MATH350

Differential Equations I
(MATH 350)
Dr. Joseph K. Ansong

(Dept. of Mathematics, University of Ghana, Legon)
LECTURE NOTES
Math 350 : Differential Equations I J.K.A
2 c Dr. Joseph K. Ansong

Contents
1 Definitions & Terminology 1

1.1 Introduction and Motivation . . . . . . . . . . . . . . . . . . . 1
1.2 Classifications and Terminology . . . . . . . . . . . . . . . . . 2
1.3 Solution of a differential equation . . . . . . . . . . . . . . . . 4
1.4 Applications in different fields . . . . . . . . . . . . . . . . . . 6
2 First Order Differential Equations 9

2.1 Separable Equations . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Linear First-Order Equations . . . . . . . . . . . . . . . . . . 12
2.3 Exact Equations . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 Substitutions and Transformations . . . . . . . . . . . . . . . 18
2.4.1 Homogeneous equations . . . . . . . . . . . . . . . . . 19
2.4.2 Equations of the form dy/dx = G(ax + by) . . . . . . . 22
2.4.3 Bernoulli equation . . . . . . . . . . . . . . . . . . . . 23
3 Linear Second Order Differential Equations 25

3.1 Homogeneneous Linear Equations . . . . . . . . . . . . . . . . 25
3.2 Non-homogeneous Equations . . . . . . . . . . . . . . . . . . . 27
3.2.1 The Method of undetermined coefficients . . . . . . . . 28
3.2.2 The Method of variation of parameters . . . . . . . . . 33
3.3 Equations with Variable Coefficients . . . . . . . . . . . . . . 37
3.3.1 Representation of Solutions to IVPs . . . . . . . . . . . 39
3.3.2 Reduction of Order . . . . . . . . . . . . . . . . . . . . 42
3.3.3 Cauchy-Euler (or Equidimensional) Equations . . . . . 44
4 Series Solutions of Differential Equations 49

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Power Series Solutions to Linear Differential Equations . . . . 53
4.2.1 Expanding About x = 0 . . . . . . . . . . . . . . . . . 63
4.2.2 Series Solutions about Regular Singular Points . . . . . 64
4.2.3 Method of Frobenius . . . . . . . . . . . . . . . . . . . 66
3
4.2.4 Form of a Second Linearly Independent Solution . . . . 70
5 Laplace Transforms 73
5.0.5 Table of Laplace Transforms . . . . . . . . . . . . . . . 77
5.0.6 Properties of Laplace Transforms . . . . . . . . . . . . 78
5.0.7 Inverse Laplace Transforms . . . . . . . . . . . . . . . 79
5.0.8 Solving IVPs Using Laplace Transforms . . . . . . . . . 84
5.0.9 IVPs with Non-zero Initial Conditions . . . . . . . . . 87
5.0.10 Transforms of Discontinuous Functions . . . . . . . . . 88
5.0.11 Properties . . . . . . . . . . . . . . . . . . . . . . . . . 91
6 Fourier Series 103

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Chapter 1
Definitions & Terminology
1.1 Introduction and Motivation

Mathematical models play a critial role in the sciences and engineering by
helping us gain a better understanding of real life phenomena. The mod-
els are generally simplified versions of the actual physical phenomena un-
der investigation. They are simplified because the parameters governing the
natural phenomena may is often not completely understood or may be too
complicated to be represented mathematically. The development of mathe-
matical models usually result in an equation or a set of equations specifying
how an unknown function (say φ(t)) changes with respect to a variable (say
t). Such an equation is referred to as a differential equation. For exam-
ple, the variation in the population of mosquitoes in a certain village may be
represented by the differential equation
dM
= kM, (1.1)
dt
where M is the number (population) of mosquitos at time t, and k is a known
constant (obtained from observational or experimental data). Fortunately for
us, equation (1.1) can be solved easily using integration techniques learned
in your Calculus class. Re-writing the equation and integrating results in
dM
= kdt,
Z M Z
dM
= kdt,
M
ln(M ) = kt + C1 ,
M = ekt+C1 = eC1 ekt
M (t) = Cekt . (1.2)
1
The integration constant C can be determined if we know the initial popu-

lation of mosquitoes. Equation (1.2) then helps us to predict the population
of mosquitoes at a future time. For instance, if k = −5, and if initially (at
t = 0) the population of mosquitoes is 1000, then we have
1000 = Ce5×0 = C,
=⇒ M (t) = 1000e−5t . (1.3)
Thus, our simple model may now be used to predict the decline in the pop-
ulation of mosquitoes in the village. A differential equation (e.g. equation
1.1) with the initial value specified (e.g. M = 1000 at t = 0; as in the above
example) is called an Initial Value Problem (IVP).
Notice that the solution to a differential equation is not a number but a
function (see equation 1.2), and contains arbitrary “constants of integration”.
The presence of these arbitrary constants implies that there is generally no
unique solution to a differential equation. Thus, equation (1.2) is referred
to as the general solution since there are infinitely many solutions; there
is a solution for every C. Three of such solutions are displayed in Figure
1.1. The solution for a particular value of C (as in equation 1.3) is called a
particular solution to the differential equation.
Figure 1.1: Three solutions of M (t) = Cekt for k = −1.
1.2 Classifications and Terminology

The unknown function in a differential equation (e.g. M in equation 1.1)
is called the dependent variable and the others (e.g. t in equation 1.1)

is called the independent variable. It is usually clear from the equation

which variable is dependent and which is independent.
Definition 1. [Differential Equation]

A differential equation is an equation involving the rate of change of a
quantity. That is, it involves a function and its derivatives.
An ordinary differential equation (ODE) is a differential equation in

which only the derivatives of the unknown function with respect to one inde-
pendent variable appear in the equation. A partial differential equation
(PDE) is a differential equation in which partial derivatives of the unknown
function with respect to at least two independent variables appear in the
equation. Examples of ODEs are
dy
+ kt = 10,
dt
d2 y dy
2
+ a + by = x2 . (1.4)
dx dx
The following are some examples of PDEs
∂u2 ∂v
+ = 0,
∂x2 ∂x
∂φ ∂φ
+b + b = 0.
dx ∂y
The order of a differential equation is the order of the highest derivative
appearing in the equation. Examples of first order equations are
dx dy
+ x = 0, and + 3xy = 0,
dt dx
and the following are second order equations:
3
d2 y d2 x dx
2
+ cy = 0, and 2
+ = 0.
dx dt dt
The constants appearing in a differential equation are called coefficients or

parameters. For example the constant k appearing in equation (1.1), and
the constants a and b in equation (1.4).
A first order ODE is called explicit if it can be written in the form
dy
= f (x, y), (1.5)
dx

for a real-valued function f of two variables. Otherwise it is said to be

implicit. An example of a first order implicit differential equation is an
equation of the form
dy
F x, y, = 0,
dx
where F is a continuous function. If the above equation can be solved for
dy/dx (or if dy/dx can be isolated), then we may obtain an explicit differential
equation of the form in (1.5).
The general form for an nth-order differential equation can be expressed
in the form
dn y

dy
F x, y, , · · · , n = 0, (1.6)
dx dx
where it is assumed that the equation holds for all x in an open interval
(a < x < b, where a or b could be infinite). If the highest-order term can be
isolated, then we can write (1.6) in the explicit form
dn y dn−1 y

dy
= f x, y, , · · · , n−1 = 0. (1.7)
dxn dx dx
Equation (1.7) is often preferred to (1.6) for computational and theoretical

purposes.
1.3 Solution of a differential equation
Definition 2. [Explicit Solution]

A function φ(x) that when substituted for y in equation (1.5) satisfies
the equation for all x in an interval I is called an explicit solution to
the equation on I.
Example 1. Show that φ(x) = x − 1 + ce−x is an explicit solution to the

linear equation
dy
+ y = x,
dx
where c is a constant.
The point is to show that y = φ(x) satisfies the given equation.

Solution. Let y = φ(x) = x − 1 + ce−x

dy
=⇒ = 1 − ce−x
dx
Note that the equations above are defined for all real x. Substituting into
the differential equation gives
dy
+ y = (1 − ce−x ) + (x − 1 + ce−x )
dx
= x. (1.8)
Hence, φ(x) = x − 1 + ce−x is an explicit solution for all real x.

Example 2. Show that φ(x) = x2 − x−1 is an explicit solution to the linear
equation
d2 y 2
− y=0 (1.9)
dx2 x2
Solution. Let
dy
y = φ(x) = x2 − x−1 =⇒ = 2x + x−2
dx
d2 y
=⇒ = 2 − 2x−3 .
dx2
Note that these expressions are defined for all x 6= 0. Substituting them into
(1.9), we get
d2 y 2 2
2
− 2 y = (2 − 2x−3 ) − 2 (x2 − x−1 )
dx x x
= (2 − 2x−3 ) − (2 − 2x−3 )
= 0.
Since this is true for any x 6= 0, the function φ(x) = x2 − x−1 is an explicit
solution to (1.9) on (−∞, 0) and (0, ∞).
Z x√
Example 3. Check whether the function y = 1 + t3 dt + C, −1 <
0
x < ∞, is a solution to the linear equation
dy √
= 1 + x3
dx
on the given interval, where C is a constant.

Solution. Applying the fundamental theorem of calculus, we can differenti-

ate Z x√
y= 1 + t3 dt + C,
0
to get
√ √
Z x
dy d
= 1 + t dt + 0 = 1 + x3 .
3
dx dx 0
So yes, the given function is a solution to the differential equation.
Definition 3 (Implicit Solution). A relation G(x, y) = 0 is said to be an
implicit solution to equation (1.6) on the interval I if it defines one or more
explicit solutions on I.
Example 4. Verify that
x + y + exy = 0,
is an implicit solution to the equation
dy
(1 + xexy ) + 1 + yexy = 0.
dx
Solution. We assume that a function y(x) exists and also differentiable (by
the so-called implicit function theorem). Applying the technique of im-
plicit differentiation, we differentiate both sides of x + y + exy = 0 to get:

d xy dy xy dy
(x + y + e ) = 1 + +e y+x
dx dx dx
dy dy
=1+ + yexy + xexy
dx dx
dy
= (1 + xexy ) + 1 + yexy = 0,
dx
where the last expression is identical to the given differential equation. Thus,
x + y + exy = 0 is an implicit solution on some interval.
1.4 Applications in different fields

In this section, we state some differential equations that arise in a variety of
subject areas.
1. Population model: Let p(t) be the population of a species at time t.
The growth of the population is governed by the logistic model (or
differential equation):
dp p(p − 1)
= k1 p − k2 (1.10)
dt 2

or
dp
= −Ap(p − p1 ), p(0) = p0 , (1.11)
dt
where A = k2 /2 and p1 = 2k1 /k2 + 1 are constants to be determined.

Note that if k2 = 0, the logistic model reduces to the simple (or
Malthusian) model we introduced earlier for the growth of mosquitoes
(equation 1.1).
2. Electricity: The time variation of electric charge Q in an electric

circuit with inductance L, resistance R, capacitance C, and with an
externally applied voltage (or electromotive force) is governed by the
differential equation
d2 Q dQ 1
L 2
+R + Q = E(t). (1.12)
dt dt C
3. Psychology: One model in psychology of the learning of a task by an

individual learner, p, is given by
dy 2p h i
= √ y 3/2 (1 − y)3/2 , (1.13)
dt n
where y represents the learner’s skill level as a function of time, t, and

n is the nature of the task to be learned. Both p and n are constants.
4. Black-Scholes equation: The value V (x, t) of an option to by or sell

a stock with price x at time t is governed by the partial differential
equation
∂V 1 ∂ 2V ∂V
+ σ 2 x2 2 + rx − rV (1.14)
∂t 2 ∂x ∂t
where the constant r is a prevailing risk-free interest rate, and σ 2 is

a measure of the volatility of the investor’s return on this particular
stock.
5. Predator-Prey model: In ecology, the population dynamics of com-

peting species in which one is a predator with population x2 , and the
other its prey with population x1 is governed by the Lotka-Volterra

system of equations
dx1
= Ax1 − Bx1 x2 ,
dt
dx2
= −Cx2 + Dx1 x2 ,
dt
where A, B, C, D are positive constants.

Chapter 2
First Order Differential

Equations
2.1 Separable Equations
Definition 4 (Separable Equations). A first order differential equa-

tion
dy
= f (x, y)
dx
is said to be separable if f (x, y) can be written as the product of a
function g(x) that depends on only x and a function h(y) that depends
on only y such that
dy
= h(y)g(x) (2.1)
dx
Method of Solution
To solve
dy
= h(y)g(x), (2.2)
dx
separate variables to get
dy
= g(x)dx, h(y) 6= 0.
h(y)
9
Now integrate both sides:

Z Z
dy
= g(x)dx, h(y) 6= 0.
h(y)
The result of the integration above yields the solution to the differential
equation.
Example 5. Solve the following differential equations
dy 1
= 3
dx xy
Solution.
dy 1
= 3.
dx xy
dx
=⇒ y 3 dy = ,
x
Z Z
3 dx
=⇒ y dy = ,
x
1
=⇒ y 4 = ln(x) + C1 ,
4
4
=⇒ y = [ln(x) + 4C1 ] = [ln(x) + C],
y = [ln(x) + C]1/4
Example 6. Solve the following differential equations
dy
= y(2 + sin x)
dx
(2.3)
Solution.
dy
= y(2 + sin x),
dx
Z Z
dy
=⇒ = (2 + sin x)dx,
y
=⇒ ln(y) = 2x − cos(x) + C1 ,
=⇒ y = e2x−cos x+C1 = eC1 e2x−cos(x) ,
∴ y = Ke2x−cos(x) .
Example 7. Solve
(1 + x)dy − ydx = 0

Solution.
(1 + x)dy − ydx = 0
dy 1
=⇒ = dx
y 1+x
Z Z
dy 1
=⇒ = dx
y 1+x
=⇒ ln |y| = ln |1 + x| + C1
y
=⇒ ln | | = C1
1+x
y
=⇒ = ±eC1 = C
1+x
=⇒ y = C(1 + x),
where C is an arbitrary constant.
Example 8. Solve
dy
= y2 − 9
dx
Solution. Z Z
dy
2
= dx, y 6= ±3
y −9
Now, we use partial fractions to write:
1 1 A B
= ≡ + ,
y2 −9 (y − 3)(y + 3) y−3 y+3
=⇒ 1 = A(y + 3) + B(y − 3),

1
y=3: 6A = 1, =⇒ A =
6
1
y = −3 : 1 = −6B, =⇒ B = −
6
1 1/6 −1/6
= +
y2 −9 y−3 y+3
So the integral becomes
Z
−1/6
Z
1/6
+ dy = dx
y−3 y+3
1 1
ln |y − 3| − ln |y + 3| = x + C1
6 6
y−3
ln = 6x + C2 , C2 = 6C1
y+3
y−3
= ±eC2 e6x = Ce6x
y+3
Thus,
1 + Ce6x

y=3 .
1 − Ce6x
2.2 Linear First-Order Equations

A linear first order differential equation is an equation of the form
dy
a1 (x) + a0 (x)y = b(x).
dx
To solve the differential equation, we first write it in a standard form
dy
+ P (x)y = Q(x). (2.4)
dx
Now suppose that equation (2.4) could be simplified by multiplying by some
function, say µ(x), such that
dy
µ(x) + µ(x)P (x)y = µ(x)Q(x) (2.5)
dx
and that
d
[µ(x)y] = µ(x)Q(x). (2.6)
dx
Then it will be easy to integrate the equation above and obtain a solution
to the differential equation. Notice that equations (2.5) and (2.6) imply that
we can compute the function µ(x) :
µ0 (x) = µ(x)P (x)
Z Z
dµ
=⇒ = P (x)dx
µ
R
=⇒ µ(x) = e P (x)dx . (2.7)
So from equation (2.6) we get the solution
Z
µ(x)y = µ(x)Q(x)dx
Z
1
∴y= µ(x)Q(x)dx, (2.8)
µ(x)

where µ(x) is given by equation (2.7). The function µ(x) is called the inte-
grating factor of the differential equation.
(2.9)
Below is a summary of the approach outlined above.

a) Write the linear equation in a standard form
dy
+ P (x)y = Q(x). (2.10)
dx
b) Calculate the integrating factor:

R
P (x)dx
µ(x) = e
c) Multiply (2.10) by µ(x) such that

d
[µ(x)y] = µ(x)Q(x)
dx
d) Integrate the equation above to get

Z
1
y= µ(x)Q(x)dx
µ(x)
Example 9. Solve the following IVP

dy 4
+ 4y − e−x = 0, y(0) = .
dx 3
Solution.
dy
+ 4y − e−x = 0
dx
dy
=⇒ + 4y = e−x
dx
R
4dx
µ(x) = e = e4x
where we have ignored the constant of integration because it cancels out in
the subsequent steps. Thus
dy
e4x + 4xe4x = e3x
dx
d 4x
=⇒ e y = e3x
dx
Z
4x 1
=⇒ e y = e3x dx = e3x + C
3
1
=⇒ y = e−x + Ce−4x
3
Applying the initial condition, we have
4 4 1
x = 0, y = , =⇒ = + C, =⇒ C = 1
3 3 3
1
∴ y = e−x + e−4x .
3
Example 10. Solve the differential equation
dy
− 4y = x6 ex
x
dx
Solution. Re-write in a standard form to get
dy y
− 4 = x5 e x , x 6= 0
dx x
− x4 dx −4
R
µ=e = e−4 ln |x| = eln x = x−4
dy
=⇒ x−4 − 4x−5 y = xex
dx
d −4
=⇒ x y = xex
dx Z
=⇒ x−4 y = xex dx
Using integration by parts, we let

Z
I= xex dx
=⇒ I = xex − ex + C = ex (x − 1) + C
Thus
x−4 y = ex (x − 1) + C
∴ y = x4 ex (x − 1) + Cx4 .
EXERCISE 1. Solve
dy
1) − y = e3x
dx
dr
2) + r tan θ = sec θ
dθ
dy y
3) = + 2x + 1
dx x

2.3 Exact Equations
Definition 5. Consider the total differential of a continuous function

F (x, y) such that
∂F ∂F
dF (x, y) = dx + dy = M (x, y)dx + N (x, y)dy (2.11)
∂x ∂y
in a rectangle R, where
∂F ∂F
M (x, y) = , N (x, y) = . (2.12)
∂x ∂y
Equation (2.11) is an exact equation in R if
dF (x, y) = M (x, y)dx + N (x, y)dy = 0. (2.13)
Remark. Equation (2.13) shows that the solution to the exact equation is
the level curves
F (x, y) = C
where C is an arbitrary constant. The following theorem gives a simple test
to determine if a given differential form, M (x, y)dx + N (x, y)dy, is exact.
Theorem 1 (Test for Exactness). Suppose the first partial deriva-

tives of M (x, y) and N (x, y) are continuous in a rectangle R. Then the
M (x, y)dx + N (x, y)dy = 0
is an exact equation in R if and only if the compatibility condition

∂M ∂N
(x, y) = (x, y) (2.14)
∂y ∂x
holds for all (x, y) in R.
Example 11. Determine whether the following equation is exact. If it is,

then solve it.
(2xy + 3)dx + (x2 − 1)dy = 0.

