MTH220

Ordinary Differential Equations

FACULTY OF BUSINESS, JUSTICE & BEHAVIOURAL SCIENCES

Study Guide
202330 (version 2.0)
Last updated: Thursday 11th May, 2023
Ordinary Differential Equations

MTH220 Study Guide

Faculty of Business, Justice & Behavioural Sciences

Written by
L Lobb, B Wilkes, D Demskoy
Produced by School of Computing and Mathematics, Charles Sturt University, Albury -
Bathurst - Wagga Wagga, New South Wales, Australia.
First Published March 2021

©harles Sturt University

Previously published material in this book is copied on behalf of Charles Sturt University
pursuant to Part VB of the Commonwealth Copyright Act 1968.
Contents

Chapter 0: Review of integration techniques . . . . . . . . . . . . . . . . . . . . . 3

0.1 Differentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.2 Integration techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
0.3 Hyperbolic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Chapter 1: First order differential equations . . . . . . . . . . . . . . . . . . . . . 17

1.1 Historical background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.3 Exact differential equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.4 Separable differential equations . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.5 Integrating factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.6 First order linear differential equations . . . . . . . . . . . . . . . . . . . . . 40
1.7 Bernoulli differential equations . . . . . . . . . . . . . . . . . . . . . . . . . . 44
1.8 First order homogeneous differential equations . . . . . . . . . . . . . . . . . 45
1.9 Applications of first order differential equations . . . . . . . . . . . . . . . . 47
1.10 Tutorial 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Chapter 2: Linear differential equations . . . . . . . . . . . . . . . . . . . . . . . 67

2.1 General linear differential equations . . . . . . . . . . . . . . . . . . . . . . . 68
2.2 Solution of linear D.E.s with constant coefficients . . . . . . . . . . . . . . . 69
2.3 Homogeneous solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.4 Method of undetermined coefficients . . . . . . . . . . . . . . . . . . . . . . 73
2.5 Method of variation of parameters . . . . . . . . . . . . . . . . . . . . . . . . 79
2.6 D-operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.7 Shift theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
2.8 Finding yp when R is a polynomial . . . . . . . . . . . . . . . . . . . . . . . . 88
2.9 Applications of linear differential equations . . . . . . . . . . . . . . . . . . . 90
2.10 Tutorial 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

Chapter 3: Infinite series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.2 Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.3 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.4 Series with positive terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.5 Alternating series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.6 Absolute convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.7 Power series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.8 Tutorial 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

Chapter 4: Series solution of differential equations . . . . . . . . . . . . . . . . . 129

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
i
 4.2 Informal series approach to the solution of differential equations . . . . . . 130
4.3 Power series solution about an ordinary point . . . . . . . . . . . . . . . . . . 140
4.4 Series solution about a regular singular point . . . . . . . . . . . . . . . . . . 148
4.5 Summary of approach to the solution of D.E.s by use of series . . . . . . . . 154
4.6 Tutorial 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

Chapter 5: Gamma function (special functions) . . . . . . . . . . . . . . . . . . . 157

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
5.2 Definition of the Gamma function . . . . . . . . . . . . . . . . . . . . . . . . 157
5.3 Tutorial 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

Chapter 6: Laplace transforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
6.2 Definition of Laplace transform . . . . . . . . . . . . . . . . . . . . . . . . . . 165
6.3 Inversion of transforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
6.4 Transforms of derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.5 Solving initial value differential equations . . . . . . . . . . . . . . . . . . . 174
6.6 Transforms of integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
6.7 Shifting theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
6.8 Tutorial 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

Chapter 7: Solutions to selected tutorial questions . . . . . . . . . . . . . . . . . 185

7.1 Tutorial 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.2 Tutorial 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
7.3 Tutorial 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
7.4 Tutorial 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
7.5 Tutorial 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270

Chapter 8: Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

ii
 1

Comments on individual Study Guide chapters

Chapter 0 contains a review of hyperbolic functions and integration techniques which you
studied in the previous calculus subjects (MTH101 and MTH102). We use these techniques
extensively in MTH220. You can skip this chapter if you recently completed MTH101 and
MTH102.
First Order Differential Equations are considered in Chapter 1. Your main task here
is to learn to classify the type of differential equation (exact, separable, linear, Bernoulli,
homogeneous), being considered so that you can select and then apply (correctly!) the ap-
propriate method. Pay particular attention to the form the equation should be put in so that
you can determine its classification. In the applications to Biology and Electrical Circuits,
you are not expected to master all the scientific concepts. Instead, you are expected to be
able to carry out the appropriate mathematical procedures for problems whose formulation
and model will be provided.
Chapter 2 considers higher order DEs of one particular type - linear - that are common in
applications and can be (reasonably) easily solved using standard methods. Your task here
is to learn how to carry out these standard methods virtually as a “recipe” which must be
applied correctly and in the right sequence lest the solution is a flop!
Chapter 3 on Infinite Series is one that some students find difficult, so try to de-vote enough
concentrated time to it! Sequences are not really important. Series and their existence
(“convergence” or “divergence”) are the essential topic. Again, as for integrals and DEs, the
question of classification of the type of series is vital before you apply the appropriate test
to ascertain whether the series “makes sense” i.e. has a fixed finite value (i.e. converges).
Note: Interestingly, the “existence” (or not) of a value can often be more easily determined
than the finding of the actual value. [Numerical methods studied later in your course often
are needed for finding the value (or sum) of a series.] The Section on power series is very
important, especially the work on Taylor series which are essential tools in opening up
numerical methods and computational methods.
In Chapter 4, series are then applied to the solution of certain differential equations which
cannot normally be solved by the methods developed in Chapters 1 and 2. Chapter 5 should
be examined carefully so as to see the “pattern” in the method of solution. (Note carefully
what work is required for the examination and what is for information and later reference,
if needed.)
Chapter 5 considers some Special Functions which are widely used in later mathematical
studies. Try to examine the “patterns” that develop in this work.
Chapter 6 considers Laplace Transforms which offer an alternative way of solving Differ-
ential Equations. Their applications may be crucial in your later studies. Now you should
be ready to tackle the next stage of your mathematical development. Relax and enjoy the
show - oh yes, and do some work, too!
2
Chapter 0

Review of integration techniques

0.1 Differentials
If y = f (x) then
dy
= f ′ (x)
dx
and we could write
dy = f ′ (x)dx
but
dy = d[f (x)]
so we have the rule
d[f (x)] = f ′ (x)dx
Example 1

d
d[sin x] = (sin x)dx = cos xdx
dx
Example 2

d (x3 ) = 3x2 dx

Function of a function

Suppose y = f (z) where z = g(x). Then by the chain rule

dy df dz dz
= = f ′ (z)
dx dz dx dx
Then
′ dz ′
∫ f (z) dx dx = ∫ f (z)dz = f (z) + C.
Example 1

2 2
∫ 2x cos x dx = ∫ cos x ⋅ 2xdx.
3
4 CHAPTER 0. REVIEW OF INTEGRATION TECHNIQUES

Let z = x2 then dz = 2xdx giving

2 2
∫ 2x cos x dx = ∫ cos zdz = sin z + C = sin x + C

Example 2
ln x 1
∫ dx = ∫ ln x × dx
x x
and we note that
1
d(ln x) = dx
x
giving
ln x 1
∫ dx = ∫ ln xd[ln x] = (ln x)2 + C.
x 2

Changing variables

Similar to previous, but any variable formula can be tried. There may be more than one
way to work the integral. Trying to find the easiest substitution can be interesting. Given

∫ f (x)dx

if we try x = g(z) then dx = d[g(z)] = g ′ (z)dz giving

′
∫ f (x)dx = ∫ f [g(z)]g (z)dz

Example 1

Calculate
1
∫ √ dx
x+1
Try substitution z = x + 1 or x = z − 1 then
1 1
dx = dz; √ = √ = z −1/2
x+1 z
giving

1 −1/2 z −1/2+1 z 1/2 √ √

∫ √ dx = ∫ z dz = + C = + C = 2 z + C = 2 x+1+C
x+1 −1/2 + 1 1/2

We could also try x + 1 = z 2 so x = z 2 − 1 giving dx = 2zdz and

1 1 1
√ =√ =
x+1 z2 z
giving
1 1 √
∫ √ dx = ∫ ⋅ 2zdz = ∫ 2dz = 2z + C = 2 x + 1 + C
x+1 z
as above.
Note: the limits of integration can also be changed accordingly.
0.2. INTEGRATION TECHNIQUES 5

Example 2

Suppose the previous integral was actually

3 1
∫0 √ dx
x+1
then we have √ √ √
[2 x + 1]30 = 2 3 + 1 − 2 1
√
=2 4−2×1=2×2−2=4−2=2
Alternatively using z = x + 1 we could have adjusted the limits as z(0) = 0 + 1 = 1 and
z(3) = 3 + 1 = 4 giving
√ √ √
[2 z]41 = 2 4 − 2 1 = 2 × 2 − 2 × 1 = 4 − 2 = 2

again.
Example 3

Calculate π
2
∫0 sin x cos xdx.

Try z = sin x then dz = d[sin x] = cos xdx; z(0) = sin 0 = 0 and z ( π2 ) = sin π2 = 1 giving
π
2
1 1 1 1 1 1
∫0 sin x cos xdx = ∫ zdz = [ z 2 ] = × 12 − × 02 = .
0 2 0 2 2 2
Note: We can check this using sin 2x = 2 sin x cos x. Then
π π π
2 1 2 1 2

∫0 sin x cos xdx = ∫ sin 2xdx = [− cos 2x]

2 0 4 0

1 1 1 1 1 1 1
= − cos π + cos 0 = − × (−1) + × 1 = + =
4 4 4 4 4 4 2
as expected.

0.2 Integration techniques

Integration by parts

Consider the product rule

d dv du
(uv) = u + v
dx dx dx
we can write this as
d(uv) = udv + vdu
Integrating this gives

∫ d(uv) = uv(by definition) = ∫ udv + ∫ vdu

6 CHAPTER 0. REVIEW OF INTEGRATION TECHNIQUES

and we can rearrange this as

∫ udv = uv − ∫ vdu.
This is known as integration by parts.

Example 1
Calculate
x
∫ xe dx.
Notice that the u in the first integral appears on the RHS differentiated. The method can
be repeated. This means that if we choose u as a polynomial (if it is possible) we will
eventually obtain a du of a constant (giving 0 ) and the method will terminate.
So we choose dv = ex dx then v = ∫ ex dx = ex ; u = x giving du = dx and we obtain

x x x x x
∫ xe dx = ∫ udv = uv − ∫ vdu = xe − ∫ e dx = xe − e + C

Let’s check this:

d
(xex − ex + C) = xex + ex − ex + 0 = xex . Amazing!
dx
Note: we could have used differentials and written the by-parts step as
x x x x x x x x
∫ xe dx = ∫ xd [e ] = xe − ∫ e d[x] = xe − ∫ e dx = xe − e + C

Example 2

Calculate
∫ ln xdx.
This is already in the form ∫ udv so we simply choose ln x = u and dv = dx so v = x and
d 1
du = (ln x)dx = dx
dx x
giving
1
∫ ln xdx = uv − ∫ vdu = (ln x) × x − ∫ x × x dx = x ln x − ∫ 1dx = x ln x − x + C

Check:
d 1
(x ln x − x + C) = ln x + x × − 1 + 0 = ln x + 1 − 1 = ln x. Amazing!
dx x
Note. Using differentials we could have done the by-parts step as

∫ ln xdx = x ln x − ∫ xd[ln x]

and so on.

Example 3 . Regenerating the integral on the RHS.

Consider
x
∫ e sin xdx.
0.2. INTEGRATION TECHNIQUES 7

Try u = sin x and dv = ex dx.

Then
v = ∫ ex dx = ex
giving
x x x x x
∫ e sin xdx = e sin x − ∫ e d[sin x] = e sin x − ∫ e cos xdx

= ex sin x − ∫ cos xd [ex ] = ex sin x − ex cos x + ∫ ex d[cos x]

= ex sin x − ex cos x − ∫ ex sin xdx
The integral has appeared again on the RHS. We then have

I = ex sin x − ex cos x − I

So
2I = ex (sin x − cos x)
giving
ex
I = ∫ ex sin xdx = (sin x − cos x).
2
Check:
d ex ex ex
[ (sin x − cos x)] = (sin x − cos x + cos x + sin x) = × 2 sin x = ex sin x. Yes!
dx 2 2 2

Powers of trigonometric functions

By parts, it can be shown that

n sinn−1 x cos x n − 1
∫ sin xdx = − + sinn−2 xdx
n n ∫
and
n cosn−1 x sin x n − 1
∫ cos xdx = + cosn−2 xdx
n n ∫
Mixtures:
n m
∫ sin x cos xdx
If n and/or m are odd; simply let z = cos x or z = sin x if n or m is odd. If n and m are both
even, then use the identities
1
sin2 x = (1 − cos 2x)
2
and
1
cos2 x = (1 + cos 2x).
2
Example 1

Calculate
3
∫ sin xdx.
Using n = 3 gives

n sin3−1 x cos x 3 − 1
∫ sin xdx = − + sin3−2 xdx
3 3 ∫
8 CHAPTER 0. REVIEW OF INTEGRATION TECHNIQUES

sin2 x cos x 2 1 2
=− + ∫ sin xdx = − sin2 x cos x − cos x + C
3 3 3 3
Example 2
2 3
∫ sin x cos xdx
3 is odd, so try z = sin x and dz = d[sin x] = cos xdx giving

2 3 2 2 2 2 2 4
∫ sin x cos xdx = ∫ sin x cos x cos xdx = ∫ z (1 − z ) dz = ∫ z − z dz

1 1 1 1
= z 3 − z 5 + C = sin3 x − sin5 x + C
3 5 3 5
Example 3

2
2 4 1 1
∫ sin x cos xdx = ∫ 2 (1 − cos 2x) [ (1 + cos 2x)] dx
2
1 1
= ∫ (1 − cos2 2x) (1 + cos 2x)dx = ∫ sin2 2x(1 + cos 2x)dx
8 8
1 1 1 1 1
= ∫ sin2 2x + sin2 2x cos 2xdx = ∫ (1 − cos 4x)dx + ∫ sin2 2x × d[sin 2x]
8 8 2 8 2
1 1 1
= x− sin 4x + sin3 2x + C
16 64 48

√
Integrals involving a2 ± x2 , a2 ± x2

These integrals can often be evaluated using a trigonometric substitution.

