a+ 1 x − a
= + +c.
− a+ 1 − a

x 1
Thus y = − + cxa is the required solution.
1− a a

Let us look at another example in which the role of x and y has been interchanged.

dx
Example 3: Solve y Iny + x – Iny = 0.
dy

dx
Solution: This equation is of first degree in x and . Hence it is a linear
dy
equation with y as independent variable and x as dependent variable.

Then given equation can be written as

dx x 1
+ = , …(14)
dy yIny y
I.F. = e ∫ 1/(y in y) dy
∴

= eIn(Iny)
= Iny

Multiplying Eqn. (14) by Iny, we get

dx 1 1
Iny dy + x= y Iny,
y
d 1
i.e., (x Iny) = y Iny.
dy

Integrating the above equation w.r.t.y, we get

1
xIny = ∫
y Iny dy + c,
(Iny)2
= + c, c is a constant,
2
or 2x Iny = (Iny)2 + c1, is the required solution where c1 = 2c.

let us consider another example.

Example 4: solve the equation ydx + (3x – xy + 2) dy = 0

79
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: Since the product y dy occurs here, the equation is not linear in
dependent variable y. it is, however, linear if we treat variable y as independent
variable and x as dependent variable. Therefore, we arrange the terms as
y dx + (3 – y) x dy = - 2 dy,
and write it in the standard form
dx 3  2
dy +  − 1 x = - , for y ≠ 0 …(15)
y  y
3 
Now, ∫  − 1 dy = 3 In|y|,
y 
So that an integrating factor for Eqn. (15) is
e(3In|y|-y) = e-y e3In|y|
= e In|y| e − y
3

= |y|3e-y
If follows that for y > 0, y3e-y is an integrating factor and for y > 0, -y3 e-y serves as
an integrating factor for the given equation. In either case, we are led to the exact
equation

Y3 e-y dx + y2 (3 – y e-y x dy = - 2y2e-y dy,

i.e., d(xy3y-y) = - 2y2e-y dy.

Integrating the above equation w.r.t.y, we get

Xy3e-y = - 2 ∫ y2 e-y dy
= 2y2 e-y + 4y e-y + c (Integrating by parts)

Thus, we can express the required solution as

xy3 = 2y3 + 4y + 4 + cey, where c is an arbitrary constant.

You may try the following exercises.

We have seen that general solution of a linear non-homogeneous differential Eqn.

(3) is given by Eqn. (8), which involves integrals. We remark that an equation y' =
f(x, y) is said to be solve when its solution is reduced to the expression of the form
∫ h(x) dx or ∫ ø(y) dy for some h(x) and ø(y) even if it is impossible to evaluate

these integrals in terms of known functions. Further, the reduction of the solutions
from one form to a simpler form may require as much labour as the solving of the
equations. In solution (8) of Eqn. (3), e − ∫ Pdx ∫ Q(x) e ∫ Pdx dx is the particular
integral of Eqn. (3) and the evaluation of this integral will depend on the form of
Q(x). this evaluation may sometimes turn out to be a tedious task. But, there are
other methods by which particular integral in some scases can be obtained without
carrying rigorous integration. We shall briefly disucss these methods now. As

80
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

these methods are more helpful for higher order differential equations, we shall
discuss them in greater detail in Block 2.

3.2.1 Method of Undetermined Coefficients

This method is applicable when in Eqn. (3), i.e.,

dy
dx
+ P(x)y = Q(x),

P(x) is a constant and Q(x) is any of the following forms:

i) an exponential
ii) A polynomial in x
iii) of the form cos β x or sin β x
iv) a linear combination of i), ii) and iii) above.

The general procedure is to assume the particular solution with arbitrary or

unknown constants and then determine the constants.

We know that on differentiating functions such as e αx ( α constant), xr (r > 0 is an

integer), sin β x or cos β x ( β constant), we again obtain an exponential, a
polynomial or a function which is a linear combination of sine or cosine function.
Hence if the non-homogeneous term Q(x) in Eqn. (3) is in any of the forms (i) –
(iv), above, then we can choose the particular integral accordingly to be a suitable
combination of the terms n(i) – (iv).

We now take up different cases according to the forms of Q(x).

Case 1: Q(x) = k emx, k and m are real constants, that is, Q(x) is an exponential
function. In this case, we prove the result in the form of the following theorem.

Theorem 3: If a, k and m are real constants, then a partuclar solution of

dy
dx
+ ay = k emx
is given by
 k
 e mx if m ≠ − a
yp(x) =  (a + m)
 kxe mx if m = − a

81
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Proof: in this case, since Q(x) is an exponential function, we assume y p(x) = remx
to be a particular solution of Eqn. (16), where r is some constant to be determined.
Now yp(x) must satisfy Eqn. (16).

Thus, we get
Rm emx + ar emx = kemx
1
Or r = if m ≠ - a.
a+ m

1
Therefore, yp(x) = emx if m ≠ - a.
a+ m
In case m + a = 0, i.e., m = - a, then you may verify that y p(x) = kx emx satisfies
Eqn. (16). The reasoning for this sort solution will be given when we discuss this
method in detail in Block 2. However, we illustrate this case by the following
example.

Example 5: Solve y' – y = 2ex

Solution: On comparing the given equation with Eqn. (16), we find that a = - 1, k
= 2, and m = 1
Also, m + a = -1 + 1 = 0 ⇒ m = -a.
∴ By Theorem 3, a particular integral is 2xex.
Further, I.F. = e − ∫ Pdx = e − ∫ dx = e = x .(∴ P = − 1)
Therefore, required solution, following relation (8), is
y = P.I + c ex,
i.e., y = 2xex + c ex.
You may now try this exercise.
n
Case II: Q(x) = ∑ax
i= 0
i
i

That is, Q(x) is a polynomial of degree n. In this Eqn. (3) reduces to

n
dy
dx
+ ay = ∑ a x i= 0
i
i
…(17)
If a = 0 in Eqn. (17), then particular solution is
n
ai
Yp(x) = ∑ i+ 1 x
i= 0
i+ 1
, which follows by direct integration.
If in Eqn. (17), a ≠ 0, then we assume
n
yp(x) = ∑ Px
i= 0
i
i
(Q(x) being a polynomial in this case),

82
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

And determine real numbers P0, P1, …., Pn so that particular solution yp(x) satisfies
Eqn. (17).
Substituting this value of yp(x) in Eqn. (17) (with y replaced by yp(x)), we have
n n n

∑ iPi x i − 1 +
i= 1
∑ aPi x i =
i= 0
∑ax
i= 0
i
i
(a ≠ 0) …(18)

Equating the coefficients of like power of x on both sides of Eqn. (18), we get

Coeff. Of xi: (I +1)Pi+1 + aPi = ai for I = 0, 1, 2, …., (n-1)

Coeff. Of xn: aPn = aa …(19)

Since Q(x) is a polynomial of degree n, thus an ≠ 0 and we can solve EQn. (19)
for P0 P1,….Pn. from Eqn. (19), we get

Pn = an/a
 na n  1
Pn-1 =  a n − 1 −  ,
 a a
n− 1 n 1
Pn-2 = an-2- (an-1- an) , and so on.
a a a

We illustrate this method with the help of following example.

dy
Example 6: Find the particular solution of + 2y = 2x2 + 3.
dx

Solution: We note that, in this case, Q(x) is a polynomial of degree 2. Assume a

particular solution of the form
n
yp(x) = ∑ Px
i= 0
i
i
= P0 + P1x + P1x2.
Substitution of yP(x) in the given equation yields
(P1 + 2P2x) + 2(P0+P1x+P2x2) = 2x2 + 3. …(20)
Equating the coefficients of like powers of x on both side of Eqn. (20), we get
Coeff. of x2: 2P2 = 2 or P2 = 1.
Coeff. of x2: 2P2 = 2 or P2 = 1.
Coeff. of x0: P1 + 2P0 = 3 or P0 = 2.

Hence, required particular solution is

yP(x) = x2 –x + 2.
And now an exercise for you.
Case III: Q(x) = sin βx or cos βx or a βx + b cos βx

83
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Where β , a and b are real constants.

In all these cases, we assume a particular solution of the form c sin βx

+ d cos βx
.

On substituting this solution in the given equation and equating the coefficients of
sin βx and cos βx on both sides, we determine the constants c and d.

Let us illustrate this case by an example.

Example 7: Find the particular integral of

dy
dx
+ y = cos 3x

solution: Here Q(x) = cos 3x.

Hence, any particular solution of the given differential equation must be a

combination of sin 3x and cos 3x. let the particular solution be
yP(x) = c cos 3x + d sin 3x
On substituting this value of yP(x) in the given equation, we get
(-3c sin 3x + 3x cos 3x) + (c cos 3x + d sin 3x) = cos 3x …(21)
comparing the coefficients of cos 3x and sin 3x on both sides of Eqn. (21), we get

c + 3d = 1 and d – 3c = 0
1 3
or c = and d =
10 10
Hence, the particular solution is
1
yp(x) = 10
(3 sin 3x + cos 3x)
we now take up an example which is a combination of all the three cases discussed
above.

Example 8: Compute the general solution of

dy
dx
+ y = ex + x + sin x …(22)

Solution: Here Q(x) = Q1(x) + Q2(x) + Q3(x),

With Q1(x) = ex, Q2(x) = x and Q3(x) = sin x.
You may recall Theorem 2; if y1, y2 and y3 are particular solutions of
dy
dx
+ y = ex …(23)
dy
dx
+y=x …(24)
and

84
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy
dx
+ y = sin x …(25)
respectively, then yp = y1 + y2 + y3 is a particular solution of the given equation.
Consider Eqn. (23). Let the particular solution be
Y1 = rex.
Substituting this in Eqn. (23), we get
1
rex +rex = ex ⇒ r =
2
1
∴ y1 = ex …(26)
2

For Eqn. (24) we assume the particular solution as

y2 = a1x + a0.
Substituting this in Eqn. (24), we get
a1 + a1x + a0 = x …(27)
comparing coefficients of like powers of x on both side of Eqn. (27), we get
a0 + a1 = 0
a1 = 1, a0 = -1, a1 = 1
⇒

Hence y2 = x-1 …(28)

In the case of Eqn (25), assume particular solutions as

y3 = c sin x + d cos x.
Substituting this in Eqn. (25), we get
c cos x – d sin x + c sin x + d cos x = sin x …(29)

On equating the coefficients of sin x and cos c on both sides of Eqn. (29), we get
c–d=1
1 1
c+d=0 ⇒
c= 2
and d = - 2

1
∴

y3 = 2
(sin x – cos x) …(30)

Hence, particular solution of Eqn. (22) can be obtained from Eqn. (26), (28) and
(30) as

1 1
yp(x) = y1 + y2 + y3 = 2
ex + x -1 + 2
(sin x – cos x)

The solution of homogeneous pat of Eqn. (22), i.e.

dy
dx
+y=0
is given by

85
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1 dy
+1=0
y dx

Integrating the above equation, we get

In y + x = In α , for some constant , α

y
i.e., = e-x
α
or y = α
e-x

Hence complete solution of Eqn. (22) is given by

1 1
y= α
e-x + 2
ex + x-1 + 2
(sin x – cos x)

How about trying an exercise now?

We thus studied the method of undetermined coefficients for finding the particular
integral of the non-homogeneous linear differential EQn. (3). We saw that this
method would be applicable only for a certain class of differential equations –
those for which P(x) is a constant and Q(x) assumes either of the forms e αx , xr, sin
βx
or cos βx , or their combinations. We shall, now, study a method that carries
no such restrictions.

3.2.2 Method of Variation of Parameters

Consider the non-homogeneous linear Eqn. (3), namely,

dy
dx
+ P(x)y(x) = Q(x).
the homogeneous equation corresponding to the above linear equation is
dy
dx
+ P(x) y(x) = 0.

Further we know , from Eqn. (8), that the solution yh(x) of the homogeneous linear
equation is given by
yh(x) = e − ∫ Pdx , α
..(31)
where is a constant.
α

In tis method we assume that , in Eqn. (31), is not a constant but a function of
α

x. that is, we vary with x and assume that the resulting function
α

y(x) = (x) e α − ∫ Pdx

…(32)
is a solution of Eqn. (3). That is, we try to determine (x) such that y given by
α

Eqn. (32) solves Eqn. (3). In other words, we determine a necessary condition on
α
(x) so that y defined by relation (32), is a solution of Eqn. (3).

86
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

On combining Eqns. (3) and (32), we get

d
[ α
(x) e − ∫ P( x ) dx ] + P(x) [ α
(x) e − ∫ P( x ) dx ] = Q(x).
dx
i.e., α
(x) = Q(x) e ] + − ∫ Pdx
'(x) e − ∫ P( x ) dx + P(x)
α α
(x) e − ∫ P( x ) dx = Q(x),
i.e., α
'(x) = Q(x) e ∫ P( x ) dx
Integrating w.r.t.x, we get
α
(x) = β + ∫ Q(x) e ∫ P( x ) dx dx …(33)

where β is a constant of integration.

Substituting the value of (x) from eQn. (33) in relation (32), the solution to
α

Eqn. (3) can be expressed as

y(x) = β e − ∫ P( x ) dx + e − ∫ P( x ) dx ∫ Q(x) e − ∫ P( x ) dx dx

You may note here that the solution obtained above is same as the one given by
Eqn. (8) that has been obtained directly. Further, the method of variation of
parameter neither simplifies any integration/solution nor provides any other form
of the solution for first order first degree differential equation. It only provides an
alternative approache to arrive at the general solution in this case. However, as we
shall see later in Block 2, this method turns out to be quite powerful in discussing
equations of higher order.

Using the method of undetermined coefficients/variation of parameters or

otherwise you may now try this exercise.

Now let us discuss some properties of linear homogeneous differential equations,

which give us some insight into qualitative theorem rather than|quantitative
solutions.

3.3 Properties of the Solution of Linear Homogeneous Differential

Equation

In this section we shall discuss certain properties enjoyed by linear homogeneous

differential equation. We start with a very important property called
superposition principle.

Theorem 4: (Superposition Principle)

If y1 and y2 are any two solution of the linear homogeneous Eqn. (9), i.e.,
dy
dx
+ P(x) y(x) = 0.
Then y1 + y2 and cy1 are also solutions of Eqn. (9), where c is a constant.
Proof: Since y1 and y2 are both solutions of Eqn. (9), therefore

87
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy1
+ P(x)y1 = 0 …(34)
dx
and
dy 2
+ P(x)y2 = 0, …(35)
dx
Let h(x) = y1 + y2
dh
∴
= dy1 + dy 2 ,
dx dx dx
= - P(x)y1 – P(x)y2, (using Eqn. (34) and (35))
= - P(x) (y1 + y2),
= - P(x) h(x),
dh
i.e., + P(x) h(x) = 0, which shows that h(x) = y1 + y2 is indeed a solution of
dx
Eqn. (9).
Next, multiplying Eqn. (34) by c (a constant), we get
dy 1
c + cP(x)y1 = 0,
dx
d
i.e. dx
(cy1) + P(x). (cy1) = 0,
which shows that (cy1) is also a solution of Eqn. (9).

In many branches of sciences, y1 + y2 is called superposition of y1 and y2 and hence

the name superposition principle for Theorem 4.

The conclusions of Theorem 4 can be reframed as – the set of real (or complex)
solution of Eqn. (9) forms a ral (or complec) vector space (ref. Block 1, MTE-02,
acourse on linear algebra).

Do you think Theorem 4, holds for non-homogeneous linear equation? Consider

the non-homogeneous equation y' = 2x.

The functions (1 + x2) and (2 + x2), x ∈ R are two solution of

y' = 2x …(36)

d
their sum (2+x2) + (2+x2) = 3 + 2x2, x ∈R
does not satisfy Eqn. (36) since dx
(3
+ 2x2) = 4x ≠ 2x, ∀ x ∈R
.

Thus, Theorem 4 need not be true for a non-homogeneous linear equation does it
work for a non-linear equation? Let us look at
y' = - y2 …(37)

88
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1 1
which has solutions y1(x) = and y2 = on an interval
(1 + x) 1+ 2 x
I = [0, ∞ [. This is true because
1 1
y1'(x) = - 2 = - y2 ,
(1 + x)
and
2 .2
y2'(x) = - = - y 22 ,
(1 + 2 x) 2
Let y = y1 + y2. Here y is well defined on I. Also, by simple computation, we have
1 2 .2 (8 x 2 + 12 x + 5)
y'(x) = - - =- …(38)
(1 + x) 2 (1 + 2 x) 2 (1 + x) 2 (1 + 2 x) 2

on the other hand,

2
(16 x 2 + 24 x + 9)
- y (x) = - …(39)
(1 + x) 2 (1 + 2 x) 2

from relations (38) and (39), it is clear that y = y1 + y2 is not a solution of (37).

In the next exercise we ask you to show an example of a non-linear equation

whose solution is y1 but cy1 is not a solution, i.e., the later part of Theorem 4 need
not be true for a non-linear equation.

Mostly we study the real solutions of Eqn. (1). You may recall that the functions a
and b in Eqn. (1) (defined on I) are assume to be valued. The reason for restricting
the study to real solutions will be clear from the following theorem.

Theorem 5: If y = p + iq is a complex valued function defined on I, which

dy
satisfies Eqn. (2), that is, a(x) = b(x)y(x), then the real part p of y and the
dx
imaginary part q of y are also solutions of Eqn. (2) on I.
(recall here that a and b are ral valued continuous functions)

proof: By definition y = p+iq and so y' = p' + iq'. Since y satisfies Eqn.s (2), we
have a(x) {p'(x) + iq'(x)} = b(x) {p(x) + iq(x)} …(40)

since a and b are real valued, on equating the ral and imaginary parts in Eqn. (4),
we get
a(x)p (x) = b(x)p(x),
and
a(x)q' (x) = b(x)q(x),
which show s that p and q are solution of Eqn. (2) on I.

89
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Theorem 5 is also true for higher order linear homogeneous equations which will
be discussed in our later blocks and the proof is virtually on the same lines. But
the theorem may fail if we replace Eqn. (2) by an arbitrary non-linear equation or a
linear non-homogeneous equation. for instance, consider the first order non-linear
equation
yy' = -2x3 …(41)

the function y(x) = ix2, x is a complex valued solution of Eqn. (41), since
∈R

y'(x) = 2ix
and y(x)y'(x) = (ix) (ix2 = -2x3.

The real par p of y is the zero function. i.e., p(x) = 0. But p is not a solution of
Eqn. (41) (since 2x3 0 for all x
≠
). ∈R

The following exercise shows that Theorem 5 may fail in the case of non-
homogeneous linear equations.

We shall now be giving another interesting result concerning linear homogeneous

dy
equation a(x) dx
= b(x)y(x), which can also be written as
y' = g(x)y, …(42)

b(x)
where g(x) = , is a real valued continuous function defined on I. Result
a(x)
which we are going to state is a consequence of the uniqueness of solutions of
initial value problem for linear equations.

Theorem 6: Let y be a solution of the Eqn. (42) on the interval such that y(x 1) = 0
for some x1 in I. Then y = 0 on I.

Proof: Consider the initial value problem

y' = g(x) y,
y(x1) = 0

By hypothesis, y is a solution of Eqn. (42). But the function z, defined by z9x) =

0 for all x I), also satisfies Eqn. (42) (because z'(x) = 0, g(x)z(x) = 0 and z(x 1) = 0).
By the uniqueness theorem for the initial value problem for linear equation (refer
Theorem 1, Unit 1), we have z = y or in other words, y(x) = 0 for x ∈ I . This
completes the proof. Just as we have seen in the case of Theorem 4 and 5,
Theorem 6 may not be true for non-linear or linear non-homogeneous equations.
Consider, for instance, the following non-linear differential equation.
y' = 2√y, x ∈ [0, ∞

[ …(43)

90
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Let c>0. we define the function y on [0, ∞

[ by
 0 if 0 ≤ x < c
y(x) = 
 2(x − c) if c ≤ | x < ∞
2

from the definition of y, we have

 0 if 0 ≤ x < c
y'(x) = 
 2(x − c) = 2 y(x), if c ≤ | x < ∞
(Note that y is differentiable at x = c and, in fact, its right as well as left derivative
is zero at x = c).

 0 = 2 y ( x), 0 ≤ | x < c
y'(x) = 
 2(x − c) = 2 y(x) if c ≤ | x < ∞
which shows that y satisfies Eqn (43) for all x > 0. We notice, here, that y
vanishes on the interval [0, c[ and yet y is a non zero function on [0, [; which
∞

shows that the conclusions. For example, y(x) = cos x + sin, x is a solution of
∈R

Theorem 6 may not be true for non-linear equations.

similarly, we can show that Teorem 6 is not valid for linear non-homogeneous
equation. for example, y(x) = cos x + sin x, x is a solution of
∈R

y' = y – 2sin x, x .
∈R
…(44)

π π
But, y(x) is zero at many points (like x = - , - +2 π
, …)
4 4

and assume both negative and positive values. Yet y is a non-linear function
which solves Eqn. (44).

Did you notice that in Theorem 6 we did not take a general linear homogeneous
equation? we only considered linear homogeneous initial value problem. Why?

Well, consider the linear homogeneous equation

sin x y'(x) = cos x y(x) …(45)

The function y(x) = sin x, x is a non-zer solution of Eqn. (45), which vanishes
∈R

at many points of R (like x = 0, ± π , ± 2 π ,…..) and also changes sign.

You may now try this exercise.

In Theorem 4 to 6, we have given some properties of linear homogeneous

equations and corresponding initial value problems. But, none of the results stated
asserts the existence of solutions of linear equations or corresponding initial value

91
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

problem. Such results are called qualitative properties of solution of linear

equations and their corresponding initial value problems.

Sometimes equations which are not linear can be reduced to the linear form by
suitable transformations of the variables.

In the next section we shall take up such equations.

3.4 Equations Reducible to Linear Equations

Let us consider an equation of the form

dy
f'(y) dx
+ P(x), f(y) = Q(x) …(46)
where f'(y) is the differential coefficient of f(y).

an interesting feature of Eqn. (46) is that it is a non-linear differential equation of

the first order that can be reduced to the linear form by putting v = f(y). with this
substitution Eqn. (46) reduced to

dy  dy dy 
+ P(x)v = Q(x),  = f' (y)  ,
dx
 dx dx 
which is a linear equation with v as dependent variable and x as independent
variable.

A very important and famous equation of this form, about which we have already
mentioned in Sec 3.2, is known as Bernoulli’s Equation, named after James
Bernoulli, who studied it in 1695 for finding its solution. The equation is or the
form.
dy
dx
+ Py = Qy", …(47)
where P an Q are functions of x above and n is neither zero nor one. Dividing Eqn
(47) by yn, we get
dy
y-n dx
+ P y1-n = Q …(48)

in the year 1696, Leibniz pointed out that Eqn. (48) can be reduced to a linear
equation by takeny1-n as the new dependent variable.

On putting v = y1-n in Eqn. (48), it reduces to

1 dy
+ P v = 0. …(49)
1− n dx

92
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

which is a linear differential in v and x. Eqn. (49) can now be solve by the known
methods.

Note that when n= 0 Eqn. (47)is a linear non-homogeneous equation and when n =
1, Eqn. (47) is a linear homogeneous equation. we now illustrate this method with
the help of a few examples.

dy tany
Example 9: Solve - = (1 + x)ex sec y …(50)
dx 1+ x

Solution: Dividing Eqn. (50) by sec y, we get

dy siny
cos y = = (1+x)ex …(51)
dx 1+ x
If we put sin y = f(y), then f'(y) = cos y and hence Eqn. (51) is of the form

dy 1
f'(y) - f(y) = (1+x+ex …(52)
dx 1+ x
which is of the type (46). To reduce it to linear from, we put
v = f(y) = sin y
Then Eqn (52) reduces to
dv 1
- v = (1+x)ex
dx 1 + x

1 1
it is a linear equation with I.F. = e − ∫ 1+ x dx = e-In(1+x) =
1+ x

Multiplying the above equation by I.F., we get

d  1  1 (1+x) ex.
 v.  =
dx  1 + x  1+ x

Integrating w.r.t.x, we have

1
v = ex + c , c being a constant
1+ x
i.e., v = )1+x)ex + c(1+x).
Substituting sin y for v, the required solution of the given Eqn. (50) is sin y =
(1+x)ex + c(1+x)
Let us look at another example in which n is neither 0 nor 1.