Solution. Let
M = (2xy + 3), N = (x2 − 1)
∂M ∂N
=⇒ = 2x, and = 2x
∂y ∂x
∂M ∂N
=⇒ = 2x =
∂y ∂x
Therefore, the equation is exact. Now
∂F
M= = 2xy + 3
∂x
Z
=⇒ F = (2xy + 3)dx = x2 y + 3x + g(y)
where g(y) is any arbitrary function of y. The “constant” of integration in

this case is a function of y, g(y), because F (x, y) is a function of both x and
y and we integrated with respect to x. We need to solve for g(y) by using
the fact that ∂F/∂y = N :
∂F
= x2 + g 0 (y) = x2 − 1,
∂y
where we first differentiated the previous equation and equated the result to
N . Thus
g 0 (y) = −1, =⇒ g(y) = −y
Thus, we get
F = x2 y + 3x − y.
So the implicit solution, F (x, y) = C, to the equation is given by
x2 y + 3x − y = C
where C is an arbitrary constant. In this case, we may solve explicitly for y

to get
C − 3x
y= 2 .
x −1
Remark. (1) In most cases it is not necessary to look for an explicit so-
lution. The implicit solution F (x, y) = C suffices.
(2) It is not necessary to include a constant of integration after solving, for

example, g 0 (y) = −1. This is because any additional constant would
finally be combined with C to get another constant.

(3) One good check of your solution procedure, in the present case, is to
ensure that the resulting expression for g(y) is a function of only y.
If not, then there is something wrong with your algebra or solution
approach.
(4) An alternative approach to the solution is to start from
∂F
= x2 − 1,
∂y
such that Z
F = (x2 − 1)dy = x2 y − y + h(x),
where the “constant” of integration in this case, h(x), must be a func-

tion of only x since F (x, y) is a function of x and y and we integrated
with respect to y. Now
∂F
= M =⇒ 2xy + h0 (x) = 2xy + 3
∂x
=⇒ h0 (x) = 3, =⇒ h(x) = 3x.
Note how the equation for h(x) contains only the variable x, telling us
that we are on the right track. Hence, we get
F (x, y) = x2 y − y + 3x
which is the same as the expression for F obtained using the first
approach. So the solution becomes x2 y − y + 3x = C as expected.
Example 12. Solve the initial value problem


 yexy − y1 dx + xexy +
 x
y2
dy = 0

y(1) = 1.

Solution. We first check whether the differential form is exact. Let

xy 1 xy x
M = ye − , N = xe + 2 .
y y
So we have
∂M 1
= exy + xyexy + 2
∂y y

and
∂N 1
= exy + xyexy + 2 .
∂x y
Thus, the equation is exact since
∂M ∂N
= .
∂y ∂x
Let
∂F 1
M= = yexy −
∂x y
Z
1 x
=⇒ F (x, y) = yexy − dx = exy − + g(y).
y y
Now
∂F x x
N= =⇒ xexy + 2 + g 0 (y) = xexy + 2 ,
∂y y y
0
=⇒ g (y) = 0, =⇒ g(y) = C1 ,
where C1 is a constant. Thus, the solution F (x, y) = C2 becomes
x
exy − + C1 = C2
y
x
∴ exy − =C
y
where C = C2 − C1 is a constant. Using the initial condition y(1) = 1 (i.e.
x = 1, y = 1), we have
1
e(1)(1) − =C =⇒ C = e − 1.
1
Hence, the solution becomes
x
exy − = e − 1.
y
2.4 Substitutions and Transformations

So far we have studied separable, linear and exact equations. Some differ-
ential equations do not fall into any these categories. However, after ap-
plying some transformations and substitutions, it is possible to turn some
them into a form that we know how to solve using the previous methods.
We will look at some these transformations in this section. Specifically, we
will learn about homogeneous equations, then equations of the form
dy/dx = G(ax + by), and finally Bernoulli equations.

2.4.1 Homogeneous equations
Definition 6. Consider the differential equation

dy
= f (x, y). (2.15)
dx
If f (x, y) can be expressed as a ratio y/x alone, then equation (2.15) is
said to be a homogeneous equation.
Remark (Test). If
f (xt, yt) = f (x, y) for all t 6= 0
then the equation is homogeneous.
Solution approach
From (2.15) we have
dy y
=G ,
dx x
where G(y/x) is the transformed version of f (x, y). Let
y dy
v= =⇒ = G(v).
x dx
We can differentiate y = vx using the product rule to get
dy dv
=v+x = G(v),
dx dx
which is now a separable equation. To see this, we re-write the equation as
dv
x = G(v) − v
dx
dv dx
=⇒ = .
G(v) − v x
We then integrate both sides of the equation:
Z Z
dv dx
= ,
G(v) − v x
to get v. Once v is obtained, we can use the transformation v = y/x to get

an implicit or explicit solution in terms of x and y.

Example 13. Determine whether the differential equation is homogeneous.

If it is, then solve it using the method for homogeneous first order differential
equations.
dy y2
= .
dx xy + y 2
Solution. Let
y2
f (x, y) =
xy + y 2
Then
(ty)2 t2 y 2 y2
f (xt, ty) = = = = f (x, y).
(tx)(ty) + (ty)2 t2 xy + t2 y 2 xy + y 2
So the equation is homogeneous. We can express the right hand side as ratio
y/x alone, by dividing both the numerator and the denominator by x2 to get
dy (y/x)2
=
dx (y/x) + (y/x)2
Let v = y/x, then

dy v2
= .
dx v + v2
Differentiating y = vx using the product rule results in
dy dv v2
=v+x =
dx dx v + v2
dv v
=⇒ v + x =
dx 1+v
dv v −v 2
=⇒ x = −v =
dx 1+v 1+v
Z Z
1+v 1
=⇒ dv = − dx
v2 x
Z
1 1
=⇒ + dv = − ln |x| + C
v2 v
1
=⇒ − + ln |v| = − ln |x| + C
v
Substitute v = y/x into the equation above to get
x y
− + ln = − ln |x| + C
y x

x
=⇒ − + ln |y| − ln |x| = − ln |x| + C
y
The term − ln |x| cancels out, so we get
x
− + ln |y| = C
y
=⇒ −x + y ln |y| = Cy
∴ y ln |y| = Cy + x.
(y 2 − xy)dx + x2 dy = 0.
Solution. It is straightforward to check that the equation is not exact. Now

re-writing the equation, we get
dy xy − y 2 y y 2
= = − ,
dx x2 x x
so we see that the equation is homogeneous. Let v = y/x
dy
=⇒ = v − v2.
dx
Differentiating y = vx using the product rule and equating the result to the
equation above, we get
dv
v+x = v − v2
dx
dv
=⇒ x = −v 2
dx
Z Z
1 1
=⇒ − 2
dv = dx
v x
1 1
=⇒ = ln |x| + C =⇒ v =
v ln |x| + C
But y = vx, so we get
x
y= .
ln |x| + C

2.4.2 Equations of the form dy/dx = G(ax + by)

Solution Approach
Consider a differential equation of the form
dy
= G(ax + by). (2.16)
dx
Make the substitution
z = ax + by (2.17)
dz dy dy 1 dz a
=⇒ =a+b =⇒ = − (2.18)
dx dx dx b dx b
So from (2.16) we get
1 dz a
− = G(z)
b dx b
dz dz
=⇒ − a = bG(z) =⇒ = a + bG(z),
dx dx
which is now a separable equation. So separating variables and integrating,
we get
Z Z
dz
= dx.
a + bG(z)
After solving for z, we then plug it into (2.17) for the solution.
dy
= (x + y + 2)2
dx
Solution. Let
z =x+y+2
dz dy dy dz
=⇒ =1+ =⇒ = − 1.
dx dx dx dx
Substituting into the differential equation yields
dz dz
− 1 = z 2 =⇒ = 1 + z2
dx dx
Z Z
1
dz = dx
1 + z2
=⇒ tan−1 z = x + C,
=⇒ z = tan (x + C).
=⇒ x + y + 2 = tan (x + C)
∴ y = tan (x + C) − x − 2.

2.4.3 Bernoulli equation
Definition 7. A first-order differential equation that can be written in

the form
dy
+ P (x)y = Q(x)y n , (2.19)
dx
where P (x) and Q(x) are continuous on an interval (a, b) and n is a real
number is called a Bernoulli equation.
Solution Approach
Note that when n = 0 or 1, the equation is either linear or separable and can
easily be solved. For other values of n, first divide through by y n to get the
equation
dy
y −n + P (x)y 1−n = Q(x),
dx
and let
v = y 1−n .
dv dy
=⇒ = (1 − n)y −n
dx dx
1 dv dy
=⇒ = y −n , n 6= 1
1 − n dx dx
Substituting into the modified differential equation, we get
1 dv
+ P (x)v = Q(x)
1 − n dx
dv
=⇒ + (1 − n)P (x)v = (1 − n)Q(x)
dx
Note that the above equation is now a linear equation in terms of v since
(1 − n) is just a real number. To see this, you can let P1 (x) = (1 − n)P (x)
and Q1 (x) = (1 − n)Q(x) and get the linear first order equation:
dv
+ P1 (x)v = Q1 (x).
dx
Example 16. Solve the following equation
dy y
+ = x2 y 2
dx x

Solution.
dy y
+ = x2 y 2
dx x
The y 2 on the right-hand side tells us that we are dealing with a Bernoulli
equation. Dividing through by y 2 , we get
dy 1 −1
y −2 + y = x2
dx x
Let
v = y −1
dv dy
=⇒ = −y −2
dx dx
dv 1 dv 1
=⇒ − + v = x2 =⇒ − v = −x2 .
dx x dx x
Note that the final equation is a linear equation with integrating factor
R 1
µ = e− 1/x
= e− ln |x| = eln |1/x| = .
x
Multiplying through by the integrating factor and re-arranging, we get

d 1
=⇒ v = −x
dx x
Z
1 1
=⇒ v = − xdx = − x2 + C.
x 2
1
=⇒ v = − x3 + Cx
2
Since v = y −1 , we get
1 2
y= 1 3 = 3 , C1 = 2C.
2
x + Cx x + C1 x

Chapter 3
Linear Second Order

Differential Equations
Second order differential equations arise in many applications such as the

vibrations of mass-spring oscillators. This chapter presents the analytical
techniques for solving such equations.
3.1 Homogeneneous Linear Equations

Consider the equation
ay 00 (t) + by 0 (t) + cy(t) = f (t). (3.1)
If f (t) ≡ 0, the resulting equation is called a homogeneous ordinary differ-
ential equation such that
ay 00 (t) + by 0 (t) + cy(t) = 0. (3.2)
The function f (t) is called the “ nonhomogeneity” of the general equation.
The following approach is used to obtain a general solution to (3.2):
a) Assume a solution of the form y = ert and substitute into (3.2) to get
the characteristic or auxilliary equation:
ar2 + br + c = 0. (3.3)
b) Find the roots of (3.3) to get

√
−b +b2 − 4ac
r1 = ,
√2a
−b − b2 − 4ac
r2 = .
2a
25
(i) If b2 − 4ac > 0, the roots r1 and r2 are real and distinct. Then the
solutions are y1 = er1 t and y2 = er2 t . Combining the solutions gives the
general solution
y = C1 er1 t + C2 er2 t
where C1 and C2 are unknown constants, often determined from given
conditions on the equation.
(ii) If b2 − 4ac = 0, then r1 and r2 are real repeated roots, and the
solutions are given by y1 = er1 t and y2 = ter1 t . The general solution is
then given by
y = C1 er1 t + C2 ter1 t
(iii) If b2 − 4ac < 0, then r1 and r2 are complex conjugate roots such
that r1 = α + iβ and r2 = α − iβ. The solutions are y1 = eαt cos(βt),
and y2 = eαt sin(βt), and the general solution is given by
y = C1 eαt cos(βt) + C2 eαt sin(βt).
Example 17. Find the general soltution to the following equations

a) y 00 − y 0 − 2y = 0
b) y 00 + 8y 0 + 16y = 0
c) y 00 − 6y 0 + 10y = 0
Solution. (a)
y 00 − y 0 − 2y = 0
Substitute y = ert into the equation to get
r2 − r − 2 = 0
=⇒ (r − 2)(r + 1) = 0
=⇒ r = 2, −1
So we get
y = C1 e2t + C2 e−t .
(b)
y 00 + 8y 0 + 16y = 0
=⇒ r2 + 8r + 16 = 0, =⇒ (r + 4)2 = 0,
=⇒ r = −4, −4,
are repeated roots.
∴ y = C1 e−4t + C2 te−4t .

(c)
y 00 − 6y 0 + 10y = 0
=⇒ r2 − 6r + 10 = 0,
√
6 ± 36 − 40 6 ± 2i
=⇒ r = = =3±i
2 2
=⇒ α = 3, β = 1.
Thus, the general solution is given by
y = C1 e3t cos(t) + C2 e3t sin(t).
Example 18. Solve the initial value problem:
y 00 + 2y 0 − 8y = 0,
y(0) = 3, y 0 (0) = −12.
Solution.
y 00 + 2y 0 − 8y = 0,
=⇒ r2 + 2r − 8 = 0, =⇒ (r + 4)(r − 2) = 0
=⇒ r = −4, 2
=⇒ y = C1 e−4t + C2 e2t
Applying the initial conditions:
y(0) = 3 =⇒ C1 + C2 = 3
y 0 (0) = −12 =⇒ −4C1 + 2C2 = −12

=⇒ 2C1 − C2 = 6
Solving the two equations C1 + C2 = 3 and 2C1 − C2 = 6 simultaneously, we
get
C1 = 3, and C2 = 0.
∴ y = 3e−4t .
3.2 Non-homogeneous Equations

If f (t) 6= 0 in equation (3.1), the equation is said to be non-homogeneous.
There are two main analytical methods for solving non-homogeneous equa-
tions: The method of undetermined coefficients and the method of variation
of parameters.

3.2.1 The Method of undetermined coefficients

By this method, we first determine the solution to the homogeneous form
of the equation (say yh ) as discussed in the previous section. Depending on
the form of the non-homogeneous term f (t), there are various techniques of
obtaining the particular solution (say yp ) to the nonhomogeneous part of
the equation as explained below. Since the differential equation is linear, the
general solution, y, is obtained by summing the homogeneous and particular
solutions to get
y = yh + yp ,
by employing the so called superposition principle.
Solution Approach
(a) To find a particular solution to the differential equation
ay 00 + by 0 + cy = ktm ert , (3.4)
assume the following form for the particular solution:
yp (t) = ts (Am tm + · · · + A1 t + A0 )ert (3.5)
with
(i) s = 0 if r is not a root of the characteristic equation

(ii) s = 1 if r is a root of the characteristic equation
(iii) s = 2 if r is a double root of the characteristic equation
(b) To find a particular solution to the differential equation of the form
ay 00 + by 0 + cy = ktm eαt cos(βt)
or
ay 00 + by 0 + cy = ktm eαt sin(βt),
assume the following form for the particular solution:
yp (t) = ts (Am tm + · · · + A1 t + A0 )eαt cos(βt)+

ts (Bm tm + · · · + B1 t + B0 )eαt sin(βt) (3.6)
with
(i) s = 0 if α + iβ is not a root of the characteristic equation.