Identities: 1 − sin2 θ = cos2 θ; 1 + tan2 θ = sec2 θ; sec2 θ − 1 = tan2 θ

Example 1

Calculate
dx
∫ √ 2 .
a − x2
Try x = a sin θ then dx = a cos θdθ giving
a cos θdθ a cos θdθ a cos θdθ cos θdθ
I =∫ √ =∫ √ =∫ √ =∫
a2 − (a sin θ)2 a2 − a2 sin2 θ a 1 − sin2 θ cos θ

x x
= ∫ dθ = θ + C = sin−1 ( ) + C = arcsin ( ) + C.
a a
Example 2

Calculate
dx
∫ a2 + x 2 .
Try x = a tan θ then dx = a sec2 θdθ giving
a sec2 θdθ a sec2 θdθ sec2 θdθ 1 1
I =∫ 2 2 2 = ∫ 2 2 = ∫ 2
= ∫ dθ = θ + C
a + a tan θ a (1 + tan θ) a sec θ a a
0.2. INTEGRATION TECHNIQUES 9

1 x 1 x
= arctan ( ) + C = arctan ( ) + C
a a a a

Example 3

Calculate
xdx
∫ √ 2 .
x −4
Try x = 2 sec θ then dx = 2 sec θ tan θdθ giving

2 sec θ ⋅ 2 sec θ tan θdθ sec2 θ tan θdθ 4 sec2 θ tan θdθ
I =∫ √ = 4∫ √ =√ ∫
(2 sec θ)2 − 4 4 (sec2 θ − 1) 4 tan θ
4 √ √
= ∫ sec2 θdθ = 2 tan θ + C = 2 tan2 θ + C = 2 sec2 θ − 1 + C
2√
x 2 √
= 2 ( ) − 1 + C = x2 − 4 + C
2

Integrals involving ax2 + bx + c

Many integrals of this type can be evaluated by completing the square

Example
dx dx dx
∫ x2 + 2x + 5 = ∫ x2 + 2x + 1 + 4 = ∫ (x + 1)2 + 22

This is of the general form a1 arctan ( ua ) + C giving

1 x+1
arctan ( ) + C(a = 2)
2 2

Integration by partial fractions

(x)
Given a rational function fg(x) where the degree of f (x) is less than g(x) [else divide] and
the factors of g(x) are known, we break the fraction into parts as follows:

f (x) A1 A2 Am
= + + ... +
(x − r1 ) (x − r2 ) ⋯ (x − rm ) x − r1 x − r2 x − rm

For repeated roots

f (x)
Consider (x−r)n . Now f (x) must be smaller in degree so we write f (x) as a polynomial
about r :
f (x) = A1 (x − r)n−1 + A1 (x − r)n−2 + . . . + An .
This can always be done. Then we obtain

A1 (x − r)n−1 + A2 (x − r)n−2 + ⋯ + An A1 (x − r)n−1 A2 (x − r)n−2 An

n
= n
+ n
+ ... +
(x − r) (x − r) (x − r) (x − r)n
A1 A2 An
= + 2
+ ... + .
x − r (x − r) (x − r)n
10 CHAPTER 0. REVIEW OF INTEGRATION TECHNIQUES

Quadratic terms
These theoretically will reduce to

A1 A2
+
x − r1 x − r2
but this may be messy if we have something like
1
x2 + 4x + 6
which does not have real factors. Instead, add
A1 A2 A1 (x − r2 ) + A2 (x − r1 ) (A1 + A2 ) x + (−A1 r2 − A2 r1 )
+ = =
x − r1 x − r2 (x − r1 ) (x − r2 ) (x − r1 ) (x − r2 )

which suggests trying Ax + B in the numerator. Similarly, for a cubic term, we would try
Ax2 + Bx + C in the numerator. And so on.

Example 1

Calculate
3x − 4 3x − 4
∫ x2 − 3x + 2 dx = ∫ (x − 1)(x − 2) dx.

Try
3x − 4 A B
= + .
(x − 1)(x − 2) x − 1 x − 2
This gives 3x − 4 = A(x − 2) + B(x − 1). By equating coefficients

3 = A + B and − 4 = −2A − B

Adding: 3 − 4 = −1 = A + B − 2A − B = −A and A = 1 giving B = 2.

Note.
Setting x = 2 gives
(3x − 4)∣x=2 = B(x − 1)∣x=2
or
3x − 4 3×2−4 6−4
B= ∣ = = =2
x − 1 x=2 2−1 1
Similarly we would find that
3x − 4 3 × 1 − 4 3 − 4 −1
A= ∣ = = = =1
x − 2 x=1 1−2 −1 −1
as before. However, this gives a neat method which works well for single linear factors.
Use the coefficient method or number substitution for other factors.
Further
3x − 4 1 2
= +
x2 − 3x + 2 x − 1 x − 2
giving
1 2
I =∫ ( + ) dx = ln(x − 1) + 2 ln(x − 2) + C
x−1 x−2
0.2. INTEGRATION TECHNIQUES 11

Example 2

Calculate
x2 − x + 4
∫ (x − 1)(x + 1)2 dx.

Try
x2 − x + 4 A B C
2
= + + .
(x − 1)(x + 1) x − 1 x + 1 (x + 1)2
For the single linear factor x − 1 we can use the quick method:

x2 − x + 4 12 − 1 + 4 1 + 3 4
A= ∣ = = 2 = =1
(x + 1)2 x=1 (1 + 1)2 2 4

Finding B and C is a bit trickier.

x2 − x + 4 = A(x + 1)2 + B(x + 1)(x − 1) + C(x − 1)

= (x + 1)2 + B (x2 − 1) + C(x − 1) = x2 + 2x + 1 + Bx2 − B + Cx − C

which simplifies to give
−3x + 3 = Bx2 + Cx − B − C.
We see immediately that B must be zero and C must be -3 giving

x2 − x + 4 1 3 3
∫ (x − 1)(x + 1)2 dx = ∫ ( x − 1 − (x + 1)2 ) dx = ln(x − 1) + x + 1 + C.

Example 3

Calculate
x2 − 4
∫ (x + 2) (x2 + 5) dx.

We try
x2 − 4 A Bx + C
2
= + 2
(x + 2) (x + 5) x + 2 x +5
Again,
x2 − 4 (−2)2 − 4 4 − 4
A= ∣ = = =0
x2 + 5 x=−2 (−2)2 + 5 4 + 5
Then we have x2 − 4 = (Bx + C)(x + 2) giving straight away B = 1, 2C = −4 so C = −2
and we obtain
1
x−2 × 2x dx
I =∫ 2 dx = ∫ 2 2 dx − 2 ∫ √
x +5 x +5 x2 + ( 5)2

1 d (x2 + 5) 2 x
= ∫ 2
− √ arctan ( √ ) + C
2 x +5 5 5

1 2 x
= ln (x2 + 5) − √ arctan ( √ ) + C.
2 5 5
12 CHAPTER 0. REVIEW OF INTEGRATION TECHNIQUES

Rational functions of sin x, cos x

Try
x
z = tan
2
which leads to the substitution formulae:
2z 1 − z2 2dz
sin x = 2
, cos x = 2
, dx = .
1+z 1+z 1 + z2
Example

dx 1 2dz 2dz
∫ 2 + cos x = ∫ 2 + (1 − z 2 ) / (1 + z 2 ) ⋅ 1 + z 2 = ∫ 2 (1 + z 2 ) + 1 − z 2

2dz 2dz 2dz 2 z

=∫ 2 2
=∫ 2 =∫ √ = √ arctan ( √ ) + C
2 + 2z + 1 − z z +3 2
z + ( 3) 2 3 3
⎡ tan ( x ) ⎤
2 ⎢ ⎥
= √ arctan ⎢⎢ √ 2 ⎥⎥ + C
3 ⎢
⎣ 3 ⎥⎦
(a bit messy but that’s probably all you can do with it!)

Infinite limits
∞ R
∫a f (x)dx = lim ∫ f (x)dx
R→∞ a

If the limit exists, and is finite, the improper integral is said to converge, otherwise it is
divergent.

Example 1

∞ dx R dx R π
∫0 2
= lim ∫ 2
= lim [tan−1 x]0 = lim tan−1 R − 0 =
1+x R→∞ 0 1+x R→∞ R→∞ 2
and the integral is convergent.

Example 2
∞
dx R dx
∫1 = lim ∫ = lim [ln x]R
1 = lim ln R = ∞
x R→∞ 1 x R→∞ R→∞

and the integral is divergent.

0.3 Hyperbolic functions

Definitions
eu − e−u eu + e−u sinh u eu − e−u
sinh u = , cosh u = , tanh u = = ;
2 2 cosh u eu + e−u
1 cosh u eu + e−u
coth u = = = ;
tanh u sinh u eu − e−u
0.3. HYPERBOLIC FUNCTIONS 13

1 2
sech u = = u −u ;
cosh u e + e
1 2
cosech u = = u −u .
sinh u e − e
Connection with trigonometric functions
From Euler’s formula

eiθ = cos θ + i sin θ

and
e−iθ = cos θ − i sin θ

giving
eiθ − e−iθ
sinh iθ = = i sin θ
2
and
eiθ + e−iθ
cosh iθ = = cos θ
2
and
sinh iθ i sin θ
tanh iθ = = = i tan θ
cosh iθ cos θ
So
1
sin θ = sinh iθ; cos θ = cosh iθ
i
and
1
tan θ = tanh iθ.
i

Identities
From
sin2 θ + cos2 θ = 1

we find
− sinh2 iθ + cosh2 iθ = 1

or
cosh2 u − sinh2 u = 1

Similarly
1 − tanh2 u = sech2 u

1 − coth2 u = − cosech2 u

sinh(x + y) = sinh x cosh y + cosh x sinh y

cosh(x + y) = cosh x cosh y + sinh x sinh y

14 CHAPTER 0. REVIEW OF INTEGRATION TECHNIQUES

Derivatives
d
(sinh x) = cosh x
dx
d
(cosh x) = sinh x
dx
d
(tanh x) = sech2 x
dx
d
(coth x) = − cosech2 x
dx
d
(sech x) = − sech x tanh x
dx
d
(cosech x) = − cosech x coth x
dx
Inverse hyperbolic functions
√
sinh−1 x = ln (x + x2 + 1) −∞<x<∞
√
cosh−1 x = ln (x + x2 − 1) ∣x∣ > 1
1 1+x
tanh−1 x = ln ( ) ∣x∣ < 1
2 1−x
1
sech−1 x = cosh−1 ( ) 0 < x ≤ 1
x
1
cosech−1 x = sinh−1 ( ) x ≠ 0
x
1
coth−1 x = tanh−1 ( ) ∣x∣ > 1
x
Derivatives of inverse hyperbolic functions
d 1
(sinh−1 x) = √
dx 1 + x2
d 1
(cosh−1 x) = √
dx x2 − 1
d 1
(tanh−1 x) = ∣x∣ < 1
dx 1 − x2
d 1
(coth−1 x) = ∣x∣ > 1
dx 1 − x2
d
(sech−1 x) = √
−1
dx x 1 − x2
d
(cosech−1 x) = √
−1
dx ∣x∣ 1 + x2
Example

(done before using x = 2 sec θ)

xdx
∫ √ 2
x −4
Now let’s try an hyperbolic substitution which should work also. Trying x = 2 cosh u so
dx = 2 sinh udu, then we obtain
2 cosh u ⋅ 2 sinh udu cosh u sinh udu 4 cosh u sinh udu
I =∫ √ = 4∫ √ =√ ∫
2
(2 cosh u) − 4 4 (cosh2 u − 1) 4 sinh u
0.3. HYPERBOLIC FUNCTIONS 15

4 √ √
= ∫ cosh udu = 2 sinh u + C = 2 sinh u + C = 2 cosh2 u − 1 + C
2
2
√
x 2 √
= 2 ( ) − 1 + C = x2 − 4 + C
2
as before.
16 CHAPTER 0. REVIEW OF INTEGRATION TECHNIQUES
Chapter 1

First order differential equations

Chapter objectives

In this chapter you will learn to:

• distinguish between exact and non-exact differential equations;

• use integrating factors to solve exact differential equations;

• solve separable and first order homogeneous differential equations;

• solve first order linear and Bernoulli differential equations; and

• apply differential equations to certain real-life problems.

Chapter summary

1. A brief history of differential equations

2. Exact differential equations

3. Separable differential equations

4. Integrating factors

5. First order linear differential equations

6. Bernoulli differential equations

7. First order homogeneous differential equations

8. Applications of first order differential equations

17
18 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

1.1 Historical background

The origin of the branch of mathematics known as differential equations dates back at least
to 1671 and Newton’s classification of first order ordinary differential equations into three
classes. Sir Isaac Newton (1642-1727) called these equations fluxional equations instead of
differential equations. He assumed that the solution of these equations could be expressed
as an infinite series, and he successively determined the coefficients in a manner similar to
the technique employed today. However, he did not consider the convergence of the series.
The notation .y for the derivative of y with respect to the independent variable was also
introduced by Newton.
Gottfried Leibniz (1646-1716) invented the differential notation dy and the symbol ∫ for
integration. To our knowledge he first used these notations in conjunction in 1675, when
he wrote:
1 2
∫ ydy = 2 y .
In 1676 Leibniz used the term “differential equation” to denote a relationship between two
differentials dx and dy. Thus was christened the branch of mathematics that deals with
equations which involve differentials or derivatives. The word integral was introduced into
mathematics in 1690 by Jacques Bernoulli (1654-1705). In 1691 Leibniz discovered the
technique of separation of variables and in 1692 he reduced the linear homogeneous first
order differential equation to quadratures. Jean Bernoulli (1667-1748) introduced the con-
cept of an integrating factor in 1694 and the technique of changing the dependent variable.
By the end of the seventeenth century the techniques which are usually employed when at-
tempting to solve first order ordinary differential equations were known. These techniques
often prove to be inadequate.
However, early in the development of the study of differential equations, it was believed that
elementary functions would be sufficient for the representation of solutions of differential
equations arising from geometry and mechanics. Thus, early attempts at solving differential
equations were directed toward finding explicit solutions or reducing the solution to a finite
number of quadratures. By 1723 at the latest it was recognised that even some first order
ordinary differential equations do not have solutions which can be expressed in terms of
elementary functions. As a matter of fact, if an ordinary differential equation is written
down at random, the probability of being able to write the solution in terms of known
functions or their integrals is nearly zero. This emphasises the necessity for developing
methods for obtaining approximate solutions.
In 1739 Léonard Euler (1707-1783) introduced the method of variation of parameters. Jean
Bernoulli had unsuccessfully attempted to solve the general linear homogeneous differen-
tial equations with constant coefficients. Euler gave a complete discussion of this problem
in 1743. He also devised the classical method for solving non-homogeneous linear differ-
ential equations.
No adequate discussion of differential equations as a unified topic existed prior to the lec-
tures developed and presented by Augustin-Louis Cauchy (1789-1857) in the 1820s. In
these lectures Cauchy developed the first existence and uniqueness theorems for first order
differential equations. Cauchy extended his theory to a system in n first order differential
equations in n dependent variables which was equivalent to a single nth order differential
equation. Rudolf Lipschitz (1832-1903) generalised Cauchy’s existence and uniqueness
theorems in 1876. Émile Picard (1856-1941) improved upon the theorems of Cauchy and
Lipschitz in 1893 by introducing the method of successive approximations.
1.2. INTRODUCTION 19

In the past 100 years, work on differential equations has gone on rapidly on many fronts.
The theoretical base of the topic has been strengthened; the set of classes into which dif-
ferential equations can be classified has been enlarged; the use of numerical methods has
been developed greatly, especially with the availability of the computer; the investigation
of partial differential equations has grown; the range of applications of differential equa-
tions has been extended and so on. Like all of mathematics, there has been an explosion
of knowledge. Today no one can be a master of the whole topic as could Euler, Laplace,
Cauchy.
Let us in this subject be content if we can master some of the first steps in the whole field
of Differential Equations!

