Ordinary differential equations

————————————————————————————————————

Contents
1 Introduction 2
1.1 Basic Concepts and Ideas . . . . . . . . . . . . . . . . . . . . . . 2

2 Solution Techniques 7
2.1 Variable Separable . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.1 Reducible into variable separable . . . . . . . . . . . . . . 9
2.1.2 Exact differential equations . . . . . . . . . . . . . . . . . 11
2.1.3 Reducible to exact form . . . . . . . . . . . . . . . . . . . 14
2.1.4 How to find integrating factors? . . . . . . . . . . . . . . . 15
2.1.5 Linear differential equations . . . . . . . . . . . . . . . . . 18
2.1.6 Reducible into linear form (Bernoulli’s equation) . . . . . . 20

3 Applications – First order 22

3.1 Physical applications . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.1.1 Newton’s law of cooling/heating . . . . . . . . . . . . . . . 22
3.1.2 Newton’s second law of motion . . . . . . . . . . . . . . . 23
3.1.3 Radio active decay . . . . . . . . . . . . . . . . . . . . . . 24
3.1.4 A simple electric circuits . . . . . . . . . . . . . . . . . . . 24
3.2 Geometrical applications . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 Orthogonal trajectories . . . . . . . . . . . . . . . . . . . . . . . . 28

4 Linear equations with constant coefficients: Higher order but

degree one 29
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2 The second order homogeneous equations with constant coefficients 30
4.2.1 Existence and uniqueness theorem . . . . . . . . . . . . . . 32
4.2.2 Wronskian . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5 The nth order homogeneous equations with constant coefficients 38

6 The second order nonhomogeneous equations with constant co-

efficients 38
6.1 Method of variation of parameters . . . . . . . . . . . . . . . . . 39
6.2 Method of undetermined coefficients . . . . . . . . . . . . . . . . 41

7 Variable coefficients 44
7.1 Method of reduction order . . . . . . . . . . . . . . . . . . . . . . 45
7.2 Euler Cauchy equation . . . . . . . . . . . . . . . . . . . . . . . . 47
 Dr. Dil Gurung, Kathmandu University

8 Applications – Second order 49

8.1 Spring mass balance system . . . . . . . . . . . . . . . . . . . . . 49
8.1.1 Undamped motion. Harmonic oscillation . . . . . . . . . . 49
8.1.2 Damped motion . . . . . . . . . . . . . . . . . . . . . . . . 54
8.1.3 Forced vibrations . . . . . . . . . . . . . . . . . . . . . . . 57
8.2 LRC series circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

1 Introduction
1.1 Basic Concepts and Ideas
Definition 1 Differential Equation
An equation involving one dependent variable and its derivatives with respect to
one or more independent variables is called a differential equation.

Example 1 The following are the examples of differential equations:

dy 1+y ∂z
1. dx = 1+x 6. + ∂z = c, where c is
∂x ∂y
a constant
dy
= xy + tan xy ∂u 2
2. dx 7. ∂t
= c2 ∂∂xu2 , c2 = ρkc
3. y 000 + xy 00 + y 0 + y = cos x, 8. ∂2u 2
= c2 ∂∂xu2 , c2 = Tρ
h
i3/2 ∂t2
dy 2 d2 y
4. 1 + dx = ρ dx 2 9. ρ c ∂T
∂t
= ∇ · (k ∇T ) + g
q
dy 3 dy 2


5. y = x dx + 1 + dx

There are mainly two classes of differential equations:

(i) Ordinary differential equations (ODE)
(ii) Partial differential equations (PDE)

Definition 2 Ordinary differential equation

A differential equation which involves derivatives with respect to a single indepen-
dent variable is known as an ordinary differential equation.

Equations (1) − (5) are ordinary differential equations.

Definition 3 Partial differential equation

A differential equation which contains two or more independent variables and
partial derivatives with respect to them is called a partial differential equation.

Equations (6) − (9) consist two independent variables and a dependent variable.
So these equations are partial differential equations.

Definition 4 Order of a differential equations

The order of a differential equation is defined to be the order of highest order
derivative (differential coefficient) occurring in it.

Equations (1), (2), and (6), are order one; equation (3) is of order three; equation
(4) and (6) − (9) are of order two.

2
 Dr. Dil Gurung, Kathmandu University

Definition 5 Degree of differential equations

The degree of a differential equation is the power of the highest order derivative
occurring in it after the equation has been made free from radicals and fractions.

Making the equation(4) free from radical, we get

 2  "   2 #3
d y dy
ρ2 2
= 1+
dx dx

The order of this differential equation is 2, and the power of highest order deriva-
tive is 2. So the degree is 2.

Exercise 1 Find the order and degree of the following differential equations.
 2 
dy
+ xy = ex ∂ y
+ y ∂y


1. dx 6. x2 ∂x 2 ∂t
=0
2. xy 00 + y 0 y = x 2

3. y 000 + 3(y 0 )2 + y = 0 7. L ddt2I + R dI

dt
+ 1c I = d(sin
dt
t)

2
4. (y 0 )3 = ((y 0 2
) + 1)
1/2
8. m ddt2x + C dx
dt
+ Kx = 0
∂y ∂y ∂u ∂u ∂u


2
5. x ∂x + t ∂t = x t 9. ∂x + ∂y + ∂z = 0

Definition 6 Linear and nonlinear differential equation

A differential equation is called linear if and only if the following conditions are
satisfied by that differential equations:
1. Every dependent variable and derivative involved occurs to the first degree
only.
2. The product of dependent variables, product of derivatives, and product of
dependent variables and derivatives are absent.
A differential equation is called nonlinear if it does not satisfy the above condi-
tions.

In the following differential equations, the first two are linear and the last four
are nonlinear.

1. y 00 + x2 y 0 + xy = sin x 4. y 00 + y 2 = 0
2
2. ∂u
∂t
= x ∂∂xu2 5. y 00 + yy 0 = x2
2 2
3. y 000 + (y 0 )3 + y = 0 6. ∂∂xu2 + ∂u ∂u
∂x ∂t
+ ∂∂t2u = 0

Definition 7 Homogeneous and non-homogeneous differential equations

A differential equation can be classified as either homogeneous or non-homogeneous
depending on whether that differential equations no additive term or some additive
term containing only the independent variable.

In the above set of differential equations, only (1) and (5) are non-homogeneous
differential equations, whereas the other four are homogeneous differential equa-
tions.

Formation of differential equation: General solution

We proceed to examine how the ordinary differential equation is formed and what
kind of result we get as its solution.

3
 Dr. Dil Gurung, Kathmandu University

(i) Let
f (x, y, a) = 0 (1)
be an algebraic relation between x, y containing an arbitrary constant a.
Differentiating it with respect to x, we get a relation between x, y, a and
dy
dx
, say
φ(x, y, dy/dx, a) = 0 (2)
Eliminating a between these two equations, we get a relation between x, y
dy
and dx , say ψ(x, y, dy/dx) = 0 which does not contain any arbitrary con-
stant.

This is a differential equation of first order. The relation (1) is called the
general solution of the differential equation (3). From above we conclude
that the solution of a differential equation of the first order contains only
one arbitrary constant.

Example 2 Consider an algebraic relation

y = ce2x (3)
where c is an arbitrary constant. Differentiating with respect to x, we get
dy
= 2ce2x (4)
dx
Eliminating c from (4) and (5), we obtain
dy
= 2y (5)
dx
The differential equation (5) is of first order, and thus the relation (4) is
the general solution of the differential equation (5).

(ii) Let
f (x, y, a, b) = 0 (6)
be a relation between x, y, and two arbitrary constants a and b. Differenti-
ating it with respect to x, we get, say
φ1 (x, y, dy/dx, a, b) = 0 (7)
Again differentiating we get
φ2 (x, y, d2 y/dx2 , a, b) = 0 (8)
Eliminating a and b from above equations, we get a relation between x, y, dy/dx
and d2 y/dx2 , say
ψ(x, y, dy/dx, d2 y/dx2 ) = 0 (9)
which is a differential equation of order two. Thus we expect that the
solution of a differential equation of the second order contains two arbitrary
constants, and so on. Thus, the general solution of a differential equation
contains as many arbitrary constants as the order of the equation.

4
 Dr. Dil Gurung, Kathmandu University

Example 3 Consider an algebraic equation

y = c1 ex + c2 e2x (10)

which contains two arbitrary constants. Differentiating with respect to x,

we get
dy
= c1 ex + 2c2 e2x (11)
dx
Again differentiating with respect to x, we get
d2 y
2
= c1 ex + 4c2 e2x (12)
dx
From (10) and (11), we get
dy
= y + c2 e2x (13)
dx
From (12) and (13), we obtain

d2 y dy
2
− 3 + 2y = 0 (14)
dx dx
Thus

y = c1 ex + c2 e2x (15)

is the general solution of the differential equation (14).

Definition 8 Solution of differential equation

A solution of a differential equation is a relation between the dependent and
independent variables, not involving the derivatives such that this relation and
the derivatives obtained from it satisfies the given differential equation.
dy
Example 4 y = ce2x is the solution of the differential equation dx = 2y, because
2x 2x dy
2ce = 2y = 2e . Thus the general solution of the differential equation dx = 2y
2x 2x
is y(x) = ce . For c = 1, we get the particular solution y = e . c = 1 we obtain
by assigning x = 0 and y = 1. This whole problem, we denote by the relation
dy
= 2y, y(0) = 1 (16)
dx
Definition 9 Particular Solution
A solution obtained from a general solution by giving particular values to the
constants is called a particular solution.

Example 5 y = c1 ex + c2 e2x is the general solution of the differential equation

d2 y dy
2
− 3 + 2y = 0 (17)
dx dx
If we take c1 = 1 and c2 = 0, we get y = ex as the particular solution.

5
 Dr. Dil Gurung, Kathmandu University

Definition 10 Singular Solution

A solution which can not be derived from the general solution but still is a solution
of the given differential equation is called a singular solution.

Example 6 The differential equation

 2
dy dy
−x +y =0 (18)
dx dx
has the general solution y = cx − c2 (verify!). Substitution also shows that y =
x2 /4 is also a solution. This is a singular solution because we can not obtain it
from y = cx − c2 by choosing any c.

Definition 11 Initial value problem (IVP)

An initial value problem consists of a differential equation (of any order) together
with a collection of initial conditions that must be satisfied by the solution of the
differential equation and its derivatives at the initial point.

IVP:
dy
= f (x, y), y(x0 ) = y0 (19)
dx
This is an example of initial value problem of first order differential equation. The
condition y(x0 ) = y0 is called an initial condition and x0 is called the initial
point.
Example 7
dy
(i) dx
= 2y − 3x, y(0) = 2
(ii) x00 (t) + 5x0 (t) + (sin t) x(t) = 0, x(1) = 0, x0 (1) = 7
dN
(iii) dt
= α N, N (0) = N0

Definition 12 Boundary value Problem (BVP)

A boundary value problem consists of a differential equation and a collection of
values that must be satisfied by the solution of the differential equation or its
derivatives at no further than two different points.

BVP:
dy
= f (x, y), y(x0 ) = y1 , y(x1 ) = y1
dx
dy
(i) dx + 5xy = 0, y(0) = y(1) = 2
d2 T
(ii) dx2 + kġ = 0, T (0) = Tb , ∂T


∂x x=L
= h(T − T∞ ), 0 < x < L

Theorem 1 Existence and Uniqueness Theorem Let f and ∂f ∂y

be contin-
uous in a rectangle R given by a < x < b, c < y < d that contains the point
(x0 , y0 ). Then, in an interval x0 − h < x < x0 + h contained in a < x < b, there
is a unique solution y = f (x) of the initial value problem
dy
= f (x, y), y(x0 ) = y0
dx

6
 Dr. Dil Gurung, Kathmandu University

Example 8 Does the initial value problem

dy
= x2 + y 3 , y(0) = 1
dx
have unique solution in a region around its initial point (0, 1)?
Solution Since f (x, y) = x2 + y 3 and ∂f
∂y
= 3y 2 are continuous everywhere, they
are certainly continuous at any rectangle R containing the initial point (0, 1).
Hence a unique solution to the initial value problem exists.
Example 9 Does the initial value problem
dy
= xy 1/3 , y(0) = 0
dx
have unique solution in a region around its initial point (0, 0)?
Solution ∂f ∂y
= x/3y 2/3 which is not defined, and therefore not continuous at the
initial point(0, 0).
Exercise 2 Solve the following differential equations.

1. y 0 = x2 3. y 00 = x4
2
2. y 0 = sin 3x 4. y 0 = xe−x
Exercise 3 Verify that the given function is a solution.
1. y 0 + y = x2 − 2, y = ce−x + x2 − 2x
2. y 00 + y = 0, y = a cos x + b sin x
3. y 000 + ex = 0, y = ex + ax2 + bx + c
4. y 00 + 2y 0 + 2y = 0, y = e−x (a cos x + b sin x)
5. x + yy 0 = 0, x2 + y 2 = 1
Exercise 4 Does the following initial value problem have unique solution ?
1. y 0 = 2x2 y 2 , y(1) = −1
2. y 0 = x ln y, y(1) = 1

2 Solution Techniques
A first order differential equation is of the form
dy
= f (x, y) (20)
dx
If we take
M (x, y)
f (x, y) =
N (x, y)
Then the equation (20) is sometimes conveniently written as
M (x, y) dx + N (x, y) dy = 0 (21)
where M (x, y) and N (x, y) are functions of x and y, or constants. Equations (20)
and (21) can’t be solved in every case, and hence it is not always convenient to
solve all differential equations. They can be solved easily if they belong to one or
other of the following standard forms:

7
 Dr. Dil Gurung, Kathmandu University

(i) Variable separable

(ii) Reducible into variable separable

(iii) Exact differential equations

(iv) Reducible to exact form

(v) Linear differential equations

(vi) Reducible to linear differential equations form

2.1 Variable Separable

If the differential equation, say,

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

of the first order and first degree where M (x, y) and N (x, y) are functions of x
and y, can be put in the form

f1 (x)dx + f2 (y)dy = 0

then we say that the variables are separable. Such equations can be easily solved
by integrating both sides of the equation. That is,
Z Z
f1 (x)dx + f2 (y)dy = 0

Example 10 Solve 9yy 0 + 4x = 0

Solution The differential equation can be written as

9ydy + 4xdx = 0

which is of the variable separation form.