Example 10: Solve y(axy + ex) dx – ex dy = 0

Solution: The given equation can be rearranged as
dy
ex dx
= exy + axy2,

93
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

i.e., dy
dx
- y = axy-xy2 …(53)

it is a Bernoulli’s equation with n = 2.

1
To solve it, let y1-2 = v, i.e., v = .
y
dv 1 dy
∴ =- 2
dx y dx

Consequently, Eqn. (53) reduces to

dv
- - v = axe-x …(54)
dx

it is a linear equation with I.F. = e ∫ 1dx = ex

Multiplying both sides of Eqn. (54) by I.F., we get
d
(vex) = - ax
dx
Integrating w.r.t.x, we get
V ex = - ∫ ax dx + c.
ax 2
=-
+c
2
 1
Replacing v by   , the required solution can be expressed as
 y
 ax 2 
Ex = y  c−  ,
 2 

Remark: There are many second or higher order linear equations which can be
solved easily by reducing them to linear first order equations by making some
transformation of the variables. We shall take up such equations later in Block 2
when we discuss second order equations.

You may now try the following exercises.

You may recall that in Unit 1, we discussed some physical situations expressed in
terms of differential equations. In the following section we have attempted to
solve some of them.

94
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Applications of Linear Differential Equations

Let us consider the problem discussed in Unit 1 one by one.

I. Population model

You may recall that while studying the equation for population problem we had
arrived at the initial value problem. (ref. Eqns. (32) and (33) of unit 1)
d
N(t) = k N(t)
dt
N(t0) = N0 …(55)
Since k is a constant in Eqn. (55), the first of the equations in (55) is a linear
differential equation of order one. From Sec. 3.3 we know that its solution is

N(t) = N(t0) exp (k(t-t0)) …(56)

In Eqn. (56), we normally assume that N(t0) is specified. If k is known then we

can find the solution using relation (56). In reality, it is too hard to measure k
(which gives the rate of growth). In a particular case, we can actually find the
exact value of k if we know the value of N at t1 (t1 ≠ t0). The details are shown in
the following example.

Example 11: Assuming that the rate of growth of a species is proportional to the
amount N(t) present at time t, find the value of N(t) given that N(0) = 100 and
after one unit of time, the size of the specie has grown to 200.

Solution: In this case t0 = 0, N(0) = 100. the solution of the problem is given by
N(t) = 100 exp (kt), t ≥ 0
We determine k from the additional condition N(1) = 200 (N(1) = size of
population at time t = 1).
Thus 200 = 100 exp (k) ⇒ k = In2

Hence the solution is

N(t) = 100 exp (t In2) = 100 exp (In2t)

Or N(t) = (100) 2t

In this problem the constant k has been determined from the given date.

In the following exercise we ask you to solve a similar problem.

Let us now discuss the problem of decay of radioactive material.

95
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

II. Radioactive Decay

In unit 1, (Ref. Eqn. (35)), we have seen that equation which governs the
radioactive decay of a given radioactive material is
y'(t) = k y(t) …(57)

Note: Half-life is the time needed for the material to reduce itself to half of its
original mass.

Where y(t) is the mass of the radioactive material at time t and k < 0 is a real
constant. Eqn. (57) can be used to find the half-life of the radioactive material.

In the following example we consider this problem in detail.

Example 12: A radioactive substance with a mass of 50 gms. Was found to have
a mass of 40 gms. After 30 years. Find its half-life.

Solution: The mass y(t) of the material satisfies

d
y(t) = k y(t)
dt
y(0) = 50 gms .,
y(30 = 40 gms . ...(58)
The solution of the first two equations in Eqns. (58) can be expressed as
y(t) = 50 exp (kt),
Using the third equation in Eqn. (58), we can write.
y(30) = 40 = 50 exp (30k),
or exp (30k) = 4/5,
1  4
i.e., k = In  
30  5

thus, the mass y(t) satisfies

 t 4
y(t) = 50 exp  In  …(59)
 30 5 

50
Let t1 be its half-life, i.e., after time t1 the mass reduces to = 25 gms.
2
Then y(t1) = 25 …(60)

We are required to find t1. using condition (60), Eqn. (59) reduces to

96
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

 t1 
25 = 50 exp  In 4 / 5 
 30 
or t1 In (4/5) = 30 In (1/2)
i.e. t1 = 30 (In (1/2))/In (4/5) …(61)
so after t1 years (t1 defined by Eqn. (61)), the mass of the material will be 25 gms.

Let us now deal with the temperature variations of a hot object.

III. Newton’s Law of Cooling

The temperature of a hot a hot body kept in a surrounding of constant temperature

T0 has been discussed in Unit 1 and the governing equation of the temperature T of
the body is
T'(t) = k(T(t) – T0) …(62)
(Ref. Eqn. (34) of Unit 1)

we illustrate this by the following example.

Example 13: A rod of temperature 100oC is kept in a surrounding of temperature

20oC. If the temperature of the rod was found to be 80oC after 10 minutes, find the
temperature T(t) of the rod.

Solution: We are required to solve

d
T(t) = k(T(t) – 20) …(63)
dt

Let us put y(t) = T(t) – 20. Then y'(t) = T'(t) and Eqn. (63) reduces to
d
y(t) = k y(t) ..(64)
dt

Eqn. (63) is not a linear homogeneous equation whereas Eqn. (64) is which
explains the reason for introducing y) Along with Eqn. (64), we have

a) y(0) = T(0) – 20 = 100 – 20 = 80oC,

b) y(10) = T(10) – 20 = 80 – 20 = 60oC …(65)

the solution of Eqn. (64), with the condition 65(a), is

y(t) = 80 exp (kt)
with this value of y and condition (65b), we have
y(10) = 60 = 80 exp (k. 10)
1 1
or, k = In (6/8) = In (3/4)
10 10

97
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Hence the value of y is determined by

 t 
y(t) = 80 exp  In(.75)  ,
 10 
and the temperature T is given by
 t 
T(t) = 80 exp  In(.75)  + 20
 10 

And now an exercise for you.

4.0 CONCLUSION

We now end this unit by giving a summary of what we have covered in it.

5.0 SUMMARY

In this unit we have covered the following points:

dy
1) The general form of the linear equation of the first order is + P(x)y =
dx
Q(x), (see Eqn. (3))
Where P(x) and Q(x) are continuous real-valued functions on some interval
I ⊆ R.
When Q(x) = 0 it is called homogeneous linear differential equation of
order one.
When Q(x) ≠ 0, it is called non-homogeneous (or inhomogeneous) linear
differential equation of order one.
I.F. for this equation of e ∫ P(x) dx and the general solution is given by
y = e − ∫ P(x) dx . ∫ Q(x) e ∫ P(x) dx
Here, e − ∫ P(x) dx ∫ Q(x) e ∫ P(x) dx is the particular solution of the equation.
2) The sum of the solution of linear non-homogeneous differential equation of
the form (3) and the solution of its corresponding homogeneous equation is
again a solution of the equation.
3) If in the differential equation
dy
+ P(x)y = Q(x),
dx

P(x) is a constant and Q(x) is any of the forms e αx ( α constant), xr(r > 0, an
integer), sin βx or cos βx ( β constant) or a linear combination of such
functions, then method of undetermined coefficients can be applied to find

98
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

the particular solution of the equation and the particular integrals for
different Q(x) are given by the following table

P(x) Q(x) Particular Integral

 e mx
 if m ≠ − a
a(constant emx (m constant)  m+ a
 xe mx if m = − a

 n a i i+ 1
 ∑ i + 1 x if a = 0
n

a
∑ a x (i > 0
i= 0
i
i

 n
i= 0

an integer)  ∑ Pi x i ifa ≠ 0
 i = 0
an 1 na n 
with Pn = , Pn-1 =  a n − 1 − ,
a a a 
1 n− 1 n 
Pn-2 =  a n − 2 − (a n − 1 − a a ) and so on
a a a 
a sin βx or cos βx A linear combination of
sin βx and cos βx
or A sin βx + B
cos βx ( β , A, B
constants)

4) Method of variation of parameters for finding the solution of non-

homogeneous linear differential equation
dy
+ P(x)y = Q(x)
dx
5) Some properties of the solution of linear homogeneous differential equation
dy
+ P(x)y = 0 are
dx
a) (Superposition Principle): If y1 and y2 are any two solutions of
the equation, then y1 + y2 and cy1 are also solution of the
equation, where c is a real constant.
b) If a complex valued function y = p+iq, defined on 1, is a
solution of the equation, then real part of p of y and imaginary
part q of y are also solutions of the equation on 1.
c) If y be a solution of the equation on I such that y(x1) = 0 for
some x1 in I, then y = 0 on I.
6) a) Bernoulli’s equation
dy
+ P(x)y = Q(x) yn,
dx

99
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

where P and Q are functions of x alone and n is neither zer nor one,
reduces to a linear equation by the substitution v = y1-n.
b) Equations of the type
dy
f'(y) + P(x) f(y) = Q(x)
dx
reduce to linear equations by the substitution f(y) = v.
7) The differential equations governing physical problems such as population
model, radioactive decay and Newton’s law of cooling have been solved.

6.0 TUTOR MARKED ASSIGNMENT

1. From the following equations, classify which are linear and which are non-
linear.
Also state the dependent variable in each case.

dy
a) dx
- y = xy2
b) rdy – 2ydx = (x – 2) ex dx.
di
c) - 6i = 10 sin 2t
dt
dy
d) dx
+ y = y2ex
e) ydx + (xy + x-3y) dy = 0
f) (2s – e2t) ds = 2(se2t – cos 2t) dt

2. Prove Theorem 2.

3. Solve the following equations:

dy
c) (x2 + 1) dx
+ 2xy = 4x2
dy 2
d) dx
+ y = sin x
x
dy
e) sec x dx
+ y = sin x
f) (1 + y2 dx = (tan-1y – x) dy
dy
g) (2x – 10y3) dx
+y=0

4. Solve the following equations.

ex
a) y' = y + , e∈ [1, ∞ [
x

100
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

b) y' = y + x + x3+ + x5
c) y' = y + x sin x ex + x5
d) y' + 3y = |x|, y(0) = 1.

dy
5. Solve + y = 2aex
dx

dy
6. Solve dx
= y + x2.

7. Solve the following differential equations:

dy
a) dx
- y = 6 cos 2x
dy
b) dx
+ 3y = x2 + 3e2x + 4 sin x

8. Solve the following equations:

a) y' – 2y = sin π x + cos x, y(1) = 1

π

x 2x
b) y' – y = cos 2x + e + e + x
c) y' – 3y = x2 – cos 3x + 2 (Hint: Treat 2 as 2e0x)
d) y' + y = -x – x2, y(0) = 0.
e) Y' – y = ex, y(0) = -3.

1
9. Show that y1(x) = - is not a solution of Eqn. (37)
(1 + x)

10. Show that the solution y(x) = ex + iex – (1+ x) of equation y' = y + x, for x
∈ I = R does not satisfy the hypothesis of Theorem 5.

11. Show that y(x) = sin 2x – cos 2x, x ∈R

is a solution of y' = 2y + 4 cos 2x.
why does Theorem 6 fail in this case?

12. Solve the following equations:

dy ey 1
a) dx
= −
x2 x
dy
b) dx
+ xy = x3y3
dy
c) 3ex tan y +|(1-ex) sec2y dx
=0

101
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

13. Find the solutions of

dy
a) dx
+ x sin 2y = x3cos2y
dy
b) dx
+ y = ex y3
dy
c) 2x dx
+ y(6y2 –x-1) = 0

14. A culture initially has N0 number of bacteria. At t = 1 hour, the number of

 3
bacteria is measured to be   N0. If the rate of growth is proportional to
 2
the number of bacteria present, determine the time necessary for the
number of bacteria to triple.

15. Suppose that a thermometer having a reading of 70oF inside a house is

placed outside where the air temperature is 10oF. Three minutes later it is
found that the thermometer reading is 25oF. Find the temperature reading
T(t) of the thermometer.

7.0 REFERENCE/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

Indira Gandhi National Open University School Of Sciences Mth-07

102
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

UNIT 2 DIFFERENTIAL EQUATIONS OF FIRST ORDER BUT

NOT OF FIRST DEGREE

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Equations which can be factorized
3.2 Equations which cannot be factorized
3.2.1 Equations soluable for y
3.2.2 Equations soluable for x
3.2.3 Equations in which independent variables or dependent
variable is absent
3.2.4 Equations Homogeneous in x and y
3.2.5 Equations of the first degree in x and y – clairant’s
3.2.6 Riccati’s Equation
3.3 Bernoulli Equation
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 References/Further Readings

1.0 INTRODUCTION

In unit 1, we discussed the nature of differential equations and various types of

solutions of differential equations. In Unit 2 and 3, we have given you the
methods of solving different types of differential equations of first order and first
degree. In this unit we shall consider those differential equations which are of first
order but not of first degree.

dy
If we denote by P, then the most general form of a differential equation of the
dx
first order and nth degree can be expressed in the form
Pn + P1pn-1 + P2pn-2 + …..+ Pn-1p + Pn = 0, …(1)
Where P1, P2,….., Pn are functions of x and y.

It is difficult to solve Eqn. (1) in its most general form. In this unit we shall
consider only those forms of Eqn. (1) which can be easily solved and discuss the
methods of solving such equations.

103
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

It was Isaac Newton (1642 – 1727), the English mathematician and scientist, who
classified differential equations of the first order (then known as fluxional
equations) in “Methodus Fluxionum et serierum infinitarum”, written around 1671
and published in1736. count Jacopo Riccati (1676-1754), an Italian
mathematician, was mainly responsible for introducing the ideal of Newton to
Italy. Riccati was destined to play an important part in further advancing the
theory of differential equations. In 1712, he reduced an equation of the second
order in y to an equation of first order in p. in 1723, he exhibited that under some
restricted hypotheses, the particular equation to which the name of Riccati is
attached, can be solved.

Later the French mathematician Alexis Claude Clairaut (1713-1765) introduced

the idea of differentiating the given differential equations in order to solve them.
He applied it to the equations that now bear his name and published the method in
1734. we shall also be discussing the equations introduced by Riccati and Clairaut
in this unit.

Clairaut (1713-1765)

2.0 OBJECTIVES

After studying this unit, you should be able to

• Find the solution of the differential equations which can be resolved into
rational linear factors of the first degree;
• Obtain the solution of equations solvable for y, x or p;
• Obtain the solution of the differential equations in which x or y is absent;
• Solve the equations which may be homogeneous in x and y;

104
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

• Identify and obtain the solution of Clairaut’s equation;

• Identify and obtain the solution of Riccati’s equation.

3.0 MAIN CONTENT

3.1 Equations which can be factorized

Let us consider the general form of differential equation of the first order and nth
degree given by Eqn. (1) namely,

Pn + P1pn-1 + P2pn-2 + …..+ Pn-1p + Pn = 0,

Where P1, P2,….., Pn are functions of x and y.

For this equation, we shall consider two possibilities:

a) When the left-hand side of Eqn. (1) can be resolved into rational factors of
the first degree.
b) When the left-hand side of Eqn. (1) cannot be factorized.
In this section we shall take up the first possibility.

When Eqn. 91) can be factorized into rational factors of the degree, then it can
take the form

(p – R1) (p – R2)……(p – Rn) = 0 …(2)

for some R1, R2, …., Rn, which are functions of x and y.

Eqn. 91) will be satisfied by a value of y that will make any of the factors in Eqn
(2) equal to zero. Hence, to obtain the solution of Eqn. (1), we equate each of the
factors in Eqn. (2) equal to zero. Thus, we get
P – R1 = 0, p – R2 = 0, …, p – Rn = 0 …(3)

There are n equations of first degree. Using the methods given in Unit 2 and 3 we
can now obtain the solution of the above n equations of first order and first degree.

Let us suppose that the solutions desired for Eqn. (3) are

f 1 (x, y, c1 ) = 0 
f 2 (x, y, c 2 = 0 
 …(4)
;; 
f n (x, y, c n ) = 0
where c1, c2,…., cn are the arbitrary constants of integration.

105
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Since each of the constants c1, c2,….., cn can take any one of an infinite number of
values, thus these solutions remain general even if

C1 = c2 = ….= cn = c, say.
In that case, the n solutions will be

f1(x, y, c) = 0
f2(x, y, c) = 0
f3(x, y, c) = 0
….
….
Fn(x, y, c) = 0

These n solutions can be left distinct or we can combine them into one equation,
namely,
F1(x, y, c). f2(x, y, c) ….fn(x, y, c) =

The reason of taking all c1, c2,…., cn equal in Eqn. (4) is the fact that Eqn. (2)
being of first order, its general solution can contain only one arbitrary constant.

We illustrate this method by the following examples:

Example 1: Solve p2 +px + py+xy = 0

Solution: The given equation is equivalent to

(p+x) (p+y) = 0
that is, either
p+x = 0 or, p+y = 0
In other words,
dy dy
+ x = 0, or +y=0
dx dx

the solutions of the factors are

2y = -x2 + c
and
x = -In |y| + c, for c being an arbitrary sonstant.

Therefore, the general solution of the given equation is

(2y + x2 – c).(x+In |y| - c) = 0.

Let us look at another example.

Example 2: Solve p3 (x + 2y) + 3p2(x + y) + (y + 2x)p = 0

106
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: The given equation is equivalent to

p[p2(x + 2y) + 3p(x + y) + (y + 2x)] = 0
⇒ p[p2(x + 2y) + p{(y + 2x) + (x + 2y)} + (y + 2x)] = 0
⇒ p(p + 1) [(x + 2y)p + (y + 2y)] = 0
its component equations are
p = 0, p + 1 = 0, (x + 2y)p + (y + 2x) = 0
dy
Now p = 0 ⇒ = 0, which has the solution
dx
y=c …(5)

dy
Now p+1 = 0 ⇒ +1=0
dx
i.e., dy + dx = 0
which has the solution
y+x =c …(6)
Further, (x + 2y)p + (y + 2x) = 0
⇒ (x + 2y)dy + (y + 2x)dx = 0
⇒ d(xy + x2 + y2) = 0.
Which has the solution
xy + 2 + y2 = c …(7)

Therefore, the general solution of the given equation, from Eqns. (5), (6) and (7),
is
(y –c). (y+x – c). (xy + x2 + y2 – c) = 0.

You may now try the following exercise.

As you know from algebra, every equation over Q need not have all its roots in Q,
i.e., it need not be factorizable in Q.

We now take up those equations of form (1) which cannot be factorized into
rational factors of the first degree.

3.2 Equations which cannot be Factorized

in this case, let the form of Eqn. (1) be

f(x, y, p) = 0 ….(8)

eqn. (8) is not solvable in its most general form

we shall discuss only those equations of type (8) which possess one or more of the
following properties.

107
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

i) It may be solvable for y.

ii) It may be solvable for x.
iii) It may solvable for p.
iv) Either it may not contain y or it may not contain x, that is, either x or y is
absent from the differential equation.
v) It may be homogeneous in x and y.
vi) It may be of first degree in x and y.
vii) It may be Riccati’s equation.

We now discuss these cases one by one.

3.2.1 Equation Solvable for y

Consider an equation
xp2 – yp – y = 0 …(9)

we can write Eqn. (9) in the form

y(p+1) = xp2,
xp 2
⇒ y=
p+ 1

That is, Eqn. (9) can be solvable for y in terms of x and p.

Similarly when Eqn. (8), i.e., f(x, y, p) = 0 is solvable for y, then it can be pt in the
form

y = F(x, p) …(10)

Differentiating Eqn. (10) w.r. to x, we get an equation of the form

 dp 
P= ø  x, p,  …(11)
 dx 

Eqn. (11) is in two variables x and p; and we may possibly solve and get a relation
of the type
ψ (x, p, c) = 0 …(12)
for some constant c.

It we now eliminate p between Eqns. (8) and (12), we get a relation involving x, y
and c, which is the required solution. In the cases when the elimination of p
between Eqns. (8) and (12) is not possible, we then obtain the values of x and y in
terms of p as a parameter and these together give us the required solution.

108
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

We now illustrate this method with the help of a few examples.

Example 3: The given equation is solvable for y.
Solving it for y, we get
x
y=p+ …(13)
p
Differentiating Eqn. (13) w.r. to x, we get
dp 1  1  dp
p= + + x  − 2 
dx p  p  dx
 1  dx 1
i.e.,  p−  + 2 x=1 …(14)
 p  dp p

this is linear equation of the first order if we consider p as independent variable

and x as dependent variable.

We can write Eqn (14) as,

dx 1 p
+ x= 2 …(15)
dp p(p− 1)(p+ 1) p −1

1
For Eqn. (15) e ∫ p(p 2
− 1)
dp
is an integrating factor.
1  1 1 1
Now, e ∫ p(p 2
− 1)
dp
= e ∫  2(p − 1) + 2(p + 1) − p  dp
(p 2 − 1)1 / 2 (p 2 − 1)1 / 2
= e In p =
p
the, the solution of Eqn. (15) is obtained as
(p 2 − 1)1 / 2 p (p 2 − 1)1 / 2 1
x
p
= ∫ p2 − 1 p
dp = ∫ p −1
2
dp = c + cos h-1p,
-1 2 -1/2
or x = p(c + cos h p) (p -1) …(16)

you may notice that elimination of p between Eqns. (13) and (16) is not easy.
However, by substituting for x from Eqn. (16) in Eqn. (13),
we get

y = p+(c + cosh-1p) (p2 -1)-1/2 …(17)

Eqns. (16) and (17) are two equations for x and y in terms of p. These are the
parametric equations of the solution of the given differential equat.

Let us look at another example.

Example 4: Solve y = 2px + p4x2, x > 0.

109
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: The given equation

y = 2px + p4x2 …(18)
is in itself solvable for y.

differentiating it w.r. to x, we get

dp dp
p = 2p + 2x + 2xp4 + 4x2p3
dx dx
dp
⇒ p(1 + 2xp3) + 2x (1+2xp3) = 0,
dx

dp
⇒
(1+2xp3) + (p+2x dx
)=0 ...(19)
dp
Eqn. (19) holds when either of the factors (1+2xp3) or (p + 2x dx
) is zero.
First consider the factor

dp
P+ 2x dx
=0
2 dp
⇒ + 1 =0
pdx x

Integrating the above equation w.r.t.x, we get

2 In |p| + In |x| = constant.
⇒
p2x = c, (c an arbitrary constant)
c
or p = .
x

Substituting this value of p in the given Eqn. (18), we get

Y = 2 cx+ c 2
Which is the required solution.

If we consider the factor 1 + 2xp3 = 0 in Eqn. (19), then by eliminating p between

this factor and given Eqn. (18), we get another solution. This solution will not
contain any arbitrary constant and is the singular solution of the given equation.

How about trying an exercise now?

We next consider the case when Eqn. (8) is solvable for x.

3.2.2 Equations Solvable for x

Consider an equation of the form

110
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

P3 – 4xyp + 8y2 = 0 ...(20)

It is difficult to solve Eqn. (20) for y whereas it is easy to solve it for x as a
function of y and p and write

p3 + 8 y 2
x= .
4 yp

In such cases when equation of the form (8) is solvable for x, and can be put in the
form
x = g(y,p) …(21)
then to solve it, we differentiate Eqn. (21) w.r. to y, and get an equation of the
form
1  dp 
= ø  y, p, 
p  dx 
on solving this equation we obtain a relation between p and y in the form
f(y,p,c) = 0, …(22)
where c is an arbitrary constant.

Now, we may eliminate p between Eqn (21) and (22) to obtain the solution or, x
and y may be expressed in terms of p as we have done in Sec. 4.3.1.

Remark: Note that when Eqn. (8) is solvable for y, we differentiate it w.r. to x,
whereas, when it is solvable for x, we differentiate it w.r. to y.

We illustrate this method by the following examples.

 p 
Example 5: Solve p = tan  x − .
 1 + p 2 

Solution: The given equation can be written as

p
X = tan-1p + ...(23)
1 + p2

Differentiating Eqn. (23) w.r. y, we get

1 p dp (1 + p 2 ) − p(2 p) dp
= +
p 1 + p 2 dy (1 + p 2 ) 2 dy

1 + p2 + 1 + p2 − 2 p2
=
(1 + p 2 ) 2

111
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

2 dp
=
(1 + p ) dy
2 2

2p
⇒
dy = dp
(1 + p 2 ) 2
…(24)

Note that Eqn. (24) is in variable separable form.

Integrating Eqn. (24), we get

1
y=c- , …(25)
1 + p2
c being an arbitrary constant.