1.2 Introduction

Read textbook sections 1.1 and 1.2 (Background, Solutions and initial value
problems)

Video (by Dr. R. Wood): First order ODEs. Sections 1.2-1.5

The subject of differential equations has, perhaps, wider applications than any other part
of calculus. Differential equations are used in such areas as astronomy, biology, chemistry,
economics, engineering, oceanography, psychology and physics. In mathematics itself the
subject is important and challenging: our knowledge is still far from complete.
In this subject we will study some methods for solving some of the ordinary differential
equations (henceforth abbreviated to D.E.s) which occur in real-life applications.
Ordinary differential equation.
An ordinary differential equation is an equation involving an unknown function (usually
denoted by “y”) and some of its derivatives with respect to the independent variables (usu-
ally “x”). (The other class of D.E.s consists of partial differential equations and is beyond
the scope of this subject.)
The order of a differential equation is taken to be the order of the highest derivative present
in the equation. We will examine in this Chapter only a number of types of first order D.E.s,
i.e. equations involving an unknown function, y say, its first derivative y ′ , and perhaps other
given functions of x.
(In Chapter 2 we will consider the special class of linear D.E.s of the second (and higher)
order, while in Chapter 4 we will develop ways of using series to solve certain other types
of D.E.s.)
Note: It is crucial that you realise at the start that the main problem in solving a given D.E.
is classifying the type of that D.E. Once that is done correctly, then the method of solution
will follow relatively easily. Thus, whenever a new type of D.E. is considered, be sure to
note very carefully the characteristics of that type so that you can recognise it quickly and
correctly!
The solution of a D.E.
Any relation between the variables that occur in a D.E., that satisfies the equation, is called
a solution of the D.E. A solution of a D.E. of order n can be shown always to include n
20 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

arbitrary constants. Such a solution is called a general solution.

Any solution that uses additional information (often called “initial conditions” or “boundary
conditions”) to determine specific values for the constants is called a particular solution.
Note: It is always (theoretically) possible to check a solution to a D.E. by substitution in
the original equation. Whenever we are developing a new “strange” method in these notes,
you should seek assurance that the method “works” by checking the solution it yields!
Integration
The simplest differential equation is given by
dy
= f (x),
dx
where f is a given function and we want to solve for the unknown function y. You can
already solve such a D.E. - it is, in fact, another statement of the central topic of this
subject, namely integration! The solution of this equation is given by

y = ∫ f (x)dx,

and can be found using one of the methods of integration.

For example, solve
dy
= 3x4 .
dx
We have that y = 3x4 dx and so y = 35 x5 + C. (One arbitrary constant since the D.E. is of
order one.)
dy
Note: This method will not work for the D.E.: dx = 2y.
We will show shortly that the general solution to this is y = Ce2x . For now we can check
this solution simply by noting that
dy
= 2Ce2x = 2y,
dx
as required.
Again, we can show that x2 + y 2 = 16 is the particular solution (given implicitly) of the
D.E.:
dy x
=− ,
dx y
with the initial condition that y = 4 when x = 0. (Sometimes written y(0) = 4.)
By differentiating x2 + y 2 = 16 implicitly we get that
dy
2x + 2y =0
dx
i.e.
dy x
=− .
dx y
2 2
Thus x + y = 16 is a solution of the given D.E. and clearly x = 0 and y = 4 satisfy this
equation, making it the particular solution for the given initial condition.
Solution of any given D.E. For the general solution of any given D.E., there are three
questions that arise:
1.3. EXACT DIFFERENTIAL EQUATIONS 21

1. Does a solution exist?

2. If so, is there more than one solution?

3. What method do we use to determine the solution, if one exists?

The answer to the first two questions is given by the following

Theorem. Consider the initial value problem

dy
= f (x, y), y (x0 ) = y0 .
dx
If f and ∂f /∂y are continuous functions in some rectangle

{(x, y) ∶ a < x < b, c < y < d}

that contains the point (x0 , y0 ), then the initial value problem has a unique solution in some
interval containing x0 .
Fortunately, for the specific types of “basic” (but very useful) D.E.s that we will examine in
this subject, we can always assume (correctly) that a solution exists. Thus the third question
is the only important one for us: we must develop and learn the appropriate methods for
solving such D.E.s.

1.3 Exact differential equations

Read textbook sections 2.4 (Exact equations)

Before proceeding to examine exact D.E.s, we will consider very briefly the problem of
finding the “derivative” (the partial derivative) of a function of more than one variable.
Thus far in Calculus, we have dealt only with functions of one variable. Such functions
may be given explicitly in the form of y = f (x) or implicitly by an equation of the form
F (x, y) = 0. In both cases, x is taken to be the independent variable and y is called the
dependent variable. Thus y varies as x varies, and we evaluate the rate of change of y with
respect to (w.r.t.) x, using the derivative, In general, it is possible to consider functions of
two (or more) variables which can be written explicitly in the form

z = f (x, y).

Note that here x and y are both independent variables, and that z is the dependent variable.
Thus z varies as we vary x and/or y independently, and we will again want to evaluate the
rate of change of z, but this time (separately) with respect to x or w.r.t. y. The notation
∂z
used for these two rates of change is ∂x ∂z
and ∂y (The notations: ∂f ∂f
∂x and ∂y or fx and fy are
also used sometimes.)
Some remarks
In general, functions of more than one variable are very important since real-life models
we want to use often have to take into account a number of “factors” or variables.
22 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

For example, the distance travelled by a car depends on both the time and the acceleration,
i.e. we have s = f (a, t) Similarly, the volume of gas depends on the pressure and the
temperature, i.e. V = g(P, T ) The volume of a circular cylindrical tank is a function of the
radius and the height, i.e. V = F (r, h).
Such functions are studied in detail in the subject MTH218 Multivariable Calculus. There
we will define the concepts of “limit” and “continuity” for such functions; we will try to
draw the surface represented by z = f (x, y); we will try to find the tangent plane to such
a surface; we will examine maxima/minima problems for such functions, and so on!
For now, we only need to consider the question of finding the partial derivatives of a
∂z
function given by z = f (x, y). We say that ∂x represents the rate of change of z (or “f ”)
with respect to x, keeping y constant, and define this “partial derivative of z w.r.t. x” as

∂z f (x + h, y) − f (x, y)
= lim
∂x h→0 h
Similarly,
∂z f (x, y + h) − f (x, y)
= lim
∂y h→0 h
is the “partial derivative of z w.r.t. y” and represents the rate of change of z w.r.t. y, keeping
x constant.
Of course, as usual, we do not use these definitions to evaluate partial derivatives! Instead
we simply differentiate according to all the ordinary rules for differentia-tion, keeping all
∂ ∂
the other (independent) variables constant. i.e. ∂x y = 0 and ∂y x = 0, and so on for more
than 2 variables.

Example 1

∂z ∂z
Let z = x2 + y 4 + 3. Find and .
∂x ∂y
Solution
∂z ∂ 4 ∂
Here, = 2x. (Note that y and 3 both equal 0, since both are constant w.r.t. x.)
∂x ∂x ∂x
∂z
= 4y 3 .
∂y

Example 2

Let z = xy 2 − 3x2 + x2 sin y.

∂z ∂z
Find and .
∂x ∂y
Solution
z = xy 2 − 3x2 + x2 sin y
∂z ∂ ∂ ∂
= (xy 2 ) − (3x2 ) + (x2 sin y)
∂x ∂x ∂x ∂x
= y 2 − 6x + 2x sin y.
Again, note that “y 2 ” and “sin y” are “constants” and are treated as such. (Note, too, that
partial differentiation of a function of 2 independent variables is entirely different to
implicit differentiation of a function of 1 independent variable!)
1.3. EXACT DIFFERENTIAL EQUATIONS 23

Similarly,
∂z
= 2xy + x2 cos y.
∂y

Example 3

Find
∂ x2 + y
{ }
∂x cos (x3 y)
24 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Solution

∂ ∂
∂ x2 + y cos (x3 y) ∂x (x2 + y) − (x2 + y) ∂x cos (x3 y)
( ) =
∂x cos (x3 y) cos2 (x3 y)
2x cos (x3 y) + 3x2 y (x2 + y) sin (x3 y)
=
cos2 (x3 y)

Example 4

Let
√ 1
f (u, r, t) = u2 + r2 − 3ur t + 2
t
∂f ∂f ∂f
Find , and .
∂u ∂r ∂t
Solution
Here we have a function of three independent variables and hence we can find three (first)
partial derivatives.

∂f √
∴ = 2u − 3r t
∂u
∂f √
= 2r − 3u t
∂r
∂f 3 2
= − urt− 2 − 3 .
1

∂t 2 t
Note
Whole new fields of investigation and applications open up from the consideration of such
functions and their partial derivatives. We will not make much use of these techniques
in this subject. The topic has really been introduced to demonstrate how the concept of
“derivative” and the process of “differentiation” can be “generalised”, and to enable us to
consider exact D.E.’s.

Tutorial 1 Question 1
1.3. EXACT DIFFERENTIAL EQUATIONS 25

Higher order derivatives

Higher order derivatives can also be defined for functions of more than one variable. If
z = f (x, y), then we can find two first order partial derivatives:
∂z ∂z
and
∂x ∂y
and four second order partial derivatives:
∂ 2z ∂ 2z
,
∂x2 ∂y 2
and the 2 “mixed partials”
∂ 2z ∂ ∂z
≡ ( )
∂x∂y ∂x ∂y
and
∂ 2z ∂ ∂z
≡ ( ),
∂y∂x ∂y ∂x
and so on.

Note
In Multivariable Calculus we will find that for many functions of two variables, the mixed
partial derivatives of the same order and “type” are equal. That is,
∂ 2f ∂ 2f
=
∂x∂y ∂y∂x
for such functions.

Example 1

Find the second order partial derivatives of z = x3 y − cos y + 2x − 7.

Solution
Here, z = x3 y − cos y + 2x − 7.
We must start by finding the first partial derivatives. That is,
∂z ∂z
= 3x2 y + 2; = x3 + sin y
∂x ∂y
∂ 2z ∂
= (3x2 y + 2) = 6xy
∂x2 ∂x
∂ 2z ∂
= (x3 + sin y) = cos y
∂y 2 ∂y
∂ 2z ∂
= (x3 + sin y) = 3x2
∂x∂y ∂x
∂ 2z ∂
= (3x2 y + 2) = 3x2
∂y∂x ∂y
Note that
∂ 2z ∂ 2z
= ,
∂y∂x ∂x∂y
26 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

as “predicted”.
y+1
(Carry out the same calculation for z = 3x4 y 2 + .)
x2 − 2
Total derivative
Another important definition is that of the total derivative. The total derivative of z(x, y),
a function of both x and y, is written dz, and is defined by

∂z ∂z
dz = dx + dy.
∂x ∂y

Example

If z = x3 y + y 2 + 2, then dz = 3x2 ydx + (x3 + 2y)dy.

If we consider a function C, constant with respect to x and y, then

∂C ∂C
=0= .
∂x ∂y
Thus for such a function,
dC = 0.
If we have the equation
z(x, y) = C (C a constant),
the total derivative of both sides can be found, and these must be equal.

dz = dC

or
∂z ∂z
dx + dy = 0.
∂x ∂y
Such an equation is called an exact D.E.
For example, if x3 + 4xy + 6y 2 = 4 (in this example z = x3 + 4xy + 6y 2 ), then

(3x2 + 4y)dx + (4x + 12y)dy = 0.

Conversely, consider a given exact D.E.:

(3x2 + 4y)dx + (4x + 12y)dy = 0.

The left-hand side is “seen” to be d(x3 + 4xy + 6y 2 ).

Hence, we can write the D.E. as d(x3 + 4xy + 6y 2 ) = 0, and the general solution of the exact
D.E. is
x3 + 4xy + 6y 2 = C.