By integrating on both sides, we obtain the general solution

92
= −2x2 + C ∗
2
Thus
x2 y 2
+ = C (C = C ∗ /18)
9 4
The solution represents a family of ellipse.

Figure 1:

Exercise 5 Solve the following differential equations:

8
 Dr. Dil Gurung, Kathmandu University

1. y 0 + 3x2 y 2 = 0 4. y 0 = 1 + y 2
2. y 0 = xy/2 5. y 0 = k y, where k is a constant.
3. y 0 + 3x2 y 2 = 0

Exercise 6 Solve the following initial value problems.

1. y 0 = −2xy, y(0) = 1 4. y 3 y 0 + x3 = 0, y(0) = 1

2. y 0 = −y/x, y(1) = 1
5. y 0 = 1 + 4y 2 , y(0) = 0
3. L dI
dt
+ RI = 0, I(0) = I0 ,
where L and R are constants. 6. ex y 0 = 2(1 + x)y 2 , y(0) = 1/6

2.1.1 Reducible into variable separable

Certain differential equations are not separable but can be separable by the in-
troduction of a new unknown function.
Case 1: By direct substitution
Example 11 Solve y 0 = (x + y)2 .
Solution: Put x + y = z. Differentiating with respect to x, we get
dy dz
= −1
dx dx
Substituting these values into the given equation, we get
dz
dx =
1 + z2
Integrating, we get

C + x = arctan z
z = tan(C + x)
x + y = tan(C + x)

Exercise 7 Solve the following differential equations.

dy
1. (x + y + 1)y 0 = 1 4. dx = (4x + y + 1)2
[Hint: Use 4x + y + 1 = v]
2. y 0 = sec(x + y) 5.y 0 = sin(x + y) + cos(x + y)
6.x dx + y dy + x dy−y dx
x2 +y 2
=0
3. y 0 + 1 = ex+y [Hint: Use x = r cos t, y = r sin t]
dy
Case 2: Differential Equation of the form: dx = φ(y/x).
Here φ is any differential function of y/x, for instance (y/x)3 , cos(y/x) etc.
Method: Put y/x = u. Thus y = xu, and by product differentiation
dy du
=u+x
dx dx
dy
Substituting into dx
= φ(y/x) gives

du
u+x = φ(u)
dx
9
 Dr. Dil Gurung, Kathmandu University

Thus
du
x = φ(u) − u
dx
and then separating variables, we get
du dx
=
φ(u) − u x

which is in variable separable form.

Example 12 Solve 2xyy 0 = y 2 − x2 .

Solution: We divide the given differential equation 2xy, we obtain

y2 x2
 
dy 1 y x
= − = − = φ(y/x)
dx 2xy 2xy 2 x y
dy
Put y = xu. So dx
= u + x du
dx
, we obtain
 
du 1 1
u+x = u−
dx 2 u

Hence, by simplification,

u2 + 1
 
du 1 1
x =− u+ =−
dx 2 u 2u

The separation of variables now gives that

2udu dx
= −
1 + u2 x
By integration

ln(1 + u2 ) = − ln x + ln c, where c is a constant of integration

ln(1 + u2 ) = ln c/x
Replacing u by y/x, we get

y2
 
c
ln 1 + 2 = ln
x x

x2 + y 2 c
2
=
x x
2 2
x + y = cx

Exercise 8 Solve the following differential equations.

dy
1. x2 ydx − (x3 + y 3 )dy = 0 6. (x + y) dx + (x − y) = 0
dy
2. (x2 + y 2 ) dx = 2xy 7. (x + y )dx − xy 2 dy = 0
3 3

dy y2 dy y(x−2y)
3. x dx + x = y 8. dx = x(y−2x)
dy dy
4. 2xy dx = x2 + y 2 9. dx = 1−2y−4x
1+y+2x
[Hint: Use y + 2x = v.]
dy
5. dx = 3x+2y
2x−3y
10
 Dr. Dil Gurung, Kathmandu University

2.1.2 Exact differential equations

Definition 13 Differential If a function u(x, y) has continuous partial deriva-
tives, its differential is
∂u ∂u
du = dx + dy
∂x ∂y

From this it follows that if u(x, y) = c = constant, then du = 0.

Example 13 If u = x + x2 y 3 = c, then du = (1 + 2xy 3 )dx + 3x2 y 2 dy = 0.

dy 1 + 2xy 3
=−
dx 3x2 y 2
is a differential equation that we can solve by going backward. This gives a pow-
erful solution method as follows:

The form of differential equation

(1 + 2xy 3 )dx + 3x2 y 2 dy = 0

can be considered in general of the form

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

Definition 14 Exact differential equation A differential equation of the form

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

is said to be exact if there exists a differential function u(x, y) such that

∂u ∂u
M (x, y)dx + N (x, y)dy = du = dx + dy
∂x ∂y

Theorem 2 The necessary and sufficient condition for the ordinary differential
equation of the first order, say

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

to be exact is
∂M ∂N
=
∂y ∂x
where M (x, y) and N (x, y) are functions of x and y.

Proof:
Necessary Condition: As the equation is exact, so let there exists a differential
function u(x, y) such that

M (x, y)dx + N (x, y)dy = du

Now,
∂u ∂u
du = dx + dy
∂x ∂y

11
 Dr. Dil Gurung, Kathmandu University

Therefore,
∂u ∂u
M= , and N =
∂x ∂y
Now
∂M ∂ 2u
=
∂y ∂y∂x
and
∂N ∂ 2u
=
∂x ∂x∂y
Since u has a continuous partial derivatives, so
∂ 2u ∂ 2u
= By Euler’s theorem
∂y ∂x ∂x ∂y
Hence
∂M ∂N
=
∂y ∂x
Sufficient condition: We have to show that if
∂M ∂N
=
∂y ∂x
then
M (x, y)dx + N (x, y)dy = du
for some function u(x, y).
Let M dx = P, so M = ∂P
R
∂x
. Then

∂ 2P
 
∂N ∂N ∂ ∂P
= = =
∂x ∂y ∂y∂x ∂x ∂y
∂P
Integrating with respect to x, N = ∂y
+ f (y).
Hence,
∂P ∂P
M (x, y)dx + N (x, y)dy = dx + dy + f (y)dy = d[P + F (y)]
∂x ∂y
where d[F (y)] = f (y)dy. Let P + F (y) = u(x, y). Then

M (x, y)dx + N (x, y)dy = du

Hence the theorem.

Note that
(i) Based on above theorem, we can say that a differential equation of the form

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

is exact iff
∂M ∂N
=
∂y ∂x
(ii) In order to obtain the solution of an exact differential equation, we have to
proceed as follows:

12
 Dr. Dil Gurung, Kathmandu University

(a) Integrating M with respect to x, keeping y as constant.

(b) Integrate with respect to y only those terms of N which do not contain x.

(c) Add the two expressions obtained in (a) and (b) above and equate the result
to an arbitrary constant. In other words, the solution of an exact differential
equation is
Z Z
M dx + N dy = C

with first integral is y with constant and second integral is term not having
x.
OR,
Z Z
M dx + N dy = C

with first integral is term not containing y and second integral is x constant.

Example 14 Solve (x3 + 3xy 2 )dx + (3x2 y + y 3 )dy = 0.

Solution
Step 1: Test the exactness
Comparing the given differential equation with

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

we get
M = x3 + 3xy 2 and N = 3x2 y + y 3
Now,
∂M ∂N
= 6xy =
∂y ∂x
So the given differential equation is exact.
Step 2: Solution Process: Using
Z Z
M dx + N dy = C

with first integral is y with constant and second integral is term not having x.
we get
Z Z
(x + 3xy )dx + y 3 dy = C
3 2

1 4 3 2 2 1 4
x + xy + y = C
4 2 4
1 4
x + 6x y + y 4 = C
2 2


4
OR,
Z Z
M dx + N dy = C

13
 Dr. Dil Gurung, Kathmandu University

with first integral is y with constant and second integral is term not having x.
we get
Z Z
x dx + (3x2 y + y 3 )dy = C
3

1 4 3 2 2 1 4
x + xy + y = C
4 2 4
1 4
x + 6x2 y 2 + y 4 = C


4
Exercise 9 Show that the following equations are exact and solve them.

1. 2xydx + x2 dy = 0 4. xdx + ydy + xdy−ydx

x2 +y 2
=0
2. e3θ (dr + 3rdθ) = 0 x/y x/y
5. (1 + e )dx + e (1 − x/y)dy = 0
3. (cot y + x2 )dx = x csc2 y dy 6. (x2 + y 2 )dx + 2xydy = 0, y(1) = 2

2.1.3 Reducible to exact form

Let us begin with the following example.
Example 15 Consider a differential equation
−ydx + xdy = 0 (22)
Comparing with the differential equation
M (x, y)dx + N (x, y)dy = 0, (23)
we get
M (x, y) = −y and N (x, y) = x
Now
∂M ∂N
= −1 and =1
∂y ∂x
So
∂M ∂N
6=
∂y ∂x
and thus the differential equation (22) is not exact. Next, multiplying the equation
(22) by a function F (x) = x12 , we get
−ydx + xdy y 1
2
= − 2 + dy (24)
x x x
−y
Take M = x2
and N = x1 , we obtain
∂M 1 ∂N 1
=− 2 and =− 2
∂y x ∂x x
So the differential equation (24) is exact. The equation (24) can rewrite as
−ydx + xdy y 1 y
= − 2 + dy = d =0 (25)
x2 x x x
Integrating (25) gives the general solution
y = cx,
where c is a constant.

14
 Dr. Dil Gurung, Kathmandu University

This example illustrates the simple idea of reduction non exact differential equa-
tion to exact form. All we have done was the multiplication of a given non-exact
equation, say

M (x, y)dx + N (x, y)dy = 0 (26)

by a function F, which in general will be a function of x and y. The result was

an equation

F M dx + F N dy = 0 (27)

that is exact, so we can solve it as just discussed. The function F = F (x, y) is

then called an integrating factor on non-exact differential equation (26).

Example 16 We can readily verify that the other integrating factors of the non-
exact differential equation
−ydx + xdy = 0
1
are namely , 1
y 2 xy
and 1
x2 +y 2
, because

−ydx + xdy y −ydx + xdy   y  −ydx + xdy   y 

= −d ; = −d ln ; = d arc tan
y2 x xy x x2 + y 2 x

Note: The integrating factor is not unique.

2.1.4 How to find integrating factors?

Method 1: Integrating factor found by inspection Integrating factor can
be easily found by inspection if we recognize the following differentials:
 y  xdy − ydx  
x xdy − ydx   y  xdy − ydx
d(xy) = xdy+ydx; d = ; d = ; d ln = ;
x x2 y y2 x xy
  
x ydx − xdy   y  xdy − ydx  y  2xydy − y 2 dx
d arctan = ; d arctan = ; d = ;
y x2 + y 2 x x2 + y 2 x x2
y 2 dx + 2xydy
 
1 xdy + ydx dx − dy
d 2
= 2 4
; d (arcsin(xy)) = p ; d (ln(x − y)) = ;
xy xy 1−x y2 2 x−y
 2
dx + dy x 2xydx − x2 dy
d (ln(x + y)) = ; d =
x+y y y2
Method 2: In case the non-exact differential equation M dx + N dy = 0 be
dy
reducible into the form dx = F (y/x). where M and N are functions of x and y,
and M x + N y 6= 0, then the integrating factor
1
= .
Mx + Ny

Method 3: When M x − N y 6= 0 and the equation is of the form yf1 (x, y)dx +
xf2 (x, y)dy = 0, then
1
= .
Mx − Ny

15
 Dr. Dil Gurung, Kathmandu University

is an integrating factor of the equation M dx + N dy = 0.

Method 4: In case of the differential equation M dx + N dy = 0, when
 
1 ∂M ∂N
−
N ∂y ∂x

is a function of x alone, say f (x), then the integrating factor of M dx + N dy = 0

is R
I.F. = e f (x)dx
Method 5: In case of the differential equation M dx + N dy = 0, when
 
1 ∂N ∂M
−
M ∂x ∂y

is a function of y alone, say f (y), then the integrating factor of M dx + N dy = 0

is R
I.F. = e f (y)dy
Method 6: If the equation is of the form

xa y b (mydx + nxdy) = 0

then
I.F. = xkm−a−1 y kn−b−1
where k has any value, and a, b, m and n are constants.

Example 17 Solve y(1 + xy)dx − xdy = 0.

solution We have

ydx − xdy + xy 2 dx = 0
ydx − xdy
+ xdx = 0
y2
 
x
d + xdx = 0
y
Integrating, we get
x x2
+ = k,
y 2
where k is a constant of integration.

Example 18 Solve (x3 + xy 4 )dx + 2y 3 dy = 0.

Solution The equation is of the form M dx + N dy = 0. So

M = x3 + xy 4 ; N = 2y 3

Then
∂M ∂N
= 4xy 3 ; =0
∂y ∂x
Now  
1 ∂M ∂N 1
− = 3
(4xy 3 ) = 2x,
N ∂y ∂x 2y

16
 Dr. Dil Gurung, Kathmandu University

which is a function of x only, say f (x). Therefore, the integrating factor is

2
R R
f (x)dx 2xdx
I.F. = e =e = ex

Multiplying the equation by I.F., we get

2 2
ex (x3 + xy 4 )dx + ex (2y 3 )dy = 0

which is exact(How ?). The solution is (with first integration y treated constant
and in the second integration term not having x) then given by
Z Z Z
3 x2 4 x2
x e dx + y xe dx + 0dy = 0

Upon integration, we find

2
ex (x2 + y 4 − 1) = 0
which is the required solution.