It is not possible to eliminate p between Eqns. (23) and (25). Thus, Eqns. (23) and
(25) together constitute the solution of the given equation in terms of parameter p.

Let us look at another example.

Example 6: Solve p2y + 2px = ∀ x, y and p > 0

Solution: We can write the given equation in the form

y py
X= − ...(26)
2p 2

Differentiating Eqn. (26) w.r. to y, we get

1 1 y  1  dp p y dp
= +  − 2  − + ,
p 2 p 2  p  dy 2 2 dy
1 p y dp  1
⇒
+ +  1 + 2  = 0
2 p 2 2 dy  p 
1 + p2 y  1 + p 2  dp
+   =0
2  p 2  dy
⇒

2p
1 + p2  y dp 
⇒
 1 +  = 0 …(27)
2p  p dy 

In Eqn. (27), we may have

1 + p2
=0
2p

112
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

 y dp 
or  1 +  = 0.
 p dy 
If we have first factor equals zero, then p2 = -1.

Thus real solution of the given problem is obtained when

1 1 dp
+ dy =0
y p

Here variables are separable. Integrating, we get

Iny + Inp = Inc
⇒
py = c
c
or p = …(28)
y

Eliminating p between Eqn.s (26) and (28), we get

y2 c y
X= -
2c y 2
y2 c
or x = - ,
2c 2
which is the required solution.

2p
Note that you could also have solved Example 6 by taking y = and then
1 − p2
proceeding as in sec. 4.3.1.

You may now try the following exercise.

We now consider Eqn. (8) with the property that Eqn. (8) may be solvable for p.
in that case Eqn. (8) which is of nt degree in p, in general, is reduced to n
equations of the first degree and this case has been considered in Section 4.2.

We next take up the case when Eqn. (8) may not contain either independent
variable x, or, dependent variable y explicitly.

3.2.3 Equations in which Independent Variable or Dependent Variable

is Absent

We shall consider the two cases separately.

113
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Case 1: Equations not containing the independent variable:

When Eqn. (8) does not contain independent variable explicitly then the equation
has the form

f(y,p) = 0 …(29)

For instance, consider the equation

1
y- = 0.
1 + p2

This equation does not contain x explicitly. Also, it is readily solvable for y, since
it can be written in the form

1
y= …(30)
1 + p2

Eqn. (30) can, now, be solved by the method discussed in Sec. 4.3.1. In case Eqn.
(29) is solvable for p, then we can write it in the form

dy
p= = ø(y) …(31)
dx

The integral of Eqn. (31) will, then, give us the solution of Eqn. (29).

To be more clear, let us consider the following example.

Example7: Solve y = 2p + 3p2

Solution: We have

y = 2p + 3p2
which is already in the form y = F(p). following the method discussed in sec.
4.3.1, we differentiate it w.r.t.x, so that

dp dp
p=2 dx
+ 6p dx

p dp
or =
2 + 6p dx

Here variable are separable and we have

114
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

2 
dx =  + 6  dp
p 
Integrating, we get
x = 6p + 2In |p| + c. …(33)
c being an arbitrary constant.
Since it is not possible to eliminate p from Eqns. (32) and (33), these equations
together yield the required solution in terms of the parameter p.

Let us look at another example

Example 8: Solve y2 = a2 (1 + p2) …(34)

Solution: The given equation is an equation in y and p only. It can be written as

y2
P2 = -1
a2
Solving for p, we get
y2
p= ±
a2 − 1

∴ Either p = y2 y2
or p = - 2 ,
a2 − 1 a −1
y2
Now p − 1 + 2 gives
a
a
dy = dx.
y − a2
2

Integrating the above equation, we get

A In |y + y 2 − a 2 | = x + c,
c being an arbitrary constant.

y2
Similarly, p = - − 1+ , on integration, yields
a2
a In |y + y 2 − a 2 | = -x + c (c being a constant).

Hence, the general solution of the given equation is

[a In |y + y 2 − a 2 | - x – c] [a In |y + y 2 − a 2 | + x – c] = 0

115
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Note that we solve Eqn. (34) for p. You could also have integrated it by solving it
for y.

We next consider the equations in which the dependent variable is absent.

Case II: Equations not containing the dependent variable:

In this case Eqn. (8) has the form

g(x, p) = 0 or x = F(p) …(35)

As in case 1, Eqn. (35) is either solvable for p or solvable for x. if it is solvable

for p, then it can be written as

p = ψ (x)
which, on integration, gives the solution of Eqn. (35)

If Eqn. (35) is solvable for x, then it corresponds to the case discussed in Section
4.3.2.

We give below examples to illustrate the theory.

Example 9: Solve x(1+p2) = 1

Solution: The given equation can be written as

1
x= …(36)
1 + p2

Differentiating Eqn. (36) w.r. to y, we get

1 − 2p dp
= ,
p (1 + p 2 ) 2 dy

− 2 p2
i.e., dy dp
(1 + p 2 ) 2
 −1 1 
i.e., dy = 2  +  dp
1 + p (1 + p 2 ) 2 
2

Here variables are separable. Integrating, we get

dp
y = - 2 tan-1p + 1 ∫ +c …(37)
(1 + p 2 ) 2
c is a constant.

116
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Eqns. (36) and (37) together yield the required solution with p as parameter.

Note that problem in example 9 could have also been done by solving it for p. we
illustrate this method in the next example.

Example 10: Solve p2 – 2xp + 1 = 0

Solution: The given equation is

P2 – 2xp + 1 = 0
Solving for p, we get

2 x± 4 x2 − 4
p= =x ±
x2 − 1
2
∴

Either p = x + x 2 − 1 or p = x - x2 − 1
Now p = x + x 2 − 1 , on integration yields

x2 x x2 − 1 1
y= + - In |x + x 2 − 1 | + c,
2 2 2

c being an arbitrary constant.

Similarly, p = x - x 2 − 1 yields

x2 1 1
y= - x x 2 − 1 + In |x + x 2 − 1 | + c,
2 2 2

Hence, the general solution of the given equation is

[x2 + x x 2 − 1 - In |x + x 2 − 1 | - 2y + c1] [x2 – x x 2 − 1 + In|x + x 2 − 1 |-2y + c1]
= 0.
Where c1 = 2c is an arbitrary constant.

And now some exercise for you.

We next discuss the case when Eqn. (8) may be homogeneous in x and y

3.2.4 Equations Homogeneous in x and y

In this case, Eqn. (8) can be expressed in the form

 y
Ø  p,  = 0 …(38)
 x

117
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

For solving Eqn. (38), we can proceed in two ways. In case Eqn. (38) is solvable
for p then it can be expressed as

dy  y
P= = f  …(39)
dx  x

We already know from our knowledge of unit 2 that equations of the type (39) can
be solved by using the substitution y = vx.

The second possibility, that is, when Eqn. (38) is solvable for y/x, then it can be
put in the form

y
= ψ (p) or y = x ψ
(p).
x

In this case we can proceed as in Sec. 4.3.1. Differentiating the above equation
w.r. to x, we get

dp
p= ψ
(p) + x ψ
'(p) dx

dx ψ' (p) dp
⇒
= …(40)
x p − ψ(p)
Eqn. (40 is in variable separable form. On integrating, it yields

ψ' (p)
In|x| = c + ∫ p− ψ(p)
dp
= c + ø(p), say.

The elimination of p between this equation and y = x (p) will give us the
ψ

required solution. But it is not always easy to eliminate p, so it may be retained as

the parameter.

To understand the theory, we take an example.

Example 11: Solve y2 + xyp – x2p2 = 0 ∀

x, y, p > 0.

Solution: The given equation is homogeneous in y and x and it may be written as

2
 y  y
2
p -   p-   =0 …(41)
 x  x

118
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solving Eqn. (41) for p, we get

(y/x) ± (y/x)2 + 4(y/x)2  1± 5 

p= = (y/x)  
2  2 
dy y  1± 5  dy  1± 5  y
Thus =   or =  
 2   2 
dx x dx x

dy dy
Let y = vx, then dx
=v+x dx

dv  1+ 5  dv  1− 5 
∴

v+x = v  and v + x = v
dx  2  dx  2 
dv  5 − 1 dv  −1− 5 
⇒
x =  v and x =  v
dx  2  dx  2 

Integrating, we get

In xc = In v2/(√5-1) and In xc = In v-2/(√5+1)

Or xc = (y/x) 2/(√5-1) and xc = (y/x) -2/(√5+1)
or y = x (xc)(√5-1)/2 and y = x(xc)-(√5+1)/2

Hence the general solution is

[y-x(xc) (√5-1)/2]. [y-x(xc) -(√5+1)/2] = 0

i.e., y2 – xy [(xc) (√5-1)/2 + (xc) -(√5+1)/2] + x2 (xc)-1 = 0

Now you may try the following exercise.

We next discuss the case when Eqn. (8) may be of first degree in x and y.

3.2.5 Equations of the First Degree in x and y – Clairaut’s equation

When Eqn. (8) is of first degree in x and y, it is solvable for x and y both and
hence can be put in either of the following forms.

y = xf1(p) + f2(p) …(42)

or x = yg1(p) + g2(p) …(43)

119
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Hence, we can use the methods discussed in Sec. 4.3.1 and 4.3.2 to solve
equations of the for (42) and (43), respectively.

However, if in Eqn. (42), f1(p) = p, then we get one particular form of this equation
known as Clairaut’s Equation and about which we have already mentioned in
sec. 4.1.

Thus the Clairaut’s equation is of the form

y = px + f(p) …(44)

In Eqn. (44), f(p) is a known function which contains neither x nor y explicitly.
also, note, that Eqn. (44) can be non-linear. For instance, y = px + p2 and y = x +
ep are examples of Caliraut equation. But equations y = xy + p2 or y = xp + yp2 are
not of the Clairaut’s form.

On differentiating Eqn. (44) w.r. to x, we have

p = p + p'x + f'(p)p' …(45)
⇒
p' [x + f'(p)] = 0
dp
Then either p' = dx
=0 …(46)
Or, x + f'(p) = 0 …(47)

The solution of Eqn. (46) is p = c, where c is an arbitrary constant. Thus, we can

write the general solution of Eqn. (44) as

y = cx + f(c) …(48)
Note that Eqn. (48) is an equation of a family of straight lines.

Now consider Eqn. (47). Since f(p) and f'(p) are known functions of p, Eqns. (47)
and (44) together constitute a set of parametric equations giving x and y in terms
of the parameter p.

If we can eliminate p from Eqn. (44) and (47) and if the resulting equation satisfies
Eqn. (44), we get another solution of Eqn. (44) (could be an implicit solution).
This solution does not contain an arbitrary constant and is a singular solution of
Eqn. (44).

We give you some examples to help you understand this method.

Example 12: Solve (y')2 + 4xy' – 4y = 0 …(49)

Solution: With p = y', Eqn. *49) can be written as

120
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1 2
y = px + p, …(50)
4
which is in the Clairaut’s form. Differentiating Eqn. (50) w.r. to x, we get
p
p = p + p'x + p'
2
p
⇒
p'(x + ) =0
2

then either p = 0 which gives p = c ( a constant) …(51)

p
or x + =0 …(52)
2

From Eqns. (50) and (51), we obtain

c2
y = cx +
4
as the solution of Eqn. (50). Eliminating p from Eqns. (50) and (52), we get
1
y = x(-2x) + ( -2x)2,
4
i.e., y(x) = - x2,
which contains no arbitrary constant. Since this value of y satisfies Eqn. (50), it is
the singular solution of Eqn. (50).

Let us look at another example.

1
Example 13: Solve y = xp + p

Solution: If we compare the given equation with Eqn. (44) we notice that in the
1 1
case f(p) = and f'(p) . From Eqn. (48) then the solution is given by
p p2

1
y = ax +
a
where a ( ≠ 0) is an arbitrary constant.
Also in this case, equation corresponding to Eqn. (47) is

1
0=x
p2
The elimination of p between the above equation and the given equation yields
y2 = 4x,

121
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

which is a singular solution of the given equation.

You may, now, try the following exercises.

Finally, we take up in the next solution, another non-linear equation known as

Riccati’s equation, which we mentioned in Sec. 4.1

3.2.6 Riccati’s Equation

Originally, this name was given to the first order differential equation

dy
dx
+ by2 = cmm, …(53)
where b, c and m are constants. This is known as the special Riccati equation.
Eqn. (53) is solvable in finite terms only if the exponent m is – 2 or, of the form
− 4k
for some integer k. Riccati merely discussed special cases of this equation
(2 k + 1)
without offering any solutions. Now a days Riccati’s equation is usually
understood by an equation of the form.

y' = a(x) + b(x) y + c(x)y2 …(54)

where a, b and c are given functions of x on an interval 1 (of R). Equations y' = 1
+ xy + exy2 and y' = x + x2y + sin (x)y2 are example of Riccati’s equations whereas,
equations y' = 1 + y + y3, and y' = 1 + y + 2y2 are not of Riccati’s type.

It is difficult to obtain a solution of Riccati’s Eqn. (54) containing an arbitrary

constant. But, the general solution of Eqn. (53) can be obtained if we have the
knowledge of a particular solution of Eqn. (53). This can be done as follows:

Let y1 be a solution of Eqn (53) Then we determine a function v so that y defined

bu the relation

1
y = y1 + …(55)
v
is a solution of Eqn. (54).
Differentiating Eqn. (55) w.r. to x, we get

 1
y' = y1 + v  − 
 v2 
Since y and y1 satisfy eqn. (54), we have
y1' = a(x) + b(x)y1 + c(x)y12

122
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

v' 2
and y1' - 2 = a(x) + b(x)y + c(x) y
v

subtracting the second equation from the first, we have

 1
v'  2  = b(x) (y1-y) + c(x) (y1 – y )
2 2
v 
or v' = b(x) v2(y1 – y) + c(x) (y1 –y) (y1 + y)v2 …(56)
From Eqn. (55), we have
(y – y1) v = 1 or (y1 – y) v = - 1 …(57)
Also, (y1 + y)v = (2y1 + y – y1)v = 2y1v + 1 (y – y1)v = 2y1v+1 (using Eqn (57))
Now (y12 – y2)v2 = (y1 – y)v.(y1 + y)v
= ( - 1) (2y1v + 1) = - 1- 2y1v …(58)

Substituting from Eqn (58) in Eqn (55), we get

v' = - (b(x) + 2c(x)y1)v – c(x), …(59)
which is a linear (non-homogeneous) equation for determining a function v.

the general solution of Eqn. (59) contains an arbitrary constant and the substitution
of this general solution in Eqn. (54 gives us the solution of Eqn. (53) containing an
arbitrary constant.

Let us now go through some examples to understand the above theory.

Example 14: Solve y' = - y + x2y2

Solution: On comparing this equation with Eqn (53) we find that in this case a =
0, b = -1 and c = x2 (1 = R). The given equation is a Riccati’s Equation which has
a (particular) solution y1 = 0. by using the substitution

1 1 1
y = y1 + =0+ v
= v
v
in Eqn (58), we have v' = v – x2 …(60)

Eqn. (60) is a linear first order equation with

I.F. = e ∫ − 1 dx = e-x

Hence, the general solution of Eqn. (60) is

v = - ex ∫ e-x x2+ dx + Aex
= (x2 + 2x + 2) + Aex
and the solution of the given equation is

123
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1
y= ,
Ae + x + 2 x + 2
x 2

which contains an arbitrary constant.

Let us look at another example.

Example 15: Solve y' = - 1- x2 + y2.

Solution: By inspection, we see that y1(x) = - x is a solution of the given

equation. Comparing the given equation with Eqn. (53), we get a = -1- x2, b = 0
and c = = 1.

1 1
We look for a function v, so that y = y1 + v
=-x+ v
.

In this case Eqn. (58) reduces to

v' = 2xv – 1
dv
⇒
- 2xv = -1
dx

The integrating factor for this equation is e-x2, and so it can be written as
d
[e-x2v] = - e-x2
dx
Thus, on integration, we write
e-x2v = - ∫ e dx + c
− x2

or, v = ex2 [- ∫ e dx + c]
− x2

where c is an arbitrary constant.

So, the required solution is

e− x 2
y(x) = - x +
− ∫ e− x 2 + c

Now the integral ∫e

− x2
dx cannot be evaluated in terms of elementary functions.
x

∫e
− t2
When an initial condition is specified, then integral of the form dt can be
x0

used.

You may now try the following exercises:

3.3 Bernoulli Equation

124
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Reduction of non-linear equation to linear form.

Here, we shall illustrate that certain non-linear first order differential equations
may be reduced to linear form by a suitable change of the dependent variable.

The differential equation

y' + p(x)y = g(x)ya
where ‘a’ is a real number is call the “Bernoulli Equation”.

For a = 0 and a= 1 the equation is linear, and otherwise it is non-linear.

Set {y(x)}1-a = u(x) and show that the equation assume the linear form.

U' + (1 – a) p(x) u = (1-a) g(x).

Solve the following Berboulli equation

x
1) y' + y =
y
2) y' + xy = xy-1
3) 3y' + y = *1 – 2x( y4
4) (A population model . the logistic law). Matheus’s law states that the time
rate of change of a population y(t) is proportional to x(t). this holds for
many populations as they are not too large. A more refined model is the
logistic law given by

dy
= ay – by2……… (a > 0, b> 0)
dt
where the “breaking term” = by2 has the effect that the population cannot
grow indefinitely.

Solve this Bernoulli equation. What is the limit of y(t) as t as ∞ ? For the
united state, vertalst predicted in 1845 the values a = 0.03 and b = 1.6 x 10 -4
where t is measured in years and y(t) in millions find the particular solution
satisfying y(0) = 5.3 (corresponding to the year 1800) and compare the
values of this solution with some actual values.

1800 1830 1860 1890 1920 1950 1980

5 – 3 13 31 63 105 150 230

5) Apply the suitable substitutions, reduce to linear form and solve:

a) y' cos y + x sin y = 2x

125
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

b) exy1 + 2exy + x = 0

4.0 CONCLUSION

We end this unit by giving a summary of what we have covered in it.

5.0 SUMMARY

in this unit we have covered the following:

1) the general differential equation of first order and nth degree is given by
Eqn (1), namely
pn + p1pn-1 + P2pn-2 + ….+ Pn-1p + Pn = 0
dy
where P1, P2 ........, Pn are functions of x and p = dx
.

2) If Eqn. (1) can be resolved into rational linear factors of the first order, then
Eqn. (1) takes the form
(p-R1 (p- R2) …… (p – Rn) = 0
for some R1, R2,……Rn which are functions of x and y, and
if f1(x, y, c) = 0, f2(x, y, c) = 0 ….., fn(x, y, c) = 0 are the solutions of p – R1
= 0, p – R2 = 0 ….., p – Rn = 0 respectively, then
f1(x, y, c) . f2(x, y, c)…… fn(x, y, c) = 0
is the general solution of Eqn. (1).

3) If Eqn. (1) cannot be factorized into rational linear factors, the

a) it is said to be solvable for y if we can express it as
y = F(x,p) (see Eqn. (10)).
To solve Eqn. (10), differentiate it with respect to x, and it may be
possible to solve resulting differential equation in x and p.
Elimination of p between the solution of resulting differential
equation and EQn. (10) gives the solution of Eqn. (10).

b) it is said to be solvable for x if we can express it as

x = g(x, p) (see Eqn. (21)).
To solve Eqn. (21), differentiate it w.r.t.y and it may be possible to
solve the resulting differential equation in y and p. Elimination of p
between the solution of the resulting equation and Eqn. (21) gives
the solution of Eqn. (21).

4) If Eqn. (1) does not contain independent variable or dependent variable

explicitly and can be put in the form

126
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

f(y, p) = 0 (See Eqn. (29))

or g(x, p) = 0 (see Eqn. (35))
then it may either be possible to factorize Eqn. (29) into linear factors or it
may be solvable for y.
Similarly, eqn. (35) can either be factorized or it may be solvable for x.

5) If Eqn. (1) is homogeneous in x and y then either substitution y = vx may

reduce it to separable equation or it may be put as
y = x (p), ψ

which is solvable for y or x.

6) Clairaut’s equation is an equation of first order and of any degree if it can

be put in the form
y = xp + f(p) (see Eqn. (44))
This equation is solvable for y and its solution is
y = cx + f(c)

7) Riccati’s equation is an equation of the form

dy
dx
= a(x) + b(x)y + c(x(y2 (see Eqn. (53))
where a(x), b(x) and c(x) are given functions on an interval ! of R.

The general solution of Eqn. (53) can be obtained if we know a particular

solution y1 of Eqn. (53( and then we determine a function v defined by
relation

1
y = y1 + v
, (see Eqn 54). …(see Eqn 54)
so that y is solution of Eqn. (54).

6.0 TUTOR MARKED ASSIGNMENT

1. Solve the following equations:

a) p2y + p(x – y) – x = 0
b) p2 – 5y + 6 = 0
c) 4y2p2 + 2pxy (3x+1) + 3x3 = 0
3
 dy 
d)   = ax4
 dx 
e) x + yp2 = p(1 + xy)

2. Solve the following equations:

127
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

a) y = x + a tan-1p
b) x = y + Inp
c) p3 + p = ey
d) y = p tan p + In cos p

3. Solve the following equations.

a) p2 – py + x = 0
b) x = y + a Inp
c) x = y + p2
d) y2Iny = xyp + p2

4. Solve the following equations:

a) (y')2 – 4 = 0
b) sin(y') = 0
c) (y')2 + 4y' – x2 = 0

5. Obtain the solution of the following equations:

a) exp (y' + (1 + x2)) = 1

b) (y')2 + 2 (x + y)y' + 4xy = 0
c) p2 – (3x +2y)p + 6xy = 0

6. Solve the following equations.

a) y = yp2 + 2px
b) x2p2 + 4xyp – 8y2 = 0

7. Solve the following equations:

a
a) y = xp + (a ≠
0, is a constant)
p
b) y = xp + p2
c) y = xp + p – p2

8. Solve e4x (p – 1) + e2y p2 = 0

9. Solve y = x4p2 – px

10. Solve xy (y – px) = x + py

128
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

11. Which of the following are Riccati’s equation, Clairaut’s equation or

neither.

a) y = 2xp + y2p3
b) y' = ex + ey + y2
2
c) y' = (1 + sin 2x) + y + exy2
1 + x2
d) y = 3px + 6y2p2
e) y' = sin x + sin y

12. Find a solution, containing an arbitrary constant (given a particular

solution), of the following Riccati’s equations:

a) y' = 1 – xt + y2 (y1(x) = x)
b) y' = 2 + 2x + x2 – y2 (y1(x) = 1 + x)
c) y' = 2x – x2 – x2 – x4 + y + y2 (y1(x) = x2)

13. By eliminating arbitrary constant c from the equation

cg(x) + G(x)
y=
cf(x) + F(x)
obtain the Riccati’s equation:
(gF – Gf)y' = (gG' – g'D) + (Gf' – gF' + g'F)y + (fF' – fF)y2.

14. Show that, when m = 0, Riccati’s equation

dy
dx
+ by2 = cxm

7.0 REFERENCES/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

Indira Gandhi National Open University School Of Sciences Mth-07.

MODULE 3

129
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

UNIT 1 FAMILIES OF CURVES ORTHOGONAL AND

OBLIYUE TRAYECTORIES. APPLICATION TO
MECHANICS AND ELECTRICITY

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Orthogonal Trajectories
3.2 Application to Orthogonal Trajectories
3.3 Approximate Solutions:
Direction fields, Iteration
3.3.1 Method of Direct Fields
3.3.2 Picards iteration method
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 References/Further Readings

1.0 INTRODUCTION

In this section we shall learn how to use differential equations for finding curve
that intersect given curves at right angles, a task that arises rather often in
applications.

If for each fixed real value of c the equation

(1) F(x, y, c) = 0
represents a curve in the xy-plane and if for variable c it represents
infinitely many curves, then the totality of these curves is called a one-
parameter family of curves, and c is called the parameter of the family.

2.0 OBJECTIVES

• To find family of curves

• To be able to solve differential equation of family of curves
• One should be able to apply orthogonal trajectories to (i) Electrical field. (2)
Mechanical field. (3) Temperature
• Also to be able to find approximate solutions to directions fields iteration.
Example 1: Families of curves
The equation

130
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

(2) F(x, y, c) = x + y + c = 0
represents a family of parallel straight lines; each line corresponds to
precisely one value of the parameter c. The equation

(3) F(x, y, c) = x2 + y2 – x2 = 0

Represents a family of concentric circles of radius c with center at the

origin.