(ii) s = 1 if α + iβ is a root of the characteristic equation.

In the above, the A0 s and B 0 s are constants.
Example 19. Find the particular solution of the following equations
(a) y 00 − y = −11t + 1
(b) 2y 00 + y = 9e2t
(c) y 00 − y = t sin(t)
Solution. (a)
y 00 − y = −11t + 1
Comparing this equation to the form (3.4), we find that r = 0 and
m = 1. Now the characteristic equation is given by
r2 − 1 = 0,
and the roots of the characteristic equation are given by
r = 1, −1.
So the homogeneous solution is given by
yh = C1 et + C2 e−t .
In this case s = 0 in (3.5), since r = 0 is not a root of the characteristic
equation (which we found to be 1 and −1). So assume the following
form for the particular solution:
yp = At + B.
This implies that we expect yp to satisfy the differential equation since
it’s a solution. Now
yp0 = A and yp00 = 0.
Substituting into the original equation results in
−yp = −11t + 1
=⇒ −At − B = −11t + 1
Comparing coefficients we get
A = 11, B = −1.
So the particular solution is given by
yp = 11t − 1.

Remark. (1) You may check to see if your answer is correct by sub-
stituting yp into the differential equation.
(2) In this case, we are asked to find only the particular solution.
But the general solution of the equation, if desired, is given by
y = yh + yp :
y = C1 et + C2 e−t + 11t − 1.
(b)
y 00 − y = t sin(t)
From the general form (3.6), we see that α = 0 and β = 1 so that
α ± iβ = ±i. Substituting y = ert into the equation results in
=⇒ r2 − 1 = 0 =⇒ r = ±1.
Thus, s = 0 in (3.6) and m = 1, so the form of the particular solution

is yp = (A1 t + A0 ) cos(t) + (B1 t + B0 ) sin(t). For simplicity, we avoid
using constants with subscripts and instead use the following form for
the particular solution:
yp = (At + B) cos(t) + (Ct + D) sin(t).
=⇒ yp0 = (A + D + Ct) cos(t) + (−At − B + C) sin(t)

=⇒ yp00 = (−At − B + 2C) cos(t) + (−2A − D − Ct) sin(t)
Substitute into the original equation to get
(−2At − 2B + 2C) cos(t) + (−2A − 2D − 2Ct) sin(t) = t sin(t)
Comparing coefficients yields:

1
−2C = 1 =⇒ C = −
2
−2A = 0 =⇒ A = 0
−2A − 2D = 0 =⇒ D = 0
1
−2B + 2C = 0 =⇒ B = C = −
2
Thus,
1 1
yp = − cos(t) − t sin(t).
2 2

(c)
2y 00 + y = 9e2t
i
=⇒ 2r2 + 1 = 0 =⇒ r = ± √ .
2
Let
yp = Ae2t
=⇒ yp0 = 2Ae2t
=⇒ yp00 = 4Ae2t
Substitute into original equation to get
8Ae2t + Ae2t = 9e2t
=⇒ 9A = 9 =⇒ A = 1.
Thus
yp = e2t .
Example 20. Find the form of the particular solution to
y 00 + 2y 0 − 3y = f (t)
where f (t) is
(a) 7 cos(3t) (c) t2 cos(πt) (e) 3tet
(b) 2tet sin(t) (d) 5e−3t (f) t2 et
Solution.
y 00 + 2y 0 − 3y = 0
=⇒ r2 + 2r − 3 = 0
(r + 3)(r − 1) = 0
=⇒ r = 1, −3.
(a) f (t) = 7 cos(3t)

yp = A cos(3t) + B sin(3t)
(b) f (t) = 2tet sin(t). The form of the particular solution is
yp = (At + B)et cos(t) + (Ct + D)et sin(t).

Remark. Here s = 0 in (3.6) because 1 + i is not a root of the char-

acteristic equation. The factors (At + B) and (Ct + D) are needed
because m = 1 as a result of the linear term 2t in f (t).
(c) f (t) = t2 cos(πt)
yp = (At2 + Bt + C) cos(πt) + (Dt2 + Et + F ) sin(πt).
Remark. Here s = 0 in (3.6) because 0 + iπ (or iπ) is not a root of the

characteristic equation. The factors (At2 + Bt + C) and (Dt2 + Et + F )
are needed because m = 2 as a result of the quadratic term t2 in f (t).
(d) f (t) = 5e−3t

Note that r = −3 is a root of the auxilliary equation, so
yp = Ate−3t .
Remark. In this case, s = 1 in (3.5) because −3 is a root of the

characteristic equation. The constant A is needed because m = 0 as a
result of the constant 5 in f (t).
(e) f (t) = 3tet

Note that r = 1 is a root of the auxilliary equation so
yp = t(At + B)et .
Remark. In this case, s = 1 in (3.5) because 1 is a root of the char-

acteristic equation. The factor (At + B) is needed because m = 1 as a
result of the linear term 3t in f (t).
(f) f (t) = t2 et
yp = t(At2 + Bt + C)et .
Remark. As in the previous case, s = 1 in (3.5) because 1 is a root

of the characteristic equation. The factor (At2 + Bt + C) is needed
because m = 2 as a result of the quadratic term t2 in f (t).
Example 21. Find the form of the particular solution to
y 00 − 2y 0 + y = f (t)
where f (t) is the same as those in example (20):

(a) 7 cos(3t) (c) t2 cos(πt) (e) 3tet
(b) 2tet sin(t) (d) 5e−3t (f) t2 et
3.2.2 The Method of variation of parameters

In the previous section we developed the method of undetermined coefficients
for finding particular solutions to the nonhomogeneous equations
ay 00 (t) + by 0 (t) + cy(t) = f (t), (3.7)
where f (t) must be of a particular functional form: polynomial, exponential,

sine, cosine or a combination of these. For instance, the method could not
be used to find a particular solution if f (t) = tan(t). Here, we present the
method of variation of parameters which is more general than undetermined
coefficients. To proceed, we revisit some of our previous elborations by first
solving the corresponding homogeneous equation
ay 00 (t) + by 0 (t) + cy(t) = 0, (3.8)
for the complementary solutions y1 (t) and y2 (t) to get the general solution
yh = c1 y1 + c2 y2 , (3.9)
where c1 and c2 are constants. To obtain a particular solution yp to (3.7)

by the method of variation of parameters, we assume that c1 and c2 are now
functions of the independent variable t such that c1 = w1 (t) and c2 = w2 (t)
and (3.9) becomes
yp = w1 (t)y1 + w2 (t)y2 . (3.10)
The idea by Lagrange, who invented this method, is to determine the funci-
tons w1 (t) and w2 (t) such that (3.10) is a particular solution to (3.7). Since
there are two functions to determine, we expect to have two equations (or
conditions) to solve. In other for (3.10) to satisfy (3.7), we need to find the
derivatives of yp . So from (3.10), we have
yp0 = w10 y1 + w1 y10 + w20 y2 + w2 y20
=⇒ yp0 = (w10 y1 + w20 y2 ) + (w1 y10 + w2 y20 ).

To simplify the algebra after taking a second derivative of yp , we assume that
w10 y1 + w20 y2 = 0, (3.11)

which is actually one of the equations we would need to eventually solve for
w1 and w2 . This assumption implies that
yp0 = w1 y10 + w2 y20 , (3.12)

yp00 = w10 y10 + w1 y100 + w20 y20 + w2 y200 . (3.13)
Substituting (3.10), (3.12) and (3.13) into (3.7) yields
ayp00 (t) + byp0 (t) + cyp (t) = f (t)
=⇒ a(w1 y100 + w10 y10 + w2 y200 + w20 y20 ) + b(w1 y10 + w2 y20 ) + c(w1 y1 + w2 y2 ) = f (t),
=⇒ w1 [ay100 + by10 + cy1 ] + w2 [ay200 + by20 + cy2 ] + a(w10 y10 + w20 y20 ) = f (t).
Since y1 and y2 are solutions to the homogeneous equation (3.8), the expres-
sions in square brackets vanish and so
a(w10 y10 + w20 y20 ) = f (t)

f (t)
w10 y10 + w20 y20 = . (3.14)
a
Thus, equations (3.11) and (3.14) provides the two equations needed to be
solved simultaneously for w1 and w2 . We rewrite them below for convenience:
w10 y1 + w20 y2 = 0, (3.15)

f (t)
w10 y10 + w20 y20 = . (3.16)
a
From (3.15) we have
y2 0
w10 = − w, (3.17)
y1 2
and from (3.16) we get

y2 0 f (t)
− w2 y10 + w20 y20 =
y1 a
0

y y2 f (t)
=⇒ w20 y20 − 1 =
y1 a
f (t)y1
=⇒ w20 (y1 y20 − y10 y2 ) =
a
f (t)y1
=⇒ w20 = . (3.18)
a[y1 y20 − y10 y2 ]

From (3.17) we get

y2 f (t)y1
w10 =− · ,
y1a[y1 y20 − y10 y2 ]
f (t)y2
=⇒ w10 = − . (3.19)
a[y1 y20 − y10 y2 ]
It is important to note that the expression in square bracket, W (y1 , y2 ) =

y1 y20 −y10 y2 6= 0, since it is the Wronskian of the linearly independent solutions
y1 and y2 . Integrating (3.18) and (3.19) results in
Z Z
f (t)y2 f (t)y1
w1 = − dt, and w 2 = dt. (3.20)
a[y1 y20 − y10 y2 ] a[y1 y20 − y10 y2 ]
Hence, the particular solution is given by (3.10):
yp = w1 (t)y1 + w2 (t)y2
where the variables w1 (t) and w2 (t) are given by (3.20).
Summary: Method of Variation of Parameters

In other to determine a particular solution yp to the second order differential
equation with constant coefficients:
ay 00 (t) + by 0 (t) + cy(t) = f (t),
execute the following steps:
(1) Determine the two linearly independent solutions y1 and y2 to the cor-
responding homogeneous equation ay 00 (t) + by 0 (t) + cy(t) = 0, and let
yp = w1 (t)y1 (t) + w2 (t)y2 (t).
(2) Determine w1 (t) and w2 (t) by simultaneously solving equations (3.11)

and (3.14):
w10 y1 + w20 y2 = 0

(3.21)
w10 y10 + w20 y20 = f (t)
a
to get w1 (t) and w2 (t) as in (3.20).

(3) Finally, substitute w1 (t) and w2 (t) into yp = w1 (t)y1 (t) + w2 (t)y2 (t) to
get a particular solution.
Remark. It is not advisable to memorize the formulas for w1 and w2 in

(3.20) since they can easily be derived from system (3.21).
Example 22. Solve the following equation
y 00 + y = sec(t)
Solution. Note that the method of undetermined coefficients cannot be used

here because of the nonhomogeneity f (t) = sec(t). So we employ the method
of variation of parameters. Considering the corresponding homogeneous
equation y 00 + y = 0, we find, after substituting y = ert , the characteris-
tic equation
r2 + 1 = 0, =⇒ r2 = −1, r = ±i.
So the solution to the homogeneous equation is
yh = c1 cos t + c2 sin t.
Let the particular solution be
yp = w1 (t) cos t + w2 (t) sin t
and solve the system:


 w10 cos t + w20 sin t = 0,
w10 (− sin t) + w20 (cos t) = sec t

.

1
Multiplying the first equation by sin t and the second by cos t and summing
them gives
w20 (sin2 t + cos2 t) = cos t · sec t = 1,
=⇒ w20 = 1, =⇒ w2 (t) = t,
since sin2 t + cos2 t = 1 and sec t = 1/ cos t. From the first equation in the
system, we get
sin t 0
w10 = − w = − tan t,
cos t 2
Z Z
sin t
w1 = − tan tdt = − dt = ln | cos t|.
cos t

In the integrations above, we let the constants of integration be zero because

we’re looking for just one particular solution. Substituting w1 and w2 into
the form of the particular solution, we get
yp = cos t ln | cos t| + t sin t.
The general solution becomes y = yh + yp :
y = c1 cos t + c2 sin t + cos t ln | cos t| + t sin t.
3.3 Equations with Variable Coefficients

Second order differential equations with variable coefficients are of the form
a(t)y 00 (t) + b(t)y 0 (t) + c(t)y(t) = g(t), (3.22)
where the coefficients are now functions of the independent variable t. Un-
like equations with constant coefficients which are amenable to the method
of undetermined coefficients or variation of paramaters, there are no straigh-
forward methods for constructing explicit solutions in this case. However,
there are some interesting aspects of variable-coefficient equations that we
would like to study.
There is a theorem for the existence and uniqueness of solutions for
variable-coefficient equations which we will state. We will also take a look
at special equations with variable coefficients (known as Cauchy-Euler or
Equidimensional equations) that can be solved using an approach similar
to what was done for equations with constant coefficients. Finally, we will
state the theorem on the method of “Reduction of Order” which states that,
given one solution, say y1 , to the homogeneous variable coeffient equation
a(t)y 00 (t) + b(t)y 0 (t) + c(t)y(t) = 0, one can find a second linearly independent
solution, say y2 .
To proceed, we first write (3.22) in standard form by dividing through by
a(t) to get an equation of the form
y 00 (t) + p(t)y 0 (t) + q(t)y(t) = f (t). (3.23)
Of course, this supposes that a(t) 6= 0 otherwise the whole structure of the
equation changes from a second order to a first order equation. The following
theorem guarantees the existence and uniqueness of solutions to equations
with variable coefficients:

Theorem 2 (Existence and Uniqueness of Solutions). Let the gov-

erning functions p(t), q(t), and f (t) in the equation y 00 (t) + p(t)y 0 (t) +
q(t)y(t) = f (t) be continuous on an interval I containing the point t0 .
Then for any choice of the initial values Y0 and Y1 , there exists a unique
solution y(t) on the interval I to the initial value problem
y 00 (t) + p(t)y 0 (t) + q(t)y(t) = f (t); y(t0 ) = Y0 , y 0 (t0 ) = Y1 .
The theorem may be used to determine the largest interval for which a unique
solution to the differential equation exists, as demonstrated by the following
example.
Example 23. Consider the initial value problem
√
(t − 2)y 00 (t) + y 0 (t) + t ty(t) = ln t; y(1) = 2, y 0 (1) = −3.
Use Theorem 2 to determine the largest interval for which a unique solution
exists.
Solution. Writing the equation in standard form gives
√
00 1 0 t t ln t
y (t) + y (t) + y(t) = ,
t−2 t−2 t−2
such that √
1 t t ln t
p(t) = , q(t) = , f (t) = .
t−2 t−2 t−2
We see that the functions are all continuous on the intervals 0 < t < 2
and 2 < t < ∞. However, the interval 2 < t < ∞ does not contain the
initial point t0 = 1 so we neglect it. Thus, the Theorem guarantees a unique
solution in the interval 0 < t < 2.
Remark. Note from the example above that q(t) does not exist for t < 0
and does f (t) exist for t ≤ 0. Also note that the Theorem guarantees the
existence of a unique solution but does not provide the solution nor the
interval of existence of the solution.
Lemma 1 (Linear Dependence of Solutions). If y1 (t) and y2 (t) are any
two solutions to
y 00 (t) + p(t)y 0 (t) + q(t)y(t) = 0, (3.24)
on an interval I where p(t) and q(t) are continuous and if the Wronskian
y1 y2
W (y1 , y2 )(t) = = y1 y20 − y10 y2
y10 y20
is zero at any point t of I, then y1 and y2 are linearly depedent on I.

