You are on page 1of 50

Contents

1 FIRST-ORDER DIFFERENTIAL EQUATIONS 1

1.1 Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Solution of a Differential Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 First-order Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3.1 Solution Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 SECOND-ORDER LINEAR DIFFERENTIAL EQUATIONS 12

2.1 Homogeneous Linear Ordinary Differential Equations . . . . . . . . . . . . . . . . . . 13

2.1.1 The Superposition (Linearity) Principle . . . . . . . . . . . . . . . . . . . . . 13

2.2 Solution Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.2.1 Reduction of order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.2.2 Second-Order Linear Homogeneous Differential Equations with Constant Co-


efficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

1
2.2.3 The Euler-Cauchy Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.3 Non-homogeneous Linear Differential Equations . . . . . . . . . . . . . . . . . . . . . 22

2.3.1 Method of Undetermined Coefficients . . . . . . . . . . . . . . . . . . . . . . 22

2.3.2 Method of Variation of Parameters . . . . . . . . . . . . . . . . . . . . . . . . 23

3 SERIES SOLUTIONS OF ORDINARY DIFFERENTIAL EQUATIONS 26

3.1 Power Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.1.1 Operations on Power Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.1.2 Existence of Power Series Solutions of ODEs . . . . . . . . . . . . . . . . . . 30

3.2 Ordinary Points and Singular Points of ODEs . . . . . . . . . . . . . . . . . . . . . . 30

3.2.1 Power Series Solutions Near Ordinary Points . . . . . . . . . . . . . . . . . . 32

3.2.2 Power Series Solutions Near Regular Singular Points . . . . . . . . . . . . . . 35

4 LAPLACE TRANSFORM METHOD 40

4.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.1.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.1.2 Laplace Transforms of Derivatives . . . . . . . . . . . . . . . . . . . . . . . . 45

4.1.3 Application to Differential Equations . . . . . . . . . . . . . . . . . . . . . . . 45

2
UNIVERSITY OF ZIMBABWE

DEPARTMENT OF MATHEMATICS

MT205 - ORDINARY DIFFERENTIAL EQUATIONS

Lecture Notes

Tapedzesa G

March 18, 2013


Chapter 1

FIRST-ORDER DIFFERENTIAL
EQUATIONS

1.1 Basic Concepts

A differential equation is an equation involving an unknown function and its derivatives. If


the unknown function depends on two or more independent variable, the differential equation is a
partial differential equation. The order of a differential equation is the order of the highest
derivative appearing in the differential equation. The following are examples of differential equations
involving the unknown function y:

dy
1. dx = cos x.
!2
d2 y dy
2. ey dx2 + 2 dx = 1.

d y 3 d y 2
3. 4 dx3 + sin x dx2 + 5xy = 0.

!3 !7 !3
d2 dx dx
4. dt2
+ 3x dt + x3 dt = 5t.

∂2u
5. k ∂u
∂t = ∂x2
- the heat equation.

1
∂2u ∂2u
6. ∂t2
+ ∂x2
= 0 - the Laplace equation.

Note that the last two equations are partial differential equations.

The general form of an nth order ordinary differential equation, in which y is the dependent variable
and x is the independent variable, is

F (x, y, y 0 , y 00 , · · · , y (n) ) = 0,

where the nth derivative must actually appear in this function. Lower order derivatives may be
missing, however. If F is a linear function of the variables y, y 0 , y 00 , · · · , y (n) ), that is

F (x, y, y 0 , y 00 , · · · , y (n) ) = an (x)y (n) + an−1 (x)y (n−1) + · · · + a1 (x)y 0 + a0 (x)y,

then the differential equation is said to be linear. Otherwise, the differential equation is said to
be non-linear.

1.2 Solution of a Differential Equation

A function f (x), defined on an interval I, which when substituted into the differential equation
reduces the equation to an identity is called a solution of the equation on the interval. A solution
of an ordinary differential equation, containing as many arbitrary constants is called the general
solution of the ordinary differential equation. Geometrically, the general solution of an ordinary
differential equation is a family of infinitely many solution curves, one for each set of values of
the constants. If we choose a specific set of values of the constants we obtain what is called a
particular solution of the ordinary differential equation. A particular solution does not contain
any arbitrary constants.

1.3 First-order Differential Equations

The general form of a first-order ordinary differential equation is


dy
= f (x, y), or P (x, y)dx + Q(x, y)dy = 0. (1.1)
dx

2
1.3.1 Solution Methods

Separation of variables

A first-order differential equation is one of separable variables if it can be cast in the form
dy
= g(x)h(x), or M (x)dx + N (y)dy = 0. (1.2)
dx
Whenever the variables are separable, equation (1.1) can be integrated to give
Z Z
M (x)dx + N (y)dy = c,

as a general solution, where c is an arbitrary constant.


Example 1.3.1. Solve
dy
= x(1 + y 2 ).
dx
Solution 1.3.1. Dividing throughout by (1 + y 2 ) and integrating with respect to x, we get
Z Z
1
dy = xdx.
1 + y2
Thus,
1
tan−1 y = x2 + c,
2
where c is a constant. Hence,
1
y = tan( x2 + c)
2
Exercise 1.3.1. Solve ! 2
sin y 2 + 3y + 6 dy ex
= .
x dx 2y + 3

Reduction to separable form: Homogeneous equations

Certain non-separable ordinary differential equations can be made separable by transformations


that introduce for y a new unknown function. We discuss this technique for a class of ordinary
differential equations of the form
!
y
y0 = f . (1.3)
x

3
Here, f is any (differentiable) function of xy . Such an ordinary differential equation is called a
homogeneous ordinary differential equation.

y
The form of (1.3) suggests that we st x = u; thus,
y = ux
and hence
y 0 = u0 x + u.
Substituting into (1.3) then gives
u0 x + u = f (u) or u0 x = f (u) − u,
which can be separated.
du dx
= . (1.4)
f (u) − u x
Example 1.3.2. Solve
2xyy 0 = y 2 − x2 .
Solution 1.3.2. In standard form, we have
! ! !
y 2 − x2 y x 1 y 1 1 y
y0 = = − = − y =f .
2xy 2x 2y 2 x 2 x x

Hence the differential equation is homogeneous, and so we set xy = u. we substitute y and y 0 and
then simplify as follows:
u 1
u0 x + u = − ,
2 2u
u 1 −u2 − 1
u0 x = − − = .
2 2u 2u
We can now separate variables the integrate:
2udu dx
2
=− ,
1+u x
and hence,
ln(1 + u2 ) = − ln |x| + ln c.
Taking exponents on both sides yields
c
1 + u2 = ,
x
that is, !
y c
1+ = or x2 + y 2 = cx.
x x

4
Exercise 1.3.2. Classify each of the following differential equations as either homogeneous or not
homogeneous:

1. y 0 = log x − log y.
y x2
2. y 0 = x − y .
!
y y
3. y 0 = x + sin x .

Exact ordinary differential equations

A first-order ordinary differential equation (1.1), written as


M (x, y)dx + N (x, y)dy = 0 (1.5)
is called an exact differential equation if there exists a function f (x, y) such that
df (x, y) = M (x, y)dx + N (x, y)dy, (1.6)
that is,
∂f
M (x, y) = , (1.7)
∂x
∂f
N (x, y) = . (1.8)
∂y
Then (1.5) can be written
df = 0.
By integration we immediately obtain the general solution of (1.5) in the form
f (x, y) = c. (1.9)
This is called an implicit solution, in contrast with a solution of the form y = h(x), which is called
an explicit solution. Let M and N be continuous and have continuous first partial derivatives
in a region in the xy-plane whose boundary is a closed curve without self-intersections. Then by
partial differentiation of (1.7)-(1.8) we have
∂M ∂2f
= , (1.10)
∂y ∂x∂y
∂N ∂2f
= . (1.11)
∂x ∂x∂y

5
By the assumption of continuity the two second partial derivatives are equal. Thus,
∂M ∂N
= . (1.12)
∂y ∂x

This is the formula for checking whether (1.5) is exact or not.