The general solution of a first-order differential equation involves a parameter c

and thus represents a family of curves. This yields a possibility for representing
many one-parameter families of curves by first-order differential equations. The
practical use of such representations will become obvious from our further
considerations.

Example 2: Differential equations of families of curves

By differentiating (2) we see that

y' + 1 = 0

is the differential equation of that family of straight lines. Similarly, the

differential equation of the family (3) is obtained by differentiation, 2x + 2yy' = 0,
that is,

y' = x/y.

if the equation obtained by differentiating (1) still contains the parameter c, then
we have to eliminate c from this equation by using (1). Let us illustrate this by a
simple example.

Example 3: eliminate of the parameter of a family

The differential equation of the family of parabolas

(4) y = cx2
is obtained by differentiating (4),

(5) y' = 2cx,

and by eliminating x from (5). From (4) we have c = y/x 2, and by
substituting this into (5) we find the desired result

(6) y' = 2y/x.

131
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

note that we may also proceed as follows. We solve (4) for c, finding c = y/
x2, and differentiate

3.0 MAIN CONTENT

3.1 Orthogonal Trajectories

In many engineering and other applications, a family of curves is given, and it is

required to find another family whose curves intersect each of the given curves at
right angles.14 The curves of the two families are said to be mutually orthogonal,
they form an orthogonal net, and the curves of the family to be obtained are called
the orthogonal trajectories of the given curves (and conversely); cf. fig. 1.

Let us mention some familiar examples. The meridians on the earth’s surface are
the orthogonal trajectories of the parallels. On a map the curves of steepest
descent are the orthogonal trajectories of the contour lines. In electrostatics the
equipotential lines and the lines of electric force are orthogonal trajectories of each
other. An illustrative example is shown in Fig. 2. We shall see later that
orthogonal trajectories are important in various fields of physics, for example, in
pydrodynamics and heat conduction.

Given a family of curves F(x, y, c) = 0 that can be represented by a differential

equation

(7) y' = f(x, y)

we may find the corresponding orthogonal trajectories as follows. From (7) we

see that a curve of the given family that passes through a point (x0, y0) has the slop
f(x0, y0) at this point. The slop of the orthogonal trajectory through (x 0, y0) at this
point should be the negative reciprocal of (x0, y0), that is, - 1/f(x0, y0), because this
is the condition for the tangents of the two curves at x0, y0) to be perpendicular.
Consequently, the differential equation of the orthogonal trajectories is

(8)
1
y' = −
f(x, y)

and the trajectories are obtained by solving this new differential equation.

132
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Fig. 1: Curves and their Fig. 2: Equipotential lines

orthogonal trajectories and lines of electric force
(dashed) between two concentric
cylinders
14
Remember that the angle of intersection of two curves is defined to be the angle
between the tangents of the curves at the point of intersection.

Example 4: Orthogonal trajectories

Find the orthogonal trajectories of the parabolas in Example 3.

Solution: From (6) we see that the differential equation (8) of the orthogonal
trajectories is
1 x
y' = - =−
2 y/x 2y
By separating variables and integrating we find that the orthogonal tracjectories
are the ellipses
x2
+ y2 = e* …(fig. 3)
2

Example 5: Orthogonal trajectories

Find the orthogonal trajectories of the circles

9) x2 + 2(y – c)y' = 0.

Solution: we first determine the differential equation of the given family, by

differentiating (9) with respect to x we obtain

133
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

10) 2x + 2(y – c)y' = 0.

We must eliminate c. solving (9) fir x, we have

x 2 + y2
11) c=
2y

By inserting this into (10) and simplifying we get

y2 − x 2 2 xy
x+ y' = 0 or y' = 2 2 .
2y x −y
From this and (8) we see that the differential equation of the orthogonal
trajectories is

x 2 − y2
y' = - or 2xyy' – y2 + x2 = 0.
2 xy

The orthogonal trajectories obtained by solving this equation (cf. example 1 in sec.
1.4) are the circles (fig. 4)

(x – c)2 + y2 = c

Fig. 3: Parabolas and their orthogonal Fig. 4: Circle and their orthogonal
trajectories in Example 4 trajectories (dashed) in Example 5

In the next section we motivate and discuss two methods of obtaining approximate
solutions without actually solving a given differential equation. The first method,
called the method of direction fields, can relatively easily produce a general
picture of the solutions (with limited accuracy) and is of great practical interest.
The second method, Picard’s iteration, is more theoretical; its practical value is
limited, since it involved integrations.

134
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

SELF ASSESSMENT EXERCISES

i. 4y – x + c = 0
ii. (x – c)2 + y2 = 4
iii. (x – c)2 + y2 = c2/2
iv. x2 – y2 = c

Represent the following families of curve in the form (1), stktch some of the
curves.

v. All nonvertical straight lines through the point (4, – 1).

vi. The catenaries obtained by translating the catenary y = cosh x in the
direction of the straight line y = - x.

Using differential equations, find the orthogonal trajectories of the following

curves. Graph some of the curves and the trajectories.

1) y = 2x + c
2) y = cx3

3.2 applications of Orthogonal Trajectories

3) (Electric field) If an electrical current is flowing in a wire along the z-exist,

the resulting equipotential lines in the xy-plane are concentric circles about
the origin, and the electric lines of force are the orthogonal trajectories of
these circles. Find the differential equation of these trajectories and solve
it.
4) (Electric field) Experiments show that the electric lines of force of two
opposite charges of the same strength at ) – 1, 0) and (1, 0) are the circles
through ( - 1, 0) and (1, 0). Show that these circles can be represented by
the equation x2 + (y – c)2 = 1 + c2. show that the equipotential lines
(orthogonal trajectories) are the circles (x + c*)2 + y2 = c*2 – 1, which are
dashed in Fig. 5 on the next page.

Other forms of the differential equations. Isogonal trajectories

5) Show that (8) may be written in the following form and use this result for
determining the orthogonal trajectories of the curves y = x+ c .
dx
= - f(x, y).
dy
6) Show that the orthogonal trajectories of a given family g(x, y) = c can be
obtained from the differential equation

135
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy ∂ g/ ∂ y
=
dx ∂ g/ ∂ x
7) Isogonal trajectories of a given family of curves are curves that intersect
the given curves at a constant angle θ . Show that at each point the slops m1
and m2 of the tangents to the corresponding curves satisfy the relation

m 2 − m1
= tan θ = const.
1 + m1m 2
Using this formula, find curves which intersect the circles x2 + y2 = c at an angle of
45o.

8) Using the Cauchy – Riemann equations (Prob. 43), find the orthogonal
trajectories of ex cos y = c.

3.3 Approximate Solutions: Direction Fields, Iteration

In applications, it will often be impossible or not feasible or not necessary to solve

differential equation exactly. Indeed, there are various differential equations, even
of the first order, for which one cannot obtain formulas for solutions. 15 There are
other differential equations for which such formulas can be derived, but they are
so complicated that they are practically useless. Finally, since a differential
equation is a model of a physical or other system, and in modeling we disregard
factors of minor influence in order to keep the model simple, the differential
equation will describe the given situation only approximately, and an approximate
solution will often be practically as informative as an exact solution.

Approximate solutions of differential equations can be obtained by numerical

methods. These are discussed in sec 20.1 and 20.2. at present we shall consider
the method of direction field, which is a geometric procedure, and then the so-
called Picard iteration, which gives formulas for approximate solutions.

3.3.1 Method of Direction Fields

In this method we get a rough picture of all solutions of a given differential

equation

1)
y' = f(x, y)

without actually solving the equation. The idea is quite natural and simple, as
follows.

136
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

We assume that the function f is defined in some region of the xy-plane, so that at
each point in that region it has one (and only one) value. The
15
Reference [All] in Appendix 1 includes more than 1500 important differential
equations and their solutions, arranged in systematic order and accompanied by
numbers references to original literature.

Solutions of (1) can be plotted as curves in the xy-plane. We do not know the
solutions, but we see from (1) that a solution passing through a point (x0, y0) must
have the slope f(x0, y0) at this point. This suggests the following method.

Ist Step (Isoclines). We graph some of the curves in the xy-plane along which
f(x, y) is constant. These curves

f(x, y) = k = const

are called curves of constant slope or isoclines. Here the value of k differs from
isoclines to isocline. So these are not yet the solution curves of (1), but just
auxiliary curves.

2nd Step (Direction field). Along the isocline f(x, y) = k we draw a number of
parallel short line segments (lineal elements) with slope k, which is the slop of
solution curves of (1) at any point of that isocline. This we do for all isoclines
which we graphed before. In this way we obtain a field of lineal elements, called
the direction field of (1).

3rd Step (Approximate solution curves). With the help of the lineal elements
we can now easily graph approximation curves to the (unknown) solution curves
of the given equation (1) and thus obtain a qualitatively correct picture of these
solution curves.

It suffices to illustrate the method by a simple equation that can be solved exactly,
so that we get a feeling for the accuracy of the method.

Example 1: Isoclines, direction field

Graph the direction field of the first-order differential equation

2) y' = xy

and an approximation to the solution curve through the point (1, 2). Compare with
the exact solution.

137
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: The isoclines are the equilateral hyperbolas xy = k together with the
coordinate axes. We graph some of them. Then we draw lineal elements by
sliding a triangle along a fixed ruler. The result is shown in Fig. 6, which also
shows an approximation to the solution curve passing through the point (1, 2).

By separating variables, y = ce x . The initial condition is y(1) = 2. Hence 2 =

2
/2

ce1/2, and the exact solution is

y = 2 e(x .
2
− 1) / 2

Fig. 7: Direction field of the differential equation (2)

3.3.2 Picard’s Iteration method16

This method gives approximate solutions of an initial value problem

3)
y' = f(x, y), y(x0) = y0

Which is assume to have a unique solution in some interval on the x-axist

containing x0. Picard’s existence and uniqueness theorem, which we shall discuss
in the next section. Its practical value is limited because it involves integrations
that may be complicated.

The basic idea of Picard’s method is very simple. By integration we see that (3)
may be written in the form

138
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

4)
x
y(x) = y0 + ∫
x0
f[t, y(t)] dt

Where t denotes the variable of integration. In fact, when x = x0 the integral is

zero and y = y0, so that (4) satisfies the initial condition in (3); furthermore, by
differentiating (4) we obtain the differential equation in (3).

To find approximations to the solution y(x) of (4) we proceed as follows. We

substitute the crude approximation y = y0 = const on the right; this yields the
presumably better approximation
x
y1(x) = y0 + ∫x0
f(t, y 0 ) dt .

In the next step we substitute the function y1(x) in the same way to get
x
y2(x) = y0 + ∫ x0
f[t, y1 (t)] dt

etc. The nth step of this iteration gives an approximating function

5) x
yn(x) = y0 + ∫x0
f[t, y n − 1 (t()] dt .

In this way we obtain a sequence of approximations.

y1(x), y2(x),……….. yn(x),………,

and we shall see in the next section that the conditions under which this sequence
converges to the solution y(x) of (3) are relatively general.
16
EMILE PICARD (1856 – 1941), French mathematician, professor in Paris since
1881, also known for his important contributions to complex analysis (see Sec
14.10 for his famous theorem).

An iteration method is a method that yields a sequence of approximations to an

(unknown) function, say, y1, y2 …., where the nth approximation, yn, is obtained in
the nth step by using one (or several) of the previous approximations, and the
operation performed in each step is the same. This is a practical advantage, for
instance in programming for numerical work.

139
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

In the simplest case, yn is obtained from yn-1; denoting the operation by T, we may
write

yn = T(yn-1).

Picard’s method is of this type, because (5) may be written

x
Yn)x) = T(yn-1(x)) = y0 + ∫x0
f(t, y n − 1 (t)) dt .

To illustrate the method, let us apply it to an equation we can readily solve exactly,
so that we may compare the approximations with the exact solution. The example
to be discussed will also illustrate that the question of the convergence of the
method is of practical interest.

Example 2: Picard iteration

Find approximate solution to the initial value problem

y' = 1 + y2, y(0) = 0

Solution: In this case, x0 = 0, f(x, y) = 1 + y2, and (5) becomes

x x
yn(x) = ∫
0
[1 + y 2n − 1 (t)] = x + ∫
0
y 2n − 1 (t) dt .
Starting from y0 = 0, we thus obtain (cf. Fig. 8)
x
y1(x) = x + ∫0
0 dt = x

x 1 3
y2(x) = x + ∫0
t 2dt = x +
3
x

2
 t3  x 1 2 1
y3(x) = x + ∫0  t +  dt = x + x 3 + x 5 + x 7
 3 3 15 63

etc. of course, we can obtain the exact solution of our present problem by
separating variables (see Example 2 in Sec 1.2), finding

1 3 2 5 17 7  π π
6) y(x) = tan x = x + x + x + x + ... − < x <  .
3 15 315  2 2

140
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

The first three terms of y3(x) and the series in (6) are the same. The series in (6)
converges for |x| < π /2. This illustrates that the study of convergence is of
practical importance.

The next section, the last of Chap. 1, concerns the problems of existence and
uniqueness of solutions of first-order differential equations. These problems are
of greater relevance to engineering applications than one would at first be inclined
to believe. This is so because modeling involves the discarding of minor factors,
and in more complicated situations it is often difficult to see whether some
physical factor will have a minor or major effect, so that one may not be sure
whether a model is faithful and does have a solution, or a unique solution, even
though the physical system can be expected to behave reasonably. The matter
becomes even more crycial in connection with numerical methods: make sure that
the solution exists before you try to compute it.

Fig 8: Approximate solutions in Example 2

SELF ASSESSMENT EXERCISES

Direction fields

In each case draw a good direction field. Plot several approximate solution curves.
Then solve the equation analytically and compare, to get a feeling for the accuracy
of the present method.

i. y' = - y/x
ii. y' = - x/y
iii. y' = x + y

141
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

iv. 4yy' + x = 0
v. (Verhulst population model) Draw the direction field of the differential
equation in Prob. 54 of Sec. 1.7, with a = 0.03 and b = 1.6. 10 -4 and use it to
discuss the general behavior of solutions corresponding to initial greater
and smaller than 187.5
vi. Apply Picard’s method to y' = y, y(0) = 1, and show that the successive
approximations tend to y = ex, the exact solution.
vii. In Prob. 12, compute the values y1, y2(1), y3(1) and compare
them with the exact value y(1) = e = 2.718........

Apply Picard’s method to the following initial value problems. Determine

also the exact solution. Compare.

viii. y' = xy, y(0) = 1

ix. y' = 2y, y(0) = 1
x. y' – xy = 1, y(0) = 1.

4.0 CONCLUSION

We now end this unit by giving a summary of what we have covered in it.

5.0 SUMMARY

Applications are included at various places. The unit part entirely devoted to
applications are 3.3.1 and 3.3.2. on separable and linear equation respectively.
And are applied to electrical circults and on orthogonal trajectories, that is curves
that intersect given curves at right angles.

Direction field (3.3.1) help in sketching families of solutions curves, for instance,
in order to gain an impression of their general behaviour.

Picard’s iteration method gives approximate solutions of initial value problems by

iteration.

6.0 TUTOR MARKED ASSIGNMENT

In each case draw a good direction field. Plot several approximate solution curves
then solve the equation analytically and compare, to get a feeling for the accuracy
of the present method;

y
1) y' = -
x

142
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

x
2) y’ = -
y
3) y' – x + y
4) 4yy' + x = 0
dy
5) draw the direction field of the differential equation = ay – by2 a>0, b>
dt
0. with a = 0.03 and b= 1.6 x 10-1 and use it to discuss the general behaviour
of solutions corresponding to initial conditions greater and smaller than
187.5.
dy
6) apply Picard’s method to = y, y(0) = 1 and show that the successive
dx
approximations tends to y = ex, the exact solutions.
dy
7) In the i.e. + 3y = e2x+ 6, compute the values y1(1), y2(1), y3(1) and
dx
compute them with the exact value y(1) = e = 2.718.

Apply Picard’s method to the following initial value problems. Determine also the
exact solution.
Compare

dy
8) a) = y, y(0) = 1
dx
dy
9) b) = 2y, y(0) = 1.
dx
dy
10) c) - xy = 1, y(0) = 1.
dx

7.0 REFERENCES/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

Indira Gandhi National Open University School Of Sciences Mth-07.

143
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

MODULE 4

Unit 1 Higher Order Linear Differential Equations

Unit 2 Method of Undetermined Coefficients
Unit 3 Method of Variations of Parameters

UNIT 1 HIGHER ORDER LINEAR DIFFERENTIAL

EQUATIONS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 General Equation
3.1.1 Conditions for the Existence of a unique solution
3.1.2 Linear Dependence and Independence
3.2 Elementary Properties of the Solutions
3.3 Method of Solving Homogeneous Equation with constant
Coefficients
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 References/Further Readings

1.0 INTRODUCTION

In Unit 1 we discussed the basic concepts related to ordinary differential

equations. Further in the introduction to Block 2, we have mentioned that the
governing differential equations in many physical or biological problems are not
necessarily of first order. Besides the differential equations arrived at, in
discussing the above said models may be linear or non-linear. Even among linear
differential equations, the coefficients of the differentials may be constants or a
function of an independent variable. In this unit we classify the general linear
differential equations into two broad categories:

i) homogeneous and non-homogeneous

ii) equations with constant coefficients and variable coefficients.

For a general linear differential equation with variable coefficients, we shall state
the conditions under which a unique solution can be found. Further, we shall learn

144
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

methods of finding the complete solutions of homogeneous differential equations

with constant coefficients.

2.0 OBJECTIVES

After reading this unit, you should be able to

• Identify linear differential equations with constant as well as with variable

coefficients.
• Identify homogeneous and non-homogeneous linear differential equations.
• Describe the conditions under which a unique solution of a linear differential
equation exists.
• Write the complete primitive of a given differential equation when its various
independent integrals are known.
• Classify solutions of non-homogeneous equations into complementary function
and particular integral.
• Obtain a solution for a homogeneous linear differential equation with constant
coefficients.

3.0 MAIN CONTENT

3.1 General Equation

We begin our discussion by considering the most general linear differential

equation which is of the form

dn y d n − 1y dy
a0 (x) + a1 (x) + ….+ a n-1 (x) + an (x) y = b (x) …(1)
dx n dx n − 1 dx

For, a0(x) = 0, the differential equation is of nth order. The coefficients a0(x),
a1(x), …., an(x) are functions of independent variable x. Eqn. (1) is called general
linear differential equation of nth order with variable coefficients.

In case coefficients a0(x), a1(x),….an(x) are all constants and do not depend on x,
then Eqn. (1) will be termed as general linear differential equation of nth order
d3 y d2 y
with constant coefficients. For example, equation 3
+ 3 2
+ y = x2 is a
dx dx
third order linear differential equation with constant coefficients.

Further, the right hand side of Eqn. (1), i.e., b(x) may assume one of the following
forms:

145
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

i) b(x) = 0
ii) b(x) = constant
iii) b9x) a function of x.

when b(x) = 0, Eqn. (1) is classified as the general homogeneous linear

differential equation. This is also known as the reduced equation of Eqn. (1).
For example, equation

d3 y d2 y dy
-4 + + 6y = 0
dx 3
dx 2
dx
is a third order linear differential equation. But if b(x) in Eqn. (1) is a constant or
a function of x, then Eqn. (1) is called general non-homogeneous linear
differential equation.

d4 y d3 y
Equation 4
+ 3
+ 3y = x2 + 1 is a linear non-homogeneous equation of 4th
dx dx
d2 y dy
order with constant coefficients; where equation x3 + x2 + xy = 2 is a
dx 2
dx
second order non-homogeneous linear differential equation with variable
coefficient.

Now suppose that we are required to find the solution of Eqn. (1) on some interval
1 which also satisfy, at some point x0 = ∈ 1 the conditions,
− 1)
y(x0) = y0, y'(x0) =,…., y(n-1) (x0) = y(n
0 …(2)

Note: Depending on the context, i could represent ]a, b[ ]0, ∞ [, ] - ∞ , ∞ [ and so

on.

− 1)
Where y0, y'0 , …., y(n0 are arbitrary constants, then Eqns. (1) and (2) together
constitute an initial-value problem (IVP). The values y(x0) = y0 y'(x0) = y'0 ,…..,
− 1) − 1)
y(n
0 (x0) = y(n
0 are called initial conditions.

In the case of a linear second order equation, we can interpret geometrically a

solution to the initial value problem

d2 y dy
an-2(x) + an-1(x) dx
+ an(x) y = b(x)
dx 2

y(x0) = y0, y'(0) = y'0

146
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

as a function defined on I whose graph passes through (x0, y0) such that the slope
of the curve at the point is the number y . '
0

You may note here that an equation of the form (1) may not always have a
solution. Moreover, even if its solution exists it may not be unique.

Let us now study the conditions under which the solution of Eqn. (1), if it exists
shall be unique.

3.1.1 Conditions for the Existence of a unique solution

We may write the general non-homogeneous linear differential Eqn. (1) in the
form

L(y) = [a0(x) Dn+ + a1(x) Dn-1 + … + an-1(x) D + an(x)] y = b(x) …(3)

dy 2 d2 n dn
Where D = ,D = , …, D = .
dx dx 2 dx n

The expression in the parentheses in Eqn. (3) is termed as a symbolic polynomial

or operator polynomial or simply a differential operator.
Thus we have herein introduced linear differential operator L of order n given by
the expression

L = a0(x) Dn + a1(x)Dn-1 + … + an-1(x) D + an(x) …(4)

In unit 8, we shall learn, in more details, about the differential operators and their
properties.

We now choose an interval I = [ α , β ] for α , β real and assume that the coefficients
a0(x), a1(x), …., an(x) and the function are continuous one-valued functions of x
throughout the interval and that a0(x) does not vanish at any point of the interval.

We know that the complete solution of Eqn. (3) shall involve arbitrary constants
whose number is equal to the order of the highest derivative involved in it, i.e., n
in this case. In order to obtain a unique solution of Eqn. (3), it is necessary to
specify n initial conditions in terms of constant values of

dy d n − 1y
y, , ….,
dx dx n − 1

at any point x0 of the interval [ α , β ].

147
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

We now state a theorem which gives the conditions whose fulfillment guarantee
the existence and uniqueness of the solution of Eqn. (3).

Theorem 1: If the functions a0(x), a1(x), …., an(x) and b(x) are continuous
function of x in the interval [ α , β ] and a0(x) does not vanish at any point of that
interval, then the initial Eqn. (3) admits of a unique solution of the form y = f(x),
which together with its first (n-1_ derivatives, is continuous in [ α , β ] and satisfies
the following initial conditions:

 dy   d n − 1y 
y(x0) = y0   = y ,….,  n − 1 
'
= y(n0 − 1) ,
 dx  x = x
0

 dx  x = x
0 0

where x0 is a point of the interval [ α , β ].

We shall not be proving this theorem here as it is beyond the scope of the present
course. However, if the functions a0(x), a1(x),…., an(x) are constants, we shall
give the solution of the corresponding equation in Sec. 5.4 when b(x) = 0 and in
units 6, 7 and 8 when b(x) ≠ 0.

We now illustrate this theorem with the help of a few examples.

Example 1: Show that y = 3e2x + e-2x – 3x is a unique solution of the initial value
problem

y" – 4y = 12x
y(0) = 4 h'(0) = 1.

Solution: We have y = 3e2x + e-2x – 3x, therefore,

y' = 6e2x – 2e-2x 3 and y" = 12e2x + 4e-2x
Now y" = 4y = 12e2x + 4e-2x – 4(3e2x + e-2x – 3x)
= 12e2x + 4e-2x – 12e2x – 4e-2x + 12x
= 12x
Also, y(0) = 3e2.0 + e-2.0 – 3.0 = 4
Y'(0) = 6e2.0 – 2e-2.0 – 3 = 1.

Thus, y = 3e2x + e-2x – 3x is a solution of the given initial value problem.

Moreover, the given differential equation is linear and the coefficients as well as
b(x) = 12x are continuous on any interval containing x = 0. we conclude from
Theorem 1 that the given function is the unique solution of the given initial vlue
problem.
Remember that both the requirements in Theorem 1, that is ai(x), i = 0, 1, …., n
be continuous and a0(x) 0 for every x in some interval say I are important.
≠

148
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Specifically, if a0(x) = 0 for some x in the interval then the solution of a liner
initial value problem may not be unique or may not even exist.

We now illustrate this through an example.

Example 2: Obtain the value of c for which the function

y = cx2 + x + 3

is a unique solution of the initial value problem

x2y" – 2xy' + 2y = 6,
y(0) = 3, y'(0) = 1
On the interval ]- ∞
,∞
[.