Note: There are two questions which obviously must be settled:

1. How do we know whether a D.E. is exact or not? (What test, if any, can we use?)

2. How do we then solve exact D.E.s?

1.3. EXACT DIFFERENTIAL EQUATIONS 27

If a D.E. is of the form

M (x, y)dx + N (x, y)dy = 0,
then it is exact if and only if
∂M ∂N
= .
∂y ∂x

Examples
Determine which of the following D.E.s are exact:

4xy
1. y ′ =
y − 2x2
2. y 2 dx = 2xydy
3. (x + sin y)dx + (y 3 + x cos y)dy = 0

Solutions

1. The equation
4xy
y′ =
y − 2x2
needs to be put into the form: M dx + N dy = 0.

(y − 2x2 )dy = 4xydx

4xydx − (y − 2x2 )dy = 0
4xydx + (2x2 − y)dy = 0
∂ ∂
Now (4xy) = 4x and (2x2 − y) = 4x. Hence the D.E. is exact.
∂y ∂x
2.
y 2 dx = 2xydy
Particular attention must be given to the solution of D.E.s of this form, where
∂ 2 ∂
(y ) = (2xy) = 2y
∂y ∂x
However, the D.E. is not exact. Before the test for exactness is performed, the D.E.
must be in the form M dx + N dy = 0, not in the form M dx = N dy.
The given D.E. is y 2 dx − 2xydy = 0 and
∂ ∂
(y 2 ) ≠ (−2xy)
∂y ∂x
Hence, the D.E. is not exact.
3.
(x + sin y)dx + (y 3 + x cos y)dy = 0
Now
∂ ∂ 3
(x + sin y) = cos y = (y + x cos y).
∂y ∂x
Hence, the D.E. is exact.
28 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Solution of exact D.E.s To obtain the general solution of an exact D.E., either of two
methods can be employed.
The first method involves integration, differentiation, and integration again: its use is illus-
trated in Examples 1, 2, 3.
The second method involved “the grouping of terms”: its use is illustrated in Examples 4,
5, 6.
Method 1
Recall that a D.E. of the form M (x, y)dx + N (x, y)dy = 0 is called exact if there exists a
function f (x, y) such that its total differential,
∂f ∂f
df = dx + dy,
∂x ∂y
is equal to M (x, y)dx + N (x, y)dy.
∂f ∂f
Thus, to solve an exact D.E., we simply use that function f for which ∂x = M and ∂y = N,
where M and N are given in the D.E.
(The way in which we then actually find f is demonstrated in the examples.)
Since M dx + N dy = 0 for an exact D.E., once we have found the appropriate function f ,
then we must have that df = 0, and so the solution to the exact D.E. is

f (x, y) = C, where C is an arbitrary constant.

Example 1
Solve the D.E.
(2x sin 3y)dx + (3x2 cos 3y)dy = 0.

Solution
Here,
∂ ∂
(2x sin 3y) = 6x cos 3y = (3x2 cos 3y),
∂y ∂x
and so the D.E. is exact.
f (x, y) must be found so that
∂f
= M (x, y) = 2x sin 3y
∂x
and
∂f
= 3x2 cos 3y.
∂y

Now consider
∂f
= 2x sin 3y.
∂x
To have obtained 2x sin 3y by partially differentiating f with respect to x, f must include
∂
a term in x2 sin 3y. (Then ∂x (x2 sin 3y) = 2x sin 3y as required.) We have

f (x, y) = x2 sin 3y + ϕ(y),

1.3. EXACT DIFFERENTIAL EQUATIONS 29

∂
where ϕ(y) is a “constant function” with respect to x. ( ∂x (x2 sin 3y+ϕ(y)) = 2x sin 3y+0 =
2x sin 3y.)
Now, for
∂f
f (x, y) = x2 sin 3y + ϕ(y), = 3x2 cos 3y + ϕ′ (y)
∂y
and this must equal
N (x, y) = 3x2 cos 3y.
Hence ϕ′ (y) must equal 0. i.e. ϕ(y) = a constant, c. Thus

f (x, y) = x2 sin 3y + c.

We have found the appropriate function for the exact D.E. and so the solution is given by

(x2 sin 3y + c) = K.

i.e.
x2 sin 3y = C.

Note
To provide some assurance that this method has worked, let us check that we have indeed
found the solution to the original D.E., which can be written as
dy 2 sin 3y
=− .
dx 3x cos 3y
The solution we have found gives
C
sin 3y =
x2
and so y is given explicitly by
1 C
y= arcsin ( 2 )
3 x
dy 1 1 −2C
= √ ⋅( 3 )
dx 3 1 − C 2 x
x4
C
2 2
=− √ x
3x 1 − C 2
x4

2 sin 3y
= − ,
3x cos 3y
C
since sin 3y = x2 and
√
√ C2
cos 3y = 1 − sin2 3y = 1−
x4
dy
Thus, dx , for the solution function, is equal to the required value and so the solution checks.
(The check can also be made by differentiating the solution x2 sin 3y = C implicitly with
respect to x. Try it!)
30 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Example 2

Show that

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

is an exact D.E. and hence find its solution.

Solution
Here,
∂ x+y ∂ x+y
(e + 3x2 − 2y) = ex+y − 2 while (e − 2x − 4) = ex+y − 2
∂y ∂x
Thus the test for exactness succeeds.
Now we must find f (x, y) such that

∂f
= ex+y + 3x2 − 2y
∂x
Hence
f = ex+y + x3 − 2xy + ϕ(y)
by inspection.
Next,
∂f
= ex+y − 2x + ϕ′ (y)
∂y
and this must equal
N (x, y) = ex+y − 2x + 4.
We have
ϕ′ (y) = 4
ϕ(y) = 4y + c
Hence
f = ex+y + x3 − 2xy + 4y + c,
and the solution is
ex+y + x3 − 2xy + 4y = C.

Example 3

Solve
(cos x + y sin x)dx = cos xdy,

subject to the condition that y(π) = 0.

Solution
We want to check this D.E. for exactness. Before we do so, we must write the equation in
the form
(cos x + y sin x)dx − cos xdy = 0
before deciding that
M = cos x + y sin x and N = − cos x.
1.3. EXACT DIFFERENTIAL EQUATIONS 31

Now,
∂
(cos x + y sin x) = sin x
∂y
and
∂
(− cos x) = sin x
∂x
which tells us that the equation is exact.
f must be such that
∂f
= cos x + y sin x
∂x
f = sin x − y cos x + ϕ(y)
.
∂f ∂
Now = (sin x − y cos x + ϕ(y))
∂y ∂y
= − cos x + ϕ′ (y)

This must equal (− cos x) and so ϕ′ (y) = 0 i.e. ϕ(y) = c.

f = sin x − y cos x + c

and the general solution is

sin x − y cos x = C.

The particular solution required is such that y = 0 when x = π.

sin π − 0 cos π = C.

i.e. C = 0. This particular solution is sin x − y cos x = 0. (i.e. y = tan x. Check that this is a
solution to the given D.E.)
Method 2
In solving an exact D.E. by the grouping method, we use three facts:

∂z ∂z
1. If z is a function of x alone, = 0, and dz = dx.
∂y ∂x
For example, if z = x3 , dz = 3x2 dx.
∂z ∂z
2. If z is a function of y alone, = 0, and dz = dy.
∂x ∂y
For example, if z = sin y 2 , dz = 2y cos y 2 dy.

3. If z is a function of both x and y, dz will be a function involving both the differentials

dx and dy.
For example (i)
z = x2 y 3
dz = 2xy 3 dx + 3x2 y 2 dy.
For example (ii)
z = x sin y
dz = sin ydx + x cos ydy.
32 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Example 4
Solve
(2xy 3 + 6x)dx + 3x2 y 2 dy = 0.

Solution
This is an exact D.E. since
∂ ∂
(2xy 3 + 6x) = 6xy 2 = (3x2 y 2 ).
∂y ∂x

In this example, the term 6xdx cannot be grouped with any term in y. Hence, we have the
grouping:
6xdx + (2xy 3 dx + 3x2 y 2 dy) = 0.

The two terms grouped together in the above equation were paired-off because both terms
involved both x and y.
The given D.E. can be written as:

d(3x2 ) + d(x2 y 3 ) = 0 (by inspection),

or
d(3x2 + x2 y 3 ) = 0.
Hence
3x2 + x2 y 3 = constant.
The solution of the given D.E. is

3x2 + x2 y 3 = C.

Example 5

Solve the exact D.E.

(6x2 y 2 + 6xy 2 + 12x + 4)dx + (4x3 y + 6x2 y + 6y 2 )dy = 0.

Video solution to Example 5

Solution
To solve this, we take two “passes” at the problem. First, collect together all terms involving
x alone and y alone, leaving the x-y mixtures to be paired-off appropriately on the second
pass. That is,

12xdx + 4dx + 6y 2 dy + {6x2 y 2 dx + 6xy 2 dx + 4x3 ydy + 6x2 ydy} = 0.

In the second pass the x-y terms are paired-off. Which of the terms 4x3 ydy or 6x2 ydy is to
be grouped with 6x2 y 2 dx? The answer should be obvious: a term like 6x2 y 2 dx is obtained
from differentiating a function involving x3 .
This means that the terms 4x3 ydy and 6x2 y 2 dx are to be grouped together, leaving the pair
6xy 2 dx and 6x2 ydy to be grouped together.
1.3. EXACT DIFFERENTIAL EQUATIONS 33

Hence, the given D.E. can be written:

12xdx + 4dx + 6y 2 dy + (6x2 y 2 dx + 4x3 ydy) + (6xy 2 dx + 6x2 ydy) = 0

d (6x2 ) + d(4x) + d (2y 3 ) + d (2x3 y 2 ) + d (3x2 y 2 ) = 0
d (6x2 + 4x + 2y 3 + 2x3 y 2 + 3x2 y 2 ) = 0
∴ 6x2 + 4x + 2y 3 + 2x3 y 2 + 3x2 y 2 = constant.

Hence, the solution of the given D.E. is

6x2 + 4x + 2y 3 + 2x3 y 2 + 3x2 y 2 = C.

Example 6

Solve the exact D.E.

1
(yexy + + ex + 2y sin xy cos xy)dx
x+y
1
+ (2x sin xy cos xy + xexy + + 8 cos y sin y)dy = 0
x+y

Solution
This looks to be a very complicated D.E. However, in trying to solve it, we en-counter no
problems in pairing-off terms. Terms involving exponentials go together, and so on, leading
to the grouping:

1 1
ex dx + 8 cos y sin ydy + (yexy dx + xexy dy) + ( dx + dy)
x+y x+y
+ (2y sin xy cos xydx + 2x sin xy cos xydy) = 0

Once the grouping has been completed, the solution is obtained by integration.
(It should be emphasised that, regardless of the method of solution used, essentially the
same integration is required to obtain the solution.)
In this case, the solution is derived as follows:

d (ex ) + d (4 sin2 y) + d (exy ) + d(log(x + y)) + d (sin2 xy) = 0

d (ex + 4 sin2 y + exy + log(x + y) + sin2 xy) = 0

Hence,
ex + 4 sin2 y + exy + log(x + y) + sin2 xy = C

is the solution of the given D.E.

Tutorial 1 Question 2
34 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

1.4 Separable differential equations

Read textbook section 2.2 (Separable equations)

A D.E. is separable if it can be written in the form

y ′ = f (x)g(y)

or
M (x)dx + N (y)dy = 0.
Clearly, written in the last form, a separable D.E. is an exact D.E., since
∂M (x) ∂N (y)
=0= .
∂y ∂x
Example 1

Solve:
(x3 + 4)dx + sin ydy = 0.

Solution
This D.E. can be solved using grouping of terms:
d ( 41 x4 + 4x) + d(− cos y) = 0
i.e. d ( 41 x4 + 4x − cos y) = 0
1 4
∴ 4x + 4x − cos y = C, (C being constant)
is the general solution.
Note that this same result can be obtained in the following way
(x3 + 4) dx = − sin ydy
3
∴ ∫ (x + 4) dx = ∫ (− sin y)dy
1 4
x + 4x = cos y + C
4
(only one arbitrary constant is required for this first order D.E.), or
1 4
x + 4x − cos y = C.
4

Note
This second approach is usually the most suitable for separable D.E.s once they have been
properly classified, i.e. writing the D.E. in the form

H(y)dy = K(x)dx,
1.4. SEPARABLE DIFFERENTIAL EQUATIONS 35

we can integrate both sides of the equation,

∫ H(y)dy = ∫ K(x)dx,

and hence obtain the solution of the D.E.

Example 2

Solve
dy x
=−
dx y
subject to the initial condition y(0) = 4.
Solution
Here, the D.E. can be written as
ydy = −xdx

∴ ∫ ydy = ∫ (−x)dx
1 1
i.e. y 2 = − x2 + c
2 2
Now, when x = 0, y = 4 and so
8 = 0 + c i.e. c = 8.
The required particular solution is x2 + y 2 = 16.
Similarly, by integrating both sides of
1
dy = 2dx,
y
we get
ln y = 2x + c
i.e. y = e2x+c = e2x ec
= Ce2x as the solution of
dy
= 2y.
dx

Example 3

Solve
5
dy x2 (2x3 − 3)
=
dx (2y − 1)3

Solution
We can separate this D.E. as
5
(2y − 1)3 dy = x2 (2x3 − 3) dx
3 2 3 5
∫ (2y − 1) dy = ∫ x (2x − 3) dx
1 1 6
∴ (2y − 1)4 = (2x3 − 3) + C
8 36
36 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

The solution, y, is given here implicitly. The solution can be simplified but this is not
necessary.
For example, we could write

2 6
(2y − 1)4 = (2x3 − 3) + k
9
or even, explicitly,
√
1⎛ 4 2 6 ⎞
y= 1± (2x3 − 3) + k .
2⎝ 9 ⎠

Example 4

Solve
(x2 + 1)y ′ + y 2 + 1 = 0.

Solution
Here, y ′ has been used instead of dy
dx . We can separate the variables to get

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

and then
dy dx
2
=−
1+y 1 + x2
Hence, arctan y = − arctan x + c, i.e. arctan y + arctan x = c.
The solution can be simplified, if required, as follows:

tan(arctan y + arctan x) = tan c = k.

tan(arctan y) + tan(arctan x)
∴ =k
1 − tan(arctan y) ⋅ tan(arctan x)
y+x
i.e. =k since (tan(arctan θ) = θ)
1 − xy
and even then
k−x
y= .
1 + kx

Note
Textbooks often put the solutions to D.E.s in the most “simplified” form. It is not neces-
sary for you to do so unless specifically required in the problem. You should concentrate
on solving the D.E. correctly, by first deciding on its type and then by carrying out the
appropriate method of solution accurately!

Tutorial 1 Question 3
1.5. INTEGRATING FACTORS 37

1.5 Integrating factors

Read textbook section 2.5 (Special integrating factors)

If a D.E. M (x, y)dx + N (x, y)dy = 0 is not exact, but we can find a function I(x, y) such
that
I(x, y)M (x, y)dx + I(x, y)N (x, y)dy = 0.
is exact, then we call I(x,y) and integrating factor for the D.E. The new equation can be
solved using the method developed in the section on Exact Differential Equations. For
a general first-order D.E., it may be difficult (or impossible) to find an integrating factor.
However, for the important class of linear first-order D.E.s, an integrating factor can always
be found in a standard way.