Remark. Because we are interested in linearly independent solutions to dif-

ferential equations, as expounded upon previously, Lemma 1 provides a sim-
pler approach of determining if two solutions to a variable-coefficient equation
are linearly dependent or independent. They’re linearly independent if the
Wronskian, W (y1 , y2 ) 6= 0.
3.3.1 Representation of Solutions to IVPs

As in equations with constant coefficients, the following Theorem guarantees
that two linearly independent solutions, y1 and y2 , to equation (3.24) with
given initial conditions can be combined to get a general solution.
Theorem 3. Suppose y1 (t) and y2 (t) are any two linearly independent
solutions to the IVP
y 00 (t) + p(t)y 0 (t) + q(t)y(t) = 0, y(t0 ) = Y0 , y 0 (t0 ) = Y1 , (3.25)
on an interval I containing the point t0 , then unique constants c1 and c2

can always be found such that the general solution y = c1 y1 (t) + c2 y2 (t)
satisfies the initial conditions for any Y0 and Y1 .
As stated in our introduction to this section, the biggest challenge with

variable-coefficient equations is the lack of a general method for explicitly
obtaining the solutions y1 (t) and y( t) to (3.25); apart from some special
cases. However, once these solutions are found, Theorem 3 assures us of a
general solution that will always satisfy the initial conditions. Secondly, once
these solutions have been obtained for the homogeneous equation (3.24), the
method of variation of parameters (for variable-coefficient equations) can be
used to solve the associated nonhomogeneous equation for a particular solu-
tions as stated below.
Theorem 4 (Variation of Parameters). Suppose y1 and y2 are two lin-
early independent solutions to (3.24) on an interval I, then a particular
solution to
y 00 (t) + p(t)y 0 (t) + q(t)y(t) = f (t), (3.26)
is given by
yp = w1 (t)y1 (t) + w2 (t)y2 (t)
where w1 and w2 are determined from the equations
w10 y1 + w20 y2 = 0,

(3.27)
w10 y10 + w20 y20 = f (t),

with solution
Z Z
f (t)y2 f (t)y1
w1 = − dt, and w2 = dt. (3.28)
W (y1 , y2 ) W (y1 , y2 )
Remark. As adviced before, equations (3.28) can easily be derived from

system (3.27) so it’s not advisable to memorize them. Also keep in mind
that, as opposed to equations with constant coefficients, system (3.27) is
obtained from the normalized equation (3.26).
Let’s do a few examples of finding the general solution, y = yh + yp , using

the method of variation of parameters for which the linearly independent
solutions y1 and y2 are given. After which we will tackle the special case of
Cauchy-Euler equations for which y1 and y2 can be found explicitly.
Example 24. Find a general solution to the following IVPs using variation of
parameters, given that y1 and y2 are solutions to the associated homogeneous
equation.
(a) ty 00 − (t + 1)y 0 + y = t2 , (b) ty 00 + (5t − 1)y 0 − 5y = t2 e−5t ,
y1 = e2 , y2 = t + 1. y1 = 5t − 1, y2 = e−5t .
Solution. (a)
ty 00 − (t + 1)y 0 + y = t2 , y1 = e2 , y2 = t + 1.
It is easy to check that y1 and y2 are indeed solutions to the homogeneous
equation ty 00 − (t + 1)y 0 + y = 0:
ty100 − (t + 1)y10 + y1 = tet − (t + 1)et + et

= tet − tet − et + et
= 0.
Also
ty200 − (t + 1)y20 + y2 = t(0) − (t + 1)(1) + (t + 1)

=0−t−1+t+1
= 0.
Writing the equation in standard form gives
(t + 1) 0 1
y 00 − y + y = t, t 6= 0.
t t
Thus, f (t) = t and
et t + 1
W (y1 , y2 )(t) = = et − tet − et = −tet .
et 1
Thus, from (3.28), we have

Z Z Z
f (t)y2 t(t + 1)
w1 = − dt = dt = (t + 1)e−t dt.
W (y1 , y2 ) −tet
Integrating by parts, let
dv = e−t dt
u = t + 1,
Z
=⇒ du = dt, and v = e−t dt = −e−t ,
Z
=⇒ w1 (t) = −(t + 1)e + e−t dt
−t
=⇒ w1 (t) = −te−t − e−t − e−t = −te−t − 2e−t = −e−t (t + 2).

=⇒ w1 (t) = −e−t (t + 2).
Also, we have w2 (t):
tet
Z Z
f (t)y1
w2 (t) = dt = dt = −t.
W (y1 , y2 ) −tet
Again, it’s redundant to keep the integration constants after integrating for
w1 and w2 . Finally, we get the particular solution:
yp = w1 (t)y1 (t) + w2 (t)y2 (t)

= −e−t (t + 2) · et − t(t + 1)
= −t − 2 − t2 − t = −t2 − 2t − 2.
Hence, the general solution, y = yh + yp , is given by
y = c1 et + c2 (t + 2) − (t2 + 2t + 2) .
So far, we have seen that given two linearly independent solutions y1 and
y2 to the homogeneous equation of a variable-coefficient equation, we can
use the method of variation of parameters to determine a particular solution
and hence the general solution. This is because we have no technique of
determining y1 and y2 in the general case of a variable-coefficient equation.
However, given one of the solutions, say y1 or y2 , the method of reduction of
order can be used to find the other solution. We describe this next.

3.3.2 Reduction of Order

The method of reduction of order is used to construct a second, linearly
independent solution, say y2 , from a known solution, say y1 . The following
Theorem formalizes this procedure.
Theorem 5. Suppose y1 (t) is a non-zero solution to the homogeneous

y 00 (t) + p(t)y 0 (t) + q(t)y(t) = 0, (3.29)
in an interval I. Then
R
e− p(t)dt
Z
y2 (t) = y1 (t) dt, (3.30)
[y1 (t)]2
is a second, linearly independent solution.
Proof. Since yh = cy1 (t) is a solution, we assume a second solution to be
y2 = w(t)y1 (t)
=⇒ y20 = w0 y1 + wy10
=⇒ y200 = w00 y1 + w0 y10 + w0 y10 + wy100
Substituting into (3.29) gives
(w00 y1 + w0 y10 + w0 y10 + wy100 ) + p(t)(w0 y1 + wy10 ) + q(t)wy1 = 0,
(wy100 + pwy10 + qwy1 ) + w00 y1 + (2w0 y10 + pw0 y1 ) = 0,

(y100 + py10 + qy1 )w + w00 y1 + (2y10 + py1 )w0 = 0.
The expression in the first bracket vanishes since y1 is a solution to (3.29).
Thus
w00 y1 + (2y10 + py1 )w0 = 0. (3.31)
Equation (3.31) can be reduced to a first order equation by letting
v = w0 .
=⇒ v 0 y1 + (2y10 + py1 )v = 0,
(2y10 + py1 )
Z Z
dv
=⇒ =− dt,
v y1

2y10
Z
=⇒ ln |v| = − + p dt,
y1
Z
=⇒ ln |v| = −2 ln |y1 | − p(t)dt,
Z
−2
=⇒ ln |v| = ln[y1 (t)] − p(t)dt,
R
R e− p(t)dt
=⇒ v = [y1 (t)]−2 · e− p(t)dt
=
[y1 (t)]2
But v = w0 , so we have
R
e− p(t)dt
Z
w(t) = dt.
[y1 (t)]2
Now y2 (t) = y1 (t)w(t),
R
e− p(t)dt
Z
∴ y2 (t) = y1 (t) dt.
[y1 (t)]2
Example 25. Given a non-trivial solution y1 of the following differential

equation, use reduction of order to find a second linearly independent solu-
tion.
t2 y 00 − 2ty 0 − 4y = 0, y1 = t−1 , t > 0.
2 4
y 00 − y 0 − 2 y = 0,
t t
−1 −2
so we have p(t) = −2t , q(t) = −4t .
Z − R p(t)dt Z − R (−2t−1 )dt
e e
y2 (t) = y1 (t) dt = t−1 dt,
[y1 (t)]2 t−2
Z Z
−1 2 2 ln t −1
=t t e dt = t t4 dt,
t5 1
= t−1 ·
= t4 .
5 5
Thus, the two linearly independent solutions to the equation are y1 = t−1
and y2 = t4 /5.
We next take a look at special equations called Cauchy-Euler or Euler or
Equidimensional equations for which it is possible for to determine both
linearly independent solutions y1 and y2 . In order words, they are special
variable-coefficient equations that can be solved explicitly.

3.3.3 Cauchy-Euler (or Equidimensional) Equations
Definition 8. A Cauchy-Euler (or Equidimensional) equation is an equa-

tion of the form
at2 y 00 (t) + bty 0 (t) + cy(t) = f (t), (3.32)
where a(6= 0), b and c are constants.
Remark. Writing the Equidimensional equation in standard form yields

bt 0 c f (t)
y 00 (t) + 2
y (t) + 2 y(t) = 2 ,
at at at
b 0 c f (t)
=⇒ y 00 (t) + y (t) + 2 y(t) = 2 .
at at at
We see that the governing functions are continuous for t 6= 0. Thus, Theorem
2 guarantees a unique solution for t < 0 or t > 0, subject to other restrictions
imposed by the function f (t).
In what follows, we first construct a general solution yh to the homogeneous
equation
at2 y 00 (t) + bty 0 (t) + cy(t) = 0. (3.33)
Later, using the method of variation of parameters, we would determine a

particular solution yp to the nonhomogeneous equation (3.32). Once yh and yp
have been obtained, we can easily construct the general solution y = yh + yp .
Solution Approach to the homogeneous case

Assume a solution of the form
y(t) = tr , t 6= 0, (3.34)
where r is a constant. Then
y 0 (t) = rtr−1 , (3.35)

=⇒ y 00 (t) = r(r − 1)tr−2 . (3.36)
Substituting (3.34)-(3.36) into (3.33) gives
at2 [r(r − 1)tr−2 ] + bt[rtr−1 ] + ctr = 0,

=⇒ ar(r − 1)tr + brtr + ctr = 0,

=⇒ [ar(r − 1) + br + c] tr = 0.
Since tr 6= 0, we get the associated characteristic equation:
ar(r − 1) + br + c = 0, (3.37)
such that
p
−(b − a) ± (b − a)2 − 4ac
r= , (3.38)
2a
with discriminant
p
D= (b − a)2 − 4ac. (3.39)
The solution to the characteristic equation (3.37) yields three different cases
depending on the discrimiant: real distinct roots, repeated roots, and com-
plex roots; as in the case for constant coefficient equations.
(a) For real distinct roots, r1 and r2 , the solution of the homogeneous
equation is given by
yh = c1 tr1 + c2 tr2 .
(b) For repeated roots, the solution of the homogeneous equation is
yh = c1 tr + c2 tr ln t, t > 0,
and this can be verified by directly substituting the solution into the
homogeneous equation.
(c) For a complex root, r = α + iβ (or r = α − iβ), we can write
tα+iβ = tα · tiβ = tα · eiβ ln t ,
since t = eln t . So we have
tα+iβ = tα [cos(β ln t) + i sin(β ln t)],
after employing Euler’s formula: eiθ = cos θ + i sin θ. Using the real
and imaginary parts gives the two independent solutions
y1 = tα cos(β ln t), y2 = tα sin(β ln t),
such that
yh = c1 tα cos (β ln t) + c2 tα sin (β ln t).

Example 26. Determine the general solution to the following equation
at2 y 00 + by 0 + cy = 0, t > 0,
where
(a) a = 1, b = 1, c = −1.
(b) a = 1, b = −1, c = 5.
(c) a = 1, b = 5, c = 4.
Solution. (a) t2 y 00 + y 0 − y = 0
p
−(b − a) ± (b − a)2 − 4ac
r=
2a
√
0± 4
=⇒ r = = ±1.
2
So we have two distinct roots, r1 = 1, r2 = −1 Therefore
yh = c1 t + c2 t−1 .
(b) t2 y 00 − y 0 + 5y = 0. We have the roots:

p √
2 ± 4 − 4(5) 2 ± −16 2 ± 4i
r= = = = 1 ± 2i.
2 2 2
These are complex roots, α ± iβ, with α = 1 and β = 2. Hence
yh = c1 t cos(2 ln t) + c2 t sin(2 ln t).
(c) t2 y 00 + 5y 0 + 4y = 0. We have the roots:

√
−4 ± 16 − 16
r= = −2.
2
So the roots are repeated, r = 2, 2, so we get
hh = c1 t−2 + c2 t−2 ln t.
Example 27. Solve the initial value problem

2
9t2 y 00 + 15ty 0 + y = 0; y(1) = 1, y 0 (1) = − .
3

Solution. The roots are given by

p p
−(b − a) ± (b − a)2 − 4ac −6 ± 36 − 4(9) −6 1
r= = = =−
2a 18 18 3
So we have repeated roots r = −1/3, and the general solution is given by
y = c1 t−1/3 + c2 t−1/3 ln t.
Now,
y(1) = 1, =⇒ c1 = 1.
1 1 1
y 0 = − c1 t−4/3 − c2 t−4/3 ln t + c2 t−1/3 ·
3 3 t
2 1 2
y 0 (1) = − , =⇒ − − c2 = −
3 3 3
2 1 1
=⇒ c2 = − = .
3 3 3
1
∴ y = t−1/3 + t−1/3 ln t .
3


Chapter 4
Series Solutions of Differential

Equations
This method provides a means of computing accurate approximate solutions

to differential equations near a point.
4.1 Introduction
In this section, we review a few concepts from Calculus needed to understand
the approach of series solutions.
Definition 9. The Taylor Polynomial of degree n about the point x0 ap-

proximating f (x) at x = x0 is given by
f 00 (x0 ) f 000 (x0 ) f n (x0 )

Pn (x) = f (x0 )+f 0 (x0 )(x−x0 )+ (x−x0 )2 + (x−x0 )3 +· · ·+ (x−x0 )n
2 3! n!
n
X f j (x0 )
=⇒ Pn (x) = (x − x0 )j .
j=0
j!
Motivation
Find the first few Taylor Polynomials approximating the solution around
x0 = 0:
y 00 = 3y 0 + x7/3 y

y(0) = 10, y 0 (0) = 5
49
Solution.
y 00 = 3y 0 + x7/3 y (4.1)
y(0) = 10, y 0 (0) = 5
y 00 (0) 2 y 000 (0) 3 y n (0) 2
Pn (x) = y(0) + y 0 (0)x + x + x + ··· + x. (4.2)
2! 3! n!
Now,
y 00 (0) = 3y 0 (0) + 07/3 y(0) = 3(5) + 0 = 15.
From (4.1), we have
7
y 000 = 3y 00 + x4/3 y + x7/3 y 0
3
7 4/3
=⇒ y 000 (0) = 3y 00 (0) + · 0 y(0) + 07/3 · y 0 (0) = 3(15) = 45
3
Similarly, we have
28 1/3 7 7
y (4) = 3y 000 + x y + x4/3 y 0 + x4/3 y 0 + x7/3 y 00
9 3 3
14 4/3 0 28 1/3
= 3y 000 + x7/3 y 00 + x y + x y
3 9
14 28 1/3
=⇒ y (4) (0) = 3y 000 (0) + 07/3 · y 00 (0) + · 04/3 y 0 (0) + · 0 y(0)
3 9
= 3y 000 (0) + 0 + 0 + 0 = 3(45) = 135.
Also,
7 56 14 28 28
y (5) = 3y (4) + x4/3 y 00 + x7/3 y 000 + x1/3 y 0 + x4/3 y 00 + x−2/3 y + x1/3 y 0
3 9 3 27 9
Note that at x = 0 the fifth derivative y (5) does not exist because of the term
(28/27)x−2/3 y. So the Taylor Polynomial
15 2 45 3 135 4
P4 (x) = 10 + 5x + x + x + x,
2 6 24
approximates the solution to the differential equation near x = 0.
Power Series

Definition 10. A Power Series about the point x0 is an expression of

the form
∞
X
an (x − x0 )n = a0 + a1 (x − x0 ) + a2 (x − x0 )2 + · · · (4.3)
n=0
where the an ’s are constants. Equation (4.3) converges at x = c if

N
X
lim an (c − x0 )n
N →∞
n=0
exists (as a finite number).
Theorem 6 (Radius of Convergence). For each power series of the

form (4.3), there is a number ρ (0 ≤ ρ ≤ ∞), called the radius of
convergence of the power series, such that (4.3) converges absolutely
for |x − x0 | < ρ and diverges for |x − x0 | > ρ. If the series converges for
all x, then ρ = ∞, and if it converges only at x = x0 then ρ = 0.
Theorem 7 (Ratio Test). Suppose the coefficients an are nonzero for

large values of n and
an 1
lim = L, (0 ≤ L ≤ ∞)
n→∞ an
then the radius of convergence of the power series
∞
X
an (x − x0 )n
n=0
1
is ρ = , with ρ = ∞ if L = 0 and ρ = 0 if L = ∞.
L
Example 28. Determine the interval of convergence of the power series

∞
X (−2)n
(x − 3)n
n=0
n+1

Solution.
(−2)n
an = ,
n+1
an+1 (−2)n+1 (n + 1
=⇒ lim = lim ·
n→∞ an n→∞ n+2 (−2)n
(−2)n (−2) n + 1
= lim ·
n→∞ (−2)n n+2
n+1 1 + 1/n
= 2 lim = 2 lim =2=L
n→∞ n+2 n→∞ 1 + 2/n
1
∴ρ=
2
Therefore the series converges for |x − 3| < 12 .
∞
X
Theorem 8. If an (x − x0 )n = 0 for all x in some open interval, then
n=0
each coefficient an equals zero.
∞
X
Remark. If f (x) = an (x − x0 )n , then within the radius of convergence,
n=0
we can differentiate and also integrate f (x) so that
∞
X
0
f (x) = nan (x − x0 )n−1 , |x − x0 | < ρ,
n=0
Z ∞
X an
f (x)dx = (x − x0 )n+1 + C, |x − x0 | < ρ.
n=0
n + 1
Analytic Function
Definition 11. A function f is said to be analytic at x0 if, in an open

∞
X
interval about x0 , this function is the sum of a power series an (x−x0 )n
n=0
that has a positive radius of convergence.