Solution: Since y' = 2cx + 1 and y" = 2c, it follows that

X2y" – 2xy' + 2y = x2(2c) – 2x(2cx + 1) + 2(cx2 + x + 3)
= 2cx2 – 4cx2 – 2x + 2cx2 + 2x + 6
=6
2
Also, y(0) = c.(0) + 0 + 3 = 3
and y' (0) = 2c.0 + 1 = 1

Thus, y = cx2 + x + 3 is a solution of the given problem for all values of c in the
given interval. The problem does not have a unique solution. In this case
although the given equation is liner and its coefficients and b(x) = 6 are continuous
everywhere but the coefficient of y" i.e., x2 is zero at x = 0.

You may now try the following exercise.

You might be familiar with the liner dependence and independence of a set of
functions on an interval. Before we study some elementary properties of the
solution of linear differential equations, ;et us recall these two concepts which are
basic to the study of linear differential equations.

3.1.2 Linear Dependence and Independence

We begin with the following two definitions.

Definition: A set of function y1(x), y2(x),…., yn(x) is said to be linearly

independent on an interval I if there exist constants. C1, c2, …., cn(x) not all zero,
such that

C1y1(x) + c2y2(x) +….+ cnyn(x) = 0

149
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

For every x in the interval.

Definition: A set of functions y1(x), y2(x),…., yn(x) is said to be linear

independent on an interval I, if it is not linearly dependent on the interval.

In other words, a set of functions is linearly independent on an interval if the only

constant for which

c1y1(x) + c2y2(x) + ….. + cnyn(x) = 0,

For every x in the interval, are c1 = c2 = …. = cn = 0.

It is easy to understand these definitions in the case of two functions y1(x) and
y2(x). if the functions are linearly dependent on an interval, then there exists
constants c1 and c2, both are not zero, such that for every x in the interval

c1y1(x) + c2y2(x) = 0

Since c1 ≠
0, it follows that

c2
y1(x) = - y1(x),
c1

That is, if two functions are linearly dependent, then one is a constant
multiple of the other. Conversely, if y1(x) = c2y2(x) for some constant c2, then
(- 1) y1(x) + c2y2(x) = 0
for every x on some interval. Hence the functions are linearly dependent, at least
one of the constants (namely, c1 = -1) is not zero. We thus conclude that two
functions are linearly independent when neither is a constant multiple of the
other on an interval.

Functions, y1(x) = sin2x and y2(x) = sinx cosx are linearly dependent on the
interval ]- , [ since c1 sin2x + c2 sinx cosx = 0 is satisfied for every real x
∞ ∞

with

1
C1 = and c2 = - 1.
2

In the consideration of linear dependence or linear independence, the interval on

which the functions are defined is important. We now illustrate it through an
example.

Example3: Show that the function y1(x) = x and y2(x) = |x| are

150
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

i) linearly independent on the interval ] - , ∞ ∞

[.
ii) linearly dependent on the interval ]0, [. ∞

Solution:

(i) it is clear that in the interval ] - ∞

, ∞
[ neither function is a constant
multiple of the other (see Fig. 1)

Y Y
y1 = |x |
y1 = x

o x o x

Fig. 1

Thus in order to have c1y1(x) + c2y2(x) = 0 for every real x, we must have c1 = 0
and c2 = 0.

(ii) For y1(x) = x and y2(x) = |x| in the interval ]0, [ ∞

c1x + c2 |x| = c1x + c2x = 0

is satisfied for any non zero choice of c1 and c2 for which c1 = - c2.
Thus y1(x) ands y2(x) are linearly dependent on the interval ]0, [. ∞

You may try the following exercises:

The procedure given for examining the linear dependence or independence of a set
of functions appears to be quite involved. We, therefore, outline below sufficient
condition of examining the linear independence of a set of n functions.

Suppose that y1(x), y2(x),…., yn(x) are n functions on an interval I possessing

derivative upto (n – I)th order. If the determinant.

151
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

y1 y2 ⋯ yn
y1' y2 '
⋯ y'n
W(y1(x), y2(x), …., yn(x)) = . . ⋯ .
. . ⋯ .
y1(n − 1) y(n2 − 1) ⋯ y(nn − 1)
Is not zero for at least one point in the interval I, then the functions y1(x), …., yn(x)
are linearly independent on the interval.

This provides a sufficient condition for the linear independence of n functions on

an interval. The determinant W(y1(x), y2(x),…., yn(x)) is called the Wronskian of
the functions. It is named after a Polish mathematician Josef Maria Hosene
Wornski (1778 – 1853).

The functions y1(x) = sin2x and y2(x) = 1-cos 2x, for instance are linearly
dependent on ] ,
∞
[ because
∞

sun x 1 − cos 2x
2

= 2sin2x sin 2x – 2sin x cos x + 2sin x cos x 2x

2sin x cox 2sin 2x
= sin 2x [2sin2x – 1 + cos 2x]
= sin 2x [2sin2x – 1 + cos2x – sin2x]
= sin 2x [sin2x + cos2x – 1]
=0
in example 3 we saw that y1(x) = x and y2(x) = |x| are linearly independent on ] -
∞
, ∞
[. However, we cannot compute the Wronskian as y2 is not differentiable at
x = 0.

Remember that in the above condition the non-vanishing of the Wronskian at a

point in the interval provides only a sufficient condition. In other words, if
W(y1y2,…., yn) = 0 for x in an interval, it does necessarily mean that the functions
are linearly dependent on the interval. We leave it for you to verify it yourself.

With above background in mind we are now set to study the elementary properties
of the solutions of linear differential equations.

3.2 Elementary Properties of the Solutions

The general homogeneous linear differential equation corresponding to Eqn. (3) is

L(y) = [a0(x)Dn + a1(x)Dn-1 + …+ an-1(x)] y = 0

152
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

i.e., L(y) = ∑ a (x) D

r= 0
r
n− r
y=0 …(5)

We can clearly think of the form of the solutions of linear differential equations by
making use of the following elementary theorems:

Theorem 2: If y = y1 is a solution of Eqn. (5) on an interval 1, then y = cy1 is also

its solution on I, where c is any arbitrary constant.

Proof: We know that

Dr (cy1) = cDry1
n

Also, L(cy1) = ∑ a (x) D

r= 0
r
n− r
(cy1 )
n

=c ∑ a (x) D
s= 0
s
n− s
y1

= c L(y1).
= 0 (∴ L(y1) = 0)
Thus, if y = y1 is a solution of Eqn (5), so y = cy1. for instance, the function y = x2
is a solution of the homogeneous linear equation.

X2y" – 3xy' + 4y = 0 on ]0, ∞

[.

Hence y = cx2 is also solutions. For various of c, we see that y = 3x 2, y = ex2, y = 0

… are all solutions of the equation on the given interval.

Have you notice that a homogeneous linear differential equation always

possesses the trivial solution y = 0? If not, you can check it now.

Now let us look at another property of the solutions of linear differential

equations.

Theorem 3: If y = y1, y2, …, ym are m solutions of homogeneous linear

differential Eqn. (5) on an interval 1, then y = c1y1 + c2y2 + …. + cmym is also a
solution of Eqn. (5) on 1, where c1, …., cm are arbitrary constants.

Proof: If y1 (i = I, …, m) are solutions of Eqn. (5) then

L(y1 = 0 (for i = 1, 2, …., m) …(6)
We know that
Dr [c1y1) + c2y2 + …+ cmym]
= Dr (c1y1) + Dr (c2y2) + …+ Dr (cmym)
= c1 Dr (y1) + c2 Dr (y2) + …+ cm Dr (ym)
now, L(c1y1 + c2y2 + …+ cmym)

153
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

n
= ∑ ar(x) Dn-r (c1y1 + c2y2 + … + cmym)
r= 0
n n n
= c1 ∑ ar(x) Dn-r y1 + c2 ∑ ar(x) Dn-r y2 + … + cm ∑ ar(x) Dn-r ym
r= 0 r= 0 r= 0

= c1 L(y1) + c2 L(y2) + … + cm L(ym) (using Eqn. (5))

= c1.0 + c2.0 +… + cm.0 (using Eqn. (6))
=0

Hence if y1, y2 …. ym are solutions of Eqn. (5), then y = c1y1 + c2y2 + … + cmym is
also a solution of Eqn. (5).
Theorem 3 is known as the superposition principle.

Let us now consider an example.

Example 4: Show that if y1 = x2 and y2 = x2 Inx are both solutions of the equation
x3y" = 2xy' + 4y = 0 on the interval ]0, ∞ [. Then c1x2 + c2x2 Inx is also a solution
of the equation on the given interval.

Solution: We have y = c1x2 + c2x2 Inx

Now y' = 2c1x + 2c2x Inx + c2x
y" = 2c1 + 2c2x Inx + 3c2
2 c2
y" =
x

therefore, x3y" – 2xy' + 4y

 2 c2 
= x3   - 2x (2c1x + 2c2x Inx + c2x) + 4c1x2 + 4c2x2 Inx
 x 
= 2c2x2 – 4c1x2 – 4c2x2 Inx – 2c2x2 + 4c1x2 + 4c2x2 Inx
=0

Thus, y = c1x2 + c2x2 Inx is also a solution of the equation on the interval.
Theorem 2 and 3 represent properties that non-linear differential equations, in
general, do not possess. This will become more clear to you after you have done
the following exercises.
Let us now consider the following definition which involves a linear combination
of solutions.

Definition: Let y1, y2, …, yn be n linearly independent solutions of homogeneous

linear differential Eqn. (5) of degree n on an interval I. Then

y = c1y1(x) + c2y2(x) + … + cnyn(x),

154
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

where c1, i = 1, 2, …, n are arbitrary constants is defined to be the general

solution or the complete primitive of Eqn. (5) on I.

The above definition automatically generates our interest in knowing when n

solutions, y1, y2, …, yn of the homogeneous differential Eqn. (5) are linearly
independent. Surprisingly, the nonvanishing of the Wronskian of a set of n such
solutions on an interval I is both necessary and sufficient for linear
independence.
That is,

If y1, y2, …, yn) be n solutions of homogeneous linear nth order differential EQn.
(5) on an interval I, then the set of solutions is linearly independent on I if and
only if

W(y1, y2, …, yn) ≠ 0

For every x in the interval. Such a set y1, …., yn of n linearly independent
solutions of Eqn. (5) on I is said to be a fundamental set of solutions on the
interval.

For instance, the second order equation y" – 9y = 0 possesses two solutions

y1 = e3x and y2 = 3-3x

e3 x e− 3x
3x -3x
Since W (3 , e ) = 3 x = -6 ≠ 0
3e − 3e− 3x
For every value of x, y1 and y2 form a fundamental set of solution on ]- ∞ , ∞ [.
The general solution of the differential equation on the interval is

y = c1 e3x + c2 e-3x

so far we have discussed the properties pertaining to the solution of homogeneous

linear equations. We now turn our attention to the non-homogeneous linear
equation. To this effect, we consider a theorem due to D’ Alembert (1762 –
1765) which defines the general solution of a non-homogeneous linear equation.

Theorem 4: If y = Y0(x) is any solution of the non-homogeneous linear

differential Eqn. (3) on an interval I and if y = Y(x) is te complete primitive of the
corresponding homogeneous linear differential Eqn. (5) on the interval, then

Y0(x) + Y(x)

155
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Is the general solution of Eqn. (3) on the given interval.

Proof: Since Y0(x) is a solution of Eqn. (3),

∴ L[Y0(x)] = b(x) ...(7)

also, Y(x) is the complete primitive of Eqn. (5),

∴

L[Y0(x)] = 0 …(8)

further, in Theorem 2 and 3 above, we have seen that the operator Dr and linear
differential operator L are distributive. Thus, using relations (7) and (8), we get

L(y) = L[Y0(x) + Y(x)],

= L[Y0(x)] + L[Y(x)]
= B(x) + 0

Thus, y = Y0(x) + Y(x) is a solution of Eqn. (3).

Since y(x) = Y0(x) + Y(x) involves n arbitrary constants (due to presence of n

arbitrary constants in Y(x)), it is, therefore, the general solution of Eqn. (3).

If Y(x) is chosen as to satisfy the condition (2) and if Y 0(x), for some point x0 of
the interval I, is such that

 dY0   d 2 Y0   d = − 1 Y0 
Y0(x0) = 0 =   =  
2  =…=  
n− 1  ,
 dx  x = x 0  dx  x = x 0  dx  x= x0
Which is possible provided that b(x) is not identically zero, then the solution

y = Y0(x) + Y(x) …(9)

also satisfies the conditions.

 dy   d n− 1y 
  = 0 ,….,  n − 1 
'
y(x0) = y0, dx y = y0n-1
  x= x0  dx  x= x0
we usually refer the solution of Eqn. (5) in the form (9) as the general solution of
the non-homogeneous linear differential Eqn. (3) and it consist of two parts:

i) The complete primitive of Eqn. (5) (the corresponding homogeneous part of

Eqn. (3)) in the form

Y(x) = c1y1(x) + c2y2(x) + …. + cnyn(x).

156
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Which contains n arbitrary constants. The solution y = Y(x) of Eqn. (5) is

known as complementary function of Eqn (3). We denote the
complementary function
Y)x) by yc(x).

ii) Any solution y = Y0(x) of Eqn. (3), (which cannot be obtained by assigning
any particular value to the arbitrary constants in yc(x) is known as
particular integral of Eqn. (3). We denote Y0(x) by yp(x).
Thus, we may write Eqn. (9) in the form y(x) = yp(x) + yc(x).

You may then ask the natural question – how to find the solution y(x) of Eqn. (3)?

In the next section we give you the methods of finding the complementary
function yc(x) of the given linear equation with constant coefficients. Since the
complementary function refer to the solution of the homogeneous equation
corresponding o the given equation, we consider the general nth order
homogeneous linear differential equation with constant coefficients.

You may recall that in Sec. 5.2, we had mentioned that if the coefficients of y and
its derivatives in Eqn (1) are constants and a0 ≠ 0, then Eqn (1) is termed as linear
differential equation of nth order with constant coefficients. Further, we had
mentioned that if the right hand side of Eqn. (1) is zero, then it will be classified as
homogeneous linear differential equation for this reason

Eqn. (1) i.e. the function b(x) is also called non-homogeneous term of Eqn. (1).
Thus, the general nth order homogeneous linear differential equation with constant
coefficients may be expressed as

dn y d n− 1y dy
+ a 1 n− 1
+ …. + an-1 + any = 0 …(10)
dx n
dx dx

where the coefficients a1, a2, …., an are constants.

we would like to mention here that in writing Eqn. (10) we have taken the
dn y
coefficient of as unity. Even if it is not so, dividing throughout by the
dx n
dn y
coefficient of (which is also assumed to be constant), the equation can be
dx n
reduced to the form (10).

Let us now discuss the methods of solving EQn. (10).

157
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

3.3 Method of Solving Homogeneous Equation with Constant

Coefficients

The method of solving Eqn. (10) was given in the year 1739 by Leonhard Euler
(1707 – 1783) who was born in Basel, Switzerland and was one of the most
distinguished mathematicians of the eighteenth century.

The method is as follows:

Assume that y = emx is a solution of Eqn. (10). On replacing y and its derivatives
upto order n by emx and mn emx in Eqn. (10), we get
(mn + a1mn-1 + … + an-1 m + an) emx = 0 …(11)
since emx ≠ 0 for real values of x, Eqn. (11) is satisfied if
mn + a1mn-1 + … + an-1 m + an = 0 …(12)
Eqn. (12) is called an auxiliary equation or characteristic equation
corresponding to differential Eqn (10).

You might have observed that an auxiliary equation of a homogeneous or non-

homogeneous linear differential operator on replacing D by some finite constant m
and equating it to zero.

You may wonder why we assumed the solution of Eqn. (10) in the exponential
dy
form. This is because we know that the linear first order equation + ay = 0.
dx
Where a is a constant, has the exponential solution y = c1 e-ax on ] - ∞ , ∞ [.
Therefore, it is natural to determine whether exponential solution exist on ] - ∞ , ∞
[. For higher order equations of the form (10).

In the discussion to follow, you will be surprised to see that all solutions of Eqn.
(10) are exponential functions or constructed out of exponential functions.

Let us now consider the following examples:

Example 5: Write auxiliary equation corresponding to the differential equation

(D6 + 12D4 + 482 + 64) y = 0

Solution: Replacing D by m in the linear differential operator of the given

equation, the auxiliary equation becomes
m6 + 12m4 +48m2 + 64 = 0

Example 6: Write the characteristic equation corresponding to the differential

equation
(D2 + 2aD + b2) y = c sinwx.

158
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: On replacing D by m in the homogeneous part of the given equation

and equating it to zero, we arrive at the following characteristic equation
m1 + 2am + b2 = 0

remember that while writing the auxiliary equation for non-homogeneous

differential equation, he non-homogeneous part is neglected.
Auxiliary Eqn. (12) is a polynomial in m of degree n and, it can have at the most n
roots.

Let m1, m2, …, mn be the n roots. Then the following three possibilities arises;
I) Roots of auxiliary equation may be all real and distinct,
II) Roots of auxiliary equation may be all real, but some of the roots may be
repeated.
III) Auxiliary equation may have complex roots.

We now proceed to find the solution of Eqn. (10) for these three cases one by one.

Case 1: Auxiliary equation has real and distinct roots:

Let the roots m1, m2, …, mn of auxiliary Eqn. (12) be real and distinct.

Now suppose m = m1. Since m1 is a root of auxiliary Eqn. 912), clearly e m x is an 1

integral of Eqn. (10) and satisfies it on the interval ]- ∞ , ∞ [.

Similarly, for m = m2, e m x is a solution of Eqn. (10) and e m x and e m x are also
2 1 2

linearly independent on the interval since

e m1x e m2x
W( e m1x
,e m2x
)=
m1e m1x m 2 e m2x
= (m2 – m1) e (m + m 1 2)x ≠ 0 for m1 ≠ m2.

Now, the n roots of Eqn (12), namely m1, m2, …, mn are real and distinct solutions,
, e , …, e m x are all distinct and linearly independent solutions of Eqn. 910).
m1x
e m 2x n

Sin Eqn. (10) is of nth ordr and we have n distinct and linearly independent
solutions, therefore, we can express the complete solution of Eqn (10) as
y = c1 e + c2 e + … + cn e ,
m1 x m 2x mn x
…(13)
where c1, c2, …, cn are arbitrary constants.

We now illustrate this case with the help of a few examples.

d2y dy
Example 7: Solve 2 + 5 - 12y = 0
dx 2 dx

159
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: The given equation can be written as

(2D2 + 5D – 12) y = 0
the auxiliary equation is
2m2 + 5m – 12 = 0
⇒ (2m – 3) (m + 4) = 0
⇒
m = 3/2, - 4

here the roots are real and distinct.

Hence complete solution of the given differential equation is y = c1e(3/2)x + c2e-4x,

where c1 and c2 are arbitrary constants.

Let us look at another example.

d2y
Example 8: If dx 2
- a2y = 0, show that y = Acosh ax + Bsinh ax is the complete
solution.

Solution: The auxiliary equation corresponding to the given differential equation

is
m2 – a2 = 0
⇒
(m – a) (m + a) = 0
⇒
m = a, - a.

roots being real and distinct, the general solution of the given equation is
y = c1eax + c2e-ax
From the definition of hyperbolic functions, we know that
1 ax
Cosh ax = (e + e-ax …(14)
2
1 ax
and sinh ax = (e – e-ax) …(15)
2
adding relations (14) and (15), we get
eax = cosh ax + sinh ax
Substrating relation (15) from (14), we get
e-ax = coshax – sinhax
the general solution of given differential equation can thus be written as
y = c1 (cosh ax + sinh ax) c2 (cosh ax – sinh ax)
⇒
y = (c1 + c2) cosh ax +(c1 – c2) sinh ax
⇒
A cosh ax + B sinh ax,
where A = c1 + c2 and B = c1 – c2 are two arbitrary constants.

160
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

We now consider an initial value problem.

Example 9: Solve the equation

d2x
- 4x = 0
dx 2

dx
with the conditions that when t = 0, x = 0 and = 3.
dt

Solution: The auxiliary equation corresponding to the given equation is

m2 – 4 = 0
⇒ (m – 2) (m + 2) = 0
⇒
m = 2, -2

hence the general solution of the differential equation is

x = c1e-2t.
we now apply the given conditions at t = 0.we have
dx
dt
= 2c1e-2t
Condition that x = 0 when t = 0 gives
0 = c1 + c2,
dx
and the condition that dt
= 3 when t = 0 gives
3 = 2c1 – 2c2
From the two equations for c1 and c2, we conclude tht
3 3
C1 = and c2 = - 4 . Therefore,
4
3
X= 4
(e2t – e-2t)
Which can also be put in the form
x = sinh 2t.
Now you may try the following exercises.

We now take up the case when the roots of auxiliary equation are all real but some
of them are repeated.

161
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Case II: Auxiliary Equation has real and repeat roots:

Let two roots of auxiliary Eqn. (12) be equal, say m`1 = m2. then solution (13) of
Eqn. (10) becomes
y = (c1 + c2) e m x + c3 3m x + … + cn e m x
1 3 n

Since (c1 + c2) can be replaced by a single constant, this solution will have (n -1)
arbitrary constants.

We know that the general or complete solution of an nth order linear differential
equation must contain n arbitary constants; hence the above solution having (n -1 )
arbitrary constants is not the general solution.

To obtain general solution in this case let us reqrite Eqn. (10) in the form
L1 (y) = (Dn + a1D.. + an-1 D + an0 y = 0 …(16)
d n dn
Where D = and D = and L1 is a linear differential operator.
dx dx n

If m1, m2, …, mn are the roots of auxiliary equation corresponding to, Eqn. (16),
then Eqn.(16) can be written as

D – m1) (D – m2) … (D – mn) y = 0 (17)

It is clear that when all the n roots m1, m2, …, mn are real and distinct the
complete solution of Eqns. (16) or (17) is constituted by the solutions of the n
equations.

(D – m1) y = 0, (D – m2) y =0 …. (D – mn) y = 0

in case the two roots are equal say m1 = m2, then Eqn. (17) takes the form
(D =- m1)2 (D – m3) … (D - mn) y = 0
and then solutions corresponding to two equatl roots are the solutions of
(D – m1)2 y = 0
⇒
(D – m1) [(D – m1) y] = 0 …(18)
Let (D – m1) y = v …(19)
Then Eqn. (18) reduces to
(D – m1) V = 0
dV
⇒
- m1V = 0
dx

it is a linear differential equation of the first order and its solution (ref. Sec. 3.3 of
unit 3) is
V = c1 e m1x

162
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

With this value of V, Eqn. (19) becomes

(D – m1) y = c1 e m1x

which is again a linear differential equation of the first order and is solution is
y = e (c + c1x),
m1x

c1, c2 being constants.

Similarly, the solution of Eqn. (17) corresponding to three equal roots say m1 = m2
= m3, are the solutions of
(D – m1)3y = 0
⇒
(D – m1) [(D = m1)2y] = 0

te (D – m1)2 y = z in the above equation. solving the equation for z and putting the
value of z obtained in the above equation, we have
(D – m1)2y = c1 e m1x

Substituting gain (D – m1) y = t and proceeding as before, we get

(D – m1) y = e m x (c2 + c1x)
1

The solution of above linear differential equation of first order is

c 2 
y = e m x  1 x + c 2 x + c3 
1

2 

thus, it is clear that if a root m1 of Eqn. (16) is repeated r times, then solution
corresponding to this root will be of the form

y = e m x (A1 + A2x + A3x2 + … + Arxr-1)

1

and the general solution of Eqn. (16) will then be

m x
+ … + An e
mx m x
y = e (A1 + A2x + A3x2 + … + Arxr-1) + Ar+1 e …(20)
1 r+ 1 n

We now illustrate the above discussion with the help of a few examples.

d4 y d 3y d 2y dy
Example 10: Solve - -9 - 11 dx
-4y=0
dx 4 dx3 dx 2

Solution: The given differential equation can be written as

(D4 – m3 – 9D2 – 11D – 4) y = 0

Auxiliary equation of the given equation is
M4 – m3 – 9m2 – 11m – 4 = 0
⇒
(m + 1)3 (m – 4) = 0
⇒
m = - 1, - 1, - 1, 4

163
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

here the root – 1 is repeatred three times and root 4 is distinct. Hence, using Eqn.
(20), the general solution of the given differential equation is
y = (A + Bx + Cx2) e-x + De4x,
where A, B, C and D are arbitrary constants.