Theorem 3

1. If M (x, y)dx + N (x, y)dy = 0 be a nonexact differential equation and

∂M ∂N
∂y
− ∂x
= h(u) where u = xy
Ny − Mx
then R
h(u)du
I.F. = e

2. If M (x, y)dx + N (x, y)dy = 0 be a nonexact differential equation and

∂M ∂N
∂y
− ∂x
= h(v) where u = x + y
N −M
then R
h(v)dv
I.F. = e

Exercise 10 Show that the given function is an integrating factor and solve.
1. sin y dx + cos y dy = 0, ex
2. y dx + [y + tan(x + y) dy] = 0, cos(x + y)
3. (a + 1) y dx + (b + 1) x dy = 0, xa y b
4. (2y + x y) dx + 2 x dy = 0, 1/xy
5. 2 cos y dx − tan 2x sin y dy = 0, cos 2x

Exercise 11 Solve the following differential equations.

1. 2xy dx + 3x2 dy = 0
2. (x2 + y 2 ) dx − 2xy dy = 0
3. 2cosh x cos y dx − sinh x sin y dy = 0
y3 2
4. y + 3 + x2 dx + 14 (x + xy 2 ) dy = 0
5. (xy 3 + y) dx + 2(x2 y 2 + x + y 4 ) dy = 0

17
 Dr. Dil Gurung, Kathmandu University

2.1.5 Linear differential equations

Definition 15 Linear differential equation of the first order A differential
equation of the form
dy
+ p(x) y = q(x) (28)
dx
where p(x) and q(x) are functions of x alone or constants, is called a linear
differential equation of the first order in y.

If the right hand side q(x) is zero for all x in the interval in which we consider
the equation, the equation is said to be homogeneous; otherwise it is said to be
non-homogeneous.
How to find solution ?
Homogeneous The homogeneous linear first order differential equation
dy
+ p(x)y = 0 (29)
dx
can be found simply by separating the variables. That is,
dy
= −p(x) dx
x
Integrating both sides, we get
Z
ln |y| = − p(x)dx + c∗

where c∗ is the constant of integration. By taking exponential on both sides

∗
R
y(x) = Ce− p(x)dx
, c = ±ec when y > or < 0

For, in the differential equation y 0 − y = 0, we have p(x) = 1. So the general

solution is R
y(x) = ce −1dx = c e−x .
Nonhomogeneous The nonhomogeneous linear first order differential equation
is
dy
+ p(x)y = q(x), (q(x) 6= 0) (30)
dx
R
p(x)dx
Multiplying on both sides of the given equation by e , we get
dy R p(x)dx R R
e + p(x)e p(x)dx y = q(x)e p(x)dx
dx
d  R p(x)dx  R
ye = q(x)e p(x)dx
dx  
R R
d ye p(x)dx = q(x)e p(x)dx dx

Integrating both sides, we get

R R
p(x)dx p(x)dx
ye = q(x)e dx + c

18
 Dr. Dil Gurung, Kathmandu University

where c is a constant of integration.

Therefore, Z 
R R
− p(x)dx p(x)dx
y(x) = e q(x)e dx + c

This is the general solution of the equation (30) in the form of an integral. Thus,
we can summarize the result as
Theorem 4 Consider the equation
y 0 + p(x) y = q(x)
where p(x) and q(x) are continuous functions of x on an interval I. If c is a
constant, then the general solution of this equation is given by
R
Z R 
− p(x)dx p(x) dx
y(x) = e e q(x) dx + c

The function yh (x) given by

R
yh (x) = c e− p(x) dx

is a solution of the homogeneous equation

y 0 + p(x) y = 0
Moreover, if x0 is a point in I, then the solution of this equation is
R
Z x R
− p(x) dx
yp (x) = e e p(t) dt q(t) dt
x0

We remark that yp (x) is the particular solution of the equation y 0 +p(x) y =

q(x) (the case c = 0) and yh (x) is the general solution of the homogeneous
equation y 0 + p(x) y = 0. The function y(x) = yh (x) + yp (x) is the general
solution of the equation y 0 + p(x) y = q(x).
Exercise 12 Show that the integrating factor of the differential equation
dy
+ p(x) y = q(x)
dx
is R
p(x)dx
e .
Working Process
dy
Step 1: First bring the equation in the standard form dx + p(x)y = q(x), that is,
dy
the coefficient of dx should be unity.
Step 2: Find an integrating factor, I.F. = ep(x)dx .
Step 3: If we multiply the given equation by I.F. and integrate, then
Z
y(x) × I.F. = q(x) × I.F.dx + c

Note: If the linear differential equation of the first order in x is

dx
+ p(y) x = q(y)
dy
then I.F. is R
e− p(y)dy
.

19
 Dr. Dil Gurung, Kathmandu University
dy
Example 19 Solve (1 − x2 ) dx − xy = 1.
Solution We have
dy x 1
− 2
=
dx 1 − x 1 − x2
which is in the standard form of linear first order differential equation. So, inte-
grating factor is
−
R x
dx 1 2 √
I.F. = e 1−x2 = e 2 ln(1−x ) = 1 − x2 .

The solution is then

√ Z
1 √
y(x) × 1− x2 = 2
× 1 − x2 + c
1−x
Z
1
= √ +c
1 − x2
= arcsin x + c

Therefore,
arcsin x c
y(x) = √ +√ .
1 − x2 1 − x2
Exercise 13 Solve the following differential equations.
dy
1. sec x dx − y = sin x.
dy
2. sec x cos x dx + y (x sin x + cos x) = 1.
dy
3. (x + 3y + 2) dx = 1.
dy
4. x dx = 2y + x3 ex
dy
5. dx − (1 + 3/x)y = x + 2, y(1) = e − 1.
3
dy e−2x
6. dx
+ 6x2 y = x2
, y(1) = 0.

2.1.6 Reducible into linear form (Bernoulli’s equation)

The differential equation of the form
dy
+ p(x) y = q(x) y n , n is any real number (31)
dx
where p and q are functions of x alone or constants, is called a Bernoulli’s equa-
tion.
Solution Method Set

u(x) = [y(x)]1−n (32)

to reduce the equation (31) into linear. Differentiating (32) with respect to x, we
obtain
du dy
= (1 − n)y −n
dx dx
−n
Multiplying (31) by (1 − n)y , we obtain
dy
(1 − n)y −n + (1 − n)y −n p(x)y = (1 − n)y −n y −n q(x)
dx
dy
(1 − n)y −n + (1 − n)y 1−n p(x)y = (1 − n)q(x) (33)
dx
20
 Dr. Dil Gurung, Kathmandu University

Using (32) and (33) in (31), we get

du
+ (1 − n)u(x)p(x) = (1 − n)q(x)
dx
which is in the standard linear differential equation of the first order in u and
thus can be solved as usual. Thus the general solution of Bernoulli’s equation is
R
Z R
1−n (1−n)p(x)dx
y e = (1 − n) q(x) e (1−n)p(x)dx dx + C

Note:
(i) The credit for the solution of this equation goes to Leibnitz.
(ii) If n = 0 or n = 1, the equation (31) is linear.
(iii) If n = 0, the equation (31) is linear non-homogeneous and if n = 1, the
equation (31) is linear homogeneous.
dy
Example 20 Solve dx = ay − by 2 , (a, b are positive constants.) This equation
is called logistic equation or Verhulst equation. The equation was derived by
P.F.Verhulst (Belgian Mathematics) during logistic population growth model in
1838.

Solution: Let
dy
= ay − by 2 (34)
dx
This equation is a Bernoulli’s differential equation of the form
dy
+ py = qy n (35)
dx
with n = 2. So let

u(x) = y 1−2 = 1/y (36)

Differentiating (36) with respect to x, we get

du dy
= −y −2 (37)
dx dx
Multiplying (34) by the factor −y −2 , we get

dy
−y −2 − a(y −2 )y = −by 2 (−y −2 )
dx
Using (36) and (37), we get

du
+ au(x) = b (38)
dx
which is a linear first order differential equation. So,
R
adx
I.F. = e = eax

21
 Dr. Dil Gurung, Kathmandu University

Multiplying (38) by I.F., and integrating, we get

Z
u(x) × I.F. = b × I.F.dx + c, c is a constant of integration
Z
ax
u(x)e = beax dx + c
b ax
=
e +c
a
b
u(x) = + ce−ax
a
Thus the general solution is
1 b
u= = + ce−ax (39)
y a
1 1
y= = b (40)
u a
+ ce−ax
dy
Note: (i) If y(0) = a/b, then dx = 0. Therefore y(x) remains unchanged.
dy
(ii) If y(0) > a/b, then dx < 0, and so y(x) decreases and approaches to a/b
whenever x → ∞.
dy
(iii) If y(0) < a/b, then dx > 0, and so y(x) increases and approaches to a/b
whenever x → ∞.
dy
(iv) If y(0) < 0, then dx < 0, and y(x) → ∞ as x → ∞.
Remark This equation can be solved using variable separation method.

Exercise 14 Reduce to linear form and solve the following equations.

1. y 0 + 2y = y 2 5. 2xyy 0 +(x−1)y 2 = x2 ex (y 2 = z)
2. y 0 + y = −x/y
6. (1 − x2 )y 0 + xy = xy 2
3. y 0 + 13 y = 13 (1 − 2x)y 4
4. y 0 + xy = xy −1 7. xy 0 + y = y 2 ln x

3 Applications – First order

3.1 Physical applications
3.1.1 Newton’s law of cooling/heating
It states that the rate of change of the temperature of an object is proportional
to the temperature difference between the body and its surrounding medium.

22
 Dr. Dil Gurung, Kathmandu University

Let T (t) denote the temperature of the object at time t and let Ts denote
the temperature of the surroundings (Ts is assumed to be constant throughout).
Then Newton’s law of cooling/heating is formulated as
dT
= k(T − Ts ), k < 0 (41)
dt
where k is the proportionality constant. This is first order first degree equation,
and using variable separation method, the solution is
T (t) = Ts + Cekt (42)
Here k < 0 makes sense because the temperature difference must decrease as time
increases. As t → ∞, ekt → 0, and so T → Ts .
Example 21 Suppose that you turn off the heat in your home at night 2 hours
before you go to bed; call this time t = 0. If the temperature T at t = 0 is 66◦ F
and at the time you go bed (t = 2) has dropped to 63◦ F, what temperature can
you expect in the morning, say, 8 hours later (t = 10)? Of course, this process
of cooling off will depend on the outside temperature Ts , which we assume to be
constant at 32◦ F.
Solution: Using Ts = 32◦ F in Eq. (42)
T (t) = 32 + Cekt
At t = 0, T (0) = 66 ◦ F, we obtain C = 34. So
T (t) = 32 + 34 ekt
1 31
Using T (2) = 63, we obtain k = 2
ln 34 = −0.046187. Thus
T (t) = 32 + 34 e−0.046187 t
Hence,
T (10) = 32 + 34 e−0.046187·10 = 53.4◦ F.
Exercise 15 A cup of fast food coffee is 180◦ F when freshly poured. After 2
minutes in a room at 70◦ F, the coffee has cooled to 165◦ F . Find the temperature
at any time t and find the time at which the coffee has cooled to 120◦ F. [Ans:
≈ 10.76 minutes]

3.1.2 Newton’s second law of motion

Force = Mass × Acceleration

F = ma
dv dv
= m , a = (acceleration)
dt dt
d2 x dx
= m 2, v= (velocity)
dt dt
From this, for a constant acceleration a, we obtain
v(t) = a t + v0 , v0 = initial velocity
and
1 2
x(t) = a t + v0 t + x0 , x0 = initial displacement
2

23
 Dr. Dil Gurung, Kathmandu University

Example 22 (Air plane take off )

An airplane taking off from a landing field has a run of 2 kilometers. If the plane
starts with a speed of 10 meter per seconds, moves with constant acceleration, and
makes the run in 50 seconds, with what speed does it take off?
Solution: Here, the initial displacement x0 = 0, distance travelled at t = 50
seconds = 2000 meters, and initial velocity v0 = 10 meter/second. Using the
model equation
1 2
x(t) = a t + v0 t + x0
2
1
= a (50)2 + 10 × 50 + 0
2
⇒ a = 1.2, meter/second 2

Then

v(t) = a t + v0
v(t = 50 second) = 1.2 × 50 + 10 = 70 meter/second = 294 km/hr.

3.1.3 Radio active decay

Model equation
dA
= k A, k<0
dt
k is called decay constant.
Half life period:
Solution:
A(t) = A0 ekt , where A(t = 0) = A0
Taking A0 /2 for A(t) at t1/2 , we obtain

ln 2 0.6931
t1/2 = − ≈−
k k
Exercise 16 If you have 50 grams of C 14 today, how much will be left in 100
years? Given that t1/2 of C 14 is 5730 years.
Solution: We have
ln 2 ln 2
k=− =− ≈ −0.000120968
t1/2 5730

After 100 years:

ln 2
A(t) = 50 e− 5730 ×100 ≈ 49.3988 grams

3.1.4 A simple electric circuits

An electric circuit containing a resistor, an inductor or a capacitor in series with
a source of electromotive force (emf) constitute a linear first order first degree
differential equation.