Example

Consider
−y dy −y
( )+ = 0, i.e. ( ) dx + dy = 0.
x dx x
This equation is not exact, since
∂ −y −1 ∂
( )= while (1) = 0.
∂y x x ∂x

However, if each term in the equation is multiplied by 1/x, the equation becomes
−y 1
( ) dx + ( ) dy = 0.
x2 x
This D.E. is exact since
∂ −y −1 ∂ 1 −1
( 2 ) = 2 and ( )= 2.
∂y x x ∂x x x
Hence, to solve the new D.E., we must find f such that
∂f −y y
= 2 , i.e. f = + ϕ(y).
∂x x x
Then,
∂f 1
= + ϕ′ (y)
∂y x
which must equal 1/x and so ϕ](y) = 0, i.e. ϕ(y) = c.
y
∴ f= + c.
x
Thus the solution of the exact D.E. is y/x = C or y = Cx. Hence, the solution of the given
D.E. is also y = Cx. (check this!)
38 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Notes

1. The equation
−y dy
( )+
x dx
can also be considered as a separable D.E., since
dy dx
=
y x
∴ ln y = ln x + c = ln x + ln C = ln Cx
∴ y = Cx again.

2. The integrating factor 1/x is not unique. Show that

x
x2 + y 2
is also an integrating factor of the given equation and hence solve the new exact D.E.
to get y = Cx as the solution again.

Method
A simplified method for finding integrating factors follows. Consider the DE:

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

which is not exact. We want to find an integrating factor R such that

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

is now exact. So we require

∂ ∂
(R ⋅ M ) = (R ⋅ N ) [test for exactness]
∂y ∂x
Expanding gives
∂R ∂M ∂R ∂N
M +R = N +R .
∂y ∂y ∂x ∂x
This can be very difficult to solve for R, but if we look for solutions of R as a single
variable function, the working is a lot easier.
For R = R(x) we obtain:
∂M dR ∂N
R = N +R
∂y dx ∂x
giving
dR ∂M /∂y − ∂N /∂x
= R.
dx N
Now if the fractional part is a function of x we can write:
dR
= P (x) ⋅ R
dx
which has the simple solution
R(x) = e∫ P dx .
1.5. INTEGRATING FACTORS 39

Similarly if we try R as a function of y we obtain

R(y) = e∫ P dy ,

where
∂M /∂y − ∂N /∂x ∂N /∂x − ∂M /∂y
P (x) = or P (y) = .
N M

Example

Find an integrating factor for the following DE and hence find the general solution:

(1 − y)dx + (x2 y 2 + x) dy = 0.

We have M = 1–y and N = x2 y 2 + x. Then

∂M ∂N
= −1 and = 2xy 2 + 1
∂y ∂x
so the DE is not exact. We then look at
∂M /∂y − ∂N /∂x −1 − (2xy 2 + 1) −1 − 2xy 2 − 1 −2xy 2 − 2
P (x) = = = = 2 2
N x2 y 2 + x x2 y 2 + x x y +x

−2 (xy 2 + 1) −2
= =
x (xy 2 + 1) x
which is indeed a function of x and we proceed:
−2
∫ P dx = ∫ dx = −2 ln x = ln (x−2 )
x
giving
−2 ) 1
R = e∫ P dx = eln(x = x−2 = .
x2
We could also look at
∂N /∂x − ∂M /∂y 2xy 2 + 1 − (−1) 2xy 2 + 2
P (y) = = =
M 1−y 1−y
which does not simplify any further nor is a function of y alone, including also x. So
we have only been able to find one integrating factor as a function of x. Sometimes a
differential equation will give two possible integrating factors, but not in this example.
Continuing we now have, using R from P (x):
1 1
2
(1 − y)dx + 2 (x2 y 2 + x) dy = 0
x x
giving
1−y 1
2
dx + (y 2 + ) dy = 0.
x x
Check that
∂M −1 ∂N 1
= 2 and =− 2
∂y x ∂x x
and the DE is now exact.
40 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Solving the DE as exact we have

∂f 1 y 1 y
M= = 2 − 2 giving f (x, y) = − + + f1 (y).
∂x x x x x
Also
∂f 1 1 y
N= = y2 + giving f (x, y) = y 3 + + f2 (x).
∂y x 3 x
Matching gives
1 1
f1 (y) = y 3 and f2 (x) = −
3 x
and we obtain
1 y 1
f (x, y) = y 3 + − .
3 x x
The solution to the DE is f (x, y) = C or
1 3 y 1
y + − = C.
3 x x
Check follows. From the previous line we obtain the differential form:
dy ydx dx
y 2 dy + − 2 + 2 = 0 and multiplying by x2 gives
x x x
2 2
x y dy + xdy − ydx + dx = 0 and grouping gives
(1 − y)dx + (x2 y 2 + x) dy = 0 as before.

Tutorial 1 Question 4

1.6 First order linear differential equations

Read textbook section 2.3 (Linear equations)

Video (by Dr. R. Wood): First order linear ODEs

Linear and non-linear
The term linear is used to describe D.E.s in which y and its derivative occur only to the
first power. For example,

dy d2 y dy
= 3y + sin x or − 2 + 3y = ex
dx dx2 dx
are both linear, while

dy dy 2
= 3y 2 + sin x or ( ) = 3y + ex
dx dx
1.6. FIRST ORDER LINEAR DIFFERENTIAL EQUATIONS 41

are both non-linear.

Class
The class of linear D.E.s is very important. In this section we will examine only linear
D.E.s of the first order. Any such equation can always be classified by writing it in the
form
dy
+ P (x)y = Q(x).
dx
Notes

1. The method we will develop will work for any first order linear D.E.. The equation
may also be exact or separable (or some other type). In such cases we have a choice
of methods of solution!

2. In Chapter 2 we will consider second order (and higher order) linear D.E.s with
constant coefficients, which can be written in the form

d2 y dy
a2 + a1 + a0 y = F (x)
dx dx
where a0 , a1 , a2 are constants.
(Such D.E.s can be solved by a very clever “trick”!)
However, in Chapter 4 we will consider the solution of such second order linear D.E.s
where the coefficients a0 , a1 , a2 are functions of x.
(Such D.E.s can be solved by the use of series!)

Method of solving first order L.D.E.s

First order linear D.E.s, as noted above, can always be written in the form

y ′ + P (x)y = Q(x).

Such D.E.s are not (usually) exact, but can be made exact by multiplying the D.E. by an
appropriate integrating factor. One way to derive this integrating factor is as follows:
Multiply the D.E. by R(x), an I.F. which is assumed to depend on x alone and which must
be found, giving
Ry ′ + RP y = RQ,
where the left-hand side of the D.E. must be the derivative of Ry.

d
i.e., Ry ′ + RP y = (Ry) = Ry ′ + R′ y
dx
∴ RP y = R′ y (subtracting Ry ′ from both sides)
R′
∴ = P (x)
R
∴ ln R = ∫ P (x)dx( integrating both sides w.r.t. x)

Hence,
R = e∫ P (x)dx .
42 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

is the required integrating factor.

The general solution can then be found by solving
d
(ye∫ P (x)dx ) = e∫ P (x)dx Q(x)
dx
by integration.
Solution of such first order linear D.E.s is best accomplished by carrying out the above in
a series of steps:

Step 1: Find ∫ P (x)dx

Step 2: Find e∫ P (x)dx .

Step 3: Multiply every term in the D.E. by the integrating factor, and hence write the D.E.
in the form
d ∫ P (x)dx
(e ⋅ y) = e∫ P (x)dx ⋅ Q(x).
dx
Step 4: Integrate both sides of the D.E., i.e.

e∫ P (x)dx ⋅ y = ∫ {e∫ P (x)dx ⋅ Q(x)} dx.

Step 5: Hence find y.

A more concise step-method is to consider d(Ry) = RQdx then

∫ d(Ry) = Ry = ∫ RQdx + C.

So an alternative set of steps can be given as:

Step 1: find R = e∫ P (x)dx .

Step 2: evaluate ∫ RQdx

Step 3: answer given by y = R−1 ∫ RQdx + CR−1 .

Example 1

Solve
dy
sin x + 4y cos x = cot x
dx
Solution
The D.E. can be re-written as
dy cos x cot x
+ (4 )y =
dx sin x sin x
which is in first order linear form.
1.6. FIRST ORDER LINEAR DIFFERENTIAL EQUATIONS 43

Step 1:
cos x 4
∫ 4 sin x dx = 4 ln sin x = ln(sin x)

Step 2:
4
eln(sin x) = (sin x)4
(Note the important use of a property of the logarithm and exponential functions.)
Step 3: By multiplying through by the I.F., the D.E. can be written as
dy cot x
(sin x)4 + (4 sin3 x cos x) y = (sin x)4 = cos x sin2 x
dx sin x
d
i.e of the form {y(sin x)4 } = cos x sin2 x
dx
Step 4:
y sin4 x = ∫ cos x sin2 xdx
1 3
= sin x + C
3
Step 5: Thus the solution is given by
1
y= csc x + C csc4 x
3

Example 2

Solve
dy
− y = ex ,
dx
subject to the condition y(0) = 2.
Solution
(This D.E. can be solved by the methods developed in Chapters 2 and 3: try it again after
completing those chapters!)
For now, we recognise that this is a linear first order D.E. with P (x) = −1 and Q(x) = ex .

Step 1: ∫ (−1)dx = −x
Step 2: I.F. is e(−x) = e−x
Step 3: (slightly modified): Multiply through by the I.F.
dy
∴ e−x − ye−x = ex ⋅ e−x = e0 = 1
dx
Now, the left-hand side, as expected, can be written as
d −x
(e ⋅ y)
dx
(Check that this is true by carrying out this differentiation!)
Therefore the D.E. is now of the form
d −x
(e ⋅ y) = 1
dx
44 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Step 4: Now integrate both sides of the D.E. i.e.,

e−x ⋅ y = ∫ dx
=x+C
Thus the general solution is given by
1
y= (x + C) = xex + Cex
e−x
Now, the particular solution required has y = 2 when x = 0.
∴ 2 = 0 ⋅ e0 + C ⋅ e0
∴ C=2
The required particular solution is:

y = xex + 2ex or y = ex (x + 2)

Tutorial 1 Question 5

1.7 Bernoulli differential equations

Read textbook section 2.6 (Substitutions and transformations)

Video (by Dr. R. Wood): Bernoulli differential equations

Bernoulli D.E.s are examples of non-linear D.E.s which can be solved using a suitable
change of variable. This change of variable always transforms a Bernoulli D.E. to a linear
first order D.E. A Bernoulli D.E. is of the form:

y ′ + P (x)y = Q(x)y n , n ≠ 0 or 1

To solve this D.E. the change of variable

z = y 1−n

is made. The method of solution is demonstrated in the following example.

Example

Find the solution of the D.E.

1 1
y′ + y = 2 y3
x x
A useful rewriting of this D.E. is
1 1
y −3 y ′ + y −2 = 2
x x
1.8. FIRST ORDER HOMOGENEOUS DIFFERENTIAL EQUATIONS 45

Let z = y −2 , then z ′ = −2y −3 y ′ . Substituting these into the D.E. gives

1 1 1
− z′ + z = 2
2 x x
′ 2 2
∴ z − z=− 2
x x
This is a linear first order D.E. with an integrating factor
1
e∫ (− x )dx = e−2 ln x = eln x =
2 −2

x2
Multiplying the D.E. by this integrating factor gives

d 1 2
( 2 ⋅ z) = − 4
dx x x
1 2 −3
∴ z = x +C
x2 3
1 1 2 −3
i.e., = x +C
x2 y 2 3
Hence,
2
y −2 = x−1 + Cx2
3
is the solution of the given Bernoulli D.E.

Tutorial 1 Question 6

1.8 First order homogeneous differential equations

Read textbook section 2.6 (Substitutions and transformations)

Video (by Dr. R. Wood): Homogeneous differential equations

A first order homogeneous D.E. is of the form
y
y′ = f ( )
x
y
This type of D.E. can be solved using the substitution v = x. When this substitution is
made, the given D.E. always reduces to a separable D.E.

Example 1

Find the general solution of the D.E.

(x2 + y 2 ) dx − 2xydy = 0
46 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

The D.E. can be written as

y 2
′ x2 + y 2 1 + ( x )
y = = ,
2xy 2 ( xy )
and is thus homogeneous.
y
Let v = . Then y = xv, and y ′ = xv ′ + v. Substituting into the D.E. gives
x
1 + v2
(xv ′ + v) =
2v
1 + v2 1 − v2
xv ′ = −v =
2v 2v
2v ′ 1
∴ v = .
1 − v2 x
This is a separable D.E. which can be integrated!

− ln (1 − v 2 ) = ln x + C
i.e., ln x (1 − v 2 ) = −C

x (1 − v 2 ) = (e−C ) = A, say
y2
∴ x (1 − 2 ) = A
x
2 2
Hence, x − y = Ax is the solution of the given D.E.

Example 2

Solve
xy ′ = x + y.
Solution We can write this D.E. in the form
dy y
=1+
dx x
which we can then classify as a homogeneous type.
y dy dv
Let v = , then, y = xv and = x + v.
x dx dx
The D.E. is transformed to
dv
x +v =1+v
dx
dv
∴ x =1
dx
dx
i.e. dv = , which is a separable D.E.
x
∴ v = ln ∣x∣ + C
y
i.e. = ln ∣x∣ + C
x
Thus the required solution is
y = x ln ∣x∣ + Cx.
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 47

Notes

1. The word “homogeneous” is widely used to describe another completely different

class of differential equations! This “confusion” should not affect the classification
or solution of either type of D.E.

2. There are many other types of D.E.s which, by an appropriate substitution as in the
Bernoulli or Homogeneous cases can be transformed into a known type of D.E. with
a “standard” method of solution. Such other transformations are beyond the scope of
this subject, but are very important in many applications of D.E.s.