Let us consider another example.

Example 11: Find the complete solution of

(D4 – 8D2 + 16) y = 0

Solution: In this case the auxiliary equation is

m4 – 8m2 + 16 = 0
⇒
(m2 – 4)2 = 0
⇒
(m- 2)2 (m + 2)2 = 0
⇒
m = 2,2, - 2, - 2

here 2 and -2 are both repeated. Therefore, the method of repeated real roots will
be separately applied to each repeated root. Hence the complete solution of the
given differential equation is

y = (A + Bx) e2x + (C + Dx) e-2x.

and now some exercises for you.

Now we shall discuss the case when the auxiliary equation may have complex
roots.

Case III: Auxiliary Equation has complex roots:

If the roots of auxiliary Eqn. (12) are not all real, then some or, may be, all the
roots are complex. We know from the theory of equations that if all the
coefficients of a polynomial equation are real, then its complex roots occur in
conjugate pairs. In Eqn. (12), all the coefficients are assumed to be real constants
and hence complex roots, if any, must occur in conjugate pairs.

Let one such pair of complex roots of Eqn 912) be m1 = α - i β , where α

and β
2
are real and i = - 1. Formally, there is no difference between this case and case I,
and hence the corresponding terms of solution are

y = c1 e( α + iβ ) x + c2 e( α − iβ ) x
= e [c [e + c 2e ]
iβ x − iβ x
α x
1 …(21)

164
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

however, in practice we would prefer to work with real functions instead pf

complex exponentials. To achieve this, we make use of the Euler’s formula,
namely,
eiθ = cos θ i sin θ and e − iθ = cos
θ
- i sin ,
θ

where θ
is any real number. Using these results, the expression (21), which is
the part pf the solution corresponding to complex roots, becomes

eα x [c1 (cos β x + i sin β x) + c2 (cos β x – i sin β

x)]
= eα x [(c1 + c2) cos β x + (c1 – c2) i sin β x]

Since c1 + c2 are arbitrary constants, we may write

A = c1 + c2 and B = i(c1 – c2),

So that A and B are again arbitrary constants, though not real. Expression (21)
now takes the form

eα x [Acos β x + Bsin β
x]
…(22)

Further, if the complex root is repeated, then the complex conjugate root will also
be repeated and the corresponding terms in the solution can be written, using the
form (20), as
e x ( α + iβ ) (c1 + c2x) + e x ( α − iβ ) (c3 + c4x)
Proceeding as above and writing
A = c1 + c3 B = i(c1 – c3) C = c2 + c4 D = i(c2 – c4),
The above expression can be written as
eα x [(A + Cx) cos β x + (B + Dx) sin β x]
…(23)

in the case of multiple repetition of complex roots, the results are obtained
analogous to those in the case of multiple repetition of real roots.

We now illustrate this case of complex roots with the help of a dew examples.

Example 12: For the differential equation

d4y
dx4
- m4y = 0, show that its solution can be expressed in the form
y = c1 cos mx + c2sin mx + c3 cosh mx + c4 sinh mx.

165
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: The given differential equation can be expressed as

(D4 – m4) y = 0
In this case since m is used as a constant in the given differential equation, we can
replace D by some other letter, λ say.
So, the auxiliary equation is
( 4 – m4) = 0
λ

⇒
( 2 – m2) ( 2 + m2) = 0
λ λ

⇒
= m, - m, ± im
λ

Now the solution corresponding to roots + m and – m can be obtained as we have

done in Example 8 and write it as

C3 cosh mx + c4 sinh mx
Solution corresponding to imaginary roots + im and –im will be
Aeimx + Be-imx
Which can be written as

A(cos mx + i sin mx + B(cos mx – i sin mx)

= (A + B) cos mx + i (A – B) sin mx
= c1 cosmx + c2 sinmx

where c1 = (A + B) and c2 = i(A – B) are constants.

Hence the general solution of the given differential equation is

Y = c1 cos mx + c2 sin mx + c3 cosh mx + c4 sinh mx

Let us look at another example.

d4y d3y d2y dy

Example 13: Solve dx4
-4 dx3
+8 dx2
-8 dx
+ 4y = 0

Solution: in this case is

M4 – 4m3 + 8m2 – 8m + 4 = 0
⇒
(m2 – 2m + 2)2 = 0
⇒
[m – (1 + i)]2 [m – (1 – 9)]2 = 0
⇒
m = 1 + i, 1 + i, 1 – i.

Roots are complex and repeated in this case.

Hence the general solution can be written as
Y= (c1 + xc2) e(1+i)x + (c3 + xc4) e(1-i)x
= ex [(c1 + xc2) eix + (c3 + xc4) e-ix]
= ex [(c1 + xc2) (cosx + i sinx) + (c3 + xc4) (cosx – i cinx)]

166
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

= ex {[(c1 + xc2) + x(c2 + c4)] cosx + i [(c1 – c3) + x(c2 – c4)] sinx}
= ex [(A + Bx) cosx + (C + Dx) sinx]
where A = (c1 + c3), B = (c2 + c4), C = i (c1 – c3) and D = i )c2 – c4) are all
constants.
You may now try the following exercise.

4.0 CONCLUSION

We now end this unit by given a summary of what we have covered in it.

5.0 SUMMARY

in this unit we have covered the following.

1) The general linear differential equation with dependent variable y and

independent variable x is termed as an equation.
a) with variable coefficients if the coefficients of y and its derivatives
are functions of x.
b) with constant coefficients if the coefficients of y and its derivatives
are all constants.
c) homogeneous if the terms other than those of y and its derivatives
are absent.
d) non-homogeneous if the terms other than those of y and derivatives
are present and are constants or functions of independent variable x.

2) A solution of general linear differential equation exists and is unique if

conditions of Theorem 1 are satisfied.

3) A set of functions y1(x), y2(x), …, yn(x) defined on an interval I is linearly

dependent if for constants c1, c2, .., cn not all zero, we have for every x in I,
c1y1(x) + c2y2(x) + …, + cnyn(x) = 0.
4) A set of functions y1(x), y2(x),…, yn(x) on I is linearly independent on I if it
is not linearly dependent on I.
5) If y = y1 is a solution of homogeneous linear differential equation on I, so
is y = c1 on I, where c is arbitrary constant.
6) If y = y1, y2, …, ym are solutions of linear homogeneous differential
equation on I, so is y = c1y1 + c2y2 + … + cmym on I, where c1, c2, …, cm are
arbitrary constants.
7) If y1, y2,…, yn are linearly independent solutions of an nth order
homogeneous linear differential equation on an interval I,
then
y = c1y1 + c2y2 + … + cnyn
(where c1,c2,…, cn being arbitrary constants)

167
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

is defined as the complete primitive of the given equation on I.

8) For a non-homogeneous equation
a) the complete primitive of the corresponding homogeneous part is
called it complementary function.
b) particular solution of the non-homogeneous part involving no
arbitrary constant is called its particular integral.
c) Complementary function and particular integral constitute its general
solution.
9) Solution y, of an nth order linear differential equation
dn y d n − 1y dy
n
+ a1 n− 1
+ … + an-1 dx + an y = 0
dx dx
with constant coefficients a1, …., an-1, an having n roots m1, m2, …, mn, when
a) roots are real and distinct, is
y = c1 e m x + c2 e m x + … + cn e m x
1 2 n

b) roots are real and repeated, say m1 = m2 = …, = mr, is

y = (c1+ c2x + … + crxr-1) e m r + cr+1 e m x + …., + cn e m x .
1 1+ 1 n

c) roots are complex and one such pair is ± i β , is

α

y = eα x (c1 cos β x + c2 sin β x)

corresponding to that pair of roots.

6.0 TUTOR MARKED ASSIGNMENT

1
1. Verify if the function y = sin 4x is a unique solution of the initial value
4
problem

y' + 16y = 0
y(0) = 0, y'(0) = 1.

2. In the following problems verify that the given function y1 and y2 are the
solutions of the corresponding equations. Decide whether the set {y1, y2}
of solutions is linearly dependent or independent.

a) y" – y = 0, y1 = ex and y2 = e-x over - <x<∞ ∞

 π
b) y" + 9y = 0, y1 = cos3x and y2 = cos  3x +  over - ∞
<x< ∞
.
 2
x x
c) y" - 2y' + y = 0 y1 = e and y2 = xe over - <x< ∞ ∞
.

3. Construct an example to show that a set of functions could be linearly

independent on some interval and yet have a vanishing Wronskian.

168
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

4. Verify that y = 1/x is a solution of the non-linear differential equation y" =

2y3 on the interval ]0, ∞ [, but the constant multiple y = c/x is not a solution
of the equation when c ≠ 0, and c ≠ ± 1.

5. Functions y1 = 1 and y2 = Inx are solutions of the non-linear differential

equation y" + (y')2 = 0 on the interval ]0, [. Then

a) is y1 + y2 a solution of the equation?

b) is c1y1 + c2y2, a solution of the equation, where c1 and c2 are arbitrary
constants?

6. Solve the following equations:

d3y d2 y dy
a) - 6 + 11 - 6y = 0
dx 3 dx 2 dx
d 2y dy
b) 9 dx 2
+ 18 dx
- 16y = 0

d3y d2y dy
c) dx3
+2 dx2
-5 dx
- 6y = 0

7. In the following equations find the solution y for x = 1:

a) (d2 – 2d – 3) Y = 0; when x = 0, y = 4 and y' = 0

b) (D3 – 4D) y = 0, when x = 0, y = 0 y' = 0 and y" = 2.

8. Find the complete primitive of the following equations:

d 3y d 2y dy
a) dx3
- dx 2
-8 dx
+ 12y = 0

d4y d3y d2y dy

b) dx4
-2 dx3
-3 dx2
+4 dx
+ 4Y = 0

d 3y d 2y dy
c) dx3
+ dx 2
- dx
-y=0

d3y d2y dy
d) dx3
-3 dx2
+3 dx
-y=0

9. Find the general solution of the following equations subject to the

conditions mentioned alongside:

a) (D2 + 4d + 4)y = 0; when x = 0, y = 1 and y' = - 1

169
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

b) (D3 – 3D – 2) y = 0; when x = 0, y' = 9 and y" = 0

c) (D4 + 3D3 + 2D2) y = 0; when x = 0, y = 0, y' = 4, y" = - 6, y" = 14.

10. Find the general solution of the following Equations:

d 2y dy 2 2
a) dx 2
-2 α

dx
+( α
+ β
)y=0

d4y
b) dx4
+ a4y = 0

c) d 2y dy
dx 2
+8 dx
+ 25y = 0

7.0 REFERENCES/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

Indira Gandhi National Open University School Of Sciences Mth-07.

170
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

UNIT 2 METHOD OF UNDETERMINED COEFFICIENTS

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Types of Non-Homogeneous Terms for which the Method is
Applicable
3.1.1 Non-homogeneous Term is an Polynomial
3.1.2 Non-homogeneous Term is an exponential function
3.1.3 Non-homogeneous Term is a sine or a cosine Function
3.1.4 Non-homogeneous Term is a product of an exponential and a
polynomial
3.1.5 Non-homogeneous Term is a product of a polynomial, an
Exponential and a Sinusoidal function
3.2 Observations and constraints of the method
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 References/Further Readings

1.0 INTRODUCTION

In Unit 5, we learnt that in order to find the complete integral of a general non-
homogeneous linear differential equation, namely

dn y d n − 1y dy
L(y) = a0 + a 1 n− 1
+ ….+ an-1 + any = b(x) …(1)
dx n
dx dx

where ao, a,, ... , a, are constants, it is necessary to find a general solution of the
corresponding homogeneous equation that is, the complementary function and
then add to it any particular solution of Eqn. (1). In Sec. S.3 we discussed the
methods of determining complementary function of linear differential equations
with constant coefficients having auxiliary equations with different types of roots.
But how do we find a particular solution of these equations? We shall now be
considering this problem in this unit.

Variety of methods exist for finding particular integral of a non-homogeneous

linear differential equations. The simplest of these methods is the method of
undetermined coefficients. Basically, this method consists in making a guess as to
the form of trial solution and then determine the coefficients involved in the trial

171
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

solution so that it actually satisfies the given equation. You may recall that we had
touched upon this method in Sec. 3.3 of Unit 3 for finding the particular integral
of non-homogeneous linear differential equations of the first order having constant
coefficients. In this unit we shall be discussing this method in general for finding
the particular integral of second and higher order linear differential equations with
constant coefficients.

2.0 OBJECTIVES

After studying this unit you should be able to

• identify the types of non-homogeneous terms for which method of

undetermined coefficients can be successfully applied.
• write the form of trial solutions when non-homogeneous terms are
polynomials, exponential functions or their combinations.
• describe the constraints of this method.

3.0 MAIN CONTENT

3.1 Types of Non-Homogeneous Terms for Which the Method is

Applicable

The method of undetermined coefficients, as we have already mentioned in Sec.

6.1, is a procedure for finding particular integral y^ in a general solution y(x) =
yc(x) °f equations of the form (1). The success of this method is based on our
ability to guess the probable form of particular solutions.

We know that the result of differentiating functions such as xr (r > 0, an integer) an

exponential function eα x ( α constant) or sinmx or cosmx (m constant) is again a
polynomial, an exponential or a linear combination of sine or cosine functions
respectively. Hence, if the non-homogeneous term b(x) in Eqn. (1) is a polynomial
an exponential function, or a sine or cosine function then we can choose the
particular integral to be a suitable combination of polynomial, an exponential. I
sinusoidal function with a number of undetermined constants. These constant! I
then be determined so that the trial solution satisfies the given equation.

Note: A function which is a combination of a sine function (or cosine function)

with an exponential function and/or a polynomial is a sinusoidal function.

Thus the types of non-homogeneous term for which the method of undetermined
coefficients is successfully applicable are

172
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

i) polynomials
ii) exponential functions
iii) sine or cosine functions
iv) a combination of the terms of types (i), (ii) and (iii) above.

We shall now discuss the method of undetermined coefficients to find the

particular integral for these various types of non-homogeneous terms one by one.

3.1.2 Non-homogeneous term is an Exponential Function:

Let us suppose that the non-homogeneous term b(x) in Eqn. (1) is an exponential
function of the form eα x ( α a constant).

In other words, suppose we have to solve an equation.

dn y d n − 1y dy
L(y) = a0 + a1 n− 1
+ … + an-1 + any = eα x …(6)
dx n
dx dx

The appropriate form of the trial solution can be taken as

yp(x) = A eα x …(7)
α x
provided e is not a solution of the homogeneous differential equation
corresponding to Eqn. (1) (i.e., α is not a root of the auxiliary equation).

If α is a root of Eqn. (6), then the choice (7) would not give us any information for
determining the value of A. in that case, we can take y p(x) = Ax eα x as the trial
solution. If α is r-times repeated root of the auxiliary equation, then the suitable
form of the trial solution for determining particular integral will be

yp(x) = Axr eα x …(8)

substituting this value of yp in Eqn. (6) and equating coefficients of eα x on both

sides, we can find the value of undermined coefficient A and thus find the
particular integral (8).

For a better understanding of whatever we have discussed above let us take up a

few examples.

Example 3: Find the general solution of the differential equation

d2 y dy
+3 + 2y = 3ex.
dx 2
dx

173
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

solution: Auxiliary equation is

(m + 1) (m + 2) = 0
⇒ m = -1, -2,
∴C.F. = c1e-x + c2e-2x
Since ex is not a part of the complementary function, hence trial solution for
finding a particular integral can be taken as
yp(x) = Aex
Substituting this value of yp in the given differential equation,
we get
2Aex + 3Aex + Aex = 3ex
⇒ 6Aex = 3ex
Equating coefficient of ex on both sides, we get
1
6A = 3, ⇒ A =
2
hence
1
P. I = ex
2
∴The general solution fo the given differential equation is
y = c1e-x + c2e-2x + ---- ex
let us consider another example which illustrate the case of repeated roots of an
auxiliary equation.

d3 y d2 y dy
Example 4: solve - 3 + 3 - y = 12ex
dx 3
dx 2
dx

Solution: Auxiliary equation is

(m – 1)3 = 0
⇒ m = 1, 1, 1.
∴C.F. = (c1 + c2x + c3x2) ex
since non-homogeneous term of the given differential equation is ex which is
present in the complementary function and moreover I is 3-times repeated root of
the auxiliary equation, we take the form of trial solution to be
yp(x) = Ax3ex.

Note that in the selection of the trial solution yp(x) no smaller power of x will give
us the particular integral. Moreover, it is not similar to any term of
complementary function of the given equation.
On substituting this value pf yp in the given differential equation, we get
- Ax33x + 3A [x3ex + 3x2ex] – 3A [x3ex + 6x2ex 6xex]
+ A [x3ex + 9x2ex + 18xex + 6ex] = 12ex

174
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Equating coefficients of ex on both sides, we get

6A = 12, ⇒ a = 2.
Thus, P.I. = 2x3ex
∴The general solution of the given differential equation is
y = (c1 + c2x + c3x2) ex + 2x3ex
And now an exercise for you.

You may also come across the situation when b(x) in Eqn. (1) is a sum of two or
more functions. Suppose b(x) = b1(x) + b2(x); then from the superposition
principle we have the P.I. yp(x) of L(y) = b(x) to the equal to yp = yp1 + yp2,
where yp1 is a P.I. of L(y) = b1(x) and yp2 is a P.I. of L(y) = b2(x). This enables us
to decompose the problem of solving linear equation L(y) = b(x) into simpler
problem an example.

Example 5: find a general solution of

d2 y dy
- 2 - + y = ex + 4.
dx 2
dx
Solution: Auxiliary equation is
M2 – 2m + 1 = 0
⇒ (m – 1)2 = 0
⇒ m = 1, 1.
∴C.F. = (c1 + xc2) ex
To find the particular solution we first consider equation
+ y = ex …(9)

1 is a repeated root of the auxiliary equation, we consider the trial solution

yp1 = Ax2ex

on substituting yp1 in Eqn. (9), we find that

(2Aex + 4xAex + x2Aex) – 2(2xAex + x2Aex) + Ax2ex = ex

comparing the coefficient of ex on both sides, we have

2Aex = ex
1
⇒A =
2
2
x x
∴yp1 e
2
Now consider the equation

175
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

d2 y dy
-2 y=4 ...(10)
dx 2
dx

Since no-homogeneous term is a constant, we try yp2 = A and find that A = 4

satisfies (10). Hence a particular solution of the given equation is

x2 x
yp = yp1 + yp2 = e =4
2
A general solution will them be
x2 x
y = c1ex + c2xex + 4 + e
2

The term b(x) can be a combination of many more terms like this. We may have
b(x) = x + ex, b(x) = x + x3, b(x) = 3 + x2 etc. In these cases, we can obtain
particular integral using I and II discussed above and by finding yp1 and yp2 as we
have done in Example 5.

We shall give you the general method of finding P.I. when we discuss cases IV
and V.

You may try these exercises.

We can now take up the case when b(x) in Eqn. (1) is either a sine or a cosine
function.

3.1.3 Non-homogeneous Term is a Sine or a Cosine Function

After going through I and II above and attempting the exercises given so far, you
know how to handle b(x) when it is polynomial, an exponential function or a
combination of both. Now can you say how this case is handled when b(x) is a
sine or a cosine function?

We know that the linear differential operator when applied to sin β x or cos β x will
yield a linear combination of sin β x and cos β x . Therefore, if the non-
homogeneous term b(x) of differential Eqn. (1) is of the form

B(x) = α 1 sin β x or α 2 cos β x or α 1 sin β x + α 2 cos β x

We can take the trial solution in the form

yp(x) = Acos β x + Bsin β x …(11)

176
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

provide ± iβ are not roots of the auxiliary equation corresponding to the given
differential equation.

If ± iβ are r-times repeated roots of the auxiliary equation, then we can take the
form of trial solution to be
yp(x) = xr(Acos β x + Bsin β x ) …(12)

We then substitute the value of yp(x) in the form (11) or (12), whichever is
applicable in Eqn. (1) and equate the coefficients of sin β x and cos β x on both
sides of the resulting equation. This gives us equations for obtaining the values of
A and B in terms of known quantities. Knowing the values of A and B, particular
integral of Eqn. (1) is obtained from relations (11) or (12).

We now illustrate this theory with the help of a few examples.

Example 6: Find the general solution of

d4 y d2 y
- 2 + y = sinx
dx 4 dx 2

Solution: Auxiliary Equation is

(m4 – 2m2 + 1) = 0
⇒ (m2 – 1)2 = 0
⇒ m = 1, 1, - 1, -1
∴C.F. = (c1 + c2x) ex + (c3 + c4x) e-x

Since ------ i is not a root of the auxiliary equation, that is, term sinx does not
appear in the complementary function, we can take the trial solution in the form
yp(x) = Asinx + Bcosx.
Substituting this value of yp in the given differential equation, we get
(Asinx + Bcosx) – 2(-Asinx – Bcosx) + (Asinx + Bcosx) = sinx
⇒ 4Asinx + 4Bcosx = sinx
Equating coefficients of sinx and cosx on both sides, we get

1
4A = 1 ⇒ A =
4
and 4B = 0 ⇒ B = 0
1
Thus, yp(x) = sinx
4
and the complete solution of the differential equation is

177
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1
y =(c1 + c2x) ex + (c3 + c4x) e-x + sinx.
4

Let us look at another example.

Example 7: Solve the initial value problem

d2 y
+ y = 2cosx, y(0) = 1, y'(0) = 0
dx 2

Solution: The auxiliary equation is

m2 + 1 = 0
⇒m = ± i
⇒ C.F. = c1cosx + c2sinx

Now since ± i is a root of the auxiliary equation i.e., cosx itself appears in the
complementary function, we take the form of the trial solution as
yp(x) = x (Asinx + Bcosx)
Substituting the value of yp(x) in the equation, we get
2(Acosx – Bsinx) + x( - Asinx – Bcosx) + x(Asinx + Bcosx) = 2cosx
⇒ 2Acosx – 2Bsinx = 2cosx

Comparing the coefficients of sinx and cosx on both sides, we get

2A = 2 ⇒ A = 1 and B = 0.
Therefore,
yp(x) = xsinx
and the general solution is
y(x) = c1cosx + c2sinx + xsinx
we now use initial conditions to determine c1 and c2
Now y(0) = 1 gives c1 = 1
And y'(0) = 0 gives c2 = 0
Thus, y(0) = cosx + xsinx

You may now try the following exercises.

In the example considered so far, did you notice that the function b(x) itself
suggested the form of the particular solution yp(x)? in fact, we can expand the list
of functions b(x) for which the method of undetermined coefficients can be
applied to include products of these functions as well. We now discuss such cases.
3.1.4 Non-homogeneous Term is a Product of an Exponential and a
Polynomial

178
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Let us suppose that b(x) is of the form

b(x) = eα x [b0xk + b1xk-1 + …. + bk-1x + bk] = eα x Pk(x)
with this form of b(x), Eqn. (1) reduces to
dn y d n − 1y dy
a0 + a1 n− 1
+…. + an-1 + an = eα x [b0xk + b1xk-1 +
dx n
dx dx
… + bk-1x + bk] …(13)

We now take the trial solution in the form

yp(x) = eα x [A0xk + A1xk-1 +….+ Ak-1x + Ak] …(14)
provide α is not a root of the auxiliary equation corresponding to Eqn. (13). If α
is a root of the auxiliary equation, say, it is r-times repeated root of the auxiliary
equation then we modify the trial solution as
yp(x) = xr eα x [A0xk + A1xk-1 +…+ Ak-1x + Ak] …(15)

Remember that in Eqn. (15) no smaller power of x will yield a particular integral.
Here r is the smallest positive integer for which every term in the trial solution
(15) will differ from every term occurring in the complementary function
corresponding to Equ (13).