The following are examples of analytic functions with their corresponding

power series representations:
∞
x x 2 x3 X xn
e =1+x+ + + ··· =
2! 3! n=0
n!
∞
x3 x5 X (−1)n
sin x = x − + − ··· = x(2n+1)
3! 5! n=0
(2n + 1)!
∞
x2 x4 X (−1)n
cos x = 1 − + − ··· = x2n
2! 4! n=0
(2n)!
But ln x is analytic for x > 0:

∞
(x − 1)2 (x − 1)3 X (−1)(n−1)
ln x = (x − 1) − + − ··· = (x − 1)n .
2 3 n=1
n
4.2 Power Series Solutions to Linear Differ-

ential Equations
In this section, we try to find the power series solutions to linear differential
equations with polynomial coefficients. Consider the equation
a2 (x)y 00 + a1 (x)y 0 + a0 y = 0. (4.4)
Writing the equation in standard form gives
y 00 + p(x)y 0 + q(x)y = 0, (4.5)
where
a1 (x) a0 (x)
p(x) = , q(x) = .
a2 (x) a2 (x)
Definition 12 (Ordinary and Singular Points). A point x0 is called

an ordinary point of (4.4) if p = a1 /a2 and q = a0 /a2 are analytic at
x0 . If x0 is not an ordinary point, it is called a singular point of the
equation.
Example 29. Determine all the singular points of
xy 00 + x(1 − x)−1 y 0 + (sin x)y = 0

Solution.
x sin x
y 00 + y0 + y = 0.
x(1 − x) x
Let
x 1 sin x
p(x) = = , q(x) = .
x(1 − x) 1−x x
Thus, p(x) is analytic except at x = 1. The point x = 0 is called a removable
singularity of p(x) since we can cancel out x to get p = 1/(1 − x). Note
that q(x) is a ratio of analytic functions, so we write
x3 5
x− 3!
+ x5! − · · · x2 x4
q(x) = =1− + − ···
x 3! 4!
Therefore q(x) is analytic for all x. Thus, the only singular point of the
equation is x = 1.
Example 30. Find a power series solution about x = 0 to
y 0 + 2xy = 0. (4.6)
Solution.
Remark. Note that we can solve the differential equation by separation of

variables:
dy
= −2xdx, =⇒ ln |y| = −x2 + C1 , C1 = constant,
y
2
∴ y = Ce−x , C = constant,
For a power series solution, note that the coefficient of y, that is 2x, is
analytic everywhere, so x = 0 is an ordinary point. Thus, we expect a power
series solution of the form
∞
X
y(x) = an xn = a0 + a1 x + a2 x2 + · · · (4.7)
n=0
∞
X
0
=⇒ y (x) = nan xn−1 = a1 + 2a2 x + 3a3 x2 + · · · · · · (4.8)
n=1
Substituting (4.7)-(4.8) into (4.6) yields

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

∞
X ∞
X
n−1
=⇒ nan x + 2an xn+1 = 0.
n=1 n=0
We can let both series have the same power of x by setting k = n − 1 in the
first equation and k = n + 1 in the second to get
∞
X ∞
X
k
(k + 1)ak+1 x + 2ak−1 xk = 0.
k=0 k=1
Now, to be able to combine the two series, we let both of them start from
the same value of k such that
X∞ ∞
X
k
a1 + (k + 1)ak+1 x + 2ak−1 xk = 0,
k=1 k=1
∞
X
=⇒ a1 + [(k + 1)ak+1 + 2ak−1 ] xk = 0.
k=1
Now, each coefficient must vanish. Thus
a1 = 0
and we get the recurrence relation from the second term on the left-hand
side:
(k + 1)ak+1 + 2ak−1 , k ≥ 1
2
=⇒ ak+1 = − ak−1 , k ≥ 1
k+1
Substituting values of k into the recurrence relation gives the following:
2
k=1: a2 = − a0 = −a0
2
2
k=2: a3 = − a1 =0
3
2 1
k=3: a4 = − a2 = a0
4 2
2
k=4: a5 = − a3 =0
5
2 1 1 1
k=5: a6 = − a4 = − · a0 = − a0
6 3 2 3!
2
k=6: a7 = − a5 =0
7
2 1 1 1
k=7: a8 = − a6 = − · a0 = a0
8 4 3! 4!
2
k=8: a9 = − a7 =0
9

From the relation above, notice that

(−1)n
a2n = a0 , n = 1, 2, · · ·
n!
a2n+1 = 0, n = 0, 1, 2, · · ·
which are the even and odd terms of (4.7) respectively. Substituting them
into (4.7) gives
1
y(x) = a0 − a0 x2 + a0 x 4 + · · ·
2!
∞
X (−1)n 2n
∴ y(x) = a0 x , (4.9)
n=0
n!
where a0 is an arbitrary constant (as expected since the equation is first

order).
∞
x
X xn
Remark. (a) e =
n=0
n!
∞ ∞
2
X (−x2 )n X (−1)n x2n
=⇒ e−x = =
n=0
n! n=0
n!
This implies that

2
y(x) = a0 e−x
as found by the method of separation of variables.
(b) Applying the ratio test to (4.9) gives
an+1 (−1)n+1 n!
lim = lim ·
n→∞ an n→∞ (n + 1)! (−1)n
(−1)(−1)n n! (−1)
= lim · = lim =0=L
n→∞ (n + 1) · n! (−1)n n→∞ n + 1
1
∴ ρ = = ∞.
L
Thus (4.9) converges for all x.
(c) Let a0 = 1. Then (4.9) yields

1 1
y(x) = 1 − x2 + x4 − x6 + · · ·
2 6
Figure XX shows the actual solution to the differential equation to-

gether with various partial sums of the power series solution. We ob-
serve that by using more terms we get better approximations to the
solution around x = 0. However, partial sums diverge as x → ±∞
since they are polynomials, but the actual solution approaches zero as
x → ±∞.
Example 31. Find a power series solution of the differential equation about
x = 2:
y 00 − y = 0. (4.10)
Solution. We seek a solution of the form

∞
X
y(x) = an (x − 2)n (4.11)
n=0
∞
X ∞
X
0 n−1
=⇒ y (x) = nan (x − 2) = nan (x − 2)n−1
n=0 n=1
X∞ ∞
X
=⇒ y 00 (x) = n(n − 1)an (x − 2)n−2 = n(n − 1)an (x − 2)n−2 (4.12)
n=1 n=2
Substituting (4.11)-(4.12) into (4.10), we get

∞
X ∞
X
n(n − 1)an (x − 2)n−2 − an (x − 2)n = 0
n=2 n=0
Letting k = n − 2 in the first series and k = n in the second gives

∞
X ∞
X
k
(k + 2)(k + 1)ak+2 (x − 2) − ak (x − 2)k = 0
k=0 k=0
∞
X
=⇒ [(k + 2)(k + 1)ak+2 − ak ] (x − 2)k = 0
k=0
The coefficient of each power vanishes, resulting in
(k + 2)(k + 1)ak+2 − ak = 0
1
=⇒ ak+2 = ak , k ≥ 0.
(k + 2)(k + 1)

Substituting values of k into the recurrence relation above gives:

1
k=0: a2 = a0
2·1
1
k = 1 : a3 = a1
3·2·1
1 1
k = 2 : a4 = a2 = a0
4·3 4·3·2·1
1 1
k = 3 : a5 = a3 = a1
5·4 5·4·3·2·1
1 1
k = 4 : a6 = a4 = a0
6·5 6·5·4·3·2·1
.. .. .... ..
. = . .. = .
From the relations above, we can establish the following two recurrence re-
lations:
1
a2n = a0 , n = 1, 2, · · ·
(2n)!
1
a(2n+1) = a1 , n = 1, 2, · · ·
(2n + 1)!
From (4.11) we have
y(x) = a0 + a1 (x − 2) + a2 (x − 2)2 + a3 (x − 2)3 + a4 (x − 2)4 + · · ·

= a0 + a2 (x − 2)2 + a4 (x − 2)4 + · · · +

a1 (x − 2) + a3 (x − 2)3 + a5 (x − 2)5 + · · ·

(4.13)
∞
X 1 ∞
X 1
= a0 (x − 2)2n + a1 (x − 2)2n+1
n=0
(2n)! n=0
(2n + 1)!
∞
X 1 2n 1 2n+1
∴ y(x) = (x − 2) a0 + (x − 2) a1
n=0
(2n)! (2n + 1)!
Remark. We first note again that the differential equation could have been
solved much easily using our previous methods. Secondly, given initial con-
ditions; say y(2) = 1 and y 0 (2) = −1, we can determine a0 and a1 . From
(4.13) we have
y(2) = a0 = 1,
and
y 0 (x) = 2a2 (x − 2) + · · · + a1 + 3a3 (x − 2)2 + · · ·
=⇒ y 0 (2) = a1 = −1.

Thus,
∞
X 1 2n 1 2n+1
y(x) = (x − 2) − (x − 2) .
n=0
(2n)! (2n + 1)!
Example 32. Find the first few terms in a power series solution about x = 0
for a general solution to
(1 + x2 )y 00 − y 0 + y = 0. (4.14)
Solution. Note that x = 0 is an ordinary point of the equation so we look

for a solution of the form
∞
X
y(x) = an x n (4.15)
n=0
∞
X ∞
X
0 n−1
=⇒ y (x) = nan x = nan xn−1 (4.16)
n=0 n=1
X∞ ∞
X
00 n−2
=⇒ y (x) = n(n − 1)an x = n(n − 1)an xn−2 (4.17)
n=1 n=2
Substitute (4.15)-(4.17) into (4.14) to get

∞
X ∞
X ∞
X
2 n−2 n−1
(1 + x ) n(n − 1)an x − nan x + an x n = 0
n=2 n=1 n=0
∞
X ∞
X ∞
X ∞
X
=⇒ n(n − 1)an xn−2 + n(n − 1)an xn − nan xn−1 + an x n = 0
n=2 n=2 n=1 n=0
∞
X ∞
X ∞
X ∞
X
k k k
=⇒ (k+2)(k+1)ak+2 x + k(k−1)ak x − (k+1)ak+1 x + ak x k = 0
k=0 k=2 k=0 k=0
=⇒ 2a2 + 6a3 x − a1 − 2a2 x + a0 + a1 x+

∞
X
[(k + 2)(k + 1)ak+2 + k(k − 1)ak − (k + 1)ak+1 + ak ] xk = 0.
k=2
=⇒ (2a2 − a1 + a0 ) + (6a3 − 2a2 + a1 )x+

∞
X
(k + 2)(k + 1)ak+2 + (k + 1)ak+1 + (k 2 − k + 1)ak = 0.

k=2

Thus, we have
2a2 − a1 + a0 = 0 (4.18)
6a3 − 2a2 + a1 = 0 (4.19)
(k + 2)(k + 1)ak+2 + (k + 1)ak+1 + (k 2 − k + 1)ak = 0 (4.20)
From (4.18) and (4.19) we get the relations

a1 − a0
a2 = ,
2
2a2 − a1 a1 − a0 − a1 a0
a3 = = =− ,
6 6 6
and (4.20) gives the recurrence relation
(k + 1)ak+1 − (k 2 − k + 1)ak
ak+2 = , k ≥ 2.
(k + 2)(k + 1)
So we have
3 −a6 0 − 3 a1 −a

3a3 − 3a2 2
0
k=2: a4 = =
4·3 12
−a0
− 3a1 −3a

2 2
0
−a0 − 3a1 + 3a0
= =
12 24
2a0 − 3a1
=
24
−3a1
4 2a024 − 7 −a6 0

4a4 − 7a3
k=3: a5 = =
5·4 20
2a0 − 3a1 + 7a0 9a0 − 3a1 3a0 − a1
= = =
120 120 40
3a0 −a1 2a0 −3a1

5a5 − 13a4 5 40
− 13 24
k=4: a6 = =
6 · 5 30
3a0 −a1 26a0 −39a1

8
− 24
=
30
9a0 − 3a1 − 26a0 + 39a1
=
720
36a1 − 17a0
=
720
From (4.15) we have
y(x) = a0 + a1 x + a2 x2 + a3 x3 + a4 x4 + a5 x5 + a6 x6 + · · ·


a1 − a0 2 a0 3 2a0 − 3a1
=⇒ y(x) = a0 + a1 x + x − x + x4 +
2 6 24

3a0 − a1 5 36a1 − 17a0
x + x6 + · · · (4.21)
40 720
Remark. Given initial conditions, we can determine a0 and a1 , and (4.21)
gives approximations to the solution of the differential equation about x = 0.
But how useful are the partial sums; for example when x = 0.3 or x = 2.4?
Note that we don’t have a general form for an so we cannot use the well-
known tests to find the radius of convergence. The following theorem helps
us to determine the radius of convergence in cases like this.
Theorem 9. Suppose x0 is an ordinary point for the differential equation
y 00 (x) + p(x)y 0 (x) + q(x)y(x) = 0. (4.22)
Then (4.22) has two linearly independent analytic solutions of the form
∞
X
y(x) = an (x − x0 )n . (4.23)
n=0
Moreover, the radius of convergence of any power series solution of the

form given by (4.23) is at least as large as the distance from x0 to the
nearest singular point (real or complex-valued) of equation (4.22).
Remark. (a) Recall that the distance between any two complex numbers
z1 = a + ib and z2 = c + id is given by
p
|z1 z2 | = (a − c)2 + (b − d)2 .
(b) We are now in a position to address the concerns raised after the pre-
vious example as demonstrated in the example below.
Example 33. Find the minimum value for the radius of convergence of a
power series solution about the point x0 :
(1 + x2 )y 00 − y 0 + y = 0, x0 = 0.

1 1
y 00 − 2
y0 + y = 0.
1+x 1 + x2

Singular points occur when
1 + x2 = 0, x = ±i.
Now, x0 = 0 is an ordinary point and the distance from x = 0 to either x = ±i

is 1. So the radius of convergence is at least 1. Thus, using the Theorem
9, the partial sums of the series solution will converge to the solution for
−1 < x < 1.
Example 34. Find the minimum value for the radius of convergence of a
power series solution about the point x0 :
(a) 2y 00 + xy 0 + y = 0; x0 = 0.
(b) (x2 − 5x + 6)y 00 − 3xy 0 − y = 0; x0 = 0.
(c) (x + 1)y 00 − 3xy 0 + 2y = 0; x0 = 1.
(d) (1 + x + x2 )y 00 − 3y 0 = 0; x0 = 1.
x 1
Solution. (a) y 00 + y 0 + y = 0, x0 = 0.
2 2
Now,
1 1
=⇒ p(x) = x, q(x) = ,
2 2
are analytic with no singular points. So the distance between the ordi-
nary point x = 0 and the nearest singular point is infinite. Thus, the
radius of convergence is infinite.
(b) (x2 − 5x + 6)y 00 − 3xy 0 − y = 0, x0 = 0.
3x 1
y 00 − y0 − 2 y = 0; x = 0.
(x2 − 5x + 6) (x − 5x + 6)
Singular points occur for
x2 − 5x + 6 = (x − 2)(x − 3) = 0
=⇒ x = 2 or x = 3.
Now x = 0 is an ordinary point. The distance from 0 to 2 is 2, and
from 0 to 3 is 3. Since 2 is the nearest singular point to 0, Theorem 9
implies that the radius of convergence is at least 2.

(c) (x + 1)y 00 − 3xy 0 + 2y = 0; x0 = 1.
3x 0 2
y 00 − y + = 0; x = 1.
x+1 x+1
The Singular point is x = −1. The radius of convergence is given by
p
ρ = (−1 − 1)1 = 2.
(d) (1 + x + x2 )y 00 − 3y 0 = 0; x0 = 1.
3
y 00 − y 0 = 0; x0 = 1.
(1 + x + x2 )
Singular points occur for
1 + x + x2 = x2 + x + 1 = 0
√ √
−1 ± 1 − 4 1 3
=⇒ x = =− ± i
2 2 2
Now x0 = 1 is an ordinary point and the distance, ρ, from x = 1 to
each of the singular points is given by
√ !2 r
v
u 2
u 1 3 9 3
ρ=t 1+ + = +
2 2 4 4
12 √
r
=⇒ ρ = = 3.
4
√
So the radius of convergence is ρ = 3.
4.2.1 Expanding About x = 0

It is easier to expand power series about x = 0. A shift in variable enables
every power series to be expanded about x = 0 as shown below.
Example 35. Find the first few terms in a power series exapansion about
x = 1 for a general solution to
2y 00 + xy 0 + y = 0.
Determine the radius of convergence of the series.