24
 Dr. Dil Gurung, Kathmandu University

• An electromotive force, E (unit: volts) usually a battery or generator, drives

an electric charge Q (unit: Coloumb) and produces a current I (unit: am-
peres). The current is defined as the rate of flow of the charge, and we can
write
dQ
I=
dt
• A resistor of resistance R (unit: ohms) is a component of the circuit that
opposes the current, dissipating the energy in the form of heat. It produces
a drop in voltage given by Ohm’s law

ER = RI

• An inductor of inductance L (unit: henrys) oppose any change in current

by producing a voltage drop of
dI
EL = L
dt

• A capacitor of capacitance C (unit: farads) stores charge. In so doing, it

resists the flow of further charge, causing a drop in the voltage of
1
EC = Q
C
Kirchhoff ’s laws
The physical principles governing electrical circuits were formulated by G. R.
Kirchhoff, a German physicist, in 1859. They are the following:
1. Kirchhoff ’s current law The algebraic sum of the currents flowing into
any junction point must be zero.
2. Kirchhoff ’s voltage law The algebraic sum of the instantaneous changes
in potential (voltage drops) around any closed loop must be zero.
RL-circuit
Consider the RL-circuit consisting the resistor and the inductor causing voltage
drop of ER , and EL respectively. The emf, however, provides a voltage E (that
is a voltage drop of −E). Thus Kirchhoff’s voltage law yields

EL + ER − E = 0
EL + ER = E
dI
L + RI = E
dt
which is a first order linear differential equation. Its solution is then
Z
−(R/L) t E(t) (R/L) t
I(t) = e e dt + Ke−(R/L)t (43)
L
For instance, if E(t) = constant = E0 and I(0) = I0 , then above solution becomes
 
E0 E0
I(t) = + I0 − e−Rt/L (44)
R R

25
 Dr. Dil Gurung, Kathmandu University

As t → ∞, the exponential term in (44) tends to zero, and I(t) → E0 /R. Thus
E0

−Rt/L
we call the E0 /R term in (44) the steady state solution and I0 − R e
term the transient part of the solution.

An another case, E(t) = E0 sin ωt, I0 = 0, then (43) gives

 
E0 ω L −Rt/L R
I(t) = 2 e + sin ωt − cos ωt (45)
R + (ω L)2 ωL

As t → ∞, the exponential term in (45) tends to zero, and we are left with the
steady solution
 
E0 ω L R
I(t) = 2 sin ωt − cos ωt
R + (ω L)2 ω L

Example 23 An induction of 2 henrys and a resistance of 10 ohms are connected

in series with an emf E = 100 sin 60t. If the current is zero when t = 0, what is
the current at the end of 0.1 seconds ?
solution: Using L = 2 H, R = 10 Ω, ω = 60 and E0 = 100 into (45), we obtain
24 −5t 2 sin 60t − 24 cos 60t
I(t) = e +
|29{z } | 29
{z }
transient part steady solution

Thus
24 −0.5 2 sin 6 − 24 cos 6
I(0.1) = e + ≈ −0.31 amperes
29 29

RC-circuit
Consider the RC-circuit consisting the resistor and the capacitor causing voltage
drop of ER , and EC respectively. The emf, however, provides a voltage E (that
is a voltage drop of −E). Thus Kirchhoff’s voltage law yields

ER + EC − E = 0
ER + EC = E
1
RI + Q = E
C
dQ
Using the fact I = dt
, we obtain the linear first order differential equation for
charge Q :
dQ 1
R + Q=E
dt C
Its solution is then
Z
−(t/RC) E(t) (t/RC)
Q(t) = e e dt + Ke−(t/RC)
R
Differentiating the equation
1
RI + Q=E
C

26
 Dr. Dil Gurung, Kathmandu University

with respect to time t, we obtain

dI 1 dQ dE
R + =
dt C dt dt
dI 1 dE
R + I =
dt C dt
which is also a first order linear differential equation. Its solution is then
Z
−(t/RC) 1 0
I(t) = e E (t) e(t/RC) dt + Ke−(t/RC)
C

Exercise 17 If a resistance of 2000 Ω and a capacitance of 5 × 10−6 farad are

connected in series with an emf of 100 volts, what is the current at t = 0.1 second
if I(0) = 0.01 ampere ?
Solution: Q(t) = 20001
+ ce−100t , I(t) = 1001 −100t
e . Thus I(0.1) = 4.54 × 10−7
amperes.

Exercise 18 Find the current I(t) in the RC-circuit assuming that E = 100
volts, C = 0.25 farad, r is a variable according to R = (200 − t) ohms when
0 ≤ t ≤ 200 sec, R = 0 when t > 200 sec, and I(0) = 1 ampere.

3.2 Geometrical applications

We note the following important formulae:

1. Equation of the tangent to y = f (x) at the point (x, y) is:

dy
Y −y = (X − x)
dx
X, Y being the current coordinates.

2. Equation of the normal to y = f (x) at the point (x, y) is

dx
Y −y =− (X − x)
dy

3. Length of the tangent, that is, the part of the tangent between the point of
contact and the x-axis is q
dy 2


y 1 + dx
dy
dx

4. Length of the normal s  2

dy
y 1+
dx

5. Length of the cartesian subtangent

dx
y
dy

27
 Dr. Dil Gurung, Kathmandu University

6. Length of the cartesian subnormal

dy
y
dx

Example 24 Find the curve through the point (1, 1) in the xy-plane having at
each of its points the slope −y/x.
Solution: We have
dy y
=−
dx x
Solving this differential equation, we obtain
c
y=
x
This gives the family of hyperbolas. For the curve to pass through (1, 1), we must
have y = 1 when x = 1. From this condition, we obtain c = 1. This gives the
particular solution y = x1 .

Exercise 19 Show that the equations of the curves for which the cartesian sub-
tangent is constant can be put in the form y = c ex/a , where a, c are arbitrary
constants.

Exercise 20 Find the equation of the curve for which the cartesian sub-tangent
varies as the reciprocal of the square of the abscissa.

3.3 Orthogonal trajectories

Definition 16 Let there be a family of curves f (x, y, c) = 0, where c is a variable
constant. Then the curves, which cut these everywhere at right angles, are called
orthogonal trajectories.

Rules:
dy
Step 1: Find dx of the given equation.
Step 2: Eliminate the constant.
dy
Step 3: Replace dx by − dx
dy
.
Step 4: Solve the new differential equation.
Example 25 Find the orthogonal trajectories of the curves y = cx2 .
Solution:
dy
Step 1: Find dx .
Here,
dy
= 2cx
dx
Step 2: Eliminating the constant c.
dy y 2y
=2 2x=
dx x x
dy
Step 3: Replacing dx
by − dx
dy
.
dx 2y
− =
dy x

28
 Dr. Dil Gurung, Kathmandu University

Step 4: Solving the differential equation:

Using variable separable, and integrating

x2 y 2
+ =λ
2 1
which gives a family of ellipses.

Exercise 21 Find the orthogonal trajectories of the following family of curves,

where c is an arbitrary constant:

1. xy = c 4. x2 + (y − c)2 = c2
2
2. y = c ex /2 5. c y 2 = x2
3. y2 = 4 c x 6. x2/3 + y 2/3 = c2/3

4 Linear equations with constant coefficients:

Higher order but degree one
4.1 Introduction
A linear differential equation of order n but of degree one is of the form

y (n) + a1 (x) y (n−1) + a2 (x) y (n−2) + · · · + an−1 (x) y (1) + an (x) y = b(x) (46)

where a1 (x), a2 (x), · · · , an (x) are functions of x or constants.

• If all the ai (x), 1 ≤ i ≤ n are constant, then the equation (46) is called a
linear differential equation of order one with constant coefficients, other-
wise it is said to be with variable coefficients. Thus a linear differential
equation with constant coefficients is

y (n) + a1 y (n−1) + a2 y (n−2) + · · · + an−1 y (1) + an y = b(x) (47)

where a1 , a2 , · · · , an are constants.

• It will be convenient to denote the differential expressions on the left of the

equality (47) by L(y). Thus

L(y) = y (n) + a1 y (n−1) + a2 y (n−2) + · · · + an−1 y (1) + an y

and the equation (47) becomes simply

L(y) = b(x)

Here L is an operator, called a differential operator which operates on

functions which have n derivatives on an interval I.

• If b(x) = 0 for all x in I the corresponding equation L(y) = 0 is called a

homogeneous equation, whereas if b(x) 6= 0 for some x in I, L(y) = b(x)
is called nonhomogemeous equation.

29
 Dr. Dil Gurung, Kathmandu University

Theorem 5 (Existence and uniqueness theorem) Let a1 (x), a2 (x), · · · , an (x)

and b(x) be continuous on an interval I, and let c1 , c2 , · · · , cn be n given constants.
Then there exists a unique solution y(x) that satisfy the linear differential equation
(46) on I, and the n initial conditions
y(x0 ) = c1 , y 0 (x0 ) = c2 , · · · , y (n) (x0 ) = cn
for some x0 in I.
Theorem 6 (Superposition principle or Linearity principle) If y1 , y2 , · · · , yn
are n solutions of L(y) = 0, and c1 , c2 , · · · , cn are n constants, then the function
y(x) = c1 y1 (x) + c2 y2 (x) + · · · + cn yn (x)
is also a solution of L(y) = 0.
Note that the theorem holds for homogeneous linear ODEs only but does not
hold for nonhomogeneous linear or nolinear ODEs.

4.2 The second order homogeneous equations with con-

stant coefficients
Here we are concerned with the equation
L(y) = y 00 + a1 y 0 + a2 y = 0 (48)
where a1 and a2 are constants.
Recall that the first order differential equation
y0 + a y = 0
has a solution e−a x . The constant −a is a solution of the equation λ + a =
0. Since differentiating an exponential eλ x any number of times, where λ is a
constant, always yields a constant times eλ x , it is reasonable to expect that for
some appropriate constant λ, eλ x will be a solution of the equation (48). Thus
we find
L(eλ x ) = (λ2 + a1 λ + a2 )eλ x
and eλ x will be a solution of L(y) = 0, that is, L(eλ x ) = 0, if λ satisfies
λ2 + a1 λ + a2 = 0
We let
p(λ) = λ2 + a1 λ + a2
and call p the characteristic polynomial of L, or of the equation (48), and
p(λ) = 0 is called the characteristic equation. The characteristic equation
p(λ) = 0 being quadratic in λ, it has two roots,
p
−a1 + a21 − 4a2
λ1,2 =
2
√
and the nature of the roots depend on the discriminant D = a1 − 4a2 , that is,
the roots λ1,2 may be real and unequal, real and equal, and complex conjugate.
We now formulate the situations as in a theorem.

30
 Dr. Dil Gurung, Kathmandu University

Theorem 7 Let a1 , a2 be constants, and consider the equation

L(y) = y 00 + a1 y 0 + a2 y = 0
and let λ1 , λ2 be the roots of the characteristic equation
p(λ) = λ2 + a1 λ + a2 = 0
i. If λ1 and λ2 are real and unequal, then the functions defined by
y1 (x) = eλ1 x and y2 (x) = eλ2 x (49)
are the solution of L(y) = 0.
ii. If λ1 and λ2 are real and repeated, that is, if λ = λ1 = λ2 , then the function
defined by
y1 (x) = eλ x and y2 (x) = x eλ x (50)
are the solution of L(y) = 0.
iii. λ1 = α+i β and λ2 = α−i β are complex conjugate roots, then the functions
defined by
y1 (x) = eα x cos β x and y2 (x) = eα x sin β x (51)
are the solution of L(y) = 0.
Definition 17 Linearly dependent and independent
1. Two functions y1 and y2 defined on an interval I are said to be linearly
dependent on I if there exist two constants c1 , c2 , not both zero, such that
c1 y1 (x) + c2 y2 (x) = 0
for all x in I.
2. The functions y1 , y2 are said to be linearly independent on I if they are
not linearly dependent there. Thus y1 , y2 are linearly independent on I if
the only constants c1 , c2 such that
c1 y1 (x) + c2 y2 (x) = 0
for all x in I are the constants c1 = 0, c2 = 0.
Example 26 Show that the functions defined by y1 (x) = eλ1 x and y2 (x) = eλ2 x ,
where λ1 and λ2 are real and unequal, are linearly independent on any interval I.
Solution: Suppose
c1 eλ1 x + c2 eλ2 x = 0 (52)
for all x in I. Then, multiplying by e−λ1 x , we obtain
c1 + c2 e(λ2 −λ1 ) x = 0
and differentiating there results
c2 (λ2 − λ1 ) e(λ2 −λ1 ) x = 0
Since λ1 6= λ2 , and e(λ2 −λ1 ) x is never zero, this implies c2 = 0. But if c2 = 0, the
relation (52) gives c1 eλ1 x = 0, and so c1 = 0.
Exercise 22 Show that the functions defined by (50) and (51) are linearly inde-
pendent for all x on any interval I.

31
 Dr. Dil Gurung, Kathmandu University

4.2.1 Existence and uniqueness theorem

Theorem 8 For any real x0 , and constants α, β, there exists a unique solution
y of the initial value problem

L(y) = y 00 + a1 y 0 + a2 y = 0, y(x0 ) = α, y 0 (x0 ) = β

on −∞ < x < ∞.

Corollary 1 Let y1 , y2 be the two LI solutions of L(y) = y 00 + a1 y 0 + a2 y = 0. If

c1 , c2 are any two constants the function

y(x) = c1 y1 (x) + c2 y2 (x)

is a solution of

L(y) = y 00 + a1 y 0 + a2 y = 0 on − ∞ < x < ∞

Conversely, if y is any solution of

L(y) = y 00 + a1 y 0 + a2 y = 0 on − ∞ < x < ∞

there are unique constants c1 , c2 such that

y(x) = c1 y1 (x) + c2 y2 (x)

Definition 18 (General solution) If y1 and y2 are two linearly independent

solutions of the equation (48) on an open interval I, then the solution of the form

y(x) = c1 y1 (x) + c2 y2 (x) (53)

on the interval I where c1 , c2 are constants, is called a general solution of the

equation (48).