In order to determine the constants A0, A1,…., Ak we substitute yp(x) in the form
(14) or (15) as the case may be in eqn. (13) and then compare the coefficients of
eα x on both sides. For a better understanding of whatever we have discussed
above, let us consider an example.

d 3 y dy
Example 8: Solve - = xe-x
dx 3
dx

Solution: Auxiliary equation is

m3 – m = 0
⇒ m(m2 –1) = 0
⇒ m = 0, - 1, 1
∴C.F. = c1 + c2e-x + c3ex
Here the non-homogeneous term is xe-x appears in the complementary function.
Further, ( -1) is a non-repeated root of the auxiliary equation. Thus, we take the
form of trial solution as
yp(x) = x [B + Ax] e-x = Ax2e-x = Ax2e-x + Bxe-x
substituting this value of yp in the given differential equation, we get
- A [-x2ex + 2xe-x] + A [-x2e-x + 6xe-x – 6e-x] – B (-xe-x + e-x) +
B(-xe-x + 3e-x) = xe-x

Comparing the coefficients of xe-x and e-x on both sides, we get

179
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1
4A = 1 ⇒ A =
4

3
and –6A + 2B = 0 ⇒ B =
4
1 2 -x 3 -x e − x 2
Hence yp(x) = x e + xe = (x + ex)
4 4 4
And the general solution is

e− x 2
-x
y = c1 + c2e + (x + 3x).
4

You may now try the following exercise.

Lastly, we take up the case when b(x) is a product of a polynomial, an exponential

function and a sinusoidal function.

3.1.5 Non-homogeneous Term is a Product of a Polynomial, an

Exponential and a Sinusoidal function

Let us suppose that the non-homogeneous term b9x) in Eqn. (1) has one of the
following two forms:

b(x) = eα x Pk(x) sin β x or b(x) eα x Pk(x) cos β x , …(16)

where Pk(x), as given by Eqn. (2), is apolynomial of degree k or less and α and β
are any real numbers. You may recall Euler’s formula and write
e( α + iβ ) x = eα x cos β x + i eα x sin β x
or, equivalently, we have
eα x cos β x = Real ( e( α + iβ ) x )
e ( α + iβ ) x + e ( α − iβ ) x
=
2
β
and e sin x = Imaginary ( e( α + iβ ) x )
α x

e ( α + iβ ) x − e ( α − iβ ) x
=
2i

Hence b(x) in Eqn. (16) reduces to

 e ( α + iβ ) x − e ( α − iβ ) x 
b(x) = (b0x + b1x +….+ bk) 
k k-i

 2i 

180
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

 e ( α + iβ ) x + e ( α − iβ ) x 
k
or b(x) = (b0x + b1xk-1
+ … + bk)  
 2 

In either of the above two cases, we take the trial solution in the form
yp(x) = (A0xk + A1xk-1 + …+ Ak) ( e( α + iβ ) x )
(B0xk + B1xk-1 + …+ Bk) ( e( α − iβ ) x )
or, equivalently,
yp(x) = (A0xk + A1xk-1+ …+ Ak) eα x cos β x +
(B0xk + B1xk-1 + …+ Bk) eα x sin β x,
provided α ± iβ is not a root of the auxiliary equation.

If ( α ± iβ ) is r-times repeated root of the auxiliary equation, we can then modify

the trial solution by multiplying it by xr. We then substitute the trial solution in
Eqn. (1) and equate the coefficients of like terms on both sides to determine A0,
A1, …., Ak and B0, B1, …., Bk. Substituting these values of undetermined
coefficients in the trial solution, we get the particular integral.

Let us now illustrate the above case with the help of a few examples.

Example 9: find the appropriate form of trial solution for the differential equation

d4 y d3 y d2 y
+2 3 +2 = 3ex + 2xe-x + e-x sinx
dx 4 dx dx 2

Solution: Auxiliary equation is

m4 + 2m3 + 2m2 = 0
⇒ m2(m2 + 2m + 2) = 0
⇒ m = 0, 0, - 1 ± i
∴C.F. = c1 + c2x + e-x (c3sinx + c4cosx)

Here the non-homogeneous term is 3ex + 2xe-x + e-x sinx

Since the term e-x sinx also appear in C.F., the appropriate form of the trial
solution is

yp = Aex + (Bx + C) e-x + xe-x (Dcosx + Esinx).

We now take up an example in which b(x) is a product of a polynomial an

exponential and a sinusoidal function.

Example 10: Write down the form of the trial solution for the equation

181
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

d2 y dy
+2 5y = x2e-x sinx
dx 2
dx

Solution: the auxiliary equation is

m2 + 2m + 5 = 0
⇒ m = - 1 ± 2i
The roots are not equal to –1 ± i. Hence the for of trial solution is
yp = (A0x2 + A1x + A2) e-xcosx + (B0x2 + B1x + B2) e-xsinx

Note that the form of trial solution taken in Case v above is the most general form.
This is because the trial solutions taken in Cases I –IV are particular forms of Case
V.

And now some exercise for you.

After going through the Cases I – V above and attempting the exercises given, you
must have understood the method of undetermined coefficient quite well. Did you
make certain observations about the method? Let us now summarize the
observations and constraints of this method.

3.2 Observations and Constraints of the Method

1) Method is straight forward in application.

2) It can be used by any learner who is not familiar with more elegant
techniques of finding the solutions f the differential equations such as
inverse operators and variation of parameters, which involve integrations
and which we shall be discussing in the subsequent units.
3) Success of this method depends to a certain extent on the ability to guess an
appropriate form of the trial solution.
4) If the non-homogeneous term is complicated and the trial solution involves
a large number of terms, then determination of coefficients in the trial
solution becomes laborious.
5) This method is not a general method of finding the particular solution of
differential equations. It is applicable to linear non-homogeneous equations
with constant coefficients and with restricted forms of the no-
homogeneous terms.

4.0 CONCLUSION

We now end this unit by giving a summary of what we have covered in it.

182
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

5.0 SUMMARY

In this unit, we have covered the following:

1) Method of undetermined coefficients is applicable if

a) The equation is a linear equation with constant coefficients.
b) The non-homogeneous term is either a polynomial, an exponential
function, a sinusoidal function or a product of these functions.
2) The results giving trial solutions corresponding to different non-
homogeneous terms in the equation L(y) = b(x), where the equation L(y) =
0 has r-times repeated roots are summarized in the following table.

Non-homogeneous term b(x) Trial solution, yp(x)

k k-1 x+b r k-1
Pp(x) = b0x + b1x + … + bk-1 k x (A0x +…+ Ak)
eα x xr(A eα x )
 sin β x
 xr(Asin β x + Bcos β x )
 cos β x
eα x Pk(x) xr eα x (A0xk +….+ Ak)
 sin β x x r [(A 0 x k + ... + A k )eα x sin β x
e Pk(x) 
α x

 cos β x + (B0 x k + ...) + Bx )eα x cos β x]

3) Observations and constrains of the method.

6.0 TUTOR MARKED ASSIGNMENT

1. Find a form of particular integral of the following equations

d2 y dy
a) 2
+ + y = x2 + 1
dx dx
d y d3 y d 2 y
4
dy
b) 4
- 3 - 2
+ = x2
dx dx dx dx

2. Determine the general solution of the following equations.

d2 y dy
a) +3 + 2y = 4x2
dx 2
dx
3
dy dy
b) +4 =x
dx 3
dx

183
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

3. Find a particular integral of the following differential equations.

d3 y dy
a) 3
- 4 = e-2x
dx dx
3 2
dy dy dy
b) 3
- 2
+ + 1 = e-x
dx dx dx

4. Find a general solution of the following differential equations:

d2 y dy
a) -3 + 2y = x2 (ex + e-x)
dx 2
dx
2
dy
b) 2 + 8y = x3 + e2x
dx 2
5. Solve the following initial value problems:

d2 y
a) 2
- y = e2x. y(0) = -1, y'(0) = 1.
dx
d2 y dy
b) -4 + 4y + e2x = 0, y(0) = y'(0) = 0.
dx 2
dx

6. Solve the following equations:

d4 y d2 y
a) +4 = sin2x
dx 4 dx 2
d 3 y dy
b) - = 2cosx
dx 3 dx

7. Solve the following initial value problems:

d2 y
a) + 4y = sinx, y(0) = 2, y'(0) = -1
dx 2
d 2 y dy −7 1
b) - - 2y = cosx – sin2x, y(0) = , y'(0) =
dx 2
dx 20 5

8. Solve the following equations:

d2 y 3x
a) 2
+ 9y = x2 e
dx

184
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

d2 y dy
b) 2
-4 + 4y = 4x e2x
dx dx
9. Write the form of the trial solution for each of the following:

d2 y dy
a) +2 + 3y = xcos3x – sin3x
dx 2 dx
d2 y dy
b) +2 + 5y = xe-xcos2x
dx 2 dx
d2 y dy
c) -5 + 6y = xexcos2x
dx 2 dx
d2 y
d) + y = x2 sinx
dx 2
d2 y dy
e) -4 + 5y = xe2xsinx
dx 2 dx

10. Find the general solution of the following equations.

d3 y dy
a) - 4 = x + 3cosx + e-2x
dx 3
dx
d y d y d2 y
4 3

b) 4
- 3 - 2
= x2 + 4 + xsinx
dx dx dx
dy3
dy
c) + = x3 + cosx
dx 3
dx

7.0 REFERENCES/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

Indira Gandhi National Open University School Of Sciences Mth-07

UNIT 3 METHOD OF VARIATION OF PARAMETERS

CONTENTS

185
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Variation of Parameter
3.2 Reduction of order
3.3 Euler’s Equations
4.0 Conclusion
5.0 Summary
6.0 Tutor Marked Assignment
7.0 Reference/Further Readings

1.0 INTRODUCTION

In unit 6, we discussed the method of undetermined coefficients for determining

particular solution of the differential equation with constant coefficients when its
non-homogeneous term is of a particular form (viz, a polynomial, an exponential,
a sinusoidal function etc).

In this unit we familiarize you with an alternative approach fir determining a

particular solution that can be applied even when the coefficients of the
differential Equation are functions of the independent variable and the non-
homogeneous term may not be of a particular form. Such an approach is due to
Joseph Louis Lagrange (1736 – 1813) and is termed as variation of parameters.
Even though the approach is quite general bit is limited in its scope in the sense
that it can be utilized in situations where the fundamental solution set for the
reduced equation is known. Also, it can be used for first and higher order
equations alike though its appreciation can be well understood for the later set of
equations. The method requires for its applicability the complete knowledge of
fundamental solution set of the reduced equation and for equations with variable
coefficients the determination of this set may be extremely difficult. In the case of
linear differential equations with variable coefficients, at times, it may not be
possible to find all linearly independent solutions of the reduced but at least one or
more may be obtainable. For such situations Jean le Rond d’Alembert (1717 –
1783), a French mathematician and a physicist, developed a method that is often
called the method of reduction of order. When one or more solutions of reduced
equation are known that D’Alembert’s method can be used to derive an equation
of order lower than that of a given equation and obtain the rest of the solutions of a
reduced equation as well as the particular integral of the non-homogeneous term.
We shall be discussing the method of reduction of order in Sec. 7.3 of the unit.
For some particular forms of the second order linear differential equations with
variable coefficients, we have also listed some rule by which one integral of the
homogeneous equation an be guessed.

186
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

However, there exist linear differential equation with variable coefficients if

second and higher order for which we may not be able to guess any integral of its
complementary function. But, among such equations is a class of equations
known as Euler’s equation or homogeneous linear differential equations, where,
by certain substitution, it is possible to find all the integrals of its complementary
function. In Sec. 7.4, we shall be discussing the method of solving Euler’s
equations and those equations which are reducible to Euler’s form.

2.0 OBJECTIVES

After reading this unit you should be able to

• Use the method of variation of parameters to find particular integral of non-

homogeneous linear differential equations with constant or variable
coefficients.
• Use the method of reduction of order to find the complete integral of the linear
non-homogeneous equation of second order when one integral of the
corresponding homogeneous equation is known.
• Write down one integral for second order linear homogeneous differential
equation with variable coefficients in certain cases merely through inspection.
• Solve Euler’s equations.

3.0 MAIN CONTENT

3.1 Variation of Parameters

Let us not discuss the details of the method by considering the non-homogeneous
second order linear equation.

L[y] = y" + a1(x) y' +a2(x)y = b(x), …(1)

Where we have taken the coefficients of y" to be 1 and a1(x), a2(x), and b(x) are
defined and continuous on some interval J. Let [y1(x), y2(x)] be a fundamental
solution set for the corresponding homogeneous equation

L[y] = 0 …(2)
Then we know that the general solution of (2) is given by
Yc(x) = c1y1(x) + c2y2(x), …(3)

Where c1 and c2 are constants. To find a particular solution of the non-

homogeneous equation, the idea associated with the method of variation of

187
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

parameters is to replace the constants in Eqn. (3) by function of x. That is, we

seek a solution of Eqn. (1) of the form
yp(x) = u1(x)y1(x) + u2(x)y2(x), …(4)
where u1(x) and u2(x) are unknown functions to be determined. Since we have
introduced two unknowns, we need two equation involving these functions for
their determination.

In other words, we impose two conditions which the functions u1 and u2 must
satisy in order that relation (4) is a solution of Eqn. (1). We call these conditions
the auxiliary conditions. These conditions are imposed in such a way that the
calculations are simplified. Let us see how this is done.

Now if relation (4) is a solution of Eqn. (1), then it must satisfy it. Thus, first we
compute y'p(x) and y"p(x) from Eqn. (4).

'
y p = (u 1' y1 + u '2 y2) + (u1y 1' +u2y '2 ) …(5)

To simplified the computation and to avoid second order derivatives for the
"
unknown u1, u2 in the expression for y p , let us choose the first auxiliary condition
as
u'1y1 + u y2 = 0 '
2 …(6)
Thus relation (5) becomes
y = u 1 y + u2 y
'
p
'
1
'
2

…(7)
and
y =u y
'
p + u y 1" + u y +u2y "2
'
1
'
1
'
1
'
2
'
2

…(8)
Substituting in Eqn (1), the expressions for yp, y and y as given b Eqn. (4), '
p
"
p

(7) and (8), respectively, we get

b(x) = L[yp]
= (u y + u1 y 1" + u y
'
1 + u2y ) + a1( u1 y1' + u 2 y'2 ) + a2( u y1 + u y2)
'
1
'
2
'
2
"
2 1 2

'
= ( u1 y + u 2 y 2 ) + u ( y1 + a y +a y ) + u ( y 2 + a y 2 2 2
' ' ' " ' " '
1 1 1
1 2 1 2 1 +a y )

=( + )+
u'1 L[y1] +
y1' L[y2] u '2 y '2 u1 u2 …(9)
since y1 and y2 are the solution of the homogeneous equation, we have
L[y1] = L[y2] = 0

Thus Eqn. (9) becomes

u y'
1+ u y = b(x) '
1
' '
2 2

…(10)

188
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

whicg is the seond auxiliary condition.

Now if we can find u1 and y2 satisfying Eqns. (16) and (10), viz.,
y1u1' + y 2 u '2 = 0 
 …(11)
y1' u1' + y'2 u '2 = b(x) 

then yp given by Eqn. (4) will be a particular solution of Eqn. (1). In order to
determined u1, u2 we first solve the linear system of Eqns (11) for u and u '2 .
'
1

Algebraic manipulations yield

− b(x)y 2 (x) ' b(x) y1 (x)

u'1 (x) = , u 2 (x) = , …(12)
W(y1 , y 2 ) W(y1 , y 2 )
where
y1 y2
W(y1, y2) = = y1 y'2 - y2 y'1
y1' '
y2

is the Wronskian of y1(x) and y2 (x).

Note that this Wronskian is never zero on J, because {y1, y2} is a fundamental
solution set.
On integrating u (x) and u given by Eqn. (12), we obtain
'
1
'
2

− b(x) y 2 (x) b(x) y1 (x)

yp(x) = ∫ W(y1 , y 2 )
dx, u2(x) = ∫
W(y1 , y 2 )
dx …(13)

Hence
− b(x) y 2 (x) b(x) y1 (x)
yp(x) = y1(x) ∫ W(y1 , y 2 )
dx + y2(x) ∫
W(y1 , y 2 )
dx ...(14)

is a particular integral of Eqn. (1).

We now sum up the various steps involved in determining a particular solution of

Eqn. (1).

Step 1: Find a fundamental solution set {y1(x), y2(x)} for the corresponding
homogeneous equation.

Step II: Assume the particular integral of Eqn (1) in the form
yp(x) = u1(x) y1(x) + u2(x) y2(x)
and determine u1(x) and u2(x) by using the formula (13) directly or by first solving
the system of Eqns. (11) for ------- and then integrating.

189
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Setp III: Substitute u1(x) and u2(x) into the expression for yp(x) to obtain a
particular solution.

We now illustrate these steps with the help of the following examples.

Example 1: Determine the general solution of the differential equation

d2 y π
+ y = secx, 0 < x <
dx 2
2

Solution: Step I: The auxiliary equation corresponding to the given equation is

m2 + 1 = 0
⇒m = ± i
and the two solutions of the reduced equation are
y1(x) = cosx
and
y2(x) = sinx.
Hence the complementary function is given by
Yc(x) = c1cosx + c2sinx.

Step II: To find particular integral, we write

yp(x) = u1(x) cosx + u2(x)sinx …(15)
dy du du 2
∴ p = [-u1(x)sinx + u2(x)cosx] + 1 cosx + sinx
dx dx dx

Let us take the first auxiliary condition as

du1 du 2
cosx + sinx = 0 …(16)
dx dx

So that
dy p
= - u1(x)sinx + u2(x)cosx
dx

Differentiating the above equation once again, we get

dy p du du 2
= - u1(x)cosx = u2(x)sinx – sin 1 + cosx …(17)
dx dx dx
Since yp(x) must satisfy the given equation, we substitute in the given equatin the
expression for yp and ----- fom Eqns. (15) and (17), respectively, and obtain
du du 2
- sinx 1 + xosx = secx …(18)
dx dx

190
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

du1 du 2
On solving Eqns. (16) and (18) for and , we get
dx dx
du1 du 2
= - tanx, = 1,
dx dx
which on integration yields
u1(x) = In(cosx) and u2(x) = x

Step III: Substituting the values of u1(x) and u2(x) in Eqn. (15) we obtain a
particular solution of the given equation in the form
yp(x) = cosx In(cosx) + xsinx
and the general solution is
y = c1cosx + c2sinx + xsinx + cosxIn(cosx)

Note that in Eqn. (1) we have taken the coefficients of y" to be 1. if the given
equation is of the form a0(x)y" + a1(x)y' + a2(x)y = b(x), then before applying the
method it must be put in the form y" + p(x)y' + q(x)y = g(x)y = g(x) as we have
done in the following example.

Example 2: Find the general solution of

d 2 y 1 dy
(1 – x2) - = f(x) …(19)
dx 2 x dx

Solution: Step 1: We first rewrite the given equation in the form

d 2y 1 dy f (x)
- =
dx 2
x(1 − x ) dx
2
(1 − x 2 )
The corresponding homogeneous equation is
d2y 1 dy
- =0 …(20)
dx 2
x(1 − x ) dx
2

dy dy
This is a first order equation in dx
. To solve this we put dx
= p. Then Eqn. (20)
reduces to
dp 1
- p=0
dx x(1 − x )
2

1 dx
or dp = …(21)
p x(1 − x 2 )

Now Eq. (21) is in variable separable form and can be expressed as

dp  1 x 
= + dx
p  x 1 − x 
2

Integrating we get

191
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

c1x
P=
1 − x2
dy c1x
or = …(22)
dx 1 − x2

Integrating Eqn. (22), once again, we get the solution of Eqn. (20) in the form
Yc(x) = - 1 − x 2 + c2 …(23)
where c1 and c2 are arbitrary constants.

Step II: For the given differential equation, assume a particular solution in the
form
yp(x) = u1(x) 1 − x + u2(x) 2

dy −x  2 du1 du 
∴ p = u1 +  1− x + 2
dx 1− x 2
 dx dx 

We choose the first auxiliary condition as

du1 du2
1 − x2
dx
+ dx
=0 …(24)
Then
dyp −x
= u1
dx 1 − x2
d 2 yp 1 1 du1
and =- u1 -
dx 2
(1 − x 2 )3/ 2 1− x 2 dx

' "
Substituting, from above, the expressions for y p and y p in Eqn. (19), we get
du1
-x 1 − x2
dx
= f(x) …(25)
as our second auxiliary condition.
Solving Eqns. 924) and (25) for u and u and integrating, we get
'
1
'
2

f (x) f (x)
U1(x) = - ∫ dx and u2(x) = ∫ dx
x 1− x 2
x
Step III: The expressions for u1(x) and u2(x) when substituted in yp(x) gives a
particular integral in the form
f (x) f (x)
Yp(x) = - 1 − x ∫ dx + ∫
2
dx
x 1− x 2
x

Hence a general integral o fthe given differential equation is

192
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

f (x) f (x)
Y = - c1 1 − x2 + c2 - 1 − x2
1− x
∫x 2
dx + ∫ x
dx

You may now try the following exercises.

If you have carefully gone through Example 1 and 2 above, and also attempted 1.
and 2., you will find that the results of second order non-homogeneous linear
differential equations can be put in the form of the following theorem.

Theorem 1: If the functions a0(x), a1(x), a2(x) and b(x) are continuous on some
interval J and if y1 and y2 are the linearly independent solutions of the
homogeneous equations associated with the differential equation
d 2y dy
A0(x) dx 2
+ a1(x) dx
+ a2(x)y = b(x), …(26)
then a particular solution of Eqn. (26) is given by
y 2 (x) b(x) y1 (x) b(x)
yp(x) = -y1(x) ∫ dx + y2(x) ∫ dx ...(27)
a 0 (x) W(y1 , y 2 ) a 0 (x) W(y1 , y 2 )
where w(y1, y2) is the Wronskian of y1(x) and y2(x).

Remark: In using the method of variation of parameters for finding a particular

integral of a given equation, it is advisable to choose a particular integral yp(x) =
u1(x) y1(x) + u2(x) y2(x), and then proceed to find u1(x) and u2(x) as we have done
in Examples 1 and 2 above. It is usually avoided to memorise formulas given by
Eqns. (13) or (27). But since the procedure involved is somewhat long and
complicated and moreover, it may not always be easy or even possible to evaluate
the integrals involved, these formulas turn out to be useful. In such cases, the
formulas for yp(x) provide a starting point for the numerical evaluation of yp(x).

The method of variation of parameters which we have discussed for non-

homogeneous second order equations can be easily generalized to nth order
equations of the form

dn y d n − 1y
A0(x) + a1 (x) + …+ an(x)y = b(x)
dx n dx n − 1
where a0(x), a1(x),……, an(x), b(x) are continuous in some interval J. The learner
interested into the details of the method for a higher order equation may refer to
the Appendix at the end of the Unit. We shall not be giving the details at this
stage but, however, illustrate it through an example.
Example 3: Find the general solution of

d3 y d2 y dy
- 6 + 11 dx
- 6y = e2x
dx 3 dx 2

193
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: Step I: The auxiliary equation corresponding to the given equation is

m3 – 6m2 + 11m – 6 = 0
⇒ (m – 1) (m2 – 5m + 6) = 0
⇒ (m – 1) (m – 2) (m – 1) = 0
thus the linearly independent solutions are
y1(x) = ex, y2(x) = e2x , y3(x) = e3x ,
and the complementary function is given by
yc(x) = c1ex + c2e2x + c3e3x …(28)

Step II: To find particular integral, we write

yp(x) = u1(x)ex + u2(x)e2x + u3(x)e3x …(29)
dyp +( ex + e2x + u ' e3x) + (u1 ex + 2u2e2x + 3u3e3x)
∴ u1' u'2
3
dx

Let the first auxiliary condition be

'
u ex + u e2x + u e3x) = 0
'
1
'
2 …(30)
3

Thus
y'p = u1ex + 2u2e2x + 3u3e3x
and
y"p = u ex + 2 u e2x + 3 u 3' e3x) + (u1ex + 4u2e2x + 9u3e3x)
'
1
'
2

Let us choose the second condition as

u'1ex + 2 u e2x + 3 u 3' e3x = 0'

2 …(31)
Then
y"p = u1ex + 4u2e2x + 9u3e3x
∴
y"p = (u u1' ex + 4 u'2 e2x + 9 u'3 e3x) + (u1ex + 8u2e2x + 27u3e3x)
' " "'
Substituting the values of yp, y p , y p and y p in the given equation, we get

( u ex + 4 u e2x + 9 u e3x) + (u1ex + 8u2e2x + 27u3e3x)

'
1
'
2
'
3

- 6(u1ex + 4u2e2x + 9u3e3x) + 11(u1ex + 2u2ex + 2u2e2x + 3u3e3x)

- 6(u1ex + u2e2x + u3e3x) = e2x …(32)
⇒ u ex + 4 u e2x + 9 u e3x = e2x,
'
1
'
2
'
3

which is our third auxiliary condition

Thus, we get the system of equations

194
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

u1' e x + u '2e 2x + u 3' e3x = 0 


u1' e x + 2u '2e 2x + 3u 3' e3x = 0  …(33)
u1' e x + 4u '2e 2x + 9u 3' e3x = e 2x 
Solving Eqns. (33) for u1' , u '2 and u 3' , we get
1 1
u1' = ex, u '2 = - 1 and u 3' = e-x
2 2
Integrating, we get
1 1
u1 = ex, u2 = - x and u3 = e-x
2 3

Step III: We get a particular integral in the form

1 1
yp(x) = e2x – xe2x - e2x = - xe2x,
2 2
and the general solution is
y = c1ex + c2e2x + c3e2x – c3e3x – xe2x,

You may now try this exercise.