If (1.5) is exact, the function f (x, y) can be found in by integrating (1.7):


Z
f (x, y) = M dx + k(y). (1.13)

In the above integration, y is to be regarded as a constant, and k(y) plays the role of a constant of
integration. Differentiation of (1.13) with respect to y gives
!
R
∂ M dx
∂f dk
= + . (1.14)
∂y ∂y dy
dk
Solving (1.8) and (1.14) simultaneously we get dy , and hence k(y) upon integration.

Example 1.3.3. Solve

cos(x + y)dx + (3y 2 + 2y + cos(x + y))dy = 0.

Solution 1.3.3. The equation is of the form (1.5) with

M (x, y) = cos(x + y), and N (x, y) = 3y 2 + 2y + cos(x + y).

Thus,
∂M ∂N
= − sin(x + y) = .
∂y ∂x
The equation is, therefore, exact. The solution is f (x, y) = c (c is a constant) where
Z Z
f (x, y) = M dx + k(y) = cos(x + y)dx + k(y) = sin(x + y) + k(y).

Differentiating with respect to y and using (1.14) we obtain

∂f dk
= cos(x + y) + = N (x, y) = 3y 2 + 2y + cos(x + y).
∂y dy
Hence,
dk
= 3y 2 + 2y.
dy

6
By integration,
k = y 3 + y 2 + c1 .
Thus,
f (x, y) = sin(x + y) + y 3 + y 2 = c
is the general solution.
Exercise 1.3.3. Solve the initial value problem
(cos y sinh x + 1)dx − sin y cosh xdy = 0, y(1) = 2.

Linear first-order ODEs

A first-order ODE is said to be linear if it can be written in the form


y 0 + p(x)y = r(x). (1.15)
The defining feature of this equation is that it is linear in both the unknown function y and its
derivative y 0 = dy
x , whereas p and r may be any given functions of x.

Exercise 1.3.4. Investigate whether the differential equation


y 0 cos x + y sin x = x (1.16)
is linear or not.

If r(x) = 0 then the equation is variables separable. In the case that r(x) 6= 0, the differential
equation (1.15) has a pleasant property; namely, it has an integrating factor depending only on x.
To find a solution we go through the following steps:

1. Put the differential equation in standard form (1.15)


y 0 + p(x)y = r(x).

2. Establish the integrating factor R


p(x)dx
e .

3. Multiply the differential equation (throughout) by the integrating factor to obtain


!
R dy R R
e p(x)dx + p(x)ye p(x)dx = r(x)e p(x)dx .
dx

7
4. By the product rule the equation simplifies to
!
d R
p(x)dx
R
ye = r(x)e p(x)dx ,
dx

which upon integration both sides gives the desired solution.

Example 1.3.4. Solve the differential equation


!
dy
x + 4y = x6 .
dx

Solution 1.3.4. In standard form, the differential equation is


dy 4
+ y = x5 .
dx x
we, therefore, find the integrating factor to be
4
R
dx
e x = x4 .

Multiplying the differential equation by x4 we obtain


dy
x4 + 4x3 y = x9 .
dx
Thus,
d 4
(x y) = x9 .
dx
Separation of variables and integration with respect to x gives the solution
1 6 c
y= x + 4.
10 x
Exercise 1.3.5. 1. Solve the differential equation
dy
cot x + 2y = tan2 x.
dx

2. Solve the initial value problem

y 0 + y tan x = sin 2x, y(0) = 1.

8
Bernoulli Equation: Reduction to linear form

Numerous applications can be modeled by ODEs that are nonlinear but can be transformed to
linear ODEs. One of the most useful ones of these is the Bernoulli equation, which has the form
dy
+ p(x)y = r(x)y n , (1.17)
dx
where n is any real number. If n = 1 or n = 1, equation (1.17) is linear. Otherwise it is nonlinear.
Then we use the substitution

u = y 1−n . (1.18)

We differentiate this and substitute y and y 0 from (1.17), obtaining the linear ODE

u0 + (1 − n)p(x)u = (1 − n)r(x). (1.19)

Exercise 1.3.6. Verify (1.19).

Solving the linear equation (1.19) gives u as a function of x. To get the required solution y we use
back-substitution u = y 1−n .
Example 1.3.5. Solve the following logistic equation (also known as Verhulst equation):

y 0 = Ay − By 2 .

Solution 1.3.5. Put in standard form, the equation becomes

y 0 − Ay = −By 2 .

This is a Bernoulli equation with n = 2, so we make a substitution u = y 1−2 = y −1 . Differentiating


and substituting for y 0 gives
u0 = B − Au.
Hence we have obtained a linear ODE

u0 + Au = B,

whose general solution is


B
u = ce−At + .
A
Since u = y −1 , this gives the required solution as
1 1
y= = −1At B
.
u ce + A

9
There are several other differential equations which can be transformed by certain substitutions
into either separable variables, exact or a general linear first-order ODE. In fact, all first-order
differential equations of the form
dy
= F (a1 x + a2 y + a3 ), (1.20)
dx
where a1 , a2 , a3 are constants can be transformed to separable variable type by introducing the
substitution

u = a1 x + a2 y + a3 . (1.21)

Exercise 1.3.7. Solve the following differential equations using a suitable substitutions:

dy
1. dx = (x + y)2 ,
dy
2. dx = (2x + y − 3)3 + 1,
dy 1−x−y
3. dx = x+y ,

dy
4. dx = sin(x + 2y + 3).

Numerical solutions

Sometimes, it is difficult or not worthy finding an exact (or analytic) solution of a first-order
differential equation. In such situations it is wise to employ numerical methods. we shall consider
some of the techniques of finding an approximate solution of a differential equation of the form
(1.1)
y 0 = f (x, y),
with y(x0 ) = y0 on an interval Ω ∈ R2 .

1. The Taylor series method

This technique is based on the assumption that the required solution can be expanded in
a Taylor series about a point xk , k = 0, 1, · · · . Given y0 = y(x0 ), we can find successive
estimates of yk = y(xk ) for k = 0, 1, · · · using the following scheme

h2 00 hn (n)
y(xk ) = y(xk ) + hy 0 (xk ) + y (xk ) + · · · + y (xk ), (1.22)
2! n!

10
where h = xk+1 − xk is the step size in the variable x and y 00 , . . . , y (n) can be obtained by
successively differentiating the given differential equation (1.1). The scheme (1.22) is known
as Taylor’s formula of order n.

Example 1.3.6. Use Taylor’s method of order 4 to approximate the solution of the differential
equation
y 0 = y + ex on [0, 2], with y(0) = 1.
Use

(a) h = 1.
(b) h = 0.5.
(c) h = 0.25.

Compare your results with the exact solution.

Solution 1.3.6. The Taylor’s formula of order 4 is

h2 00 h3 h4
y(xk ) = y(xk ) + hy 0 (xk ) + y (xk ) + y 000 (xk ) + y (iv) (xk )
2! 3! 4!
1 1 1
= yk + h(yk + exk ) + h2 (yk + 2exk ) + h3 (yk + 3exk ) + h4 (yk + 4exk )
2 6 24
h2 h3 h4 h3 h4 xk
= (1 + h + + + )yk + (h + h2 + + )e .
2 6 24 2 6
Use the automated formula to complete the solution.

2. Euler’s method

3. Runge-Kutta’s methods

11
Chapter 2

SECOND-ORDER LINEAR
DIFFERENTIAL EQUATIONS

A second-order ordinary differential equation is said to be linear if it can be written in the form

y 00 + p(x)y 0 + q(x)y = r(x). (2.1)

The characteristic feature of this equation is that it is linear in the unknown function y and its
derivatives, whereas p(x) and q(x) as well as r(x) may be any given functions of x.

If r(x) ≡ 0 for all x considered, then (2.1) becomes simply

y 00 + p(x)y 0 + q(x)y = 0 (2.2)

and is said to be homogeneous. Otherwise, if r(x) 6≡ 0, then (2.1) is said to be non-homogeneous.