Tutorial 1 Question 7

1.9 Applications of first order differential equations

Read textbook sections 3.1, 3.2 (Mathematical modelling, Compartmental

Analysis)

Video (by Dr. R. Wood): Applications of 1st order ODEs

Introduction
There are many applications of D.E.s to problems of mechanics, electrical circuits, rockets,
chemistry, economics, suspension bridges, bending of beams, vibrating springs and so on
and so on! We now consider applications of first order linear differential equations to some
aspects of biology and population models. The notes here are based on source material
from the texts Differential equations and their applications, by M. Braun, Spring-Verlag,
1978 and Mathematical Ideas in Biology, by J.Maynard Smith, C.U.P., 1971.
Note
You do NOT have to master the scientific theories or techniques where D.E.s arise, but you
must learn to solve the relevant D.E.s.
The growth of tumours
“Free-living” dividing cells grow at a rate proportional to the volume of dividing cells
present at that time. For example, bacteria and early embryonic cell growth obey this law.
Hence if V (t) denotes the volume of dividing cells at time t, we have

dV
= λV,
dt
where λ is some constant. This is a linear first order differential equation

dV
− λV = 0.
dt
48 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

The integrating factor is

e∫ −λdt = e−λt
so that
dV
e−λt − λe−λt V = 0
dt
d −λt
(e V ) = 0
dt
and integrating gives e−λt V = C, where C is some constant. If at time t = 0 we have the
volume V = V0 we find the constant C to be C = V0 . Hence
e−λt V = V0 , and so
V = V0 eλt
is the solution. Hence the volume of cells grows exponentially with time.
The time for the volume to double is a constant under this process. This follows since at
t = t1 we have V = V1 so that
V1 = V0 eλt1
The time T at which this doubles gives then

2V1 = V0 eλT

Hence
V0 eλT = 2V0 eλt1
eλT = 2eλt1

λT = ln (2eλt1 ) = ln 2 + ln eλt1

λT = ln 2 + λt1
ln 2
T − t1 =
λ
Hence the time to double its volume (T − t1 ) is ln 2/λ.
Solid tumours, however, differ from “free-living” dividing cells in their growth. As the
tumour grows the time for its volume to double becomes longer and longer. That is, as it
gets bigger its rate of growth decreases.
Several researchers have found that the data for tumour growth is fitted extremely well by
the following exponential function:
λ
V = V0 exp { (1 − exp(−αt))}
α
where α and λ are positive constants. This is usually referred to as the Gompertzian func-
tion. You will notice that at t = 0, we have
λ λ
V = V0 exp { (1 − exp(0))} = V0 exp { (1 − 1)}
α α
= V0 exp{0} = V0

and that as t → ∞, V → V0 eλ/α . Consequently a solid tumour usually has a limiting size
(fortunately). Ordinary exponential growth has no limit as t → ∞, so how can this growth
behaviour of tumours be explained?
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 49

We can gain some insight into this question by finding out the differential equation which
is satisfied by the Gompertzian function. Differentiating both sides of the equation gives
dV λ d λ
= V0 exp { (1 − exp(−αt))} { (1 − exp(−αt))}
dt α dt α
λ λαe−αt
= V0 exp { (1 − exp(−αt))} ⋅ [0 + ]
α α
λ
= λe−αt V0 exp { (1 − exp(−αt))}
α
and so
dV
= λe−αt V
dt
is the differential equation governing tumour growth. Notice that at t = 0 we have dV
dt = λV ,
dV −αt
a simple exponential growth model, but as t → ∞, we have dt → 0 (since e → 0) which
reflects the fact that the rate of growth is slowing down.
Two different theories have been put forward to explain the dynamics of tumour growth. It
has been suggested that the retarding of tumour growth is due to either:

i. decrease in the reproductive rate of the cells, the proportion of reproducing cells re-
maining constant, and

ii. a drop in the proportion of reproducing cells, the reproductive rate remaining constant.

We could re-write the above tumour growth equation as

dV
= λe−αt V = λ∗ V
dt
where λ∗ = λe−αt denotes the reproductive rate which decreases with time. This would
reinforce theory (i) above.
However, we could also write
dV
= λe−αt V = λV ∗
dt
where V ∗ = e−αt V denotes the volume of reproducing cells which decreases with time,
leaving the rate λ constant. This would reinforce theory (ii) above.
The second theory seems to be more popular. It is suggested that in fact a necrotic region
develops in the centre of some types of tumour. The necrotic core (where reproductive rate
has slowed right down or stopped) develops because the supply of blood and nutrients is
mostly to the surface of the tumour. As the tumour be-comes larger the supply of nutrients
to the core by diffusion alone gets more and more difficult.
Whichever theory is closest to reality, the equation
dV
= λe−αt V
dt
describes the growth behaviour.
The Malthusian law of population growth
Populations only change at any time by integer amounts. If p(t) denotes the population
of a certain species at time t, and if p(t) is large, then we can conveniently assume that
50 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

p is a continuous (and differentiable) function of t so that the rate of change dp/dt has
meaning. In the simplest case, if the growth rate, a, is constant for each individual (i.e.
each individual reproduces a new individuals per unit time), then the differential equation
for population growth is simply
dp
= ap
dt
where a is a constant and p = p(t). This is a first order linear differential equation as seen
in the last section and so if p(t0 ) = p0 we can solve to get the solution

p(t) = p0 ea(t−t0 ) .

The above model for population growth was postulated by the English economist Malthus
(1766-1834). His model has been applied to human populations to explain the rapid in-
crease over recent years. The population of the world has been increasing at about 2%
per year and so the time to double can be determined from the above equation. In fact if
p(t) = 2p0 , then
p(t) = 2p0 = p0 ea(t−t0 ) , where a = 0.02
Hence
e0.02(t−t0 ) = 2
0.02 (t − t0 ) = ln 2
and
t − t0 = ln 2/0 ⋅ 02 ≈ 34.7 years.
In fact it has been observed that the world’s population has been doubling every 35 years.
The model, however, cannot be realistic in the long term.
Example
The population of the world in 1976 was 4 billion. Given a growth rate of 2%, find the
population predicted by the Malthusian model in the years 2000 and 2510 and 2670. Fur-
thermore, if the area of the earth’s surface is 170 000 billion square metres, how much room
will people have in these years?
Solution
dp
The equation dt = ap has solution

p = p0 ea(t−t0 )

Hence
i. t0 = 1976, t = 2000, a = 0.02, p0 = 4 billion
p = 4e0.02(2000−1976) = 4e0.48 ≈ 6.5 billion.
Each person then has 170000 ÷ 6.5 = 26000 square metres of room.

ii. t0 = 1976, t = 2510, a = 0.02, p0 = 4 billion

p = 4e0.02(2510−1976) = 4e10⋅68 ≈ 174000 billion.

Each person would then have 170000 ÷ 174000 ≈ 1 square metre of room (a bit cramped!).

iii. p = 4e0.02(2670−1976) = 4e13.88 ≈ 4000000 billion.

How much room then? We would all be standing at least two deep on each other’s shoul-
ders.
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 51

So Malthus’ model is obviously not accurate for large values of p. Should we then discard
it? Consider the next example taken from an actual experiment.
Example
A certain species of rodent has a reproduction rate of 40% per month. Initially there were
2 rodents present (one of each sex) and after 2, 6 and 10 months respectively there were
observed 5, 20 and 109 rodents present.
How well does this compare with the Malthus model?
Solution
dp
Since dt = 0.4p, p(0) = 2 we have the solution p = 2e0.4t where t is in months.
As you can easily verify, at t = 2, 6 and 10 respectively, the values of p are 4.5, 22.0 and
109.2, which shows close agreement with the predictions of the Malthus model. Hence
the Malthus model is accurate in certain cases. In fact it seems to be very useful provided
that the population p does not get too large. Obviously as p increases, other factors such
as competition for resources become important and the model is then too simple to explain
actual growth behaviour. This will be considered in the next section.
The logistic equation (Getting more complicated!)
The Dutch biomathematician Verhulst introduced in 1837 an improved version of the pop-
ulation model which included a competition term. He suggested the equation

dp
= ap − bp2 = (a − bp)p
dt
where a is still the birth rate per individual and where bp represents the deathrate per in-
dividual, b being a constant. You will notice that as p increases, the deathrate increases in
proportion, which is much as you might expect due to competition, finite resources, limited
food supply and so on. The term −bp2 on the right-hand side has a levelling out effect on
the solution - as we shall see. In practice, the constant b is small compared with a. Also
when p is small, the term −bp2 will be negligible compared with ap, and so the equation
resembles the Malthus model. However, when p is large, the term −bp2 will be quite signif-
icant and serves to slow down the growth in population. The above equation is called the
Logistic Equation.
Fortunately, we can even solve the Logistic equation, since it is of separable type.
Separating the variables we get
dp
= dt
p(a − bp)
so that
dp
∫ p(a − bp) = ∫ dt
1
To integrate p(a−bp) we use partial fractions.
Let
1 A B
= +
p(a − bp) p a − bp
so that 1 = A(a − bp) + Bp.
At p = 0 we have 1 = Aa, giving A = 1/a.
52 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Further, at p = a/b we have 1 = Ba/b, giving B = b/a.

Hence
1 1 b
∫ ( + ) dp = ∫ dt
a p a − bp
1
[ln p − ln(a − bp)] = t + C
a
or
1 p
ln ( ) = t + C.
a a − bp
If we take p = p0 at t = 0 we get the constant C to be
1 p0
C= ln
a a − bp0
Hence
1 p 1 p0
ln ( ) = t + ln ( )
a a − bp a a − bp0
p a − bp0
ln [ × ] = at
p0 a − bp
and taking exponentials of both sides
p a − bp0
= eat
p0 a − bp
p (a − bp0 ) = eat p0 (a − bp)
p (a − bp0 ) = eat p0 a − eat p0 bp
p [a − bp0 + bp0 eat ] = eat p0 a
ap0 eat
p=
a − bp0 + bp0 eat
Dividing numerator and denominator by eat then gives
ap0
p=
bp0 + (a − bp0 ) e−at
the Logistic Growth function.
ap0 a
Notice that as t → ∞, p → = the limiting size of the population. It should also be
bp0 b
noted that this value is independent of the initial value p0 .
The following figure shows the shape of the curve for p.
It can easily be verified (by the second derivative test) that the inflexion point occurs at
a
p = 2b , i.e. at half the limiting population.
We have a very good idea of the value of a, the birth rate - it has been estimated to be
about 0.029. But how do we estimate a value for b - the competition constant? In 1961 the
population of the earth was 3.06 billion and it was found to be increasing at the rate of 2%
per year. Hence from the logistic equation, since
1 dp
= a − bp
p dt
we have 0.02 = 0.029 − b(3.06)109 which gives
b = 2.94 × 10−12 .
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 53

dp dp
(Note: Since is the total rate of change of the population, then ÷ p is the rate of
dt dt
change per individual which in this case is 2%.
Hence from above, the limiting population
a 0.029
= ≈ 10 billion.
b 2.94 × 10−12

Verhulst applied his methods to a study of populations in European countries with some
interesting results. In 1845 he predicted a maximum population of 6,600,000 for Belgium
and a maximum population of 40,000,000 for France. By 1930, the population of France
was steady around this figure, but Belgium already had over 8 million people. Perhaps the
answer is that Belgium had the Congo and at the same time was enjoying a remarkable rise
in industry. This allowed them to support a much greater population.
Example
In 1790 the population of the U.S. was 3.9 million, the birth rate was 3.134% and the
competition constant b was 1.5887 × 10−10 . The population in 1800 was in fact 5.308
million. How close is this to the figure predicted by the logistic model?
Solution
The logistic function found above is
ap0 a
p= =
bp0 + (a − bp0 ) e−at b + (ap−1
0 − b) e
−at

and so, using the above information, we get

(0.03134)
p=
(1.5887)10−10 0.03134
+ [( 3.9×106 ) − 1.5887 × 10
−10 ] e−0.03134t

313, 400, 000

=
1.5887 + 78.7703e−0.03134t
and so when t = 1800 − 1790 = 10, p ≈ 5.297 million which is very close to the actual figure
of 5.308 million.
54 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Applications of differential equations to electrical circuits

Read textbook section 3.5 (Electrical circuits)

Note At this point, ignore any reference in the text to second order DEs. We will cover
these in the next chapter. Preliminary electrical theory In this section we apply some of
the methods for solving differential equations (as studies in earlier chapters) to the solution
of some electrical circuit problems. The notes are based partly on material from the book
Applied Differential Equations by M. Spiegel, Prentice-Hall, 1967, which incidentally is an
excellent reference book for you if you wish to brush-up on standard methods for solving
differential equations.
The current I in an electrical circuit is said to be the time rate of change of the quantity
of electricity Q (or charge) flowing from one part of the circuit to another. Hence we have
dQ
I= .
dt
A very simple electric circuit would consist of a source of electromotive force, E (such
as a battery or generator), and a resistor R (such as an electric light element or heating
element or electric motor. The electromotive force (or voltage) produces the current I by
forcing charge Q to flow in a similar manner to the way in which a pump causes water to
flow in a circuit of pipe. The diagram below illustrates a simple circuit,

the direction of the current flow I being as shown. The symbols shown on the circuit are
used for resistors R and sources of electromotive force E. Note that the wire is assumed to
have no resistance. If it had any at all we could include it in R.
Two other electrical elements will be considered, the inductor and the capacitor. The in-
ductor L is essentially a coil of wire which has the property that it opposes any change in
current I. If a steady current is flowing around the circuit, the inductor will have no effect.
But if the current increases or decreases, the inductor will resist such change. Hence it can
be seen that the inductor has an inertia effect on current - much like trying to speed up or
slow down a moving train, the mass of the train and its inertia resists such a change.
The capacitor C stores electrical energy (or charge) in the circuit. A capacitor consists of
parallel plates separated by an insulating medium. Charge then collects on the plates and
can be stored.
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 55

The circuit below shows the symbols used for all of these elements:

We can measure the voltage drop (like a change in pressure) across any of these elements
using voltmeter. It is found that the following three laws can be verified by experiment:

1. Voltage drop across a resistor Er is directly proportional to the current I passing through
the resistor. Hence Er ∝ I and defining the constant of proportionality to be R - the
resistance, we get
Er = RI.