Clearly the method of variation of parameters has an advantage over the method of
undermined coefficients in the sense that it always yields a particular solution yp
provided all the solutions of the corresponding homogeneous equation are known.
Moreover, its application is not restricted to particular forms of the non-
homogeneous term. In the next section we will discuss a technique which is very
similar to the method of variation of parameter.

3.2 Reduction of Order

For a given nth order linear homogeneous differential equation, if one nontrivial
solution is known, then the method of reduction of order, as the name suggests,
reduces the equation to an (n- 1)th order equation. Thus, if we can find in some
way, one or more linearly independent solutions of the reduced equation, we can
accordingly reduce the order of the given differential equation. In other words, if
p independent solutions of a homogeneous linear corresponding to an nth order
equation are known, where p < n, then the technique can be used to obtain a linear
equation of order (n- p). This fact is particularly interesting when n = 2, since the
resulting first order equation ca always be solved by the methods we have done in
Block 1. That is, if we know one solution of the homogeneous linear differential
equation of the second order, we can solve the non-homogeneous equation by the
method of reduction of order and obtain both a particular solution and a second

195
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

linearly independent solution of the homogeneous equation. Let us now see how
method works for a second order linear equation.

Consider a second order non-homogeneous equation of the form (1), viz.,

d2 y dy
+ a1(x) + a2(x) y = b(x),
dx 2
dx
where a1(x), a2(x) and b(x) are continuous on some interval J. Suppose that y =
y1(x) is a nontrivial solution of the corresponding homogeneous equation
d2 y dy
+ a1(x) + a2(x)y = 0 …(34)
dx 2
dx

Then y = cy1(x) is also a solution of Eqn. (34) for some constant c. We now
replace the constant c by an unknown function v(x) and take a second trial solution
in the form

y = v(x)y1(x)
Now,
y' = v'y1 + vy 1'
y" = v"y1 + 2v' y 1' + vy 1"
substituting from above the expression for y, y' and y" in the given equation, we
get
(v"y1 + 2v'y 1' + vy 1" ) + a1(v'y1 + vy 1' ) + a2vy1 = b(x)
⇒ v"y1 + v' (2y 1' + a1y1) + v(y 1" + a1y 1' + a2y1) = b(x) …(35)
Since y1 is a solution of Eqn. 934), the last term on the l.h.s. of Eqn. (35) is zero.
Therefore Eqn. (35) reduces to
v"y1 + v'(2y 1' + a1y1) = b(x) …(36)

dv
Let = p(x), so that Eqn. (36) becomes
dx
dp 2y1' + a1y1 b(x)
+ p= …(37)
dx y1 y1
This is a first order linear differential equation with integrating factor

 2y1' + a1 y1 
I.F. = EXP  ∫ dx 
 y1 
2y + a y
'

Now ∫ 1 1 1 dx = 2Iny1 + ∫ a1 (x) dx

y1
∴ I.F. = y1 e ∫ a (x )dx = y12 h(x), where h(x) = e∫ a--(x)dx
2
1

196
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Thus Eqn. (37) reduces to

y12 h(x)p(x) = c1 + ∫ b(x) y1h(x) dx

dv 1
 c + b(x) y1h(x) dx 
y1 h(x)  1 ∫
⇒ = 2 
dx

Integrating the above equation once again, we obtain

1 1
V(x) = c2 + c1 ∫ 2 dx + ∫ 2  b(x) y1h(x) dx  dx
y h(x)
1 y h(x)  ∫
1

Thus the general solution of the given equation can be expressed as

1
y = v(x)y1 (x) = c2y1(x) + c1y1(x) ∫ y h(x)
2
1

1
+ y1(x) ∫ y h(x)  ∫ b(x)
2 y1h(x) dx  dx …(38)
1

1
Note that the function y1(x) ∫ y h(x) , in the second term on the r.h.s. of Eqn. (38),
2
1

is the 2nd linearly independent solution of Eqn. (34) and the last term on the r.h.s. is
a particular integral of the given non-homogeneous equation.

We now take up an example to illustrate the theory.

Example 4: Find the general solution of

x2y" – xy' + y = x1/2, 0 < x < ∞,
Given that y1 = x is a solution of the corresponding homogeneous equation.

Solution: The given equation is

x2y" – xy' + y = x1/2 …(39)
Let us take y = xv(x) as a trial solution for Eqn. (39). So that
y' = v + xv'
v" = 2v' + xv"
Substituting for y, y' and y" from above in Eqn. (39), we obtain
x2(2v' + xv") – x(v + xv') + xv = x1/2
⇒ x3v" + x2v' = x1/2
⇒ v" + ----- v' = x-5/2 …(40)

Eqn. (40) is a linear differential equation in v'. ts integrating factor is

197
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

1
I.F. = e ∫ x dx = eInx = x
Therefore, Eqn. (40) yields
V' x = ∫ x.x-5/2 dx + c1
⇒ v' = c1 x-1 – 2x-3/2
Integrating once again, we have
v = c1 Inx + 4x-1/2 + c2
Thus,
y = xv = c1x Inx + c2x + 4x1/2
is the general solution of Eqn. 939).

And now some exercise for you.

From that above it is seen that if one solution of the second order linear
homogeneous Eqn. (34) is known, then the second linearly independent solution
and a particular integral of the associated non-homogeneous equation can be
determined.

We now give some rules, which will help you to find one integral included in the
complementary function merely by inspection.

For a homogeneous equation of the form (34) if

Rule I: 1 + a1(x) + a2(x) = 0, then y = ex is an integral of the Eqn. (34).

For instance, consider an equation
xy" – y' + (1 – x) y = x2e-x …(41)

To bring it to the form (34), we write it a

1  1− x  -x
y" - y' +   y = xe
x  x 
1 1
Thus, a1(x) = - and a2(x) = -1
x x
1 1
Now 1 + a1(x) + a2(x) = 1 - + - 1 = 0
x x
x
Thus, according to Rule 1, y = e is an integral of the equation. You can verify
your result by substituting y = ex in the given equation and check if it satisfies the
given equation.

Rule II: a1(x) + xa2(x) = 0 then y = x is an integral of the Eqn. (34)

Consider the equation,
91 – x2) y" + xy' – y = x((1 – x2)3/2
This equation can be written as

198
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

x 1
y" + y' - y = x 1 − x2
1− x 2
1 − x2
Comparing the above equation with Eqn. (34), we have
x 1
a1(x) = and a2(x) = - .
1− x 2
1 − x2
Here a1(x) + xa2(x) = 0, hence by the above rule y = x is an integral of the
homogeneous equation corresponding to the equation.

Rule III: 1 – a1(x) + a2(x) = 0, then y = e-x is an integral of the Eqn. (34)

Rule IV: 2 + 2xa1(x) + x2 a2(x) = 0, then y = x2 is an integral of the Eqn. (34).

a1 (x) a 2 (x) α
Rule V: 1 + + = 0, α > 0, then y = e x is an integral of the Eqn. (34).
α α 2

Note that in applying Rules I – V the given equation should be first put in the form
of Eqn. (34).

You may now try the following exercise.

So far you have seen that the method of variation of parameters can be used only
for those differential equation for which we know all the linearly independent
solutions of the corresponding homogeneous equation. Method of reduction of
order is helpful for finding complete solution of the second order non-
homogeneous linear equations even if one solution of the corresponding
homogeneous equation is known. There exists certain rules which, at once, give
one solution merely through an inspection, included in the complementary
function of the second order linear equations with constant coefficients. But, no
rules exist which may help to guess one or more integrals included in the
complementary function when the equation is or order higher than two and is
having variable coefficients. However, there exists a class of linear differential
equations with variable coefficients known as Euler’s equations for which it is
possible to find all the linearly independent integrals of the complementary
function. In the next section we take up the method of solving Euler’s Equations.

3.3 Euler’s Equations

Consider the following differential equations

d3 y d2 y
x3 3 + x2 2
+ 4x + 2y = ex …(42)
dx dx

199
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

d2 y dy
x + 2x + 2y = ex …(43)
dx 2
dx
3
3 d y dy
(2x – 1) + (2x – 1) - 2y = sinx …(44)
dx 3
dx

All the three equations given above are linear as the dependent variable y and its
derivative appear in their first degree and moreover there is no term involving the
product of the two. Out of the three equations, only Eqn. (42) is such that the
powers of x in the coefficients are equal to the orders of the derivatives
associated with them. This type of equation known as homogeneous linear
differential equation or Euler’s Equation. Eqn (43) is linear but not
homogeneous. Eqn. (44) is not of Euler’s from form but can be reduced to Euler’s
form by the substitution X = 2x – 1. Here we shall consider only equations of the
form (42) and (44).

The general form of Euler’s equation of nth order is

dn y d n − 1y dn− 2 y dy
xn n + P1 xn-1 n − 1 + P2xn-2 n− 2
+ ….+ Pn-1 + Pny = f(x), ...(45)
dx dx dx dx
where P1, P2, …, Pn are constants and right hand side is a constant or a function of
x alone.

Eqn. (45) can be transformed to an equation with constant coefficients by

changing the independent variable through the transformation

z = Inx or x = ez

with this substitution, we have

dy dy dz 1 dy
= . =
dx dz dx x dz
dy dy d
⇒ x = = D1y, where D1 =
dx dz dz
dy2
d  1 dy  1 dy 1 d 2 y dz
Also =   = - +
dx 2 dx  x dz  x 2 dz x dx 2 dx
1  d 2 y dy 
= 2  2−
x  dz dz 
d 2 y d 2 y dy
⇒ x2 = 2 - = (D 12 - D1) y = D1 (D1 – 1) y
dx 2
dz dz
Proceeding as above, we shall, in general, get
dn y
xn n = D1 (D1 – 1) (D1 – 2) …. (D1 - n − 1) y
dx

200
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Thus, Eqn (45) is transformed to the equation

[D1 (D1 – 1) – (D1- n − 1) + P1 D1(D1 – 1) ….( D1 n − 2) + …
… + Pn-2 D1 (D1 – 1) + Pn-1 D1 + Pn] y = f(ez) …(46)

Eqn. (46) is an equation with constant coefficients and its complementary function
can be determined by the methods given in Unit 5. for obtaining its particular
integral either the method of undetermined coefficients (as given in Unit 6 subject
to the form of f(ez)), or the method of variation of parameters can be utilized if the
solution of Eqn. (46) is
y = g(z),
then the solution of Eqn. (45) will be
y = g(Inx)
we illustrate this method by the following examples

d2 y
2 dy
Example 5: Solve x - x + y = Inx
dx 2 dx

Solution: It is Euler’s equation of order 2. To solve it, let

x = ez or z = Inx
dy dy dz 1 dy dy dy d
∴ = . = ⇒x = = D1y, where D1 =
dx dz dx x dz dx dz dz
2
dy d  1 dy  1 dy 1 2
dy dz 1  d 2
y dy 
∴   = - + =  −
dx 2 dx  x dz  x 2 dz x dz 2 dx x 2  dz 2
dz 
d2 y
⇒x 2
= D1 (D1 – 1) y
dx 2

dy d2 y
Substituting for and in the given equation, we get
dx dx 2
[D1 (D1- 1) - D1 + 1] y = z
⇒ (--- - 2 D1 + 1] y = z …(47)
A.E. is
m2 – 2m + 1 0
⇒ m = 1, 1
∴C.F. = (c1 + cz) ez
To find P.I. of Eqn. (47), let us assume that
yp = u1(z) ez + u2(z) zez …(48)
dy p ' '
∴ = u1 ez + u 2 zez +u1ez + u2 (zez + ez)
dz

As first auxiliary condition, assume that

201
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

u1' ez + u '2 zez = 0, …(49)

so that
dy p
= u1ez + u2(z + 1) ez …(50)
dz

Differentiating once again, we have

dy p
= u1' ez + u '2 (z + 1) ez + u1ez + u2ez (z + 1) + u2ez …(51)
dz

If yp(z) is a solution of Eqn. (47), it must satisfy it. Hence substituting the
expression for yp, and --- and --- from Eqn. (48), (50) and (51), respectively, in
Eqn. (47), we obtain the second auxiliary condition as

u1' ez + (z + 1) ez = z …(52)

Solving Eqn. (49) and (52) for-------, we get

u '2 ez = z and ez u1' = - z2
'
⇒ u1 = - z2 e-z and --- = ze-z

Integrating the above equation, we get

u1 = - ∫ z2 e-z dz
 2 e− z 
= - z + 2 ∫ z e − z dz 
 −1 

 2 e− z 
= + z e – 2 z
2 -z
+ ∫ e − z dz 
 −1 
2 -2 -z -z
= z e + 2z e + 2e
2 = ∫ z e dz = - ze + ∫ e dz = - z e e
and u -z -z -z -z -z

Substituting the values of u1(z) and u2(z) in Eqn. (48), a particular integral of Eqn.
(47) can be expressed in the form
yp(z) = (z2 + 2z + 2) e-z. ez + ( - z – 1) e-z zez
= (z2 + 2z + 2) – z(z + 1)
= z2 + 2z + 2 – z2 – z
=z+2
and the general solution of Eqn. (47) is
y = (c1 + c2z) ez + z z + 2

202
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Replacing z by In x, the general solution f the given equation is

y = (c1 + c2Inx). X + Inx + 2

The complementary function of Euler’s Eqn. (45) can also be found by assuming y
= xm in the homogeneous part of the equation and then finding the values of m. we
illustrate it through the following example.

d3 y 2
2 d y dy
Example 6: Solve x3 - x - 6x + 18y = 0
dx 3
dx 2
dx

Solution: Let y = xm
dy
∴ = m xm-1
dx
2
dy
= m (m – 1) (m – 2) xm-3
dx 2
Substituting the above values in the given equation, we get
[m (m – 1) (m – 2) … m(m – 1) – 6m + 18] xm = 0
⇒ (m3 – 4m2 – 3m + 18) xm = 0

thus, if
m3 – 4m2 = 3m + 18 = 0, …(53)
m
then y = x satisfies the given equation.
Equ. (53) is an algebraic equation of 3rd degree in m and its root are
M = - 2, 3, 3.
Thus, y = x-2, y = x3 and y = x3 are the solutions of the given equation. hence the
general solution of the given equation is
Y = c1x-2 + x3(c2+ c3 (Inx))
Note: Had all the roots of Eqn. (53) been real and different, the solutions
corresponding to these roots would have been independent solutions and the
general solution would have been pf the form
m m m
y = c1 x + c2 x + c3 x
1 2 3

In the case of repeadted real roots of Eqn. (53), if a root m1 is repeated r times, the
integral corresponding to root m1 is
m
[c1 + c2Inx + c3 (Inx)2 +…+ cr (Inx)r-1] x 1

Further, if Eqn. (53) had a pair of complex roots, say α ± iβ , then the
corresponding part of the complementary function would have been
x α [c1 cos ( β Inx) + c2sin ( β Inx)]
we illustrate the case of complex roots by the following example.

Example 7: Solve x2y" + xy' + 4y = 0

203
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

Solution: Substituting y = xm in the given equation, we get

[m(m – 1) + m + 4] xm = 0
Thus y = xm satisfies the given equation if
m(m – 1) + m + 4 = 0
⇒ m2 + 4 = 0
⇒ m = ± 2i

Hence, the general solution of the given equation is

y = c1 cos (2Inx) + c2sin(2Inx)

You may now try the following exercises.

Earlier we mentioned that Eqn. (44) is not Euler’s equation, but can be reduced to
Euler’s form by the substitution X = 2x – 1. We now consider such equations
which are reducible to Euler’s form.

Equations Reducible to Euler’s form

Consider the general nth order equation

dn y d n − 1y dy
(ax + b)n dx n + (ax +b)n-1P1 dx n − 1 +…+(ax + b) Pn-1 dx +Pny = f(x), ...(54)
where a, b, P1, …, Pn are all constants.
Equations of the form (54) can be reduced to Euler’s equations by substituting X
= ax + b.
With this substitution
dy dy d 2 y d2 y dn y dn y
dx = a dX , dx 2 = a2 dx 2 , …, dx n = an dX n
and Eqn. (54) reduces to the equation,

dn y d n − 1y dy
an Xn dx n + an-1 Xn-1 dx n − 1 + …+ aX Pn-1 dx + Pny = g(X), …(55)

where g is transformed form of the function f.

Eqn. (55) is now in Euler’s form and can solved by the methods given earlier.

However, Eqn. (54) can be directly reduced to an equation with constant

coefficients by substituting ax + b = ez, instead of first substituting ax + b = X and
then X = ez.

204
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

We illustrate the above theory with the help of following example.

d2 y2 dy
Example 8: Solve (3x + 2) + 3 (3x + 2) - 36y = 3x2 + 4x + 1
dx 2
dx

Solution: The given equation is an equation reducible to Euler’s equation. We

can, however, reduce it to an equation with constant coefficients by a single
substitution.

3x + 2 = ez or z = In (3x +2)

dy dy dz 1 dy ⇒ dy dy
∴ = = .3 (3x + 2) =3
dx dz dx 3x + 2 dz dx dz
dy2
d  3 dy  −32
dy 3 d 2 y dz
and = = +
dx 2 dx  3x + 2 dz  (3x + 2) 2 dx 3x + 2 dz 2 dx
32  d 2 y dy 
= −
(3x + 2) 2  dz 2 dz 

Substituting y' and y" from above in the given equation, we get

 d 2 y dy  dy 1
9  2 −  + 3.3 - 36y = [e2z – 1]
 dz dz  dz 3

d2 y 1
⇒ - 4y = (e2z – 1) …(56)
dz 2
27
A.E is
m2 – 4 = 0 ⇒ m = ± 2

Hence C.F. = yc = c1e2z + c2e-2z

To find a particular integral, we write
yp(z) = u1(z) e2z + u2(z) e-2z …(57)
dy 2z -2z 2z -2z
∴ p = u1' e + u '2 e + 2 (u1e – u2e )
dz
As the first auxiliary condition, let
u1' e2z + u '2 e-2z = 0 …(58)
so that
dy p
= 2 (u1e2z – u2e-2z) …(59)
dz
Differentiating Eqn. (59) once again, we get

205
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

dy p
= 2( u1' e2z - u '2 e-2z) + 4u1e-2z + 4u2e-2z …(60)
dz

Since yp(z) must satisfy Eqn. (56), hence on combining Eqn. (57), (59) and (60)
we get the second auxiliary condition as

1
2( u1' e2z - u '2 e-2z) = (e2z – 1)
27
1 2z
⇒ u1' e2z - u '2 e-2z = (e – 1) …(61)
54

Solving Eqns. (58) and (61) for u1' and u '2 , we get
1 1
u1' = (1 – e-2z) and u '2 = - (1 – e2z)
108 108

Integrating u1' and u '2 , we get

1  e − 2z  2z 1  e − 2z 
u1 = z+ 2  e - z− 2 
108   108  
on substituting the values of u1(z) and u2(z) in relation (57), a particular solution of
Eqn. (56) is obtained in the form.

1  e − 2z  2z 1  e − 2z  -2z
yp = z+ 2  e - z− 2  e
108   108  
1 1
= z (e2z – e-2z) +
108 108
∴ The general solution of Eqn. (56) is
1 1
y = c1e2z + c2e-2z + z [e2z – e-2z] +
108 108
and the required solution of the given equation is
c2 1  1  1
In(3x + 2)  (3x + 2) −
2
y = c1 (3x + 2)2 + 2 + 2  +
(3x + 2) 108  (3x + 2)  108

You may not try the following exercises.

4.0 CONCLUSION

We now end this unit by giving a summary of what we have covered in it.

206
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

5.0 SUMMARY

In this unit we have studied the details concerning the following results:

1) Let y1 and y2 be the linearly independent solutions of the reduced equation

of a non-homogeneous second order linear differential equation with
constant or variable coefficients. Then on substituting y = y1u1(x) + y2u2(x)
and imposing the conditions (11), the particular integral of the given
equation can be found.

2) If y = y1(x) is one solution of the reduced equation, then on substituting

y = y1(x) v(x)
the second solution of the reduced equation and a particular integral of the
corresponding non-homogeneous equation can be determined.

3) Rules for finding one integral included ion the complementary function of
equations of the form (34) by mere inspection are given by the following
table:

Condition satisfied One integral

1 + a1(x) + a2(x) = 0 y = ex
1 – a1(x) + a2(x) = 0 y = e-x
a1(x) + xa2(x) = 0 y=x
2 + 2x a1(x) + x2 a2(x) = 0 y = x2
a (x) a 2 (x)
1+ 1 + = 0, α > 0 y = eα x
α α2

4) Differential equation with variable coefficient of the form

n− 1 n− 2
dn y n-1 d y n-2 d y dy
xn + P1 x n− 1
+ P2x n− 2
+…+ Pn-1x + Pny = f(x),
dx n
dx dx dx
where P1, P2,…, Pn are constants and in which the powers of x in the
coefficients are equal to the orders of the derivatives associated with them,
is known as Euler’s equation. This equation can be reduced to an equation
with constant coefficients by using the substitution x = ez.

6.0 TUTOR MARKED ASSIGNMENT

1. Determine a particular integral, using the method of variation of parameter

for the following differential equations:

207
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

π
a) y" + y = cosec x, 0 < x <
2
b) y" – 2y' + y xex Inx, x > 0
π
c) y" + y = tanx, 0 < x <
x

2. Find a general solution of the following differential equations, given that

the functions y1(x) and y2(x) for x > 0 are linearly independent solutions of
the corresponding homogeneous equations.

a) x2y" – 2xy' + 2y = x + 1; y1(x) = x, y2(x) = x2

1
b) x2y" + xy' – y = x2ex; y1(x) = x, y2(x) =
x
2 x
c) xy" – (x + 1)y' + y = x ; y1(x) = e , y2(x) = x + 1

3. Using the method of variation of parameters, find the general solution of

the following equations:

a) y" – y' = x2
b) y" – 2y" – y' + 2y = e3x

4. Solve the following differential equations:

a) x2y" – 2xy' + 2y = 4x2, x > 0; y1(x) = x

b) x2y" + 5xy' – 5y = x-1/2, x > 0; y1(x) = x

5. A solution of the differential equati0n

d2 y dy
X2 (1 – x2) - x3 - 2y = 0
dx 2
dx
1 − x2
Is y1 = . Use the method of reduction of order to find a general
x
solution.

6. Solve equation
d2 y 2 dy
X(xcosx – 2sinx) + (x + 2) sinx - 2 (xsinx + cosx) y = 0
dx 2 dx
Given that y = x2 is a solution.

7. Verify that y1(x) = ex is a solution of the homogeneous equation

corresponding to Eqn. (41).

208
MTH 232 ELEMENTARY DIFFERENTIAL EQUATION

8. Find an integral included in the complementary function of the following

equations, merely by inspection:

a) y" – cotx y' – (1 – cotx) y = ex sinx

b) (x sinx + cosx) y" + x (cosx) y' – y cosx = x
c) (3 – x) y" – (9 – 4x) y' + (6 – 3x) y = 0

9. Solve the following equations:

1
a) (x2D2 + 3xD) y =
x
b) (x D + xD – 1) y = xm
2 2

 3 4 2 5 2
c) D − D + 2 D− 3  y = 1
 x x x 

10. Solve the following equations.

a) [(x + a)2 D2 – 4 (x + a) D + 6] y = x
b) [(1 + x)2 D2 + (1 + x) D + 1] y = 4 cos [In (x + 1)].

7.0 REFERENCES/FURTHER READINGS

Theoretical Mechanics by Murray. R.Spiegel

Advanced Engineering Mathematics by Kreyzig

Vector Analysis and Mathematical MethodS by S.O. Ajibola

Engineering Mathematics by p d s Verma

Advanced Calculus Schaum’s outline Series by Mc Graw-Hill

Indira Gandhi National Open University School Of Sciences Mth-07

209