Solution.
2y 00 + xy 0 + y = 0. (4.24)
x = 1 is an ordinary point. So we expect a solution of the form

∞
X
y(x) = an (x − 1)n . (4.25)
n=0
The radius of convergence is infinite. To shift from x0 = 1 to say t0 = 0, let
t=x−1 =⇒ x = t + 1
Let Y (t) = y(t + 1). Applying the Chain Rule:
dY dy(t + 1) dy dx dy
= = · = ,
dt dt dx dt dx
since dx/dt = 1. Thus, we have
dy dY d2 y d2 Y
= and = .
dx dt dx2 dt2
So equation (4.24) becomes
d2 Y dY
2 2
+ (t + 1) + Y = 0,
dt dt
and we seek solutions of the form
∞
X
Y (t) = an t n .
n=0
We can now proceed as usual to get the solution as

1 2 1 3 1 2 1 3
Y (t) = a0 1 − t + t + · · · + a1 t − t − t + · · · .
4 24 4 8
Setting t = x − 1, yields the solution with the correct variables as

1 2 1 3 1 2 1 3
y(x) = a0 1 − (x − 1) + (x − 1) + · · · +a1 (x − 1) − (x − 1) − (x − 1) + · · ·
4 24 4 8
4.2.2 Series Solutions about Regular Singular Points

Definition 13. A singular point x0 of
y 00 (x) + p(x)y 0 (x) + q(x)y(x) = 0, (4.26)
is said to be a regular singular point if both (x − x0 )p(x) and (x −

x0 )2 q(x) are analytic at x0 . Otherwise x0 is called an irregular singular
point.
Example 36. Classify the singular points of the equation
(x2 − 1)2 y 00 (x) + (x + 1)y 0 (x) − y(x) = 0.
Solution.
(x + 1) 0 1
y 00 (x) + 2 2
y (x) − 2 y(x) = 0.
(x − 1) (x − 1)2
Thus,
(x + 1) (x + 1) 1
p(x) = 2 2
= 2
=
(x − 1) [(x − 1)(x + 1)] (x − 1)2 (x + 1)
−1 −1
q(x) = =
(x2− 1)2 (x − 1)2 (x + 1)2
Thus, the singular points are
x = 1 and x = −1.
For x = 1, we have
1
(x − 1)p(x) =
(x − 1)(x + 1)
which is not analytic at x = 1. So x = 1 is an irregular singular point. (Note
that we don’t need to proceed to analyze (x − 1)2 q(x)).
For x = −1, we have
1
(x + 1)p(x) = ,
(x − 1)2
which is analytic at x = −1. Now
−1
(x + 1)2 q(x) = ,
(x − 1)2
which is also analytic at x = −1. So x = −1 is a regular singular point.

4.2.3 Method of Frobenius

This is a method used to find a series solution to
y 00 (x) + p(x)y 0 (x) + q(x)y(x) = 0, (4.27)
about a regular singular point, say x0 . The idea is that, since y = xr is the
form of a solution to Cauchy-Euler equations (ax2 y 00 + bxy 0 + cy = 0, x > 0),
we expect, at a regular singular point x = 0, a solution of the form
∞
X
y(x) = W (r, x) = xr an x n , x>0
n=0
∞
X
=⇒ W (r, x) = an xn+r , (4.28)
n=0
with a0 6= 0. Now
∞
X
0
W = (n + r)an xn+r−1 , (4.29)
n=0
X∞
=⇒ W 00 = (n + r)(n + r − 1)an xn+r−2 , (4.30)
n=0
Also, since xp(x) and x2 q(x) are analytic, we can expand them as
∞
X
xp(x) = an x n ,
n=0
∞
X
=⇒ p(x) = an xn−1 , (4.31)
n=0
and
∞
X
q(x) = an xn−2 . (4.32)
n=0
Substituting (4.28)-(4.32) into (4.27) we get

∞ ∞
! ∞
!
X X X
n+r−2 n−1 n+r−1
(n + r)(n + r − 1)an x + pn x (n + r)an x +
n=0 n=0 n=0
∞
! ∞
!
X X
qn xn−2 an xn+r =0
n=0 n=0

Expanding and grouping like terms results in the equation
[r(r − 1) + p0 r + q0 ] a0 xr−2 + · · · = 0
where xr−2 is the lowest power in this case. Equating each coefficient to zero
results in the indicial equation:
r(r − 1)p0 r + q0 = 0, a0 6= 0,
as defined formerly below.
Definition 14 (Indicial Equation). If x0 is a regular singular point of

y 00 + p(x)y 0 + q(x)y = 0, then the indicial equation for x0 is
r(r − 1) + p0 r + q0 = 0,
where
p0 := lim (x − x0 )p(x) and q0 := lim (x − x0 )2 q(x).

x→x0 x→x0
The roots of the indicial equation are called exponents or indices of

the singularity x0 .
Example 37. Find the indicial equation and exponents at the singularity
x = −1 of
(x2 − 1)2 y 00 (x) + (x + 1)y 0 (x) − y(x) = 0.
Solution. We analyzed this equation in example 36, where we found that
x = −1 is a regular singular point, and
1 −1
p(x) = and q(x) = .
(x − 1)2 (x + 1) (x − 1)2 (x + 1)2
1 1
p0 = lim (x + 1)p(x) = lim 2
=
x→−1 x→−1 (x − 1) 4
−1 1
q0 = lim (x + 1)2 q(x) = lim = −
x→−1 x→−1 (x − 1)2 4
So the indicial equation r(r − 1) + p0 r + q0 = 0 becomes
1 1
r(r − 1) + r − = 0,
4 4
=⇒ 4r(r − 1) + r − 1 = 0, =⇒ 4r2 − 3r − 1 = 0

=⇒ 4r2 − 4r + r − 1 = 0, =⇒ 4r(r − 1) + (r − 1) = 0
=⇒ (r − 1)(4r + 1) = 0.
So the exponents are:
1
r = 1, − .
4
The Frobenius Theorem stated next enables us to find one series solution of
a variable-coefficient equation.
Theorem 10 (Frobenius Theorem). If x0 is a regular singular point

of
y 00 (x) + p(x)y 0 (x) + q(x)y(x) = 0, (4.33)
then there exists at least one series solution of the form

∞
X ∞
X
n
W (r, x) = (x − x0 ) an (x − x0 ) = an (x − x0 )n+r ,
n=0 n=0
where r = r1 is the larger root of the associated indicial equation. More-

over, this series converges for all x such that 0 < x − x0 < ρ, where ρ is
the distance from x0 to the nearest other singular point (real or complex)
of (4.33).
Example 38. Find the series expansion about the regular singular point
x = 0 for a solution to
(x + 2)x2 y 00 (x) − xy 0 (x) + (1 + x)y(x) = 0, x > 0. (4.34)

x (1 + x)
y 00 (x) − 2
y 0 (x) + y(x) = 0
(x + 2)x (x + 2)x2
So we have
−x −1
p(x) = 2
= .
(x + 2)x (x + 2)x
1+x
q(x) = .
(x + 2)x
Thus,
−1 1
p0 = lim xp(x) = lim =−
x→0 x→0 x+2 2

1+x 1
q0 = lim x2 q(x) = lim =
x→0 x→0 x + 2 2
The indicial equation r(r − 1) + p0 r + q0 = 0 becomes
1 1
r(r − 1) − r + = 0,
2 2
2r2 − 2r − r + 1 = 0, 2r2 − 3r + 1 = 0
=⇒ (r − 1)(2r − 1) = 0
So the exponents are
1
r = 1, and r = .
2
So we seek a solution of the form
∞
X ∞
X
r n
W (r, x) = x an x = an xn+r
n=0 n=0
To obtain the constants an , we apply Frobenius Theorem and use the larger
root r = 1 to get one solution y1 (x) = W (1, x) such that
∞
X
W (1, x) = an xn+1 (4.35)
n=0
∞
X
0
=⇒ W = (n + 1)an xn
n=0
∞
X
=⇒ W 00 = (n + 1)nan xn−1
n=0
Substitute the above expressions into (4.34) to get

∞
X ∞
X
2 n−1
(x + 2)x (n + 1)nan x −x (n + 1)an xn +
n=0 n=0
X∞
(1 + x) an xn+1 = 0
n=0
This may be written as

∞
X ∞
X
k
[(k − 1)(k − 2) + 1] ak−2 x + [2k(k − 1) − k + 1] ak−1 xk = 0
k=2 k=1

Combining the summations results in

(k 2 − 3k + 3)ak−2 + (2k − 1)(k − 1)ak−1 = 0
k 2 − 3k + 3
=⇒ ak−1 = − ak−2 , k ≥ 2.
(2k − 1)(k − 1)
Thus, we get
1
a1 = − a0
3
1
a2 = a0
10
1
a3 = − a0
30
.. ..
.=.
From (4.35) we have
W (1, x) = a0 x + a1 x2 + a2 x3 + a3 x4 + · · ·
1 1 1
=⇒ W (1, x) = a0 x − a0 x2 + a0 x3 − a0 x4 + · · ·
3 10 30
1 1 2 1 3
∴ y1 (x) = a0 x 1 − x + x − x + · · · ; x > 0
3 10 30
4.2.4 Form of a Second Linearly Independent Solution
Theorem 11. Let x0 be a regular singular point for
y 00 + p(x)y 0 + q(x)y = 0
and let r1 and r2 be the roots of the associated indicial equation, where
R(r1 ) ≥ R(r2 ).
(a) If r1 −r2 is not an integer, then there exist two linearly independent
solutions of the form
∞
X
y1 (x) = an (x − x0 )n+r1 , a0 6= 0, (4.36)
n=0
X∞
y2 (x) = bn (x − x0 )n+r2 , b0 6= 0. (4.37)
n=0
(b) If r1 = r2 , then there exist two linearly independent solutions of

the form
∞
X
y1 (x) = an (x − x0 )n+r1 , a0 6= 0, (4.38)
n=0
∞
X
y2 (x) = y1 (x) ln(x − x0 ) + bn (x − x0 )n+r2 , b0 6= 0. (4.39)
n=0
(b) If r1 − r2 is a positive integer, then there exist two linearly inde-

pendent solutions of the form
∞
X
y1 (x) = an (x − x0 )n+r1 , a0 6= 0, (4.40)
n=0
∞
X
y2 (x) = Cy1 (x) ln(x − x0 ) + bn (x − x0 )n+r2 , b0 6= 0. (4.41)
n=0
where C is a constant that could be zero.


Chapter 5
Laplace Transforms
Definition 15. Let f (t) be a function on [0, ∞). The Laplace transform
of f is the function F or L{f } defined by the integral
Z ∞
F (s) := e−st f (t)dt. (5.1)
0
The domain of F (s) is all the values of s for which the integral in (5.1)
exists.
Remark. (1) The integral in (5.1) is an improper integral so to integrate

we effect
Z ∞ Z N
−st
e f (t)dt = lim e−st f (t)dt
0 N →∞ 0
whenever the limit exists.
(2) The transform is used to convert constant coefficient differential equa-

tions from the t−domain to (simpler) algebraic equations in the x−domain.
(3) Consider the equation
ay 00 + by 0 + cy = f (t) (5.2)
Laplace transforms are more useful if f (t) is not a continuous function.

Besides, Laplace transforms are more simpler to use in solving certain
more complicated nonhomogeneous differential equations.
73
Figure 5.1: Schematic of the mixing problem
MOTIVATION: Mixing Problem

The rate of change of the amount of salt in a tank, x(t), is governed by the
dx 3
+ x = f (t), x(0) = 30,
dt 500
where

2.4km/min, 0 < t < 10, Valve A
f (t) = (5.3)
1.2km/min, t > 10, Valve B
Using the Laplace transform to solve the mixing problem is much easier than
our previous methods.
Example 39. Determine the Laplace transform of
(a) f (t) = 1, t ≥ 1.
(b) f (t) = eat , a = constant

(
0, 0 < t < 2,
(c) f (t) =
t, t > 2.
Solution. (a)
Z ∞ Z N
−st
F (s) = e f (t)dt = lim e−st · 1dt
0 N →∞ 0

N
1 1 1
= lim − e−st = lim − e−sN +
N →∞ s 0
N →∞ s s
1
∴ F (s) = , for s > 0.
s
If s ≤ 0 the integral diverges, so
1
F (s) = ∀s > 0.
s
(b)
Z ∞ Z N
−st at
F (s) = e e dt = lim e−(s−a)t dt
0 N →∞ 0
N
1 −(s−a)t 1 −(s−a)N 1
= lim − e = lim − e +
N →∞ s−a 0
N →∞ (s − a) (s − a)
1
∴ F (s) = , for s − a > 0 or s > a.
s−a
If s ≤ a the integral diverges, so the domain of F (s) is all s > a.
1
∴ L{eat } = ∀s > a.
s−a
(
0, 0 < t < 2,
(c) f (t) =
t, t > 2.
Z ∞ Z 2 Z ∞
−st −st
F (s) = e f (t)dt = e · 0dt + e−st dt
0 0 2
Z N
= lim te−st dt
N →∞ 2
We integrate by parts. Let
u = t, dv = e−st dt
1
v = − e−st
=⇒ du = dt,
s
( N Z N )
1 −st 1 −st
=⇒ F (s) = lim − te + e dt
N →∞ s 2 2 s
( N )
1 −sN 2 −2s 1 −st
= lim − N e + e + − 2e
N →∞ s s s 2


1 −sN 2 −2s 1 −sN 1 −2s
= lim − N e + e − 2e + 2e
N →∞ s s s s
After taking the limit, we get
2 1
F (s) = e−2s + 2 e−2s , s > 0.
s s

−2s 2s + 1
∴ F (s) = e , s > 0.
s2

5.0.5 Table of Laplace Transforms

Table 5.1 shows some Laplace transforms.
1
1 s
, s>0
1
eat s−a
, s>a
n!
tn , n = 1, 2, · · · sn+1
, s>0
b
sin bt s2 +b2
, s>0
s
cos bt s2 +b2
, s>0
n!
eat tn , n = 1, 2, · · · (s−a)n+1
, s>a
b
eat sin bt (s−a)2 +b2
, s>a
s−a
eat cos bt (s−a)2 +b2
, s>a
Table 5.1: A brief table of Laplace transforms
Example 40. Use the Table of Laplace transforms to compute the Laplace
transform of
(a) f (t) = te3t (d) f (t) = t2
(b) f (t) = cos 2t (e) f (t) = e−3t sin 2t
(c) f (t) = e2t cos(3t) (f) f (t) = t2 e4t
Solution. Let L{f (t)} = F (s).

(a) f (t) = te3t .
We use the relation
n!
L{eat tn } = , s > a,
(s − a)n+1
with n = 1 and a = 3. Thus
1
F (s) = .
(s − 3)2

(b) f (t) = cos 2t. We apply

s
L{cos bt} = , s>0
s 2 + b2
with b = 2. Thus
s
F (s) = .
s2 +4
(c) f (t) = e2t cos(3t).
We use
s−a
L{eat cos bt} = , s>a
(s − a)2 + b2
with a = 2 and b = 3. So we get
s−2
F (s) = , s > 2.
(s − 2)2 + 9
5.0.6 Properties of Laplace Transforms

1. L{f + g} = L{f } + L{g}
2. L{cf } = cL{f } c = constant.
3. L{eat f (t)}(s) = L{f }(s − a)
4. L{f 0 }(s) = sL{f }(s) − f (0)
5. L{f 00 }(s) = s2 L{f }(s) − sf (0) − f 0 (0)
6. L{f n }(s) = sn L{f }(s) − sn−1 f (0) − sn−2 f 0 (0) − · · · − f n−1 (0).
dn
7. L{tn f (t)}(s) = (−1)n (L{f }(s)) .
dsn
Example 41. Determine the Lapace transform of the following expressions
(a) 3t2 − e2t
(b) e−2t sin 2t + e3t t2
Solution. (a) 3t2 − e2t
L{3t2 − e2t } = L{3t2 } + L{e2t } = 3L{t2 } + L{e2t }

2! 1
=3 3 + , s>0
s s−2
6 1
= 3+ , s > 0.
s s−2

(b) e−2t sin 2t + e3t t2 . Using L{eat f (t)}(s) = L{f }(s − a), we have
2 2
L{sin 2t} = =⇒ L{e−2t sin 2t} =
s2 +4 (s + 2)2 + 4
dn
Also, employing L{tn f (t)}(s) = (−1)n (L{f }(s)) gives
dsn
2

1 2 d 1 2
L{e3t } = , 2 3t
=⇒ L{t e } = (−1) 2 = .
s−3 ds s − 3 (s − 3)3
Example 42. Determine
(a) L{t cos bt}
(b) L{t2 cos bt}
Solution. (a) L{t cos bt}. Now

s
L{cos bt} =
+ b2s2

d s
=⇒ L{t cos bt} = (−1)
ds s2 + b2
2
(s + b2 ) − s(2s)
2
s − 2s2 + b2

=− =−
(s2 + b2 )2 (s2 + b2 )2
s 2 − b2
= .
(s2 + b2 )2
(b) L{t2 cos bt}. Using the results in (a), it can be shown that
2s3 − 6sb2
L{t2 cos bt} = L{t · t cos bt} = .
(s2 + b2 )3
5.0.7 Inverse Laplace Transforms
Definition 16. Given a function F (s), if there is a function f (t) that

is continuous on [0, ∞) and satisfies L{f } = F , then f (t) is the inverse
Laplace transform of F (s):
f (t) = L−1 {F }

Properties of Inverse Laplace Transforms

1. L−1 {F1 + F2 } = L−1 {F1 } + L−1 {F2 }
1. L−1 {cF } = cL−1 {F }, c = contant.
Motivation
Given the IVP:
(
y 00 (t) − y(t) = 0
(5.4)
y(0) = 0, y 0 (0) = 1.
Let L{y} = Y (s). Then (5.4) becomes, after taking the Laplace transform,
L{y 00 } − L{y} = 0
=⇒ s2 Y (s) − sy(0) − y 0 (0) − Y (s) = 0

=⇒ s2 Y (s) − 1 − Y (s) = 0
=⇒ Y (s)[s2 − 1] = 1
1
=⇒ Y (s) = 2 .
s −1
But we want the solution of the IVP to be y(t). To get this, we need to find
the inverse transform of Y (s). Let
1 1 A B
Y (s) = = ≡ +
s2 −1 (s − 1)(s + 1) s−1 s+1
=⇒ 1 ≡ A(s + 1) + B(s − 1)
1
s = 1, =⇒ A =
2
1
s = −1, =⇒ B = −
2
1 1 1 1
∴ Y (s) = · − ·
2 s−1 2 s+1
Thus,
1 1 1 1
y(t) = L−1 {Y } = L−1 { } − L−1 { }
2 s−1 2 s+1
From the Table of Laplace transforms, we get
1 1 et − e−t
y(t) = et − e−t =
2 2 2
∴ y(t) = sinh(t).