2. Voltage drop across an inductor El is directly proportional to the rate of change of the
current I.
Hence El ∝ dIdt
and defining the constant of proportionality to be L - the inductance, we get
dI
El = L
dt

3. Voltage drop across a capacitor Ec is directly proportional to the charge Q on the capac-
itor.
Hence Ec ∝ Q and defining the constant of proportionality to be 1/C - where C is called
the capacitance, we get
Q
Ec = .
C

The following units and abbreviations will be used in this chapter:

• Time - t - seconds

• Charge - Q - coulombs

• Current - I - amps

• Voltage (or electromotive force) - E - volts

• Resistance - R - ohms

• Inductance - L - henrys

• Capacitance - C - farads.
56 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Kirchoff’s Law
This states that the algebraic sum of all the voltage drops around a closed circuit is
zero.
This is just another way of saying that the voltage supplied by the source of electromotive
force (e.g. the battery) is equal to the sum of the voltage drops in the rest of the circuit.
We will use Kirchoff’s Law in order to establish a mathematical formulation for certain
types of circuits. In fact we will get differential equations which can be solved for the
current flowing at any particular time.
Note that in the next section you will not be required to carry out all the details of sketching
circuits and setting up the D.E.s. However, you must be able to solve the resulting D.E.
First order linear differential equations for RL and RC circuits We will take the ex-
ample of the RL circuit first of all to illustrate the use of Kirchoff’s Law. The diagram of
the circuit is shown below.

Now by Kirchoff’s Law, the voltage supplied E is equal to the voltage drop across the
resistor (Er = RI) plus the voltage drop across the inductor
dI
El = L
dt
Hence,
El + Er = E
giving
dI
L + RI = E
dt
the corresponding differential equation for the RL circuit. This is an example of the well-
known linear first order differential equation.

Example 1

A battery with electromotive force 100 volts is connected in series with a 20 ohm resistor
and an inductor of 4 henrys. The switch is closed at time t = 0 and the current starts to flow.
Sketch the circuit, set up the differential equation for the current and solve it for the current
at any time t.
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 57

Solution

The differential equation becomes

dI
4 + 20I = 100
dt
or
dI
+ 5I = 25.
dt
The Integrating Factor is
e∫5 dt = e5t
so that we have
dI
e5t + 5e5t I = 25e5t
dt
d 5t ,
(e I) = 25e5t
dt
e5t I = 5e5t + K
where K is some constant to be determined by an initial condition.
Now we know that at t = 0 the switch has just been closed, so that the current has not yet
begun to move. Hence at t = 0, I = 0. This gives the value of K to be −5 and so

e5t I = 5 (e5t − 1)

and so
I = 5 (1 − e−5t ) .
Notice that at t = 0, I = 0 and as t → ∞, I → 5. A sketch of the solution is shown below.

Hence the current starts off at zero and quickly builds up towards its limiting value of 5
amps.
Note
58 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

Since the linear differential equation had constant coefficients, it could have been solved
by the method of separation of variables.

Example 2

Solve example 1 if the battery is replaced by a generator with alternating electromotive

force given by 20 sin 5t volts.
Solution
The differential equation now becomes
dI
4 + 20I = 20 sin 5t, or
dt
dI
+ 5I = 5 sin 5t
dt
The integrating factor is once again e5t so that
d st
(e I) = 5e5t sin 5t, and so
dt
est I = ∫ 5est sin 5tdt

Now integrating by parts twice the right-hand side gives (u = 5e5t , v ′ = sin 5t)

5t 5t 5t
∫ 5e sin 5tdt = −e cos 5t + ∫ 5e cos 5tdt
(u = 5e5t , v ′ = cos 5t)

5t 5t 5t 5t
∫ 5e sin 5tdt = −e cos 5t + [e sin 5t − ∫ 5e sin 5tdt]

Hence
5t 5t sin 5t − cos 5t
∫ 5e sin 5tdt = e [ 2
]+K

and so returning to the above solution,

sin 5t − cos 5t
e5t I = e5t [ ] + K,
2
1
where K = 2 from the initial condition I = 0 at t = 0. Hence,

sin 5t − cos 5t e−5t

I= +
2 2
is the final solution for the current at any time t.
Now the term e−5t /2 becomes negligible as t → ∞ and is called the transient current. The
remaining term
sin 5t − cos 5t
2
is called the steady state current.
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 59

Next we will consider the case of RC circuits, shown below.

By Kirchoff’s Law, the voltage supplied E is equal to the voltage drop across the resistor
(Er = RI) plus the voltage drop across the capacitor (Ec = Q
C ). Hence Er + Ec = E, giving

Q
RI + = E.
C
dQ
Now the current I is defined to be the rate of change of charge dt , so that

dQ 1
R + ⋅Q=E
dt C
a first order linear differential equation, this time in terms of Q instead of I.

Example 3

A rapidly discharging battery with electromotive force E = 200e−5t is connected in series

with a 20 ohm resistor and a 0.01 farad capacitor. If the charge in the circuit at t = 0 is
zero, find the charge and current at any time. Show, furthermore, that the charge reaches a
maximum, find it and find also at what time this maximum charge is reached.
Solution

The RC equation from above becomes

dQ 1
20 + Q = 200e−5t , or
dt 0.01
dQ
+ 5Q = 10e−5t
dt
60 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

The Integrating Factor is e5t , so that

dQ
e5t + 5e5t Q = 10
dt
d 5t
dt (e Q) = 10

e5t Q = 10t + K
At t = 0, Q = 0 so that K = 0.
Hence
Q = 10te−5t
the charge at any time t as represented by the sketch below.

The current at any time is readily found from

dQ
I= = 10e−5t − 50te−5t
dt
The maximum charge can be found using the derivative test. That is, Q will have its
maximum when dQ
dt = 0.

Now
dQ
= 10e−5t (1 − 5t)
dt
which is zero when t = 51 seconds. (You should show that this is a maximum and not a
minimum, but that is fairly obvious anyway.) At t = 15 , Q = (10) ( 51 ) e−1 ≈ 0.74 coulombs.
It is interesting to compare the two curves for charge and current. If we sketch

Q = 10te−5t , and
I = 10e−5t (1 − 5t), we get:

It can be seen that at t = 0 the current is high, so the charge starts to quickly build up on
the capacitor. When the current becomes negative (reverses direction) the charge will of
1.9. APPLICATIONS OF FIRST ORDER DIFFERENTIAL EQUATIONS 61

course decrease on the plate of the capacitor and eventually will discharge completely as
there is no current remaining in the circuit (the battery has gone flat!)

Tutorial 1 Question 8
62 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

1.10 Tutorial 1
Back to Chapter 1

Question 1
∂z ∂z
In each of the following, find and :
∂x ∂y

i. z = xy

ii. z = x2 y + ln y

iii. z = (x2 + y 2 )3

iv. z = sin x cos y + 12 x2 ey

Answers:

∂z ∂z
i. = y, =x
∂x ∂y
∂z ∂z 1
ii. = 2xy, = x2 +
∂x ∂y y
∂z 2 ∂z 2
iii. = 6x (x2 + y 2 ) ; = 6y (x2 + y 2 )
∂x ∂y
∂z ∂z 1
iv. = cos x cos y + xey , = − sin x sin y + x2 ey
∂x ∂y 2

Back to Chapter 1

Question 2

Part 1
Test for each of the following D.E.s for exactness, and solve, if possible:

i. 3xdx + 4ydy = 0
x−y
ii. y ′ =
x+y
x
iii. y ′ =
x+y
iv. (ye−x − sin x) dx − (e−x + 2y) dy = 0

v. (x2 + x) dy + (2xy + 1 + 2 cos x)dx = 0

1.10. TUTORIAL 1 63

Part 2
Solve each of the following, subject to the given condition:

i. 2xydx + (x2 + 1) dy = 0, y(1) = 3

ii. (x2 + 2ye2x ) y ′ + 2xy + 2y 2 e2x = 0, y(0) = 1

Part 3
Show that the D.E. ydx + (4x − y 2 )dy = 0 is not exact, but becomes exact in multiplying
every term by the integrating factor, y 3 . Hence solve the D.E.
Short answers:

Pt 1 i. 3x2 + 4y 2 = C
ii. x2 − 2xy − y 2 = C
iii. Not exact
iv. ye−x − cos x + y 2 = C
v. Not exact

Pt 2 i. x2 y + y − 6 = 0
ii. x2 y + y 2 e2x − 1 = 0

Pt 3 6xy 4 − y 6 = C

Detailed solutions to Question 2

Back to Chapter 1

Question 3

Solve the following ODEs

i. 3x (y 2 + 1) dx + y (x2 + 2) dy = 0

ii. 2ydx + e−3x dy = 0

dy x + xy 2
iii. = , y(1) = 0
dx 4y
dy
iv. x = y2 + 1
dx
v. ydx + (x3 y 2 + x3 ) dy = 0.

vi. Textbook exercise set 2.2.

Problems 11, 15, 19, 25 (8th, 8e-th editions);
Problems 8, 16 (slightly different to 15th from 8th ed), 19, 25 (9th edition)

Short answers:
64 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

3
i. (x2 + 2) (y 2 + 1) = C

ii. 2e3x + 3 ln y = C

iii. x2 − 4 ln (1 + y 2 ) = 1

iv. ln x − arctan y = C
1
v. − y 2 − 2 ln y = C.
x2
vi. see textbook or detailed solutions

Detailed solutions to Question 3

Back to Chapter 1

Question 4

Textbook exercise set 2.5. Problems 1, 3, 5, 7, 9, 11.

Detailed solutions to Question 4 (Problems 7, 9, 11)
Back to Chapter 1

Question 5

Solve each of the following linear D.E.s:

dy
i. x + 5y − 2x2 + 7 = 0
dx
ii. dy + 2x(y − 1)dx = 0

iii. xy ′ + 3y = x2

iv. y ′ + 3y = e−2x ; y(0) = 5

dy y
v. = x − ; y = 1 when x = 1
dx 2x
dy
vi. = 2y
dx
vii. Textbook exercise set 2.3.
Problems 1, 3, 5, 7, 9, 11. (8th, 8e-th editions)
Problems 1, 3, 5, 11, 12. (9th edition)

Short answers:

2 7
i. y = x2 − + Cx−5
7 5
ii. y = 1 + Ce−x
2
(Also solve this as a separable D.E.)
1.10. TUTORIAL 1 65

iii. x5 − 5x3 y = C

iv. y = e−2x + 4e−3x

2 3 1
v. y = x2 + x− 2
5 5
vi. y = Ce2x (Also separable)

vii. see textbook or detailed solutions

Detailed solutions to Question 5

Back to Chapter 1

Question 6

i. Solve the equation

z ′ − z = xz 2 .

ii. Textbook exercise set 2.6. Problems 21, 23, 25, 27.

Short answers

i.
1
z=
1 − x + Ce−x
ii. see textbook or detailed solutions

Detailed solutions to Question 6

Back to Chapter 1

Question 7

Solve each of the following homogeneous D.E.s:

dz z z 2
i. = + z(1) = 1
dx x x2
ii. xz ′ = 2x + 3z
2
′
z + x (cos xz ) π
iii. z = z(1) =
x 4

Short answers:

x
i. z =
1 − ln x
ii. z = Cx3 − x (can be solved as linear )
66 CHAPTER 1. FIRST ORDER DIFFERENTIAL EQUATIONS

z
iii. 1 + ln x = tan
x

Detailed solutions to Question 7

Back to Chapter 1

Question 8

1. Textbook exercise set 3.2. Problems 5, 9, 13.

2. The current I, in amperes, in a certain electrical circuit is modelled by the differ-

ential equation
dI
+ 2I = 10e−2t
dt
where t is the time in seconds. If I = 0 when t = 0, find I as a function of t.
(Answer: I(t) = 10te−2t .)

Detailed solutions to Question 8

Back to Chapter 1
Chapter 2

Linear differential equations

Chapter objectives

In this chapter you will learn to:

• distinguish between general linear and non-linear equations;

• find the homogeneous solution yh for linear differential equations (LDEs);

• find the particular solution yp for LDEs using various methods; and

• apply linear differential equations to certain real-life problems.

Chapter summary

1. n-th order linear differential equations

2. Solution of linear D.E.s with constant coefficients

3. The homogeneous solution yh of a linear D.E. with constant coefficients

4. A particular solution via the method of undetermined coefficients

5. A particular solution via the method of variation of parameters

6. A particular solution via the D-operator method

7. Applications of linear differential equations

67
68 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

2.1 General linear differential equations

Read textbook section 6.1 (Basic Theory of Linear Differential Equations)

Video (by Dr. R. Wood): General linear ODE’s

Linear D.E.s occur frequently in the solution of physical problems. We have already con-
sidered first order linear D.E.s in Chapter 1. In the remainder of the subject we will consider
higher order linear D.E.s but will concentrate mainly on the second order linear D.E. The
methods we will develop could also be used for the first order case but it is usually easier
to use the Integrating Factor method of Chapter 1.
Definition. An nth order linear D.E. is of the form
an (x)y (n) + an−1 (x)y (n−1) + . . . + a1 (x)y ′ + a0 (x)y = F (x)
n
i.e. ∑ ak (x)y (k) = F (x)
k=0

where an (x), an−1 (x), . . . , a1 (x), a0 (x) and F (x) are functions of x alone.
The D.E. is of nth order since the highest derivative of y is the nth derivative (that is, y (n)
or dn y/dxn . It is a linear D.E. since the function y and/or its derivatives are not multiplied
together (that is, no term of the form yy ′′′ , (y ′ ) , y ′ y ′′ , y 4 , etc. occurs). Non-linear D.E.s
2

can be very difficult to solve or may only yield to numerical methods beyond the scope of
this subject.
Examples of linear D.E.s

1. 3xy (4) + y ′′′ + (x − 1)2 y = cos x3

d2 y dy
2. − 3 + 2y = 12e4t ; y(0) = 1, y ′ (0) = 0
dt2 dt
3. y ′′ + 3y ′ + 2y = 3e−2x + x
4. (x2 − 1) y ′′ + 2xy ′ − y = 0

Note
Example 1 is beyond the scope of this subject; examples 2 and 3 can be solved using the
methods developed in this Chapter (and by a number of other important methods beyond
the scope of this subject); and example 4 is a type that will be dealt with in Chapter 5.
Examples 2 and 3 are of a type that is generally easier to deal with because they are Linear
D.E.s with constant coefficients, i.e. the ak ’s are all constants.
Examples of non-linear D.E.s

1. y ′′ + 2xy 2 = 3x (the y 2 terms makes the D.E. non-linear).