The functions p(x) and q(x) in (2.1) and (2.2) are called the coefficients of the equations. A
function
y = h(x)
is called a solution of a solution of a second-order ordinary differential equation on some open
interval I if h is defined and twice differentiable throughout that interval and is such that the
ordinary differential equation becomes an identity if we replace y and its derivatives by h(x) and
its corresponding derivatives.

12
2.1 Homogeneous Linear Ordinary Differential Equations

2.1.1 The Superposition (Linearity) Principle

Theorem 2.1.1. For a homogeneous linear differential equation (2.2), any linear combination of
two solutions on an open interval I is again a solution of (2.2) on I.

Proof. Let y1 (x) and y2 (x) be solutions of (2.2) on I. Then by substituting y = c1 y1 + c2 y2 and its
derivatives into (2.2), we get

y 00 + py 0 + qy = (c1 y1 + c2 y2 )00 + p(c1 y1 + c2 y2 )0 + q(c1 y1 + c2 y2 )

= c1 y100 + c2 y200 + c1 py10 + c2 py20 + c1 qy1 + c2 qy2


= c1 (y100 + py10 + qy1 ) + c2 (y200 + py20 + qy2 )
= 0 + 0 = 0,
because y1 and y2 are solutions of (2.2), by assumption. This shows that y = c1 y1 + c2 y2 is a
solution of (2.2) on I.

Note that this highly important result holds for homogeneous linear ODEs only but does not hold
for non-homogeneous linear and non-linear ODEs.

Recall that for a first-order ODE, an initial value problem consists of the ODE and one initial
conditiony(x0 ) = y0 . The initial condition is used to determine the arbitrary constant, usually
denoted by c, in the general solution of the ODE. That solution is called a particular solution of the
ODE. Now for a second-order homogeneous linear ODE (2.2) an initial value problem consists
of the ODE itself and two initial conditions

y(x0 ) = y0 and y 0 (x)0 ) = y00 . (2.3)

The conditions (2.3) are used to determine the two arbitrary constants c1 and c2 in a general
solution

y = c1 y1 + c2 y2 (2.4)

of the ODE, in which y1 and y2 are solutions of (2.2) that are not proportional, and c1 and c2 are
arbitrary constants. These y1 and y2 are called a basis (or a fundamental system) of solutions

13
of (2.2) on I. In fact, the two functions y1 and y2 are said to be linearly independent, which
they must be, on an interval I where they are defined if

α1 y1 (x) + α2 y2 (x) = 0 implies α1 = 0 and α2 = 0.

Otherwise, the functions are linearly dependent.

y = 0 is always a solution of (2.2). It is called the trivial solution.

2.2 Solution Methods

2.2.1 Reduction of order

Type I

If in a second-order differential equation, the dependent variable y does not appear explicitly, the
equation can then be written in the form

y 00 + p(x)y 0 = r(x). (2.5)

By setting z = y 0 we can obtain a first-order differential equation in z(x) and from its solution, the
solution of the original equation (by simply integrating).

Example 2.2.1. Reduce the following differential equation to first-order, then solve it:

2xy 00 = 3y 0 .

Solution 2.2.1. We substitute z = y 0 , so that the equation reduces to first-order as

2xz 0 = 3z.

We then separating variables then integrate:


Z Z
2dz 3dx
= ,
z x
which simplifies to give
2 ln z = 3 ln x + ln c.
z 2 = cx3 .

14
Back-substitution then gives
3 √
y 0 = Dx 2 , D= c.
Integration yields
5
y = c1 x 2 + c2 ,
2D
which is the general solution, where c1 = 5 , c2 are arbitrary constants.

Type II

Another type of equations reducible to first order is when the independent variable x does not
appear explicitly in the differential equation. That is, the differential equation is of the form

y 00 + a1 y 0 + a0 y = b0 (2.6)

where a0 , a1 , b0 are constants.

Using the chain rule we show that y 00 = z dy dz


where again z = y 0 . Substituting these in (2.6)
we obtain a first-order differential equation with y as the independent variable:
dz
z + a1 z + a0 y = b0 .
dy
Exercise 2.2.1. Complete the solution for the example above.

Type III

Suppose we obtain a solution y1 (x) (not identically zero) of a given homogeneous differential equa-
tion (2.2) by some other specified method or by guessing. Then the transformation

y = v(x)y1 (x)

reduces the equation (2.2) to a first-order homogeneous linear differential equation.

Since y = v(x)y1 (x) then y 0 = vy10 + y1 v 0 and hence y 00 = vy100 + 2v 0 y10 + y1 v 00 . Substituting
these into (2.2) we obtain

v 00 y1 + v 0 (2y10 + py1 ) + v(y100 + py10 + qy) = 0.

15
The expression in the last parenthesis is zero, since y1 is a solution of (2.2). Thus,

v 00 y1 + v 0 (2y10 + py1 ) = 0.

Let w = v 0 and so w0 = v 00 , then !


2y10
w0 + + p w = 0.
y1
Separation of variables and integration yield
Z
ln w = −2 ln y1 − pdx + c.

We take exponents to get


c − R pdx
w= e .
y12
But w = v 0 , hence the desired second solution is
Z
y2 = vy1 = y1 wdx.

It is worthy noting that the method of reduction of order works even for non-linear second-order
ODEs.

Example 2.2.2. Given that y1 (x) = x4 is a solution of the second-order differential equation

x2 y 00 − 7xy 0 + 16y = 0,

use the method of reduction of order to obtain a second linearly independent solution.

Solution 2.2.2. The second solution is of the form y = v(x)y1 = v(x)x4 , for a suitable choice of
the function v(x). We directly establish that

y 0 = x4 v 0 + 4x3 v and y 00 = x4 v 00 + 8x3 v 0 + 12x2 v.

Substituting these into the given differential equation we obtain

x2 y 00 − 7xy 0 + 16y = x2 [x4 v 00 + 8x3 v 0 + 12x2 v] − 7x[x4 v 0 + 4x3 v] + 16x4 v = 0.

Thus,
1 0
v 00 + v = 0.
x
setting w = v 0 , we obtain
1
w0 + w = 0,
x

16
which is separable. We integrate to get
c1
w= ,
x
where c1 is a constant. Recall that w = v 0 , hence,
Z Z
c1
v = wdx = dx = c1 ln |x| + ln c2 .
x
Setting c1 = 1 and c2 = 1 since we are looking for the second linearly independent solution, we
obtain
y2 (x) = vy1 = x4 ln |x|.
Exercise 2.2.2. 1. Reduce the order of the following differential equations and hence solve:
(a) y 00 = y 0 tanh x.
(b) xy 00 + y 0 = (y 0 )2 .
(c) y 00 + ey (y 0 )3 = 0.
(d) y 00 + (y 0 )3 cos y.
2. Verify that the given function is a solution of the given differential equation for all positive x
and find y2 such that y1 , y2 form a fundamental set of solutions for all positive x.
(a) x2 y 00 − 4xy 0 + 4y = 0, y1 = x.
(b) y 00 + 1 0
xy + (1 − 1 2
4 x )y = 0, y1 = cos
√ x.
x

2.2.2 Second-Order Linear Homogeneous Differential Equations with Constant


Coefficients

Consider the differential equation of the form

a2 y 00 + a1 y 0 + a0 y = 0, (2.7)

defined on the interval (−∞, ∞), where a0 , a1 , a2 are real constants.

Recall that the solution of the first-order homogeneous linear differential equation with constant
coefficients (y 0 + ky = 0) is y = ce−kx . We thus conjecture that

y = eλx (2.8)

might be a solution of the differential equation (2.7) if λ is properly chosen. Substituting (2.8) and
its respective derivatives
y 0 = λeλx and y 00 = λ2 eλx

17
into (2.7), we obtain
eλx (a2 λ2 + a1 λ + a0 ) = 0.
Since emx 6= 0 for all x ∈ (−∞, ∞), we have
a2 λ2 + a1 λ + a0 = 0. (2.9)
Equation (2.9) is called the characteristic equation or auxiliary equation of the differential
equation (2.7). The characteristic equation (2.9) may have three kinds of roots, depending on
the sign of the discriminant a2 − 4b, namely; two real roots if a2 − 4b > 0, a real double root if
a2 − 4b = 0, and complex conjugate roots if a2 − 4b < 0.