Remark. Note that the motivational example above could have been solved
much more easily with our previous methods for homogeneous differential
equations with constant coefficients. However, the example demonstrates
some of the capabilities of the Laplace transform approach. And as men-
tioned earlier, it is a much more powerful approach, for instance, in cases
where the initial conditions are discontinuous as we’ll encounter later.
Example 43. Determine L−1 {F }, where

2 4
(a) F (s) = (c) F (s) =
s2 +4 s2 +9
2 s−1
(b) F (s) = (d) F (s) =
s3 s2 − 2s + 5
Solution. Determine L−1 {F }, where

2
(a) F (s) =
s2 +4
Note that
b b
L{sin bt} = =⇒ L−1 { 2 } = sin bt.
s2 +b 2 s + b2
Let b = 2, =⇒ L−1 {F } = sin 2t.

2
(b) F (s) =
s3
n! 2!
tn → n+1
, =⇒ t2 →
s s3
=⇒ L−1 {F } = t2 .
4
(c) F (s) =
s2 +9
4 4 3
F (s) = = · 2
+9 s2
3 s +9

−1 4 −1 3 4
=⇒ L {F (s)} = L 2 2
= sin 3t.
3 s +3 3
s−1
(d) F (s) =
s2
− 2s + 5
s−1 s−1
F (s) = 2 = . By completing the square.
s − 2s + 5 (s − 1)2 + 4

Note that
s s−a
cos bt = , and eat cos bt = .
s2 + b2 (s − a)2 + b2
Thus, if we let a = 1 and b = 2, we get

−1 s−1 −1 s−1
L =L = et cos 2t.
s2 − 2s + 5 (s − 1)2 + 4
Remark. To compute L−1 of rational functions, review and apply the tech-
niques of partial fractions as demonstrated by the following examples.
Example 44. Determine L−1 {F }.
7s − 1
(a) F (s) = . Non-repeated linear factors
(s + 1)(s + 2)(s − 3)
s2 + 9s + 2
(b) F (s) = . Repeated linear factors
(s − 1)2 (s + 3)
2s2 + 10s
(c) F (s) = . Quadratic factors
(s2 − 2s + 5)(s + 1)
Solution. Determine L−1 {F }.
7s − 1
(a) F (s) = .
(s + 1)(s + 2)(s − 3)
7s − 1 A B C
= + +
(s + 1)(s + 2)(s − 3) s+1 s+2 s−3
7s − 1 ≡ A(s + 2)(s − 3) + B(s + 1)(s − 3) + C(s + 1)(s + 2)
s = −1 : −8 = −4A =⇒ A = 2
s = −2 : −15 = 5B =⇒ B = −3
s=3: 20 = 20C =⇒ C = 1.
7s − 1 2 −3 1
=⇒ = + +
(s + 1)(s + 2)(s − 3) s+1 s+2 s−3
∴ L−1 {F } = 2e−t − 3e−2t + e3t .

s2 + 9s + 2
(b) F (s) = .
(s − 1)2 (s + 3)
s2 + 9s + 2 A B C
F (s) = 2
= + 2
+
(s − 1) (s + 3) s − 1 (s − 1) s+3
=⇒ s2 + 9s + 2 = A(s − 1)(s + 3) + B(s + 3) + C(s − 1)2
s=1: 12 = 4B =⇒ B = 3
s = −3 : −16 = 16C =⇒ C = −1
s=1: 2 = −3A + 9 − 1 = −3A + 8
=⇒ −6 = −3A =⇒ A = 2.
2 3 1
=⇒ F (s) = + 2
−
s − 1 (s − 1) s+3
−1 t t −3t
∴ L {F } = 2e + 3te − e .
2s2 + 10s
(c) F (s) = .
(s2 − 2s + 5)(s + 1)
By completing the square, we have s2 − 2s + 5 = (s − 1)2 + 4. Thus
2s2 + 10s 2s2 + 10s
F (s) = =
[(s − 1)2 + 4](s + 1) [(s − 1)2 + 22 ](s + 1)
A(s − 1) + 2B C
≡ +
(s − 1)2 + 4 s+1
=⇒ 2s2 + 10s = [A(s − 1) + 2B](s + 1) + C[(s − 1)2 + 4]
s=1: 12 = 4B + 4C =⇒ 3 = B + C
s = −1 : −8 = 8C =⇒ C = −1
=⇒ 3 = B − 1 =⇒ B = 4.
s=0: 0 = −A + 8 − 5 =⇒ A = 3.
3(s − 1) + 8 1
=⇒ F (s) = 2
−
(s − 1) + 4 s + 1
(s − 1) 2 1
=⇒ F (s) = 3 2 2
+4 2 2
−
(s − 1) + 2 (s − 1) + 2 s+1
−1 t t −t
∴ L {F (s)} = 3e cos 2t + 4e sin 2t − e .
EXERCISE 2. Determine L−1 {F }.
2s + 8
1) F (s) =
(s − 1)2 + 4
7s2 + 23s + 30
2) F (s) =
(s − 2)(s2 + 2s + 5)

5.0.8 Solving IVPs Using Laplace Transforms

Solution Approach
(a) Take the Laplace transform of both sides of the differential equation.
(b) Get an equation for the Laplace transform of the solution, say y(t), and
solve for it, say Y (s).
(c) Find the inverse Laplace transform of Y (s) to get the solution, say y(t).
Example 45. Use Laplace transforms to solve the following IVPs.
(a) y 00 − 2y 0 + 5y = 0; y(0) = 2, y 0 (0) = 4.
(b) y 00 + 6y 0 + 9y = 0; y(0) = −1, y 0 (0) = 6.
Solution. (a) y 00 − 2y 0 + 5y = 0; y(0) = 2, y 0 (0) = 4.

Taking the Laplace transform of the equation gives
L{y 00 } − 2L{y 0 } + 5L{y} = 0
=⇒ [s2 Y (s) − sy(0) − y 0 (0)] − 2[sY (s) − y(0)] + 5Y (s) = 0.

=⇒ s2 Y (s) − 2s − 4 − 2sY (s) + 4 + 5Y (s) = 0
=⇒ Y (s)[s2 − 2s + 5] − 2s = 0
2s
=⇒ Y (s) = .
s2 − 2s + 5
Thus, we have
∴ y(t) = L−1 {Y (s)}
Now,
2s 2(s − 1) + 2 s−1 2
Y (s) = = =2 +
s2
− 2s + 5 2
(s − 1) + 4 (s − 1) + 4 (s − 1)2 + 4
2

−1 s−1 −1 2
=⇒ y(t) = 2L +L
(s − 1)2 + 4 (s − 1)2 + 4
∴ y(t) = 2et cos 2t + et sin 2t.

Check: We may verify that our solution is correct using our previous
technique. The roots of the characteristic equation can be found from
r2 − 2r + 5 = 0.

√
2±4 − 20 2 ± 4i
=⇒ r = = = 1 ± 2i
2 2
=⇒ y(t) = C1 et cos 2t + C2 et sin 2t
We next apply the initial conditions. Now
y 0 = C1 et cos 2t − 2C1 et sin 2t + C2 et sin 2t + 2C2 et cos 2t
y(0) = 2 =⇒ 2 = C1
y 0 (0) = 4 =⇒ 4 = C1 + 2C2 =⇒ C2 = 1
t t
∴ y(t) = 2e cos 2t + e sin 2t,
which is the same as the solution we got using Laplace transforms.
(b) y 00 + 6y 0 + 9y = 0; y(0) = −1, y 0 (0) = 6.

Taking the Laplace transform of the equation gives
L{y 00 } + 6L{y 0 } + 9L{y} = 0
=⇒ [s2 Y (s) − sy(0) − y 0 (0)] + 6[sY (s) − y(0)] + 9Y (s) = 0.

=⇒ s2 Y (s) + s − 6 + 6sY (s) + 6 + 9Y (s) = 0
=⇒ Y (s)[s2 + 6s + 9] + s = 0
−s −s
=⇒ Y (s) = 2 = .
s + 6s + 9 (s + 3)2
−(s − 3) + 3 1 3
=⇒ Y (s) = 2
=− +
(s + 3) s + 3 (s + 3)2
=⇒ L−1 {Y (s)} = −e−3t + 3te−3t
∴ y(t) = −e−3t + 3te−3t .
The next set of examples reveal some of the advantages of the Laplace trans-
form approach compared to the other techniques.
Example 46. Solve for the Laplace transform, Y (s), of y(t).
(a) y 00 + 4y = g(t), y(0) = −1, y 0 (0) = 0,
where (
t, t < 2
g(t) =
5, t > 2.
(b) y 00 − 3y 0 + 2y = cos t
y(0) = 0, y 0 (0) = −1.

Solution. (a) Let the Laplace transform of g(t) be G(s). Taking the
Laplace transform of the equation results in
L{y 00 } + 4L{y} = L{g}
=⇒ [s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = G(s). (5.5)
Now, from definition of the Laplace transform, we have

Z 2 Z ∞
−st
G(s) = te dt + 5e−st dt
0 2
Let Z 2
I1 = te−st dt
0
Integrating by parts, we let u = t, dv = e−st dt

1
=⇒ du = dt, v = − e−st
s
t −st 1 2 −st
Z
=⇒ I1 = − e + e dt
s s 0
2
t −st 1 −st
= − e − 2e
s s 0

2 −2s 1 −2s 1
= − e − 2e − − 2
s s s

2s + 1 −2s 1
=⇒ I1 = − 2
e + 2.
s s
Let Z ∞ Z N
−st
I2 = 5e dt = lim 5e−st dt
2 N →∞ 2
N
5 5 5
= lim − e−st = lim − e−N s + e−2s
N →∞ s 2
N →∞ s s
5
=⇒ I2 = e−2s , s > 0.
s
Thus,
1 2s + 1 5
G(s) = I1 + I2 = 2 − e−2s + e−2s .
s s2 s

From (5.5), we have
s2 Y (s) + s + 4Y (s) = G(s)
1 2 1 5
=⇒ Y (s)[s2 + 4] =
2
− e−2s − 2 e−2s + e−2s − s
s s s s
1 2 1 5 s
=⇒ Y (s) = 2 2 − 2 e−2s − 2 2 e−2s + 2 e−2s − 2
s (s + 4) s(s + 4) s (s + 4) s(s + 4) (s + 4)
1 3 1 s
=⇒ Y (s) = + e−2s − 2 2 e−2s − 2
s2 (s2
+ 4) s(s + 4)2 s (s + 4) (s + 4)
(5.6)
1 + (3s − 1)e−2s − s3
∴ Y (s) = . (5.7)
s2 (s2 + 4)
This question does not ask for the actual solution, y(t). But to deter-
mine the actual solution, you could find the inverse Laplace transform
of (5.7) or (5.6).
(b) This is much easier, try it.
5.0.9 IVPs with Non-zero Initial Conditions

We use the following example to illustrate how to solve IVPs with non-zero
initial conditions.
Example 47. Solve the IVP using Laplace transforms:
z 00 (t) + 5z 0 (t) − 6z(t) = 21et−1 (5.8)

z(1) = −1, z 0 (1) = 9. (5.9)
Solution. The idea is to shift the initial condition to start from t = 0. To

do that, let
y(t) = z(t + 1),
where the 1 in t + 1 is because the initial value is 1.
=⇒ y 0 (t) = z 0 (t + 1), and y 00 (t) = z 00 (t + 1).
In equation (5.8), replace t by t + 1 to get
z 00 (t + 1) + 5z 0 (t + 1) − 6z(t + 1) = 21et .

So we get a new IVP:

y 00 + 5y 0 − 6y = 21et (5.10)
y(0) = z(1) = −1, and y 0 (0) = z 0 (1) = 9.
We can now solve as usual by taking the Laplace transform of (5.10) to get
21
[s2 Y (s) − sy(0) − y 0 (0)] + 5[sY (s) − y(0)] − 6Y (s) =
s−1
21
=⇒ [s2 Y (s) + s − 9] + 5sY (s) + 5 − 6Y (s) =
s−1
21
=⇒ Y (s)[s2 + 5s − 6] + s − 4 =
s−1
21
=⇒ Y (s)(s2 + 5s − 6) = +4−s
s−1
21 4−s
=⇒ Y (s) = +
(s − 1)(s2 + 5s − 6) (s2 + 5s − 6)
21 −(s − 1) + 3
= 2
+
(s − 1)(s + 5s − 6) (s − 1)(s + 6)
21 1 3
= 2
− +
(s − 1)(s + 5s − 6) (s + 6) (s − 1)(s + 6)
21 + 3(s − 1) 1
=⇒ Y (s) = 2
−
(s − 1) (s + 6) (s + 6)
3(s + 6) 1
= 2
−
(s − 1) (s + 6) (s + 6)
3 1
= 2
−
(s − 1) (s + 6)
Taking the inverse Laplace transform gives
y(t) = 3tet − e−6 .
=⇒ z(t + 1) = 3tet − e−6t
We can now replace t by t − 1 to get
z(t) = 3(t − 1)et−1 − e−6(t−1) .
5.0.10 Transforms of Discontinuous Functions

Unit Step Function

Definition 17. The unit step function or Heaviside function u(t) is de-
fined by (
0, t < 0
u(t) =
1, t > 0.
See Figure 5.2.
Figure 5.2: The unit step function
Remark. (a) The jump can be moved to a different location, say t = a:

( (
0, t − a < 0 0, t < a
u(t − a) = =
1, t − a > 0 1, t > a
(b) The height of the jump can also be changed by multiplying by a con-
stant, say M : (
0, t < a
M u(t − a) =
M, t > a
Figure 5.3 illustrates some of these properties.
Figure 5.3:
Example 48. Write the given function, f (t), in terms of unit step functions.

(
3, t<2
(a) f (t) =
1, 2 < t < 5.


 3, t<2

1, 2<t<5
(b) f (t) =
t,
 5<t<8

 t2
10
, t>8
(
3, t < 2
Solution. (a) f (t) =
1, 2 < t < 5.
The function is sketched in Figure 5.4. We see that it is equal to 3 until
t reaches 2, then it jumps by −2 units to the value 1. Thus, it can be
written as
f (t) = 3 − 2u(t − 2).
Figure 5.4:


 3, t<2

1, 2<t<5
(b) f (t) =
t,
 5<t<8

 t2
10
, t>8
The function is sketched in Figure 5.5, and is can be written as
2
t
f (t) = 3 − 2u(t − 2) + (t − 1)u(t − 5) + − t u(t − 8).
10

Figure 5.5:
5.0.11 Properties
e−as
(1) L{u(t − a)}(s) = , s > 0.
s
e−as

−1
(2) L = u(t − a)
s
(3) L{f (t − a)u(t − a)}(s) = e−as F (s)
(4) L−1 e−as F (s) (t) = f (t − a)u(t − a)

(5) L {g(t)u(t − a)} (s) = e−as L{g(t + a)}(s).