• The two linearly independent solutions y1 , y2 are called a basis (or a fun-
damental system) of solutions of (48) on I.

– If λ1 , λ2 are the real and unequal roots of the characteristic equation

p(λ) = λ2 + a1 λ + a2 = 0 then the general solution of the equation (48)
is of the form
y(x) = c1 eλ1 x + c2 eλ2 x
where c1 , c2 are arbitrary constants.
– If λ = λ1 = λ2 are the real and equal roots of the characteristic
equation p(λ) = λ2 + a1 λ + a2 = 0 then the general solution of the
equation (48) is of the form

y(x) = (c1 + c2 x) eλ x

where c1 , c2 are arbitrary constants.

32
 Dr. Dil Gurung, Kathmandu University

– If λ1 = α + i β, λ2 = α − i β, α, β ∈ R (β > 0) are complex conjugate

roots of the characteristic equation p(λ) = λ2 + a1 λ + a2 = 0 then the
general solution of the equation (48) is of the form

y(x) = eα x [A cos β x + B sin β x]

where A, B are arbitrary constants.

• A particular solution of (48) on I is obtained if we assign specific values

of c1 and c2 in (53).

• In case of two functions y1 , y2 , linearly independent (LI) and linearly de-

pendent (LD) can be checked easily by taking ratios of these two functions.
y1 y2
– If y2
= constant or y1
= constant, then these functions are LD.
y1 y2
– If y2
= function of x or y1
= function of x, then these functions are LI.

• There is a simple test which enables us to tell whether two solutions y1 , y2

of L(y) = 0 are linearly independent or not. It involves the determinant
 
y y2 
W (y1 , y2 ) =  10 = y1 y20 − y2 y10
y1 y20 

which is called the Wronskian1 (1812) of y1 , y2 . It is a function, and its

value at x is denoted by W (y1 , y2 )(x).

4.2.2 Wronskian
In general, the Wronskian of y1 , y2 , · · · , yn is defined by
 
 y1
 0 y 2 y 3 · · · yn−1 y n


0 0 0 0
 y1
 00 y 2 y 3 · · · yn−1 y n


00 00 00 00 
 y1
W (y1 , · · · , yn ) =  y2 y3 · · · yn−1 yn 
 .. .. .. .. .. .. 
 . . . . . . 
 (n−1) (n−1) (n−1) (n−1) (n−1) 

y 1 y2 y3 · · · yn−1 yn

We now discuss the following theorems for second order linear homogeneous dif-
ferential equation, which are similarly applied to higher order linear homogeneous
differential equations.

Theorem 9 (Abel’s theorem). If y1 , y2 are two solutions of

y 00 + a1 (x) y 0 + a2 (x) y = 0 (54)

on an interval I, and a1 (x) and a2 (x) are continuous on I, then the Wronskian
of the two solution is
R
W (y1 , y2 )(x) = c e− a1 (x) dx
(55)
1
Jozef Maria Hoene Wronski (1778-1853) A Polish mathematician now remembered because
of his introduction of the functional determinant called the Wronskian.

33
 Dr. Dil Gurung, Kathmandu University

where c is a constant depending on only y1 , y2 but not on x.

Proof: Since y1 (x) and y2 (x) are solutions of (54), their second derivatives exist.
Hence,
W 0 (y1 , y2 )(x) = y1 (x)y200 (x) − y100 (x)y2 (x)
and
y100 + a1 (x)y10 + a2 (x)y1 = 0 and y200 + a2 (x)y20 + a2 (x)y2 = 0
Multiplying the first of these equations by y2 and the second by y1 and subtracting,
we obtain
y1 y200 − y2 y100 + a1 (x)(y1 y20 − y2 y10 ) = 0
which is just
W 0 + a1 (x)W = 0
This equation is a linear first order equation, and so its solution is
R
W (y1 , y2 )(x) = c e− a1 (x) dx

for some arbitrary constant c. 

Equation (55) is known as Abel’s formula. If a1 (x) = a1 a constant, then

the Abel formula is
W (y1 , y2 )(x) = c e−a1 x
and if x0 is a point in I, then Abel’s formula reduces to

W (y1 , y2 )(x) = W (y1 , y2 )(x0 ) e−a1 (x−x0 )

Corollary 2 Let y1 , y2 are solutions to the equation y 00 + a1 (x) y 0 + a2 (x) y = 0.

Then either W (y1 , y2 )(x) = 0 for all x in I or never zero in I.

This corollary gives the following result:

For any two solutions y1 , y2 of L(y) = 0,

W (y1 , y2 )(x) 6= 0 ∀x ∈ I ⇔ ∃ x0 ∈ I such that W (y1 , y2 )(x0 ) 6= 0

The importance of Abel’s theorem is given by the following theorem.

Theorem 10 Two solutions y1 , y2 of L(y) = y 00 + a1 y 0 + a2 y = 0 are LI on an

interval I if, and only if
W (y1 , y2 )(x) 6= 0
for all x in I.
Proof: (⇒) Assume y1 , y2 are LI on an interval I, and we prove that W (y1 , y2 )(x) 6=
0 for all x in I. We prove it by contradiction method. For, suppose that there is
an x0 ∈ I such that W (y1 , y2 )(x0 ) = 0. This implies that the system of two linear
homogeneous equations in two unknowns

c1 y1 (x0 ) + c2 y2 (x0 ) = 0
(56)
c1 y10 (x0 ) + c2 y20 (x0 ) = 0

34
 Dr. Dil Gurung, Kathmandu University

has a nontrivial solution c1 , c2 , that is, c1 , c2 both are not equal to zero. Let c1 , c2
be such a solution and consider the function

y(x) = c1 y1 (x) + c2 y2 (x)

Now
L(y) = 0
and from (56), we obtain

y(x0 ) = 0, and y 0 (x0 ) = 0

Thus y(x) solves the initial-value problem

L(y) = y 00 + a1 y 0 + a2 y = 0, y(x0 ) = y 0 (x0 ) = 0

But this initial-value problem has the unique solution, so necessarily y(x) = 0 for
all x in I, and thus
y(x) = c1 y1 (x) + c2 y2 (x) = 0
But this contradicts the fact that y1 , y2 are LI on I. Thus the supposition that
there was a point x0 in I such that

W (y1 , y2 )(x0 ) = 0

must be false. We have consequently proved that

W (y1 , y2 )(x0 ) 6= 0

for all x in I.
(⇐) Assume that W (y1 , y2 )(x) 6= 0 for all x in I, and let c1 , c2 be constants such
that

c1 y1 (x) + c2 y2 (x) = 0 (57)

for all x in I. Also

c1 y10 (x) + c2 y20 (x) = 0 (58)

for all x in I. The determinant of the coefficients is just W (y1 , y2 )(x) which is
not zero. Therefore, c1 = 0, c2 = 0 is the only solution of (57), (58). This proves
that y1 , y2 are LI on I. 

It is convenient to compute W (y1 , y2 ) at only one convenient point to test the

linear independence of the solutions y1 , y2 of the equation L(y) = y 00 +a1 y 0 +a2 y =
0, which is stated as in the following theorem.
Theorem 11 Let y1 , y2 be two solutions of

L(y) = y 00 + a1 y 0 + a2 y = 0

on an interval I, and let x0 be any point in I. Then y1 , y2 are LI on I if, and only
if
W (y1 , y2 )(x0 ) 6= 0

35
 Dr. Dil Gurung, Kathmandu University

Based on the concept of linearly independence we can show that any two linearly
independent solutions of L(y) = 0 determines all solutions, in the sense of the
following theorem.
Theorem 12 Let y1 , y2 be two linearly independent solutions of
L(y) = y 00 + a1 y 0 + a2 y = 0
on an interval I. Every solution y of L(y) = 0 can be written uniquely as
y(x) = c1 y1 (x) + c2 y2 (x)
where c1 , c2 are constants.
The following theorem states the fact about linearly dependence of two dif-
ferentiable functions.
Theorem 13 If y1 and y2 are differentiable functions on an interval I such that
they are LD. Then W (y1 , y2 )(x) = 0 for all x in the interval I.
The converse of above theorem for LD need not be true. That is, if W (y1 , y2 )(x) =
0 for all x in I, then y1 , y2 may not be LD. This we can illustrates with the fol-
lowing example.

Example 27 Show that x and |x| are LI even though W (x, |x|) = 0 for all
x ∈ R\{0}.
Solution: We define √
|x| = x2 , ∀x ∈ R\{0}
then
d d√ 2 2x x
|x| = x = √ =
dx dx 2 x2 |x|
√
Take y1 = x, y2 = |x| = x2 , then
W (y1 , y2 )(x) = y1 y20 − y10 y2
x √
= x − 1 · x2
|x|
x2 √ 2
= − x
|x|
√ √
= x2 − x2
= 0
Suppose the functions x and |x| are linearly dependent, then we write
ax + b|x| = 0 ∀x ∈ R, |a| + |b| =
6 0
• If x < 0, then |x| = −x and it implies that x(a − b) = 0 ⇒ a = b.
• If x > 0, then |x| = x ⇒ ax + bx = 0 ⇒ a = −b.
Thus x and |x| are linearly independent on R\{0}.

Exercise 23 If y1 6= 0 on an open interval I, and W (y1 , y2 )(x) = 0 for all x in

I, the show that yy12 = constant.

36
 Dr. Dil Gurung, Kathmandu University

4.3 Examples
Example 28 Solve the equation y 00 + y 0 + y = 0, y(0) = 1, y 0 (0) = 3.
Solution: The characteristic equation is λ2 + λ + 1 = 0. Its roots are
√ √
−1 + i 3 −1 − i 3
λ1 = and λ2 =
2 2
√
3
Then α = − 21 and β = ,
so the general solution is
2
√ √ !
3 3
y(x) = e−x/2 c1 cos x + c2 sin x
2 2
To solve the initial-value problem, we differentiate, and set x = 0, and solve the
simultaneous equations
c1 = 1
√
3 1
c2 − c1 = 3
2 2
Thus c1 = 1, c2 = √7 , and
3
√ √ !
3 7 3
y(x) = e−x/2 cos x + √ sin x
2 3 2
Example 29 Solve the equation y 00 + 2y 0 − y = 0, y(0) = 0, y 0 (0) = −1.
Solution: The characteristic equation is λ2 + 2λ − 1 = 0. Its roots are
√ √
λ1 = −1 + 2 and λ2 = −1 − 2
The general solution is then
√ √
y(x) = c1 e(−1+ 2)x
+ c2 e(−1− 2)x

To solve the initial-value problem, we differentiate, and set x = 0, and solve the
simultaneous equations
c1 + c2 = 0
√ √
(−1 + 2)c1 + (−1 − 2)c2 = −1
Thus c1 = − 2√1 2 , c2 = 1
√
2 2
, and
1 √ 1 √
y(x) = − √ e(−1+ 2)x + √ e(−1− 2)x
2 2 2 2
Exercise 24 Solve the following differential equations.
1. y 00 − 4y = 0
2. y 00 − 3y 0 + 2y = 0
3. 4y 00 + 20y 0 + 25y = 0, y(0) = 1, y 0 (0) = 2
4. y 00 + 2y 0 + y = 0, y(1) = 2/e, y 0 (1) = −3/e
5. y 00 + y = 0, y(π) = 2, y 0 (π) = −1
6. y 00 + 2y 0 + 5y = 0, y(π) = e−π , y 0 (π) = 3e−π
7. y 00 − 8y 0 + 16y = 0, y(0) = 1, y 0 (0) = 6
8. Show that the boundary value problem
y 00 + y = 0, y(0) = 0, y 0 (π) = 1
has no solutions.

37
 Dr. Dil Gurung, Kathmandu University

5 The nth order homogeneous equations with

constant coefficients
The nth order linear homogeneous equation with constant coefficient is

y (n) + a1 y (n−1) + a2 y (n−2) + · · · + an−1 y (1) + an y = 0 (59)

where a1 , a2 , · · · , an are constants. If we substitute y = eλx in (59) and then

divided by eλx , we obtain the characteristic equation as

λn + a1 λn−1 + a2 λn−1 + · · · + an−1 λ + an = 0 (60)

Eq.(60) has n roots λ1 , λ2 , · · · , λn . Some of these roots may be real and distinct,
real and equal, distinct complex conjugate pairs, or equal complex conjugate
pairs.

Example 30 Find the general solution of y 000 − 3y 00 − 10y 0 + 24y = 0.

Solution: The characteristic equation is

λ3 − 3λ2 − 10λ + 24 = (λ − 2)(λ + 3)(λ − 4)

with roots λ1 = 2, λ2 = −3 and λ3 = 4. Since the roots are real and distinct, three
linearly independent solutions are

y1 = e2x , y2 = e−3x , y3 = e4x

and the general solution is

y(x) = c1 e2x + c2 e−3x + c3 e4x

Exercise 25 Find the general solution of the differential equations:

1. y iv − 16y = 0
2. y 000 + 9y 00 + 27y 0 + 27y = 0
3. y iv + 5y 00 + 4y = 0

6 The second order nonhomogeneous equations

with constant coefficients
Theorem 14 Let b be continuous on an interval I. Every solution y of L(y) =
b(x) on I can be written as

y(x) = c1 y1 (x) + c2 y2 (x) + yp (x)

where yp (x) is a particular solution, y1 , y2 are two linearly independent solutions

of L(y) = 0, and c1 , c2 are constants. Conversely, every such y is a solution of
L(y) = b(x).
Proof: We have L(y − yp ) = L(y) − L(yp ) = b − b = 0 on I. This shows that
y − yp is a solution of the homogeneous equation L(y) = 0. Therefore, if y1 , y2 are

38
 Dr. Dil Gurung, Kathmandu University

linearly independent solutions of L(y) = 0, there are unique constants c1 , c2 such

that
y − y p = c1 y 1 + c2 y 2
It means that every solution y of L(y) = b(x) can be written in the form

y(x) = c1 y1 (x) + c2 y2 (x) + yp (x)

Conversely, if y1 , y2 be two linearly independent solution of L(y) = 0, then L(y1 ) =

L(y2 ) = 0. Note that if L(yp ) = b, and c1 , c2 are any constants, then

y(x) = c1 y1 (x) + c2 y2 (x) + yp (x)

satisfies L(y) = b.