Case I: Two distinct real roots λ1 , λ2

In this case, a basis of solutions of (2.7) on any interval is


y1 = eλ1 x , y2 = eλ2 x . (2.10)
The corresponding general solution is
y = c1 eλ1 x + c2 eλ2 x .
Example 2.2.3. Solve the initial value problem
y 00 + y 0 − 2y = 0, y(0) = 4, y 0 (0) = −5.
Solution 2.2.3. The characteristic equation is
λ2 + λ − 2 = 0.
Its roots are
λ1 = 1, λ2 = −2,
so that we obtain the general solution
y = c1 ex + c2 e−2x .
Hence,
y 0 (x) = c1 − 2c2 e−2x .
Invoking initial conditions,
y(0) = c1 + c2 = 4, y 0 (0) = c1 − 2c2 = −5,
we get c1 = 1 and c2 = 3. The solution is, therefore,
y = ex + 3e−2x .

18
Case II: Real double root λ1 = λ2 = − a2

In this case, we know that we have only one immediate solution


a
y1 = e− 2 x .
To obtain a second independent solution y2 (needed for a basis), we use the method of reduction
of order. Setting y2 (x) = v(x)y1 (x) and substituting this and its derivatives y2 = v 0 y1 + vy10 and y200
into (2.7), we have
(v 00 y1 + 2v 0 y10 + vy 00 ) + a(v 0 y1 + vy10 ) + bvy1 = 0.
Thus,
v 00 y1 + u0 (2y10 + ay1 ) + u(y100 + ay10 + by1 ) = 0.
the expression in the last parenthesis is zero, since y1 is a solution of (2.7); also 2y10 + ay1 =
ax
−ae− 2 − ay1 = 0. We are thus left with
v 00 y1 = 0.
Hence v 00 = 0. Integrating twice, we get v = c1 x + c2 . We can simple choose c1 = 1 and c2 = 0
ax
and take v = x. Then y2 = xy1 = xe− 2 . Hence in the case of a double root of the characteristic
equation, a basis of solutions of (2.7) on any interval is
ax ax
e− 2 , xe− 2 .
The corresponding general solution is
ax
y = (c1 + c2 x)e− 2 .
Example 2.2.4. Solve the initial value problem
y 00 + y 0 + 0.25y = 0, y(0) = 3.0, y 0 (0) = −3.5.
Solution 2.2.4. The characteristic equation is
λ2 + λ + 0.25 = (λ + 0.5)2 = 0.
It has a double root λ = −0.5. This gives the general solution
y = (c1 + c2 x)e−0.5x .
Hence,
y 0 = c2 e−0.5x − 0.5(c1 + c2 x)e−0.5x .
Invoking initial conditions we obtain
y(0) = c1 = 3.0, y 0 (0) = c2 − 0.5c1 = −3.5,
which gives c2 = −2. The particular solution is, therefore
y = (3 − 2x)e−0.5x .

19
Case III: Complex conjugate roots λ1 = a + iω, λ2 = a − iω

This case occurs when the discriminant is negative. We can show that a basis of solutions of (2.7)
on any interval is
y1 = eax cos ωx, y2 = eax sin ωx.
The corresponding general solution is

y = eax (c1 cos ωx + c2 sin ωx).

Example 2.2.5. Solve the initial value problem

y 00 + 0.4y 0 + 9.04y = 0, y(0) = 0, y 0 (0) = 3.

Solution 2.2.5. The characteristic equation is

λ2 + 0.4λ + 9.04 = 0.

It has the roots −0.2 ± 3i. Hence the general solution is

y = e−0.2x (c1 cos 3x + c2 sin 3x).

The first initial condition gives y(0) = c1 = 0. Thus y = c2 e−0.2x sin 3x, whose derivative is

y 0 = c2 (−0.2e−0.2x sin 3x + 3e−0.2x cos 3x).

From this and the second initial condition we obtain y 0 (0) = 3c2 = 3.Hence c2 = 1, and the solution
is
y = e−0.2x sin 3x.
Exercise 2.2.3. Show that in the special case, λ1 = −λ2 , and real, the general solution can be
written as
y = c1 cosh λ1 x + c2 sinh λ1 x.

2.2.3 The Euler-Cauchy Equations

The so-called Euler-Cauchy equations are differential equations of the form

x2 y 00 + axy 0 + by = 0, (2.11)

where a and b are constants. Such equations can be solved by making a substitution

y = xm (2.12)

20
and its derivatives
y 0 = mxm−1 , y 00 = m(m − 1)xm−2
into (2.12). This gives
x2 m(m − 1)xm−2 + axmxm−1 + bxm = 0.
Dividing throughout by xm and simplifying we obtain the auxiliary equation

m2 + (a − 1)m + b = 0. (2.13)

Hence y = xm is a solution of (2.11) if and only if m is a root of (2.13). if the roots m1 and m2 are
real and different, then the functions

y1 (x) = xm1 , y2 = xx2

constitute a basis of solutions of the differential equation (2.11) for all x for which these functions
are defined. The corresponding general solution is

y = c1 xm1 + c2 xm2 ,

where c1 , c2 are constants.

Example 2.2.6. Solve the differential equation

x2 y 00 − 1.5xy 0 − 1.5y = 0.

Solution 2.2.6. This is an Euler-Cauchy equation with auxiliary equation

m2 + 0.5m − 0.5 = 0.

The roots are m1 = 0.5 and m2 = −1. Hence a basis of solutions for all positive x is

y1 = x0.5 , y2 = x−1 .

The corresponding general solution is


√ c2
y = c1 x + , x > 0.
x
Exercise 2.2.4. Make notes on the cases where the auxiliary equation has

1. a double real root.

2. complex conjugate roots.

21
2.3 Non-homogeneous Linear Differential Equations

We now consider the methods for solving second-order non-homogeneous linear differential equa-
tions of the form

a2 (x)y 00 + a1 (x)y 0 + a0 (x)y = f (x) (2.14)

where f (x) 6= 0. A general solution of the homogeneous equation (2.14) on some open interval I is
a solution of the form
y = yc + yp .
yc = c1 y1 (x) + c2 y2 (x) is called the complementary function and yp is called the particular
integral.

The complementary function is the general solution of the corresponding homogeneous equation
on I, and the particular is any solution of (2.14) on I containing no arbitrary constants. The main
task in this section is to discuss methods for finding the particular integral, yp .

2.3.1 Method of Undetermined Coefficients

The method of undetermined coefficients applies to linear differential equations with constant co-
efficients

y 00 + ay 0 + by = f (x), (2.15)

where f (x) is either an exponential function, a power of x, a cosine or sine, or sums or products
of such functions. The key idea of the method is to assume for yp an expression similar to that
of f (x), involving unknown coefficients that need to be determined by substitution of yp into the
given differential equation.

Rules for the method of undetermined coefficients

1. If f (x) in (2.15) is any one of the functions in the first column in the table below, choose
the corresponding yp in the second column and determine its ’undetermined’ coefficients by
substituting yp and its derivatives into (2.15).

2. If f (x) is a solution of the homogeneous equation corresponding to (2.15), then multiply your
choice of yp by x, or by x2 if this is a solution again.

22
3. If f (x) is a sum and/or product of functions listed in several row in the first column of the
table, then choose for yp the sum and/or product of the functions in the corresponding rows
in the second column.