Z ∞
Proof. (3) L{f (t − a)u(t − a)}(s) = e−st f (t − a)u(t − a)dt
0
But (
0, t<a
u(t − a) =
1, t>a
Z ∞
=⇒ L{f (t − a)u(t − a)}(s) = e−st f (t − a)dt
a
Let v = t − a, =⇒ dv = dt.
Z ∞
=⇒ L{f (t − a)u(t − a)}(s) = e−s(v+a) f (v)dv
0
Z ∞
= e−as e−sv f (v)dv
0
∴ L{f (t − a)u(t − a)}(s) = e−as F (s).
(1) Let f = 1 in property (3). Since L{1} = 1/s, we get
e−as
L{u(t − a)}(t) = .
s
(5) From property (3) let g(t) = f (t − a). Then f (t) = g(t + a). Thus
L {g(t)u(t − a)} (s) = e−as L{g(t + a)}(s).
Example 49. Determine the Laplace transform of

(a) (t − 1)2 u(t − 1)
(b) t2 u(t − 2)
(c) u(t − 1) − u(t − 4).
Solution. (a) (t − 1)2 u(t − 1)
From property (3), we have
L{f (t − a)u(t − a)}(s) = e−as L{f (t)}(s) = e−as F (s)
Let
f (t − a) = (t − 1)2
and a = 1.
2! 2
=⇒ f (t) = t2 =⇒ F (s) = 3 = 3
s s
2 2e−s
=⇒ L{(t − 1)2 u(t − 1)} = e−s · 3 = 3 .
s s

(b) t2 u(t − 2)
From property (5), we have L {g(t)u(t − a)} (s) = e−as L{g(t + a)}(s).
g(t) = t2 , a = 2.
=⇒ g(t + a) = (t + 2)2 = e2 + 4t + 4
2 4 4
=⇒ L{g(t + a)} = L{g(t + 2)} = 3 + 2 +
s s s

2 4 4
=⇒ L{t2 u(t − 2)} = e−2s 3
+ 2+ .
s s s
(c) Let q = u(t − 1) − u(t − 4).
e−s e−4s
L{q} = − .
s s
Example 50. Determine the inverse Laplace transform of
e−3s e−2s − 3e−4s
(a) q = (c) q =
s2 s+2
e−2s se−3s
(b) q = (d) q =
s−1 s2 + 4s + 5
e−3s
Solution. (a) q =
s2
e−3s −3s 1
q= = e ·
s2 s2
Use the property
L−1 e−as F (s) (t) = f (t − a)u(t − a)

1
F (s) = =⇒ f (t) = t, a = 3
s2
=⇒ f (t − 3) = (t − 3)

−1 −3s 1
=⇒ L e = (t − 3)u(t − 3).
s2
e−2s
(b) q =
s−1
Use the property
L−1 e−as F (s) (t) = f (t − a)u(t − a)


1
F (s) = =⇒ f (t) = et , a=2
s−1
f (t − 2) = et−2
=⇒ L−1 {q} = et−2 u(t − 2).
e−2s − 3e−4s
(c) q =
s+2
1 1
=⇒ q = e−2s · − 3e−4s ·
s+2 s+2
Use the property
L−1 e−as F (s) (t) = f (t − a)u(t − a)

For
1 1
e−2s · , F (s) = , a=2
s+2 s+2
=⇒ f (t) = e−2t =⇒ f (t − 2) = e−2(t−2)
=⇒ f (t − 2) = e−2t+4

−1 −2s 1
=⇒ L e · = e−2t+4 u(t − 2).
s+2
For
1 1
e−4s · , F (s) = , a=4
s+2 s+2
=⇒ f (t − 4) = e−2(t−4) = e−2t+8

−1 −4s 1
=⇒ L e · = e−2t+8 u(t − 4).
s+2
∴ L−1 {q} = e−2t+4 u(t − 2) − 3e−2t+8 u(t − 4).
se−3s
(d) q =
s2 + 4s + 5
s
=⇒ q = e−3s ·
s2 + 4s + 5
But
s (s + 2) − 2 (s + 2) 1
= = −2
s2 + 4s + 5 2
(s + 2) + 1 2
(s + 2) + 1 (s + 2)2 + 1
(s + 2) 1
=⇒ q = e−3s · 2
− 2e−3s ·
(s + 2) + 1 (s + 2)2 + 1

Use the property
L−1 e−as F (s) (t) = f (t − a)u(t − a)

Let
(s + 2)
F1 (s) =
(s + 2)2 + 1
=⇒ f1 (t) = e−2t cos t =⇒ f1 (t − a) = e−2(t−3) cos(t − 3)
Let
1
F2 (s) =
(s + 2)2 + 1
=⇒ f2 (t) = e−2t sin t =⇒ f2 (t − 3) = e−2(t−3) sin(t − 3)
=⇒ L−1 {q} = e−2(t−3) cos(t − 3)u(t − 3) − 2e−2(t−3) sin(t − 3)u(t − 3)
∴ L−1 {q} = e−2(t−3) [cos(t − 3) − 2 sin(t − 3)] u(t − 3).
Example 51. Solve the IVP using Laplace transforms.
(a) y 00 + y = u(t − 3), y(0) = 0, y 0 (0) = 1.
(b) y 00 + 5y 0 + 6y = g(t), y(0) = 0, y 0 (0) = 2,

where 
0, 0 ≤ t < 1,

g(t) = t, 1 < t < 5,

1, t > 5.

Solution. (a) y 00 + y = u(t − 3), y(0) = 0, y 0 (0) = 1.

Taking the Laplace transform of the equation:
L{y 00 } + L{y} = L{u(t − 3)}
e−3s
=⇒ s2 Y (s) − sy(0) − y 0 (0) + Y (s) =

s
e−3s
=⇒ s2 Y − 1 + Y =
s
e−3s
=⇒ Y (s)(s2 + 1) = +1
s
e−3s 1 1 1
=⇒ Y (s) = 2
+ 2 = e−3s · 2
+ 2
s(s + 1) s + 1 s(s + 1) s + 1

Note that
−1 1
L 2
= sin t.
s +1
Now,
1 A Bs + C
= + 2
s2 + 1 s s +1
=⇒ 1 = A(s2 + 1) + (Bs + C)s
s=0: 1=A
s = 1 : 1 = 2A + B + C
=⇒ B + C = −1
s = −1 : 1 = 2 − (−B + C) = 2 + B − C
=⇒ B − C = −1
=⇒ 2B = −2 =⇒ B = −1
=⇒ C = −1 + 1 = 0
Thus,
1 1 s
= −
s2 + 1 s s2 + 1

−1 1
=⇒ L = 1 − cos t
s2 + 1
e−3s

−1 −1 1
=⇒ y(t) = L +L
s(s2 + 1) s2 + 1
∴ y(t) = [1 − cos(t − 3)]u(t − 3) + sin t.
(b)
y 00 + 5y 0 + 6y = g(t), (5.11)
y(0) = 0, y 0 (0) = 2,
where 
0,
 0 ≤ t < 1,
g(t) = t, 1 < t < 5,

1, t > 5,

is sketched in Figure 5.6. g(t) can be written as
g(t) = tu(t − 1) + (1 − t)u(t − 5).

Figure 5.6:
Let the Laplace transform of g(t) be G(s).
G(s) = L{g(t)} = L{tu(t − 1)} + L{(1 − t)u(t − 5)}
For L{tu(t − 1)}, we use the property
L{f (t)u(t − a)} = e−as L{f (t + a)}
f (t) = t, a=1
1 1
=⇒ L{f (t + a)} = L{(t + 1)} = 2 +
s s

1 1
=⇒ L{tu(t − 1)} = e−s +
s2 s
For L{(1 − t)u(t − 5)}, we again use the property
L{f (t)u(t − a)} = e−as L{f (t + a)}, a = 5.
Now f (t) = 1 − t=⇒ f (t + a) = 1 − (t + 5) = −4 − t.

4 1
=⇒ L{f (t + a)} = L{−4 − t} = − − 2
s s

−5s 4 1
=⇒ L{(1 − t)u(t − 5)} = e − − 2 .
s s
e−5s

−s 1 1 1
∴ G(s) = e 2
+ −4 − 2 e−5s .
s s s s

Now, from the differential equation (5.11):
[s2 Y (s) − sy(0) − y 0 (0)] + 5[sY (s) − y(0)] + 6Y (s) = G(s).
=⇒ (s2 Y − 2) + 5sY + 6Y = G(s)

=⇒ Y (s)(s2 + 5s + 6) = G(s) + 2
G(s) 2
=⇒ Y (s) = + 2
s2 + 5s + 6 s + 5s + 6
s2 + 5s + 6 = (s + 2)(s + 3)
Thus,
(s + 1) 1
Y (s) = e−s − 4e−5s +
s2 (s
+ 2)(s + 3) s(s + 2)(s + 3)
2 1
− e−5s 2
(s + 2)(s + 3) s (s + 2)(s + 3)
(s + 1) 2
Y (s) = e−s 2 + −
s (s + 2)(s + 3) (s + 2)(s + 3)

−5s 4 1
e +
s(s + 2)(s + 3) s2 (s + 2)(s + 3)

−s (s + 1) 2 −5s 4s + 1
Y (s) = e 2 + −e
s (s + 2)(s + 3) (s + 2)(s + 3) s2 (s + 2)(s + 3)
(5.12)
We next simplify each term in (5.12) using partial fractions and find
their inverse.
(s + 1) A B C D
Y1 (s) = ≡ + + +
s2 (s + 2)(s + 3) s s2 s + 2 s + 3
=⇒ s + 1 = As(s + 2)(s + 3) + B(s + 2)(s + 3) + Cs2 (s + 3) + Ds2 (s + 2)

1
s=0: 1 = 6B =⇒ B =
6
1
s = −2 : −1 = 4C =⇒ C = −
4
2
s = −3 : −2 = −9D =⇒ D =
9

1 1 2
s = 1 : 2 = 12A + 12 +4 − +3
6 4 9

2
=⇒ 2 = 12A + 2 − 1 +
3
2 1
=⇒ 12A = 1 − =
3 3
1
=⇒ A = .
36
Thus,

(s + 1) 1 1 1 1 1 1 2 1
2
= + − +
s (s + 2)(s + 3) 36 s 6 s2 4 s+2 9 s+3
1 1 1 2
=⇒ L−1 {Y1 (s)} = + t − e−2t + e−3t
36 6 4 9
Thus,

−1 (s + 1) 1 1 1 −2(t−1) 2 −3(t−1)
L e−s 2
= + (t − 1) − e + e u(t − 1)
s (s + 2)(s + 3) 36 6 4 9
(5.13)
From (5.12), we let
2 A B
≡ +
(s + 2)(s + 3) s+2 s+3
=⇒ 2 = A(s + 3) + B(s + 2)
s = −2 : 2=A
s = −3 :
2 = −B =⇒ B = −2
2 2 2
=⇒ = −
(s + 2)(s + 3) s+2 s+3

2
=⇒ L = 2e−2t − 2e−3t (5.14)
(s + 2)(s + 3)
From (5.12), we let
4s + 1 A B C D
≡ + 2+ +
s2 (s + 2)(s + 3) s s s+2 s+3
=⇒ 4s + 1 = As(s + 2)(s + 3) + B(s + 2)(s + 3) + Cs2 (s + 3) + Ds2 (s + 2)

1
s=0: 1 = 6B =⇒ B =
6
7
s = −2 : −7 = 4C =⇒ C = −
4
11
s = −3 : −11 = −9D =⇒ D =
9

1 7 11
s=1: 5 = 12A + (3)(4) − − (4) + (3)
6 4 9
11 4
=⇒ 5 = 12A + 2 − 7 + =−
3 3
4 19
=⇒ 12A = 5 + =
3 3
19
=⇒ A =
36
So,

4s + 1 19 1 1 1 7 1 11 1
= + − +
s2 (s + 2)(s + 3) 36 s 6 s2 4 s+2 9 s+3
4s + 1 19 1 7 11
=⇒ L−1 { }= + t − e−2t + e−3t
s2 (s + 2)(s + 3) 36 6 4 9
Thus,

−1 4s + 1 19 1 7 −2(t−5) 11 −3(t−5)
L e−5s 2
= + (t − 5) − e + e u(t − 5)
s (s + 2)(s + 3) 36 6 4 9
(5.15)
From (5.12) we have

−s (s + 1) 2 −5s 4s + 1
y(t) = e 2 + − e
s (s + 2)(s + 3) (s + 2)(s + 3) s2 (s + 2)(s + 3)
(5.16)
Substituting (5.13)-(5.15) into (5.16), we finally get

1 1 1 −2(t−1) 2 −3(t−1)
y(t) = + (t − 1) − e + e u(t − 1)−
36 6 4 9

19 1 7 −2(t−5) 11 −3(t−5)
+ (t − 5) − e + e u(t − 5) + e−2t − 2e−3t .
36 6 4 9
EXERCISE 3. Solve the IVP using Laplace transforms
y 00 + y = u(t − 2) − u(t − 4)
y(0) = 1, y 0 (0) = 0.

Solution. Taking the Laplace transform of the equation, we get
e−2s e−4s
[s2 Y − sy(0) − y 0 (0)] + Y = −
s s
e−2s e−4s
=⇒ s2 Y − s + Y = −
s s
1 1
=⇒ Y (s)(s2 + 1) = e−2s − e−4s + s
s s
1 1 s
Y (s) = e−2s − e−4s 2 + 2 (5.17)
s(s2 + 1) s(s + 1) (s + 1)
Let
1 A Bs + C
≡ + 2
s(s2 + 1) s s +1
=⇒ 1 = A(s2 + 1) + (Bs + C)s
s=0: 1=A
s=1: 1=2+B+C =⇒ B + C = −1
s = −1 : 1 = 2 − (−B + C) =⇒ B − C = −1
=⇒ 2B = −2 =⇒ B = −1
=⇒ C = −1 + 1 = 0
1 1 s
=⇒ 2
= − 2
s(s + 1) s s +1

1
L−1 2
= 1 − cos t
s(s + 1)

−1 −2s 1
=⇒ L e = [1 − cos(t − 2)]u(t − 2),
s(s2 + 1)
and
−1 −4s 1
L e 2
= [1 − cos(t − 4)]u(t − 4).
s(s + 1)
Also
−1 s
L = cos t
s2 + 1
Hence, from (5.17), we get
y(t) = [1 − cos(t − 2)]u(t − 2) − [1 − cos(t − 4)]u(t − 4) + cos t.


Chapter 6
Fourier Series
6.1 Introduction
103

MATH350

Uploaded by

Copyright:

Available Formats

MATH350

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MATH350

Uploaded by

Copyright:

Available Formats

Differential Equations I

Dr. Joseph K. Ansong

2 c Dr. Joseph K. Ansong

1 Definitions & Terminology 1

2 First Order Differential Equations 9

3 Linear Second Order Differential Equations 25

4 Series Solutions of Differential Equations 49

4.2.4 Form of a Second Linearly Independent Solution . . . . 70

6 Fourier Series 103

4 c Dr. Joseph K. Ansong

Definitions & Terminology

1.1 Introduction and Motivation

The integration constant C can be determined if we know the initial popu-

Figure 1.1: Three solutions of M (t) = Cekt for k = −1.

1.2 Classifications and Terminology

2 c Dr. Joseph K. Ansong

is called the independent variable. It is usually clear from the equation

Definition 1. [Differential Equation]

An ordinary differential equation (ODE) is a differential equation in

The constants appearing in a differential equation are called coefficients or

3 c Dr. Joseph K. Ansong

for a real-valued function f of two variables. Otherwise it is said to be

Equation (1.7) is often preferred to (1.6) for computational and theoretical

1.3 Solution of a differential equation

Definition 2. [Explicit Solution]

Example 1. Show that φ(x) = x − 1 + ce−x is an explicit solution to the

The point is to show that y = φ(x) satisfies the given equation.

4 c Dr. Joseph K. Ansong

Solution. Let y = φ(x) = x − 1 + ce−x

Hence, φ(x) = x − 1 + ce−x is an explicit solution for all real x.

5 c Dr. Joseph K. Ansong

Solution. Applying the fundamental theorem of calculus, we can differenti-

1.4 Applications in different fields

6 c Dr. Joseph K. Ansong

where A = k2 /2 and p1 = 2k1 /k2 + 1 are constants to be determined.

2. Electricity: The time variation of electric charge Q in an electric

3. Psychology: One model in psychology of the learning of a task by an

where y represents the learner’s skill level as a function of time, t, and

4. Black-Scholes equation: The value V (x, t) of an option to by or sell

where the constant r is a prevailing risk-free interest rate, and σ 2 is

5. Predator-Prey model: In ecology, the population dynamics of com-

7 c Dr. Joseph K. Ansong

8 c Dr. Joseph K. Ansong

First Order Differential

2.1 Separable Equations

Definition 4 (Separable Equations). A first order differential equa-

Now integrate both sides:

10 c Dr. Joseph K. Ansong

=⇒ 1 = A(y + 3) + B(y − 3),

2.2 Linear First-Order Equations

12 c Dr. Joseph K. Ansong

Below is a summary of the approach outlined above.

b) Calculate the integrating factor:

c) Multiply (2.10) by µ(x) such that

d) Integrate the equation above to get

Example 9. Solve the following IVP

Using integration by parts, we let

14 c Dr. Joseph K. Ansong

2.3 Exact Equations

Definition 5. Consider the total differential of a continuous function