Question: How to find the particular solution yp (x)?

Various methods:

1. Method of Variation of parameters.2

2. Method of undetermined coefficients.

6.1 Method of variation of parameters

Theorem 15 Let y1 , y2 be two linearly independent solutions of the equation
L(y) = 0, where L(y) = y 00 + a1 y 0 + a2 y = 0 on an open interval I. Then
the nonhomogeneous equation L(y) = b(x) has a particular solution yp (x) given
by Z Z
y2 (x) b(x) y1 (x) b(x)
yp (x) = −y1 dx + y2 dx
W (y1 , y2 )(x) W (y1 , y2 )(x)
Proof: Let us consider u, v such that the combination

yp = u y1 + v y2

will satisfy the equation L(y) = b(x). Then

yp0 = (uy10 + vy20 ) + (u0 y1 + v 0 y2 )

To simplify the expression (i.e., to avoid appearance of u00 and v 00 ), it is convenient

(but not necessary) to set
u0 y1 + v 0 y2 = 0
then
yp0 = uy10 + vy20
Differentiating once again, we obtain

yp00 = uy100 + u0 y10 + vy200 + v 0 y20

2
This method was first used by Johann Bernoulli in 1697 to solve linear equation of first
order, and then by Lagrange in 1774 to solve linear equations of second order.

39
 Dr. Dil Gurung, Kathmandu University

Substituting of the expressions for yp , yp0 and yp00 into the equation L(y) = b(x)
yields
yp00 + a1 yp0 + a2 yp = u0 y10 + v 0 y20 = b(x)
Finally the system of two linear equations

u0 y1 + v 0 y2 = 0
u0 y10 + v 0 y20 = b(x)

determine the functions u and v that we need. The determinant of the system is
 
y1 y2 
y1 y20  = W (y1 , y2 )(x) 6= 0
 0 

for all x in an open interval I, since y1 and y2 are linearly independent. Solving
the system of equations, we obtain
Z Z
y2 (x) b(x) y1 (x) b(x)
u=− and v =
W (y1 , y2 )(x) W (y1 , y2 )(x)
Hence, the particular solution yp is
Z Z
y2 (x) b(x) y1 (x) b(x)
yp (x) = −y1 dx + y2 dx
W (y1 , y2 )(x) W (y1 , y2 )(x)

Theorem 16 Let y1 , y2 , · · · , yn be n linearly independent solutions of the equa-

tion
L(y) = y (n) + a1 y (n−1) + a2 y (n−2) + · · · + an−1 y (1) + an y = 0
on an open interval I. Then the nonhomogeneous equation L(y) = b(x) has a
particular solution yp (x) given by
Z Z Z
W1 (x) b(x) W2 (x) b(x) Wn (x) b(x)
yp (x) = y1 dx + y2 dx + · · · + yn dx
W (x) W (x) W (x)

where W (x) is the Wronskian of y1 , y2 , · · · , yn and Wj (x), j = 1, · · · , n obtained

from W (x) by replacing the jth column of W (x) by the column

[0 0 ··· 0 1]T

Remark: This method is applicable when a1 , a2 , · · · , an are functions of x on an

interval I.
x
Example 31 Solve y 00 − 2y 0 + y = 12x3e by the variation of parameters.
Solution: The characteristic equation λ2 − 2λ + 1 = 0 has roots λ1,2 = 1, 1
which are real and equal. So the general solution of the homogeneous equation
L(y) = y 00 − 2y 0 + y = 0 is

yh (x) = (c1 + x c2 )ex

where c1 , c2 are constants. Thus the two linearly independent solutions of L(y) = 0
are
y1 = ex and y2 = x ex

40
 Dr. Dil Gurung, Kathmandu University

Here,
12 ex
W (y1 , y2 )(x) = e2x and b(x) =
x3
Then
xex 12ex /x3
Z Z
y2 (x) b(x) 12
u=− dx = − 2x
dx =
W (y1 , y2 )(x) e x
and
ex 12ex /x3
Z Z
y1 (x) b(x) 6
v= dx = 2x
dx = − 2
W (y1 , y2 )(x) e x
and so the particular solution is

yp = u y1 + v y2 = ex (12/x) + xex (−6/x2 ) = 6ex /x

Hence the general solution of the equation is

y(x) = yh (x) + yp (x) = (c1 + xc2 )ex + 6ex /x

Exercise 26 Use variation of parameters to solve the following differential equa-

tions.
1. y 00 + 4y = 4 sec 2x
2. y 00 − 4y 0 + 4y = 6 + e2x /x
3. y 00 − 2y 0 + y = ex sin x
4. y 00 + 2y 0 + y = 4e−x ln x
5. y 00 + y = 4 sin x, y(0) = 3, y 0 (0) = −1
6. y 00 + 2y 0 + 2y = sin 3x, y(0) = 2, y 0 (0) = 0
7. y 000 + 3y 00 + 3y 0 + y = 30e−x , y(0) = 3, y 0 (0) = −3, y 00 (0) = −47.
8. y 000 + 3y 00 + 3y 0 + y = e−x sin x, y(0) = 2, y 0 (0) = 0, y 00 (0) = −1.

6.2 Method of undetermined coefficients

This method is applicable in a linear differential equation in case of constant
coefficients, and if b(x) is a linear combination of (finite) products of the functions
of the following three types.

1. A polynomial in x.

2. An exponential function erx , where r is a constant.

3. cos qx or sin qx.

The method assumes that the solution to the equation L(y) = b(x) is of exactly
the same form as b(x). The technique requires that we replace each dependent
variable y in L(y) = b(x) with an expression of the same form as b(x) having
terms with undetermined coefficients. So the method is termed as the method of
undetermined coefficients.

Choice for yp :
The following rules are used for the choice of the particular solution yp of L(y) =
b(x).

41
 Dr. Dil Gurung, Kathmandu University

1. Basic rule: A particular solution of equation L(y) = b(x) has the form
yp (x) given by the table below:
b(x) Choice for yp (x)
A polynomial of degree n in x a0 + a1 x + a2 x2 + · · · + an xn
K erx M erx
K cos qx or K sin qx M cos qx + N sin qx

2. Modification rule: If any term of yp (x) guessed using Basic rule is a

solution of homogeneous equation L(y) = 0, use modification rule. The
rule is:
If any term of guessed yp (x) is a solution of the homogeneous equation
L(y) = 0, multiply yp (x) by x repeatedly until no term of the product xk yp (x)
is a solution of L(y) = 0. Then use the product xk yp (x) as a particular
solution of the equation L(y) = b(x).

3. Sum rule: If b(x) is a sum of functions in the first column of the above
table, choose for yp (x) the sum of functions in the corresponding lines of
second column. For example, if b(x) = x2 + e2x + sin 3x, then choose yp (x)
as yp (x) = a0 + a1 x + a2 x2 + M e2x + A sin 3x + B cos 3x.

Example 32 Solve y 00 + 2 y 0 + y = 2 x sin x using method of undetermined coef-

ficients.
Solution: The general solution of L(y) = y 00 + 2y 0 + y = 0 is

yh (x) = (c1 + x c2 ) ex

Here, b(x) = 2 x sin x, so we guess the particular solution as

yp (x) = (a0 + a1 x)(A cos x + B sin x)

= K1 cos x + K2 sin x + K3 x cos x + K4 x sin x
= (K1 + x K2 ) cos x + (K3 + x K4 ) sin x

where K1 = a0 A, K2 = a0 B, K3 = a1 A, K4 = a1 B. Using yp and its derivatives

in the equation y 00 +2 y 0 +y = 2 x sin x, we obtain K1 = 1, K2 = 1, K3 = −1, K4 =
0. Thus,
yp (x) = cos x + sin x − x cos x
Therefore, the required solution is

y(x) = yh (x) + yp (x) = (c1 + c2 x)ex + cos x + sin x − x cos x

Note: If b(x) = p(x) cos ωx or b(x) = p(x) sin ωx, where p(x) is a polynomial of
degree n, then we can choose:

yp = (a0 + a1 x + · · · + an xn ) cos ωx + (b0 + b1 x + · · · + bn xn ) sin ωx

Example 33 Find a general solution of y 00 − y = x ex + sin x.

Solution: The associated homogeneous equation is

y 00 − y = 0

42
 Dr. Dil Gurung, Kathmandu University

Its characteristic equation is

λ2 − 1 = 0 ⇒ λ = ±1

The complimentary function is then

yh (x) = c1 e−x + c2 ex

where c1 , c2 are arbitrary constants. By basic rule, to accommodate the non-

homogeneity x ex , we demand a particular solution of the form (c3 + c4 x) ex .
Since ex is a solution of the homogeneous equation, so the appropriate particular
solution to accommodate the non-homogeneity x ex is

x (c3 + c4 x) ex = c3 x ex + c4 x2 ex

under the modification rule. To accommodate the non-homogeneity sin x, we need

an undetermined coefficient form c5 sin x + c6 cos x under basic rule. Thus, the
appropriate particular solution yp under sum rule is given by

yp = c3 x ex + c4 x2 ex + c5 sin x + c6 cos x

where values of c3 through c6 in the particular solution are determined by substi-

tutions

yp00 − yp = [c3 x ex + c4 x2 ex + c5 sin x + c6 cos x]00

−[c3 x ex + c4 x2 ex + c5 sin x + c6 cos x] = x ex + sin x

This leads to c3 = − 14 , c4 = 14 , c5 = − 21 and c6 = 0. Thus a general solution is

 
−x 1 1 1
y(x) = c1 e + c2 e + x − + x ex − sin x
x
4 4 2

Example 34 Find a particular solution of 4y 00 +4y 0 +y = e−x/2 (−8+48x+144x2 ).

Solution: Substituting y = u e−x/2 and its derivatives
1 1
y 0 = u0 e−x/2 − ue−x/2 and y 00 = u00 e−x/2 − u0 e−x/2 + u e−x/2
2 4
into the equation, we obtain

4(u00 − u0 + u/4) + 4(u0 − u/2) + u = 4u00 = −8 + 48 x + 144 x2

Or,
u00 = −2 + 12 x + 36 x2
which does not contain u or u0 because e−x/2 and x e−x/2 are both solutions of the
complementary equation L(y) = 4y 00 +4y 0 +y = 0. To obtain a particular solution,
we integrate twice, taking the constant of integration to be zero, then

u0p = −2x + 6x2 + 12x3 and up = −x2 + 2x3 + 3x4 = x2 (−1 + 2x + 3x2 )

Therefore,
yp = up e−x/2 = x2 e−x/2 (−1 + 2x + 3x2 )

43
 Dr. Dil Gurung, Kathmandu University

Remark: Consider

L(y) = ay 00 + by 0 + cy = eαx g(x)

where g(x) is a polynomial of degree greater than zero, use the substitution

y = u eα x

Then the equation will turn out to be

a u00 + p0 (α) u0 + p(α) u = g(x)

where
p(λ) = a λ2 + b λ + c
is the characteristic polynomial of the complementary equation L(y) = 0. If eα x
is a solution of the complementary equation then p(λ) = 0, and so

u00 + p0 (α) u = g(x)

while if both eα x and x eα x are solutions of the complementary equation L(y) = 0,

then
p(λ) = a (λ − α)2 and p0 (λ) = 2a(λ − α)
and so p(λ) = p0 (λ) = 0 and then

a u00 = g(x)

Exercise 27 Use method of undetermined coefficients to solve the following dif-

ferential equations.
1. y 00 + 2y 0 + y = x2
2. y 00 − 3y 0 + 2y = ex sin x
3. y 00 + y = cos x
4. (D2 − 3D + 2)y = 2 x2 + 3 e2x , D ≡ dy/dx
5. y 00 + 9y = (x2 + 1) e3x
6. y 00 − y = 3e2x cos x
7. y 00 − y 0 − 2y = 3 e2x , y(0) = 0, y 0 (0) = −2.
8. y 00 + 4y 0 + 4y = 4 cos x + 3 sin x, y(0) = 1, y 0 (0) = 0.
9. y 00 − 4y 0 + 4y = 6 x e2x , y(0) = 0, y 0 (0) = 3.

7 Variable coefficients
Consider a second order homogeneous differential equation with variable coeffi-
cients

L(y) = y 00 + a1 (x)y 0 + a2 (x)y = 0 (61)

We will discuss the following methods to find the solution of the equation (61).
1. Method of reduction order.