Term in f (x) Choice of yp (x)


ke λx ceλx
n
kx (n = 0, 1, · · · ) n
cn x + cn−1 x n−1 + · · · + c1 x + c0
k cos ωx and/or sin ωx c1 cos ωx + c2 sin ωx
Exercise 2.3.1. Solve

1. y 00 + 4y = 8x2

2. y 00 − 3y 0 + 2y = ex

3. y 00 − 2y 0 + y = ex + x.

4. y 00 + 2y 0 + 8y = 16ex + sin 2x.

2.3.2 Method of Variation of Parameters

The method of undetermined coefficients is useful for obtaining particular solutions of non-homogeneous
linear differential equations only in the case where coefficients are constants and also the non-
homogeneous function was restricted to a specific class of functions.

We are now going to discuss a slightly more powerful method; which solves a rather larger class
of linear differential equations, in that the non-homogeneous function can be any function and the
coefficients need not be constants.

We consider linear differential equations of the form

y 00 + p(x)y 0 + q(x)y = f (x) (2.16)

where p(x), q(x), f (x) are assumed to be continuous on an open interval I.

Let y1 (x) and y2 (x) be the linearly independent solutions of the corresponding homogeneous part
of equation (2.16), then the complementary function is given by

yc (x) = c1 y1 + c2 y2 .

23
A particular solution of (2.16) is sought in the form

yp (x) = v1 y1 + v2 y2 (2.17)

which can be obtained from the complementary function by replacing the constants c1 and c2 by
the functions v1 and v2 , hence the name variation of parameters. Differentiation of (2.16) gives

yp0 = v10 y1 + v1 y10 + v20 y2 + v2 y20 . (2.18)

Let us conveniently impose a second arbitrary condition by setting

v10 y1 + v2 y2 = 0. (2.19)

Thus,

yp0 (x) = v1 y10 + v2 y20 . (2.20)

Differentiation of (2.20) yields

yp00 = v10 y10 + v1 y100 + v20 y20 + v2 y200 . (2.21)

Substituting (2.17), (2.20) and (2.21) into (2.16) and collecting terms containing v1 and terms
containing v2 we readily obtain

v1 (y100 + py10 + qy1 ) + v2 (y200 + py20 + qy2 ) + v10 y10 + v20 y20 = f (x).

Since y1 and y2 are solutions of the corresponding homogeneous equation, the above equation
reduces to

v10 y10 + v20 y20 = f ()x). (2.22)

Equation (2.22) together with condition (2.19) is a system of two linear algebraic equations for the
unknown functions v10 and v20 . The solution of such a system is obtained, by Cramer’s rule, as
y2 f (x) y1 f (x)
v10 = , v2 = , (2.23)
W W
where W = y1 y20 − y10 y2 is the Wronskian of y1 and y2 .

Integration of (2.23) gives


Z Z
y2 f (x) y1 f (x)
v1 = − dx, v2 = dx.
W W
Substituting these expressions into (2.17), we obtain the desired solution
Z Z
y2 f (x) y1 f (x)
yp = −y1 dx + y2 dx. (2.24)
W W

24
Here we have ignored the constants of integration since we are looking for a particular solution.

The general solution of the equation (2.16) is therefore given by


Z Z
y2 f (x) y1 f (x)
y(x) = c1 y1 + c2 y2 −y1 dx + y2 dx.
W W
Example 2.3.1. Use the method of variation of parameters to solve

y 00 + y = sec x.

Solution 2.3.1. y1 = cos x, y2 = sin x form the solution basis of the corresponding homogeneous
equation.The Wronslian, W = y1 y20 − y10 y2 = cos2 x + sin2 x = 1. Thus, the particular solution is
Z Z
y2 f (x) y1 f (x)
yp = −y1 dx + y2 dx
W W
Z Z
= − cos x sin x sec xdx + sin x cos x sec xdx

= − cos x ln | cos x| + x sin x.


Hence, the general solution is

y = (c1 ln | cos x|) cos x + (c2 + x) sin x

Exercise 2.3.2. Solve

1. y 00 + 4y 0 + 3y = 65 cos 2x,

(a) by the method of undetermined coefficients.


(b) by the method of variation of parameters.

2. y 00 + 9y = sec 3x + sin 3x,

(a) by the method of undetermined coefficients.


(b) by the method of variation of parameters.

Observe that the method of undetermined coefficients is simpler than the method of variation of
parameters.

25
Chapter 3

SERIES SOLUTIONS OF
ORDINARY DIFFERENTIAL
EQUATIONS

Many linear differential equations of second order have solutions which are not expressible in terms
of the well-known functions such as the trigonometric, logarithmic, polynomial and exponential
functions. We may obtain, however, a valid representation of many such functions by the power
series method, which gives solutions in the form of power series.

3.1 Power Series

Definition 3.1.1. A power series about a point x = x0 is an infinite series of the form

X
cn (x − x0 )n , (3.1)
n=0

where cn , n = 0, 1, · · · are constants.

26
For example,
∞ ∞
X (−1)n n
X (x − 3)n
x and
n n3
n=1 n=1
(−1)n
are power series about the points x = 0 and x = 3, respectively. In the first example cn = n
and x0 = 0, whereas, in the second, cn = n13 and x0 = 3.

Further examples of power series are the Maclaurin series:


1 X
1. = xm = 1 + x + x2 + · · · (|x| < 1).
1−x
m=0

X xm x2 x3
2. ex = =1+x+ + + · · ·.
m! 2 3!
m=0

X x2m x2 x4 x6
3. cos x = =1− + − + · · ·.
(2m)! 2! 4! 6!
n=0

X x2m+1 x3 x5 x7
4. sin x = =x− + − + · · ·.
(2m + 1)! 3! 5! 7!
n=0

Definition 3.1.2. Every power series has an interval of convergence. This is the set of all
numbers for which the series converges.

Convergence of the power series can be determined by the ratio test:



c
n+1
lim |x − x0 | = L. (3.2)
n→∞ cn

The series will converge for all values of x such that L < 1 and will diverge for L > 1, otherwise
the test is inconclusive. If the interval of convergence is finite, it has the mid-point x = x0 , so that
it is of the form

|x − x0 | < R (3.3)

and the series (3.1) converges for all values of x such that |x − x0 | < R and diverges for all x such
that |x − x0 | > R. The number R is called the radius of convergence of (3.1), and is given by

c
n
R = lim .
n→∞ cn+1

27
For a power series

X
cn (x − x0 )n
n=0

we have three possibilities for the radius of convergence:

1. If R = 0, the series converges only at its center x0 .

2. If 0 < R < ∞, the series converges, and respectively diverges for all x satisfying

|x = x0 | < R and |x = x0 | > R,

where R > 0.

3. If R is not finite, the series converges for all x.

Example 3.1.1. Find the radius of convergence of each of the following series


X (−1)n x6 x9
1. x3n = 1 − f racx3 8 + − + · · ·.
8n 64 512
n=0

X (x − 3)n
2. .
n3
n=0

(−1)n
Solution 3.1.1. 1. The power series is centered at x3 = 0, with cn = 8n and so
(−1)n+1
cn+1 = 8n+1
. The radius of convergence is given by

c
n 8n+1
R = lim = lim = 8.
n→∞ cn+1 n→∞ 8n

Hence the series converges for |x3 | < 8, that is , |x| < 2, and diverges for |x| > 2.

2. The power series is centered at x = 3. in this case, cn = n13 . The radius of convergence is,
therefore,
c
n n + 13
R = lim = lim = 1.
n→∞ cn+1 n→∞ n3
The series converges for |x − 3| < 1 or 2 < x < 4, and so diverges for (−∞, 2) ∪ (4, ∞). On
the boundaries:

28
• at x = 2, the series becomes

X (−1)n
,
n3
n=0

an alternating series which converges.


• at x = 4, the series becomes

X (1
,
n3
n=0

a p-series which converges.

Thus, the interval of convergence of the series is

2 ≤ x ≤ 4.

3.1.1 Operations on Power Series

We can differentiate, add and multiply power series as follows:

Termwise differentiation

A power series may be differentiated term-by-term. More precisely; if


X
y(x) = am (x − x0 )m
m→∞

converges for |x − x0 | < R, where R < 0, then the series obtained by differentiating term by term
also converges for those values of x. that is,
X
y 0 (x) = mam (x − x0 )m−1 (|x − x0 | < R).
m→∞

Similarly, X
y 00 (x) = m(m − 1)am (x − x0 )m−2 (|x − x0 | < R).
m→∞

29
Termwise addition

Two power series may be added term by term. More precisely; if the series
X X
am (x − x0 )m and bm (x − x0 )m
m→∞ m→∞

have positive radii of convergence and their sums are, say, f (x) and g(x), then the series
X
y(x) = (am + bm )(x − x0 )m
m→∞

converges and represents f (x) + g(x) for each x that lies in the intersection of the convergence
intervals of the original series.

Vanishing of all coefficients

If a power series has a positive radius of convergence and a sum that is identically zero throughout
its interval of convergence, then each coefficient of the series must be zero.

3.1.2 Existence of Power Series Solutions of ODEs

If the coefficients p(x) and q(x) and the function r(x) on the right side of

y 00 + p(x)y 0 + q(x)y = r(x) (3.4)

are real analytic functions (that is, have power series representations), then (3.4) has power series
solutions. Thus, polynomials, sines, cosines and exponential functions are everywhere analytic; so
too are sums and products of these functions

3.2 Ordinary Points and Singular Points of ODEs

Consider a differential equation of the form

y 00 + p(x)y 0 + q(x)y = r(x). (3.5)

30
Definition 3.2.1. The point x = x0 is an ordinary point of the differential equation (3.5) if both
p(x) and q(x) are analytic at x = x0 . If either of these functions is not analytic at x = x0 , then x0
is a singular point of (3.5).
Example 3.2.1. Identify all the singular points, if they exist, of each of the following differential
equations

1. (x2 + 4)y 00 − 6xy 0 + 3y = 0 on (−∞, ∞).

2. x(3 − x)y 00 − (3 − x)y 0 + 4xy = 0 on (−∞, ∞).


Solution 3.2.1. 1. p(x) = x−6x 3
2 +4 and q(x) = x2 +4 , which are both analytic on (−∞, ∞). So
there are no singular points for the given differential equation.

2. p(x) = −1 4
x and q(x) = 3−x , which are not analytic at x = 0 and x = 3 respectively. So x = 0
and x = 3 are the singular points of the given differential equation.

There are two types of singular points; a regular singular point and an irregular singular point.
Definition 3.2.2. A point x = x0 is called a regular singular point of a differential equation
(3.5) if not all of p(x) and q(x) are analytic but if (x − x0 )p(x) and (x − x0 )2 q(x) are analytic at
x = x0 . Otherwise the point is an irregular singular point.
Example 3.2.2. Find and classify the singular points of the equation

(x4 − 2x3 + x2 )y 00 + 2(x − 1)y 0 + x2 y = 0.

Solution 3.2.2. Observe that


2(x − 1) 1
p(x) = and q(x) = .
x4 − 2x3 + x2 x2−2x+1
Therefore singular point occur when x4 − 2x3 + x2 , that is, when x = 0 and x = 1. Now
2(x − x0 )(x − 1) (x − x0 )2
(x − x0 )p(x) = and (x − x0 )2 q(x) = .
x4 − 2x3 + x2 x2 − 2x + 1
Thus, for x0 = 0,
2
(x − x0 )p(x) = ,
−x x2
which is not analytic at x = 0. Hence, x = 0 is an irregular singular point.

For x0 = 1,
2(x − 1)2 2
(x − x0 )p(x) = 2 2
= ,
x (x − 1) x

31
which is analytic at x = 1. On the other hand,

(x − 1)2
(x − x0 )2 q(x) = = 1,
(x − 1)2

which again is analytic at x = 1. Ultimately, the point x = 1 is a regular singular point.

Exercise 3.2.1. Locate and classify the singular points of each of the following differential equations

1. (x2 − 3x)y 00 + (x + 2)y 0 + y = 0.

2. x3 (x2 − 25)(x − 2)y 00 + 3x(x − 2)y 0 + 7(x + 5)y = 0.

3.2.1 Power Series Solutions Near Ordinary Points

Suppose that the point x = x0 is an ordinary point of the differential equation (3.5), then the
differential equation has two linearly independent power series solutions of the form

X
y= cn (x − x0 )n , (3.6)
n=0

where these power series converge in some interval |x − x0 | < R about x0 , where R > 0.

Thus, near or about an ordinary point (x = x0 ) of a differential equation, we can find a series
solution valid in a region surrounding the point x = x0 with radius of convergence R. The radius
of convergence of this series is as large as the distance to the nearest point of singularity to x0 .
However, if the differential equation does not have singular points in the plane, then the power
series solution converges in the entire plane. The main task in solving any differential equation
using the power series method is to find a relation between the coefficients cn , n ≥ 2, and the two
arbitrarily constant coefficients c0 , c1 .

Example 3.2.3. Solve the differential equation

(x2 − 1)y 00 + xy 0 − y = 0, near x = 0.

Solution 3.2.3. Clearly, the point x = 0 is an ordinary point of the given differential equation.
We, therefore, seek solutions of the form

X
y= cn xn . (3.7)
n=0

32
Successive differentiation of (3.7) gives

X
y0 = ncn xn−1 (3.8)
n=0

and

X
y 00 = n(n − 1)cn xn−2 . (3.9)
n=0

Substituting these into the given differential equation we obtain


∞ ∞ ∞
! !
X X X
n(n − 1)cn xn−2 (x2 − 1) + ncn xn−1 (x) − cn xn = 0. (3.10)
n=0 n=0 n=0

This simplifies to

X ∞
X
2 n
(n − 1)cn x − n(n − 1)cn xn−2 = 0. (3.11)
n=0 n=0

We then follow the following steps.

Step I: Index shifting

Shift the powers of x−x0 , which is simply x in this case, to the smallest one present in the equation.
In this case the smallest exponent of x is n − 2, hence we shift the index of the term in the first
sum in (3.11) by replacing n by n − 2. Thus,

X ∞
X
[(n − 2)2 − 1]cn−2 xn−2 − n(n − 1)cn xn−2 = 0.
n−2=0 n=0

This simplifies to

X ∞
X
2 n−2
(n − 4n + 3)cn−2 x − n(n − 1)cn xn−2 = 0. (3.12)
n=2 n=0

33
Step II: Expressing all summations over a common range

The common range of the summations in (3.12) is [2, ∞). Writing out, explicitly, the terms that
do not belong to the common range we have

X ∞
X
2
((n − 4n + 3)cn−2 x n−2
− n(n − 1)cn xn−2 − 0 · c0 x−2 + 0 · c1 x−1 = 0.
n=2 n=2

That is,

X
0 · c0 x−2 + 0 · c1 x−1 + [(n − 3)(n − 1)cn−2 − n(n − 1)cn ]xn−2 = 0. (3.13)
n=2

Step III: Equating coefficients of like terms

0 · · · c0 = 0, 0 · · · c1 = 0 implies c0 , c1 are arbitrary constants.


Now for n ≥ 2 we have
(n − 3)(n − 1)cn−2 − n(n − 1)cn = 0.
Thus,
n−3
cn = cn−2 . (3.14)
n
The important relation (3.14) is called a recurrence relation. We use it to obtain the coefficients
cn , n ≥ 2, in terms of C0 and c1 which are arbitrary. So

c2 = − 21 c0 ,

c3 = 0,

c4 = 14 c2 = − 81 c0 ,

c5 = − 32 c3 = 0,

c6 = 21 c4 = − 16
1
c0 , etc.

The general solution is therefore given by


!
1 1 1
y = c0 1 − x2 − x4 − x6 − · · · + c1 x.
2 8 16

34
The points x = 1 and x = −1 are the only singular points of the given differential equation. The
minimum distance between x0 = 0 and these singular points, which is the radius of convergence, is
R = 1. The above solution is, therefore, valid in the interval |x| < 1.