2. Euler-Cauchy equation

44
 Dr. Dil Gurung, Kathmandu University

7.1 Method of reduction order

The method is based on the fact that one non-zero solution y1 (x) of (61) is known.
Then the other solution y2 (x) such that y1 and y2 are linearly independent, can
be found by taking

y2 = u y1 (62)

where u(x) is a function of x alone. Differentiating (61) successively

y20 = uy10 + u0 y1
y200 = uy100 + 2u0 y10 + u00 y1

Substituting y2 and its derivatives into Eq. (61), we obtain

u00 y1 + u0 (2y10 + a1 (x)y1 ) = 0

Dividing by u0 y1 , we obtain

u00 y10
= −2 − a1 (x) (63)
u0 y1

Set v = u0 in (63). Then v 0 = u00 and Eq. (63) becomes

v0 y0
= −2 1 − a1 (x) (64)
v y1
which is a separable first order equation, so the method is reduction order method.
The functions in both sides of (64) are functions of x, so we integrate with respect
to x to obtain
Z
ln v = −2 ln y1 − a1 (x) dx
 Z   Z 
1
v = exp −2 ln y1 − a1 (x) dx = 2 exp − a1 (x) dx
y1

But since v = u0 , we have

 Z 
0 1
u = 2 exp − a1 (x) dx
y1

Since the exponential is never zero, u is non constant. To find u, we perform

another integration and obtain
R

exp − a1 (x) dx
Z
y2 = y1 u = y1 dx
y12

Example 35 Verify that y1 = x is a solution of x2 y 00 − xy 0 + y = 0, x > 0, and

hence find the general solution.
Solution: It is easy to check y1 = x is a solution of the equation. Set y2 = y1 u =
x u. It follows that
y20 = xu0 + u, y200 = xu00 + 2u0

45
 Dr. Dil Gurung, Kathmandu University

Substituting in the equation, we obtain

x3 u00 + x2 u = 0
Setting v = u0 and separating variables, we have
dv dx
= −
v x
ln v = − ln x (since x > 0)
1
v =
x
and so u0 = v = x1 ⇒ du
dx
= x1 . Thus u(x) = ln x, so y2 = y1 u = x ln x and the
general solution is
y = c1 x + c2 x ln x, x>0
Exercise 28 If u is a solution of the equation L(y) = y 00 + a1 (x)y 0 + a2 (x)y = 0,
then show that the solution v of the nonhomogeneous equation L(y) = b(x) is
given by
Z  − R a1 (x)dx  Z 
e R
a1 (x) dx
v(x) = u(x) b(x) e dx + C dx
u(x)2
Use this relation to find the solution of the differential equation y 00 − xy 0 − y = 1,
taking y(x) = x as a solution of the homogeneous equation.
Exercise 29 In the following differential equations verify that y1 (x) is a solution
and find a second linearly independent solution, and the general solution.
1. y 00 − 2y + y = 0, y1 (x) = ex
2. x2 y 00 + xy 0 − 4y = 0, y1 (x) = x2
3. y 00 − 2xy 0 + 2y = 0, y1 (x) = x
4. x2 y 00 − 2xy 0 + (x2 + 2)y = 0, (x > 0) y1 (x) = x sin x
00 0
5. xy + (2x − 1)y − 2y = 0, (x > 0) y1 (x) = e−2x
00 0
6. xy + (x − 1)y − y = 0, y1 (x) = −e−x
Exercise 30 In the following nonhomogeneous differential equations verify that
u(x) is a complementary solution and hence find the general solution by method
of reduction order.
1. x2 y 00 − (x2 + 2x)y 0 + (x + 2)y = x3 ex , u(x) = x
00 0 x
2. (x + 2)y − (2x + 5)y + 2y = (x + 1)e , u(x) = e2x
Exercise 31 The Legendre equation is given by
(1 − x2 )y 00 − 2xy 0 + n(n + 1)y = 0, −1 < x < 1
For n = 1, verify that y1 (x) = x is a solution. Find other linearly independent
solution y2 and hence find the general solution.
Exercise 32 The Bessel differential equation is given by
x2 y 00 + xy 0 + (x2 − p2 )y = 0
sin
√x
For p = 1/2, verify that y1 (x) = x
is a solution for x > 0. Find a second
linearly independent solution.

46
 Dr. Dil Gurung, Kathmandu University

7.2 Euler Cauchy equation

An equation of the form

x2 y 00 + axy 0 + by = 0, x 6= 0 (65)

is called a homogeneous Euler equation or Euler Cauchy equation, where a and b

are constants.
Theorem 17 Assume y = xm for some number m, be the solution of the equation
(65).
1. Show that the characteristic equation of (65) is

m(m − 1) + am + b = 0 or m2 + (a − 1)m + b = 0

2. If the two roots m1 , m2 of the characteristic equation be real and unequal,

then show that the general solution is

y = c1 xm1 + c2 xm2

where c1 and c2 are constants.

3. If the two roots m = m1 = m2 be real and equal, then show that the general
solution is
y = (c1 + c2 ln |x|)xm
where c1 and c2 are constants.
4. If the conjugate complex roots are m1 = α + iβ and m2 = α − iβ then show
that the general solution is

y = xα [c1 cos(β ln |x|) + c2 sin(β ln |x|)]

where c1 and c2 are constants.

√
Example 36 Find the general solution of x2 y 00 + 2xy 0 − 12y = x, x 6= 0.
Solution: The associated homogeneous equation is

x2 y 00 + 2xy 0 − 12y = 0

Comparing with Eq.(65), we have a = 2, b = −12. The characteristic equation is

then

m2 + (a − 1)m + b = 0
m2 + (2 − 1)m − 12 = 0

Its roots are m1 = 3 and m2 = −4, which are real and unequal. So the compli-
mentary function is
yh (x) = c1 x3 + c2 x−4
Transforming the equation into standard form of second order non-homogeneous
linear differential equation by dividing x2 :
2 0 12
y 00 + y − 2 y = x−3/2
x x
47
 Dr. Dil Gurung, Kathmandu University

Now, Wronskian of y1 = x−4 , y2 = x3 is

 −4 
3 
 x x 7
W (y1 , y2 )(x) =  −5 2 = 2

−4 x 3x x

The integrals
Z Z 3 −3/2
y2 f (x) x x 2 9/2
dx = 2
dx = x
W (y1 , y2 )(x) 7/x 63
Z Z −4 −3/2
y1 f (x) x x 2 −5/2
dx = dx = − x
W (y1 , y2 )(x) 7/x2 35

Using variation of parameter method, the particular solution is then

Z Z
y2 f (x) y1 f (x) 4
yp (x) = −y1 dx + y2 dx = − x1/2
W (y1 , y2 )(x) W (y1 , y2 )(x) 45

The general solution is then

4 1/2
y(x) = yh (x) + yp (x) = c1 x3 + c2 x−4 − − x
45
Exercise 33 Show that the homogeneous Euler equation (65) can be transformed
into the constant coefficient equation

y 00 + (a − 1)y 0 + by = 0

by making the substitution x = et (t = ln x). Use this method to solve the equation
x2 y 00 + 7xy 0 + 5y = x.

Exercise 34 The Cauchy equation or Euler equation of the third order is

x3 y 000 + ax2 y 00 + bxy 0 + cy = 0

Show that y = xm is a solution of the equation if and only if m is a root of the

characteristic equation

m3 + (a − 3)m2 + (b − a + 2)m + c = 0

Solve the Euler equation x3 y 000 − 3x2 y 00 + 6xy 0 − 6y = 0.

Exercise 35 Solve the following Euler-Cauchy equations:

1. x2 y 00 − 3xy 0 + 4y = 0, x > 0
2. x2 y 00 + 5xy 0 + 13y = 0, √x > 0
3. x2 y 00 + 2xy 0 − 12y = x, x > 0
4. x2 y 00 − 2xy 0 + 2y = x3 cos x
5. x2 y 00 − xy 0 + y = x ln |x|
6. (x2 D2 + xD − 4I)y = x3 , y(1) = −4/5, y 0 (1) = 93/5
7. x3 y 000 − 3x2 y 00 + 6xy 0 − 6y = x4 ln x
8. x3 y 000 − x2 y 00 − 7xy 0 + 16y = 9x ln x, y(1) = 6, y 0 (1) = 18, y 00 (1) = 65
9. (x3 D3 + 5x2 D2 + 2xD − 2I)y = 7x3/2 , y(1) = 10.6, y 0 (1) = −3.6, y 00 (1) = 31.2

48
 Dr. Dil Gurung, Kathmandu University

8 Applications – Second order

8.1 Spring mass balance system
Suppose a small block of mass m is attached to one end of a spring, and the
other end of the unstrached spring is attached to a fixed wall. The block is pulled
downward, stretching the spring x0 unit and then released from the rest (the
initial velocity zero) Figure 2. The block moves up and down.

Figure 2: Mass spring balance system. (a) unstretched spring (b) system in static
equilibrium (c) system in motion.

Figure 3: Direction below the equilibrium position is positive.

This phenomena can be formulated under Hooke’s law and Newton’s second
law of motion.

8.1.1 Undamped motion. Harmonic oscillation

For the formulation of undamped motion or harmonic oscillation of the system,
the following assumptions are made:

i. All motion is along a vertical line through the center of gravity of the object,
which is treated as a point mass.

ii. There is no damping force, Fd due to medium in which the mass is moving
(such as air resistance).

49
 Dr. Dil Gurung, Kathmandu University

iii. No other forces (beyond the ones already mentioned) are applied to the
mass.
The Hooke’s law states that the spring force, Fs (acting upward) exerted on
the object (of mass m) is proportional to the difference between the length l + s
of the spring and its natural or equilibrium length l . That is

Fs = k(l + s − l) = ks

where k is a positive constant, called the spring constant. If the mass is at rest,
then it is the force of gravity, Fg = mg that pulls the spring beyond its natural
length. This means that
mg = ks
where g is the acceleration due to gravity. It is the case that Fg = Fs only as
long as the mass is left at rest.
Let us denote the position where the spring has length l + s by x = 0. Suppose
that the object is given an initial displacement x0 and an initial velocity v0 . That
is
x(0) = x0 , x0 (0) = v0
If the spring oscillation is not rest, then the spring force Fs depends on the
displacement x of the mass:
Fs = k(x + s)
As we increase x (moving downward in the same direction as the gravitational
force), the spring force acting upward increases, so the total force on the mass is

FR = Fg − Fs = mg − k(x + s) = −kx

since mg = ks. Here, the force FR is called the restoring force.

If we ignore the damping force, then by the Newton’s law of motion (Force =
mass × acceleration), we have
d2 x
F =m
dt2
Equating the two forces, we have
d2 x
m + kx = 0
dt2
The model equation is then
d2 x
m + kx = 0, x(0) = x0 , x0 (0) = v0 (66)
dt2
The general solution of (66) is

x(t) = c1 cos ω0 t + c2 sin ω0 t (67)

q
k
where ω0 = m
is called the circular frequency [rad/sec]. Using the initial
conditions, we find that
v0
c1 = x 0 , c 2 =
ω0

50
 Dr. Dil Gurung, Kathmandu University

So the particular solution of (66) is given by

x(t) = x0 cos ω0 t + (v0 /ω0 ) sin ω0 t (68)

Eq.(68) describes the free vibrations or free motion of the mechanical system. We
express the Eq.(68) as
x(t) = A sin(ω0 t + φ)
Writing it as
x(t) = A sin ω0 t cos φ + A cos ω0 t sin φ
Equating with Eq.(68), we have
v0
A sin φ = x0 , A cos φ =
ω0
and s
 2  2
v0 v0
x20 + 2
=A ⇒A= x20 +
ω0 ω0
Also
v0 x0 x0 ω0
cos φ = , sin φ = ⇒ tan φ =
Aω0 A v0
Thus we may write the Eq.(68) as

x(t) = A sin(ω0 t + φ) (69)

with
s  2
v0 x0 ω0 x0 ω 0 x0 ω0
A= x20 + and φ = tan−1 or φ = tan−1 +π or φ = tan−1 +2π
ω0 v0 v0 v0

Remark: We have to take care to determine which quadrant φ is in.

The motion of the mass in (69) is called simple harmonic motion, since it is
sinusoidal. From Eq.(69), we note the followings:

i. the mass oscillates between the extreme positions ±A. This A is called the
amplitude of the motion.
2π
ii. The sine term has period T = ω0
. This is the time required for each
complete oscillation.

iii. The natural frequency f = T1 = 2ωπ0 [Hertz = cycles/sec] of the motion is

the number of complete oscillations per unit time.

iv. The dimensionaless parameters φ is called the phase or phase angle or phase
shift, and measures the displacement of the wave from its natural position
corresponding to φ = 0. If φ > 0, then the graph of x(t) is shifted to the
left compared with φ = 0, and if φ < 0, then the graph of x(t) is shifted to
the right compared with φ = 0.

v. ω0 is called the circular frequency [rad/sec].