Exercise 3.2.2. Find the power series solution of the differential equation

(x3 − 1)y 00 + x2 y 0 + xy = 0

about the point x = 0.

3.2.2 Power Series Solutions Near Regular Singular Points

The power series of the form



X
y= cn (x − x0 )n
n=0
is a good way of representing a solution of a differential equation near an ordinary point x = x0 .
However, several second-order ODEs of considerable practical importance have coefficients that are
not analytic, but are not too bad at the same time, so that these ODEs can still be solved by
series (that is power series times a logarithm or times a fractional power of x, and so on). The so
called Frobenius method permits an extension of the power series method to find solutions near
a regular singular point.

Theorem 3.2.1 (Frobenius Method). If x0 is a regular singular point for the differential equation

y 00 + p(x)y 0 + q(x)y = 0 (3.15)

then the differential equation has at least one non-zero solution of the form

X
r
y(x) = (x − x0 ) cn (x − x0 )n , (3.16)
n=0

where the exponent r is a constant which may be real or complex, chosen so that c0 6= 0. If
(x − x0 )p(x) and (x − x0 )2 q(x) have power series which converge for |x − x0 | < R then the power
series also converges for |x − x0 | < R.

We will call a solution of the form (3.16) a generalized power series solution. Although the
theorem guarantees only one generalized power series solution, not a basis, the differential equation
(3.15) also has a second solution, such that these two solutions are linearly independent, that may
similar to (3.16), but with a different r and different coefficients, or may contain a logarithmic term.

35
Let us now illustrate the Frobenius method by solving the differential equation

2xy 00 + y 0 + y = 0

which can be written as


1 0 1
y 00 + y + y = 0. (3.17)
2x 2x
We can easily check that x = 0 is a regular singular point. We can therefore apply the Frobenius
method by assuming a solution of the form

X ∞
X
r n
y(x) = x cn (x) = cn (x)n+r , (c0 6= 0).
n=0 n=0

Then

X ∞
X
0 n+r−1 00
y (x) = (n + r)cn x , y (x) = (n + r)(n + r − 1)cn xn+r−2 .
n=0 n=0

Substitution into the given differential equation gives



X ∞
X ∞
X
2(n + r)(n + r − 1)cn xn+r−1 + (n + r)cn xn+r−1 + cn xn+r = 0,
n=0 n=0 n=0

which simplifies to
∞ ∞
" #
X X
xr (2n + 2r − 1)(n + r)cn xn−1 + cn xn = 0.
n=0 n=0

We can divide by xr to obtain



X ∞
X
(2n + 2r − 1)(n + r)cn xn−1 + cn xn = 0.
n=0 n=0

Shifting all the indices to n − 1 we have



X ∞
X
(2n + 2r − 1)(n + r)cn xn−1 + cn−1 xn−1 = 0.
n=0 n=1

Writing out all the summations over a common range,m and simplifying, we obtain

X
(2r − 1)rc0 x−1 + [(2n + 2r − 1)(n + r)cn + cn−1 ]xn−1 = 0.
n=1

36
Now, the coefficient of each power of x must be zero. In particular, we must have

(2r − 1)rc0 = 0, (2n + 2r − 1)(n + r)cn + cn−1 = 0. (3.18)

Recall that we must have c0 6= 0 hence

(2r − 1)r = 0.

This is called the indicial equation of the given differential equation. The Frobenius method
gives one of the basis solutions, which will always be of the form (3.16), where r is a root of the
indicial equation. The other solution will be of the form indicated by the indicial equation. There
are three possible cases which we state in form of a theorem as follows:
Theorem 3.2.2. Suppose that the differential equation (3.15) satisfies the assumptions in Theorem
3.2.1. Let r1 and r2 be the roots of the indicial equation. Then we have the following three cases:

Case 1. Distinct roots, Not differing by an Integer

A basis is
y1 (x) = xr1 (c0 + c1 x + c2 x2 + · · · )
and
y2 (x) = xr2 (A0 + A1 x + A2 x2 + · · · )
with coefficients obtained successively from (3.18) with r = r1 and r = r2 , respectively.

Case 2. Double root r1 = r2 = r

A basis is
y1 (x) = xr (c0 + c1 x + c2 x2 + · · · )
and
y2 (x) = y1 (x) ln x + xr (A1 x + A2 x2 + · · · ) (x > 0).

Case 3. Roots differing by an Integer

A basis is
y1 (x) = xr1 (c0 + c1 x + c2 x2 + · · · )

37
and
y2 (x) = ky1 (x) ln x + xr2 (A0 + A1 x + A2 x2 + · · · ) (x > 0),
where the roots are so denoted that r1 > r2 and k may turn out to be zero.

In our example, the indicial equation has the roots


1
r1 = 0, r2 = (distinct and not differing by an integer).
2
From (3.18) we yield the recurrence formula
1
cn = − cn−1 , for n ≥ 1.
(2n + 2r − 1)(n + r)
If r = 0, the recurrence formula becomes
1
cn = − cn−1 .
(2nr − 1)(n)
Thus,
c1 = −c0 ,
1 1
c2 = − c1 = c0 ,
2·3 2·3
1 1
c3 = − c2 = − c0 ,
5·3 (5 · 3)(3 · 2)
1 1
c4 = − c3 = c0 ,
7·4 (7 · 5 · 3)4!
and so forth. In general, we would have
1
cn = (−1)n c0 .
(2n − 1)(2n − 3) · · · 1 · n!
One of the power series solutions to the differential equation is, therefore,
" #
0 1 2 1 3 1
y1 (x) = x 1 − x + x − x + − ···
3·2 (5 · 3)3! (7 · 5 · 3)4!

1 2 1 1
=1−x+ x − x3 + − ··· .
3·2 (5 · 3)3! (7 · 5 · 3)4!
If r = 12 , the recurrence formula becomes
1
cn = − cn−1 .
n(2n + 1)

38
We find that
1
c1 = − ,
3
2 1
c2 = − c1 = c0 ,
5 2·5·3
1 1
c3 = − c2 = − c0,
3·7 3!(7 · 5 · 3)
and in general,
1
cn = (−1)n .
n!(2n + 1)(2n − 1) · · · 1
We thus obtain a second power series solution to the given differential equation:
" #
1 1 1 1
y2 (x) = x 2 1 − x + x2 − x3 + · · · .
3 2·5·3 3!(7 · 5 · 3)

The general solution to the given differential equation is a superposition of y1 (x) and y2 (x):
" # " #
1 2 1 1 √ 1 1 1
y = c1 1−x+ x − x3 + −· · · +c2 x 1− x+ x2 − x3 +· · · .
3·2 (5 · 3)3! (7 · 5 · 3)4! 3 2·5·3 3!(7 · 5 · 3)

We obtained two linearly independent power series solutions in this case, but this does not always
happen. If the roots of the indicial equation differ by an integer, we may obtain only one power
series solution. In that case, a second independent solution can then be found by referring to
Theorem 3.2.2 or by the method of reduction of order.

39
Chapter 4

LAPLACE TRANSFORM METHOD

4.1 Definitions

We have so far looked at the methods of solving linear differential equations, mainly those with
a continuous non-homogeneous function. However, in many applications, the non-homogeneous
function is not continuous but may be piecewise continuous. The Laplace transform method is an
invaluable tool for solving such problems. The process of solution consists of three steps:

1. The given ODE is transformed into an algebraic equation, known as the subsidiary equa-
tion.

2. The subsidiary equation is solved by purely algebraic manipulations.

3. The solution obtained in step 2 is transformed back, resulting in the solution of the given
problem.

The switching from calculus to algebra is called operational calculus. The Laplace transform
method is the most important operational method for solving initial value problems that arise
in engineering. The method solves problems more directly; initial value problems without first
determining a general solution, and non-homogeneous ODs without first solving the corresponding
homogeneous ODE.

40
Definition 4.1.1. The class of functions defined by
Z ∞
T {f (t)}(s) = K(s, t)f (t)dt = F (s) (4.1)
−∞

is called integral transforms.

he function K(s, t) is called the kernel of the transformation.

Various choices of K(s, t) in (4.1) lead to special transforms, each with its own properties to make
it useful in special circumstances.

Definition 4.1.2. The Laplace transform of a function f (t), t ≥ 0 is denoted by L{f (t)} and is
given by
Z ∞
L{f (t)}(s) = e−st f (t)dt = F (s), (4.2)
t=0

if the integral exists.

Furthermore, the original function f (t) in (4.2) is called the inverse Laplace transform of F (s)
and is denoted by L−1 {F (s)}, that is,

f (t) = L−1 {F (s)}.

4.1.1 Notation

1. Original functions depend on t and their transforms on s.

2. Original functions are denoted by lowercase letters and their transforms by uppercase letters,
so that F (s) denotes the transform f (t), and Y (s) denotes the transform of y(t), and so on.

Example 4.1.1. Find the Laplace transforms, F (s) = L{f (t)}(s) in each of the following cases:

1. f (t) = a, a constant function.

2. f (t) = t.

3. f (t) = eat when t ≥ 0, where a is a constant.

41
Solution 4.1.1. 1. Z ∞
a
L{a}(s) = ae−st dt = .
0 s
2. Z ∞
a
L{t}(s) = te−st dt =
0 s

−te−st 1 ∞ −st
Z
= + e dt
s s t=0

t=0
1
= .
s2
3. ∞
Z ∞
at −st at 1 −(s−a)t
L{e }(s) = e e dt = e
a−s

0
t=0
1
= .
s−a
Exercise 4.1.1. Complete the following table of Laplace transforms:

f (t) L{f (t)} = F (s)


f 0 (t) sL{f (t)} − f (0)
f 00 (t) s2 L{f (t)} − sf (0) − f 0 (0)
bt
e f (t) F (s − b)
t n

eat
cos at
sin at
cosh at
sinh at
teat
tn eat
ebt cos at
ebt sin at
ebt cosh at
ebt sinh at
t sin at
t cos at

42
Theorem 4.1.1 (Linearity of the Laplace Transform). The Laplace transform is a linear opera-
tion; that is, for any functions f (t) and g(t) whose Laplace transforms exist and any constants a
and b the Laplace transform of af (t) + bg(t) exists, and is given by

L{af (t) + bg(t)} = aL{f (t)} + bL{g(t)}.

Proof. By the definition of a Laplace transform


Z ∞
L{af (t) + bg(t)} = e−st [af (t) + bg(t)]dt
t=0
Z ∞ Z ∞
−st
=a e f (t)dt + b e−st g(t)dt
t=0 t=0
= aL{f (t)} + bL{g(t)}.

In fact, the operator L{·} is an integral operator.

Example 4.1.2. 1. Let f (t) = cosh at. Find L{f (t)}(s).

2. Let F (s) = 1
(s−a)(s−b) , a 6= b. Find L−1 {F (s)}(t).

Solution 4.1.2. 1. Since cosh at = 21 (eat + e−at ), we obtain from Example 4.1.1 and Theorem
4.1.1
L{f (t)} = L{cosh at}
!
1 1 1 1 s
= (L{eat } + L{e−at }) = + = .
2 2 s−a s+a s2 − a2

2. Since ! !
1 1 1 1
= − ,
(s − a)(s − b) a−b s−a s−b
we have from Theorem 4.1.1
( )
−1 −1 1
L {F (s)}(t) = L
(s − a)(s − b)
!" ( ) ( )#
1 1 1 eat − ebt
= L−1 − L−1 = .
a−b s−a s−b a−b

43
Exercise 4.1.2. Let
s
F (s) = , a 6= b.
(s − a)(s − b)
Find L−1 {F (s)}(t).

Theorem 4.1.2 (The First Shifting Theorem). If f (t) has the Laplace transform F (s), then eat f (t)
has the Laplace transform F (s − a). That is,

L{eat f (t)} = F (s − a)

or, if we take inverses both sides,

eat f (t) = L−1 {F (s − a)}.

Exercise 4.1.3. Prove Theorem 4.1.2.

Example 4.1.3. 1. Find L{e7t t4 }.

2. Find L−1 { s2 +4s+5


s
}.
4!
Solution 4.1.3. 1. Take f (t) = t4 , so that F (s) = s5
. Then

24
L{e7t t4 } = F (s − 7) = .
(s − 7)5

2. Since
s s s+2 2
= = − ,
s2 + 4s + 5 2
(2 + 2) + 1 (s + 2) + 1 (2 + 2)2 + 1
we readily deduce that
( ) ( ) ( )
s s+2 2
L−1 2
= L−1 − 2L−1
s + 4s + 5 (s + 2) + 1 (s + 2) + 1

= e−2t cos t − 2e−2t sin t.

Exercise 4.1.4. Find the inverse of the Laplace transform


3s − 137
L{f (t)}(s) = .
s2 + 2s + 401

44
4.1.2 Laplace Transforms of Derivatives

Since the ultimate objective is to solve differential equations using Laplace transforms, the rela-
tionship between the Laplace transform of a derivative and the Laplace transform of the original
function will play an important role. By definition
Z ∞
L{f (t)} = e−st f (t)dt
0

−f (t)e−st 1 ∞ −st 0
Z
= + e f (t)dt
s s 0


t=0
f (0) 1
= + L{f 0 (t)}.
s s
Hence,

L{f 0 (t)} = sL{f (t)} − f (0). (4.3)

Further, by definition, Z ∞
0
L{f (t)} = e−st f 0 (t)dt
0

−f 0 (t)e−st 1 ∞ −st 00
Z
= + e f (t)dt
s s 0


t=0
0
f (0) 1
= + L{f 00 (t)}.
s s
Thus,

L{f 00 (t)} = s2 L{f (t)} − sf (0) − f 0 (0). (4.4)

4.1.3 Application to Differential Equations

We shall now discuss how the Laplace transform method solves initial value problems. There are
three major steps which shall be illustrate using an example.

45
Illustrative example

Solve
y 00 − 3y 0 + 2y = 0, y(0) = −3, y 0 (0) = 5
using the Laplace transform method.

Step 1: Setting up the subsidiary equation

Let L{y(t)} = Y (s) be the Laplace transform of the solution of the differential equation. Trans-
forming the differential equation and using the linearity property of the Laplace transform operator,
we obtain
L{y 00 − 3y 0 + 2y} = L{0}.
Thus,
L{y 00 } − 3L{y 0 } + 2L{y} = 0.
Substitution according to (4.3) and (4.4) gives

s2 Y (s) − sy(0) − y 0 (0) − 3[sY (s) − y(0)] + 2Y (s) = 0.

Substituting the given initial conditions, we have

s2 Y (s) + 3s − 5 − 3sY (s) − 9 + 2Y (s) = 0.

Thus,

(s2 − 3s + 2)Y (s) = 14 − 3s. (4.5)

Equation (4.5) is called the subsidiary equation of the given differential equation.

Step 2: Solution of the subsidiary equation

Solve for Y (s) in the subsidiary equation (4.5) above, to obtain


14 − 3s 8 11
Y (s) = = − ,
s2 − 3s + 2 s−2 s−1
in partial fractions.

46
Step 3: Inversion of Y (s) to obtain y(t) = L−1 {Y (s)}(t)

Taking the inverse transform both sides we get


( )
−1 −1 8 11
L {Y (s)} = L −
s−2 s−1
() ( )
−1 8 −1 11
=L −L
s−2 s−1
( ) ( )
−1 1 −1 1
= 8L − 11L
s−2 s−1

= 8e2t − 11et .
Thus, y(t) = 8e2t − 11et is the solution of the given initial value problem.

Exercise 4.1.5. Solve the differential equation

y 00 − 2y 0 + 10y = 0, y(0) = 3, y 0 (0) = −3,

using the Laplace transform method.

47

You might also like