51
 Dr. Dil Gurung, Kathmandu University
φ
vi. The time lag, δ = ω0
.

Note:

a. If we express (67) as
x(t) = A sin(ω0 t + φ)
then
c1 c2
sin φ = , cos φ =
A A
and so

−1 c1
q tan c2
 if c1 , c2 > 0 (First quadrant)
A = c21 + c22 , φ = π + tan−1 cc21 if c2 < 0 (second or third quadrant)

2π + tan−1 cc12 if c1 < 0, c2 > 0 (fourth quadrant)


b. If we express (67) as
x(t) = A cos(ω0 t − φ)
then
c2 c1
sin φ = , cos φ =
A A
and so

−1 c2
q tan c1
 if c1 , c2 > 0 (First quadrant)
A = c21 + c22 , φ = π + tan−1 cc12 if c1 < 0 (second or third quadrant)

2π + tan−1 cc21 if c2 < 0, c1 > 0 (fourth quadrant)


Example 37 (Undamped motion. Harmonic oscillation ) If an iron ball

of weight W = 98 Newton stretches a spring 1.09 meter, how many cycles per
minute will this mass-spring system execute ? What will its motion be if we pull
down the weight an additional 16 centimeter and let it start with zero initial
velocity ?
Solution: Here an iron ball of weight W = 98N stretches the spring 1.09 m. So
by Hooke’s law

Kg m/s2
   
W 98 N
Fs = W = 1.09 k ⇒ k = = = = 89.908 ≈ 90 N/m
1.09 1.09 m m

The mass of the iron ball is

W 98
m= = = 10 Kg
g 9.8
The cyclic frequency r r
k 90
ω0 = = = 3 rads/s
m 10
Then the frequency is
ω0 3
f= = = 4.7123 ≈ 0.48 Hz = 0.48 cycles/s = 28.8 cycles/min ≈ 29 cycles/min
2π 2π

52
 Dr. Dil Gurung, Kathmandu University

From Eq.(68) and the initial conditions, x(0) = x0 = 16 cm = 0.16 m and x0 (0) =
v0 = 0, we obtain the undamped motion or harmonic oscillation of the system is
x(t) = 0.16 cos 3 t
Hence its amplitude of motion is
p
A= (0.16)2 = 0.16 m
To find phase angle, we write
x(t) = A sin(ω0 t + φ) = 0.16 sin(3 t + φ) = 0.16 sin 3 t cos φ + 0.16 cos 3 t sin φ
Thus, sin φ = 1, cos φ = 0 ⇒ φ = tan−1 10 = π2 rad.
The time lag is then
φ π
δ= = = 0.52 s
ω0 6
49
Example 38 Suppose a 98 N weight stretches a certain spring 320
m. If the
weight is pulled 13 m below the equilibrium position and released.
i. Set up the differential equation and associated conditions describing the mo-
tion.
ii. Find the position of the weight as a function of time t.
iii. Find the amplitude, period and frequency of motion.
1
iv. Determine the position, velocity and acceleration of the weight 2
sec after it
has been released.
49
Solution: Here a weight W = 98N stretches the spring 320 m. So by Hooke’s
law
Kg m/s2
   
49 W 98 N
Fs = W = k⇒k= = = = 640N/m
320 49/320 49/320 m m
The mass of the object is
W 98
m= = = 10 Kg
g 9.8
i. Using the Newton’s second law of motion, the model equation is
d2 x
m +kx = 0
dt2
d2 x
10 2 + 640 x = 0
dt
d2 x
+ 64 x = 0
dt2
At time t = 0, the mass is 31 m below the equilibrium position. We thus
have x(0) = x0 = 13 . Also, since the weight is released, that is, it has zero
velocity, at t = 0. So x0 (0) = 0. Then the model equation with associated
conditions is
d2 x
+ 64x = 0, x(0) = 1/3, x0 (0) = 0. (70)
dt2

53
 Dr. Dil Gurung, Kathmandu University

ii. The characteristic equation of the homogeneous equation is

λ2 + 64 = 0

Its roots are

λ1,2 = ±8i
The general solution is then

x(t) = A cos 8t + B sin 8t

Using initial conditions, we obtain A = 1/3 and B = 0. The solution is

then
1
x(t) = cos 8t
3
iii. From above solution form, we have
1
amplitude =
3
2π 2π π
period, T = = =
ω0 8 4
1 4
frequency, f = =
T π

iv. From the solution, we obtain

8 64
velocity, v = − sin 8t; acceleration, a = − cos 8t;
3 3
At t = 1/2,
1
x(1/2) = cos 4 = −0.21788
3
8
v(1/2) = − sin 4 = 2.01814
3
64
a(1/2) = − cos 4 = 13.9444
3

8.1.2 Damped motion

Damping force, Fd , (in case where a vertical spring moves through a liquid; hor-
izontal spring moves in frictionless surface) is proportional to the velocity of the
mass and acts in the direction opposite to the motion. Thus
dx
damping force, Fd = −c
dt
where c is a positive constant, called the damping constant. Thus in this case,
Newton’s second law of motion gives
d2 x
m = FR + Fd
dt2
dx
= −kx − c
dt
54
 Dr. Dil Gurung, Kathmandu University

Thus
d2 x dx
m 2
+c + kx = 0
dt dt
Or
d2 x c dx k
2
+ + x=0 (71)
dt m dt m
The Eq.(71) is a standard second order linear homogeneous differential equation.

Note:

a. If at some instant, x0 = dx
dt
> 0, the body is moving downward, which is
a positive direction. Hence, the damping force Fd = −cx0 always acting
against the direction of motion, must be an upward force, which means
that it must be negative, Fd = −cx0 < 0, so that −c < 0 and c > 0.

b. For an upward direction, x0 < 0 and we have a downward force, Fd =

−cx0 > 0, and hence −c < 0 and so c > 0.

The characteristic equation of (71) is

c k
λ2 + λ+ =0
m m
Its roots are
p √
−c/m ± (c/m)2 − 4k/m −c ± c2 − 4km
λ1,2 = =
2 2m
Case 1 (Overdamping)
If the damping constant c is so large that c2 > 4km, then λ1 and λ2 are distinct
real roots. The general solution is then

x(t) = c1 eλ1 t + c2 eλ2 t , λ1 < 0, λ2 < 0

This equation represents a smooth and oscillatory motion. In this case, damping
takes out energy so quickly that the body does not oscillates. Practically speak-
ing, after a sufficiently long time the mass will be at rest at the static equilibrium
position (x = 0). In this case, x(t) → 0 as t → ∞.

Case 2 (Critical damping)

Critical damping is the border case between nonoscillatory motion (Case 1) and
c
oscillations (Case 3). It occurs if c2 = 4mk. So λ = λ1 = λ2 = − 2m , the case of
double roots. Then the corresponding general solution of (71) is

x(t) = (c1 + c2 t)eλ t

c
Because eλ t = e− 2m t > 0 and c1 + c2 t has at most one positive zero (i.e., one of
c1 or c2 must be negative), the body passes through its equilibrium position at
most once, and it is clear that x(t) → 0 as t → ∞.

55
 Dr. Dil Gurung, Kathmandu University

Case 3 (Underdamping)
It occurs if the damping constant c is so small that c2 < 4mk. In this case, the
roots are complex conjugate, say, λ1 = α + iβ and λ2 = α − iβ, where α = −c/2m
and β = (4mk − c2 )/2m. Then the corresponding general solution of (71) is

x(t) = eα t (A cos β t + B sin β t) = C eα t sin(β t + φ)

which shows an oscillation with frequency

√
β 4mk − c2
f= =
2π 4πm
and so the general solution represents damped oscillations between the curves
x(t) = C eα t and x(t) = −C eα t . The factor

eλ t = e(−c/2m) t

is called the damping factor.

Example 39 In Example 38, assume that a damping force in Newton numeri-

cally equal to 200 times the instantaneous velocity is taken into account. Find
x(t) as a function of t.
Solution: Taking into account the damping force, the differential equation of 38,
we find

d2 x dx
10 2 + 200 + 640 x = 0
dt dt
d2 x dx
2
+ 20 + 64 x = 0
dt dt
The model equation with associated initial conditions is

d2 x dx
2
+ 20 + 64 x = 0, x(0) = 1/3, x0 (0) = 0. (72)
dt dt
The characteristic equation is

λ2 + 20 λ + 64 = 0

Its roots are −4 and −16 which are real and unequal. Hence

x(t) = c1 e−4t + c2 e−16t

Using initial conditions, we obtain c1 = 4/9 and c2 = −1/9. Therefore, the

required solution is
4 1
x(t) = e−4 t − e−16 t
9 9

56
 Dr. Dil Gurung, Kathmandu University

8.1.3 Forced vibrations

Suppose that, in addition to the restoring force and the damping force, the motion
of spring is affected by an external force, say, f (t), which is due to the motion
of the object to which the upper end of the spring is attached. In this case the
mass undergoes forced vibrations. Then the Newton’s second law of motion gives

d2 x
m = FR + Fd + f (t)
dt2
dx
= −kx − c + f (t)
dt
Thus
d2 x dx
m 2
+c + kx = f (t)
dt dt
Or
d2 x c dx k
2
+ + x = f (t) (73)
dt m dt m
The Eq.(73) is a standard second order non-homogeneous linear differential equa-
tion with constant coefficients describing the motion of a spring under the appli-
cation of restoring force, damping force and external force.

Example 40 Assume that a damping force in Newton numerically equal to 80

times the instantaneous velocity in m/s acts on the weight shown in Example 38
and assume that a periodic force (external) given by f (t) = 1280 cos 8 t.

i. Set up the differential equation.

ii. Find the position x of the weight as a function of time t.

Solution:

i. From the Newton’s second law of motion,

d2 x
m = FR + Fd + External force
dt2
Applying these forces and using the parameter values, we obtain

d2 x dx
10 2
+ 80 + 640 x = 1280 cos 8 t
dt dt
d2 x dx
+ 8 + 64 x = 128 cos 8 t
dt2 dt
The model equation with associated initial conditions is

d2 x dx
2
+8 + 64 x = 128 cos 8 t, x(0) = 1/3, x0 (0) = 0. (74)
dt dt

57
 Dr. Dil Gurung, Kathmandu University

ii. The characteristic equation of the associated homogeneous equation is

λ2 + 8 λ + 64 = 0
√
Its roots are −4 ± 4 3 i which are complex conjugates. Hence, the compli-
mentary solution is
−4t
h √ √ i
xh (t) = e c1 cos 4 3 t + c2 sin 4 3 t

To accommodate the non homogeneity 128 cos 8t using method of undeter-

mined coefficients, the particular solution is taken as

xp (t) = A sin 8t + B cos 8t

Using this yp into the nonhomogeneous equation, and equating the like
terms, we obtain A = 2, B = 0. So

xp (t) = 2 sin 8t

Hence the general solution of the nonhomogeneous equation is

h √ √ i
x(t) = e−4t c1 cos 4 3 t + c2 sin 4 3 t + 2 sin 8t

√
Using initial conditions, we obtain c1 = 1/3 and c2 = −11/3 3. Therefore,
the required solution is
√ √
 
−4t 1 11
x(t) = e cos 4 3 t − √ sin 4 3 t + 2| sin
{z 8t}
3 3 3
| {z } steady term
transient terms

−4t
√
−4t
h
1
11
√ i
When t → ∞, e → 0. So the terms e cos 4 3 t − 3 3 sin 4 3 t are
√
3
called transient terms and are significant only when t is near zero. These
transient terms in the solution, when they are significant, are sometimes
called the transient solution. When the transient terms are negligible,
the term 2 sin 8t remains. This is called steady state term or steady
state solution.

8.2 LRC series circuit

Consider the RLC-circuit consisting the resistor, inductor and the capacitor caus-
ing voltage drop of ER = RI, EL = L dI dt
and EC = C1 Q respectively. The elec-
tromotive force, emf, provides a voltage E(t). Thus Kirchhoff’s voltage law yields

EL + ER + EC = E
dI 1
L + RI + Q = E(t) (75)
dt C
dQ
Since I = dt
, and then

d2 Q dQ 1
L + R + Q = E(t) (76)
dt2 dt C
58
 Dr. Dil Gurung, Kathmandu University

Figure 4: LRC series circuit.

for the charge Q(t) under the assumption that the voltage E(t) is known.
If E(t) = 0, the electrical vibrations of the circuit are said to be free. The
characteristic equation of the Eq.(76) with E(t) = 0 is
1
Lλ2 + Rλ + =0
C
Its two roots are p
−R ± R2 − 4L/C
λ1,2 =
2L
and hence the Eq.(76) with E(t) = 0 has three forms of solutions depending on
the value of the discriminant R2 − 4L/C. Thus, we say that the circuit is

• overdamped if R2 − 4L/C > 0

• critically damped if R2 − 4L/C = 0

• underdamped if R2 − 4L/C < 0

dQ
Differentiating Eq.(75) with respect to t and using I = dt
, we obtain

d2 I dI 1 dE
L 2
+R + I = (77)
dt dt C dt
for the current I(t). As in spring mass system, if E(t) is periodic, that is, E(t) =
A cos ωt or A sin ωt, then the general solution of Eq.(77) consists

I(t) = Itransient current + Isteady periodic

The steady periodic current is the particular solution of (77).

Analogies between mechanical system and electrical system

59
 Dr. Dil Gurung, Kathmandu University
Mechanical system Electrical system
mx00 + cx0 + kx = f (t) LQ00 + RQ0 + C1 Q = E(t)

x : displacement Q : charge

x0 : velocity Q0 : current

x00 : acceleration Q00 : change in current

m : mass L : inductance

c : damping constant R : resistance

k : spring constant 1/C : where C is the capacitance

f (t) : forcing function E(t) : voltage source

Example 41 An RLC-circuit consists of an inductor of 1 henry, a resistor of

12 ohms, a capacitor of 0.01 farads, and a generator having voltage given by
E(t) = 24 sin 10t. Find the charge Q and the current I at time t if Q = 0 and
I = 0 at t = 0.
Solution: From Eq.(76), we have

d2 Q dQ 1
1 2
+ 12 + Q = 24 sin 10t
dt dt 0.01
d2 Q dQ
2
+ 12 + 100Q = 24 sin 10t (78)
dt dt
The roots of the associated homogeneous equation are −6 ± 8i, and so the com-
plimentary solution is

Qc = e−6t (c1 cos 8t + c2 sin 8t)

We use method of undetermined coefficients to obtain the particular solution Qp

to accommodate the nonhomogeneity 24 sin 8t. Take

Qp = A cos 10t + B sin 10t

Therefore

Q0p = −10A sin 10 t + 10B sin 10 t

Q00p = −100A cos 10 t − 100B cos 10 t

Substituting Qp , Q0p and Q00p in Eq.(78) and equating the like coefficients, we get
A = 1/5 and B = 0. Then
1
Qp = − cos 10 t
5
Therefore the general solution of Eq.(78) is
1
Q(t) = Qc + Qp = e−6t (c1 cos 8t + c2 sin 8t) − cos 10 t
5

60
 Dr. Dil Gurung, Kathmandu University

From the initial condition Q(0) = 0 and I(0) = 0 we find that c1 = 1/5 and
c2 = 3/20. Hence the charge Q is

e−6t 1
Q(t) = (4 cos 8t + 3 sin 8t) − cos 10 t
20 5
and the current I = dQ/dt is
5
I(t) = − e−6t sin 8 t + |2 sin
{z10 }t
| 2 {z } steady periodic
transient current

61