2. y ′′ + 2xyy ′ + y = 0 (the yy ′ term makes the D.E. non-linear).
2x
3. y ′′ − = x2 + 1 (this is non-linear since it can be put it the form
y
yy ′′ − (x2 + 1) y = 2x which has the non-linear term yy ′′ ).
2.2. SOLUTION OF LINEAR D.E.S WITH CONSTANT COEFFICIENTS 69

2.2 Solution of linear D.E.s with constant coefficients

Linear D.E.s with constant coefficients are of the form
dn y dn−1 y dy
an n
+ a n−1 n−1
+ . . . + a1 + a0 y = F (x)
dx dx dx
where the ak ’s are all constants.
Notes

1. Such linear D.E.s (L.D.E.s) can be written more conveniently using the differential
d d2
operator, D. Thus, D, is used for ; D2 for 2
and, in general, Dn is used for
dx dx
dn
. Hence the L.D.E. y ′′ + 3y ′ + 4y = sin x can be written as
dxn
D2 y + 3Dy + 4y = sin x

or
(D2 + 3D + 4) y = sin x
or
ϕ(D)y = sin x, where ϕ(D) = D2 + 3D + 4

2. We will not make great use of the linear operator ϕ(D) in this subject. However, it is
a very important concept in the general study of D.E.s, and opens up a whole range
of methods for the solution of L.D.E.s.

3. Consider a L.D.E. with constant coefficients of the form

ϕ(D)y = F (x)

If F (x) is actually 0, then the D.E., ϕ(D)y = 0, is called a homogeneous L.D.E.

(otherwise the L.D.E. is called inhomogeneous.)

Note that this is where the “double use” of the word “homogeneous” occurs! (In the previ-
ous Chapter we considered homogeneous first order D.E.s which were non-linear.)
The study notes will use the term “homogeneous” here too, as this is the term commonly
used in texts. There should (hopefully) be no confusion as the context should always make
the usage clear!
General and particular solutions
The method of approaching the solution of L.D.E.s depends on a crucial theorem:
Consider a L.D.E. of the form
ϕ(D)y = F (x)
Theorem
The general solution of a L.D.E. is given by y = yh + yp , where yp is a particular solution of
the given D.E. (i.e. does not involve any arbitrary constants), and yh is the general solution
of the corresponding (reduced) homogeneous D.E. ϕ(D)y = 0.
70 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

Notes

1. Of course, if the given L.D.E. is, in fact, homogeneous, then we need only to find yh ,
which is the required general solution. (Here yp = 0, in fact. Why?)
2. Clearly there are two major problems confronting us:
• How do we find yh ?
• How do we find yp ?

In the case of a linear D.E. with constant coefficients, i.e. where ϕ(D) is a polynomial,
then there is a very simple standard procedure for finding yh . (See the following section.)
However, for linear D.E.s with non-constant coefficients (i.e. where the coefficients are
functions of x, ak (x)) , then the procedure for finding yh is more complicated. (We will
examine some “elementary” cases in Chapter 5.)
The methods for finding a particular solution yp are numerous and often specialised, de-
pending on the form of the function F (x) on the right-hand side of the D.E.
Notes
In this Chapter, we will consider only three methods:

• the method of undetermined coefficients;

• the method of variation of parameters;
• and the D-operator method.

(In Chapter 6 we will see how Laplace Transforms can be used to transform a given differ-
ential equation to an algebraic equation which can then be solved by alternate methods.)

2.3 Homogeneous solution of linear DE with constant co-

efficients

Read textbook sections 4.2 (Homogeneous linear equations: the general solu-
tion), 6.2 (Homogeneous linear equations with constant coefficients)

Introduction
The solution of a homogeneous L.D.E., ϕ(D)y = 0, is aided by the following two theorems:

1. If y1 (x) and y2 (x) are solutions of this D.E., then so is the linear combination
y(x) = c1 y1 (x) + c2 y2 (x)
2. The solution to any homogeneous linear D.E. of order n consists of a linear combi-
nation of n linearly independent solutions y1 (x), y2 (x), . . . , yn (x) i.e. any solution
of ϕ(D)y = 0 (of order n ) is of the form
y(x) = c1 y1 (x) + c2 y2 (x) + . . . + cn yn (x)
where c1 , c2 , . . . , cn are arbitrary constants.
2.3. HOMOGENEOUS SOLUTION 71

Second order L.D.E. Because of the Theorem above, the problem of finding the solution
of a homogeneous L.D.E. has been reduced to the problem of finding the correct number of
linearly independent solutions. The general solution is immediately a linear combination
of these solutions. We will now develop the method for finding such solutions for L.D.E.s
with constant coefficients.
Note
We will concentrate on the case of a second-order L.D.E. with constant coefficients, but the
method is entirely general.
Consider the homogeneous L.D.E.
dn y dy
an + . . . + a 1 + a0 y = 0
dxn dx
i.e.
(an Dn + . . . + a1 D + a0 ) y = 0
where an , . . . , a1 , a0 are fixed constants.
The general solution to such a D.E. is found using the following steps.

1. Form the corresponding characteristic (or auxiliary) equation

an rn + . . . + a1 r + a0 = 0

if the D.E. is given by ϕ(D)y = 0, form the polynomial equation ϕ(r) = 0.

2. Solve the characteristic equation getting the characteristic roots r1 , r2 , . . . , rn
3. The final step depends on the nature of these n roots. There are three possibilities:
I. the n roots are real and distinct. In this case, the general solution of the D.E. is
given by
y = c1 er1 x + c2 er2 x + . . . + cn ern x
where c1 , c2 , . . . , cn are arbitrary constants.
II. a root is “repeated”; say r is a root of multiplicity k. In this case, that part of the
general solution corresponding to the root r is given by

(b1 + b2 x + . . . + bk xk−1 ) erx

where b1 , b2 , . . . , bk are arbitrary constants.

III. a pair of complex number roots of the form α ± βi occurs. In this case, that part
of the general solution corresponding to these two roots is given by

eαx (A cos βx + B sin βx)

where A and B are arbitrary constants.

Note
To solve homogeneous L.D.E.s with constant coefficients, you must learn these three steps
and the three Cases. (The three “rules” in these cases mathematically guarantee the re-
quired linear independence. The only real difficulty is finding the roots of the characteristic
equation!)
72 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

Example 1

Solve
y ′′ − 3y ′ + 2y = 0
Solution
This linear D.E. can be written in the form

(D2 − 3D + 2) y = 0

∴ Let r2 − 3r + 2 = 0 i.e. (r − 2)(r − 1) = 0, and so the roots are r = 2 and r = 1, which are
distinct (Case I).
∴ The required general solution is given by

y = c1 e2x + c2 ex

(Check by differentiation that this is indeed the solution of the given D.E.!)

Example 2

Solve
d3 y d2 y dy
3
− 4 2
+ 5 − 2y = 0
dx dx dx
Solution
The operator form is usually more “helpful”: i.e.

(D3 − 4D2 + 5D − 2) y = 0

∴ Let r3 − 4r2 + 5r − 2 = 0. After some effort, we find that the roots are r = 1, 1, 2 since the
equation can be written as (r − 2)(r − 1)2 = 0
∴ Using Cases I and II, the general solution of this D.E. is given by

y = c1 e2x + (c2 + c3 x) ex . (Check!)

Example 3

Solve
(D2 + 2D + 5) y = 0
Solution
Here we let r2 + 2r + 5 = 0. The roots of this equation are
√
−2 ± −16
r= = −1 ± 2i (Case III)
2
∴ The general solution of the D.E. is given by

y = e−x (A cos 2x + B sin 2x). (Check!)

Example 4
2.4. METHOD OF UNDETERMINED COEFFICIENTS 73

Solve
(D4 − 5D2 + 12D + 28) y = 0, given that
(D4 − 5D2 + 12D + 28) = (D + 2)2 (D2 − 4D + 7)
Solution
Here the characteristic equation is

r4 − 5r2 + 12r + 28 = (r + 2)2 (r2 − 4r + 7) = 0

√
∴ The characteristic roots are r = −2, −2, 2 ± 3i. Hence the general solution of the D.E. is
given by
√ √
y = (c1 + c2 x) e−2x + e2x (c3 cos 3x + c4 sin 3x) . (Easy!)

Tutorial 2 Question 1

2.4 Method of undetermined coefficients

Read textbook section 4.4 (Non-homogeneous equations: the method of un-

determined coefficients)

Video (by Dr. R. Wood): Method of undetermined coefficients

Introduction
To find the general solution of the L.D.E. given by ϕ(D)y = F (x), we have already noted
that we must find a particular solution yp of this equation and add it to the solution yh of
the corresponding homogeneous equation ϕ(D) = 0. (Note that yh is normally found first.)
In the last section we discovered how to find yh . In this section and the next we will discover
how to obtain a particular solution, yp .
Undetermined coefficients
There are many methods for finding yp . A general method widely used in science is the
method of undetermined coefficients. This method is relatively simple to comprehend, but
unfortunately it can become unwieldy and even incapable of use in certain cases.
The use of this method for solving ϕ(D)y = F (x) depends on the “form” of F (x). If
F (x) contains a polynomial; terms of the form ekx ; cos kx; sin kx, (where k is a constant);
or even combinations of sums and products of these functions, then the method of unde-
termined coefficients may be useful. (We will not consider very “complicated” cases. A
number of “exceptions” can also occur when we are trying to use the method. In such
cases, alternative methods might profitably be sought!)
We use the pattern from the following table to “guess” an appropriate form for yp whose
undetermined coefficient(s) must then be found by “equating coefficients” as illustrated in
the examples.
74 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

Given F (x) as: Try yp as

ekx aekx , and find a.
c a , and find a.
kx + l or kx ax + b, and find a, b
kx2 + lx + m ax2 + bx + c, and find a, b, c.
(If l or m = 0, still use the same yp )

Cubic polynomial etc. ax3 + bx2 + cx + d, and find a, b, c, d.

etc. for higher polynomials

sin kx or cos kx or both a sin kx + b cos kx, and find a, b.

Note
For any F (x) made up of a sum (or product) of the above, just try for yp the corresponding
sum (or product) of the yp forms e.g. if F (x) = (x2 − 4) e3x , try yp = (ax2 + bx + c) e3x ,
and try to find a, b, c.
Note that a trial yp = (ax2 + bx + c) dex is superfluous, since dex is already covered in the
first trial. Note further that, if the trial yp “fails to work”, i.e. if you cannot find the values
of the undermined coefficients, then it may be necessary to “adjust” yp , as shown later.

Example 1

Find a particular solution of the L.D.E.

y ′′ + 4y = 9e3x .
Solution
We can make a reasonable assumption about the “form” of yp ∶ since we are looking for a
function whose second derivative added to four times the function equals 9e3x , it is reason-
able to “guess” that yp must be of the form yp = ae3x , where a must now be determined.
∴ Let y = ae3x and substitute in the given D.E. (Here, y ′ = 3ae3x and y ′′ = 9ae3x )
∴ (9ae3x ) + 4 (ae3x ) = 9e3x
i.e. 13ae3x = 9e3x . The only way that this can be true for all x is for the “coefficients” of
e3x to be same on both sides of the equation, i.e. we equate the coefficients, getting
13a = 9
9
a=
13
Hence,
9 3x
yp = e
13
(This is a particular solution since it does not contain an arbitrary constant. It is (automati-
cally) a solution as you should check!)
Note
To find the general solution of the given D.E., we must solve y ′′ + 4y = 0, getting (easily)
yh = A cos 2x + B sin 2x.
2.4. METHOD OF UNDETERMINED COEFFICIENTS 75

∴ The general solution is given by

y = yp + yh
9 3x
i.e. y = A cos 2x + B sin 2x + 13 e

(As “required”, this solution of the given second order L.D.E contains two arbitrary con-
stants and satisfies the given D.E. Check!)

Example 2

Find yp for
(D2 + 2D − 3) y = x2 + 4
Solution
Here F (x) is a quadratic polynomial. Clearly polynomials when differentiated yield poly-
nomials. In general (apart from some “exceptions”), when F (x) is a polynomial of degree
k, we assume that yp is a polynomial of degree k whose coefficients must be determined.
∴ Here, let yp = ax2 + bx + c
∴ y ′ = 2ax + b
and y ′′ = 2a.
Now substituting in the given D.E., y ′′ + 2y ′ − 3y = x2 + 4, we get

(2a) + 2(2ax + b) − 3 (ax2 + bx + c) = x2 + 4

∴ − 3ax2 + (4a − 3b)x + (2a + 2b − 3c) = x2 + 4

Equating coefficients, we get

⎧
⎪ −3a = 1
⎪
⎪
⎨ 4a − 3b = 0
⎪
⎪
⎩ 2a + 2b − 3c = 4
⎪
Hence, a = − 31 , b = − 49 , c = − 50
27

1 4 50
y p = − x2 − x − .
3 9 27
Example 3

Solve
y ′′ + 2y ′ + y = 2 cos 2x
Solution
Here F (x) = 2 cos 2x, and so we assume a form of a sin 2x + b cos 2x for yp

i.e. let yp = a sin 2x + b cos 2x

∴ y ′ = 2a cos 2x − 2b sin 2x
and y ′′ = −4a sin 2x − 4b cos 2x
∴ (−4a sin 2x − 4b cos 2x) + 2(2a cos 2x − 2b sin 2x)
+(a sin 2x + b cos 2x) = 2 cos 2x

i.e. (−3a − 4b) sin 2x + (4a − 3b) cos 2x = 2 cos 2x.

76 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

∴ Equating coefficients, we get

−3a − 4b = 0
{
4a − 3b = 2

We must now solve these two (linear) equations simultaneously, getting

8 6
a= and b = −
25 25
Hence
8 6
yp = sin 2x − cos 2x
25 25
To find yh , we must solve
y ′′ + 2y ′ + y = 0
∴ solve r2 + 2r + 1 = 0, getting r = −1, −1

∴ yh = (c1 + c2 x) e−x

Hence the general solution of the given D.E. is

8 6
y = (c1 + c2 x) e−x + sin 2x − cos 2x.
25 25

Example 4

Find yp for
y ′′ + 3y ′ + 2y = 4e−2x
Solution
Here, we note the form of the right-hand side as 4e−2x , and so we try letting

yp = ae−2x

∴ y ′ = −2ae−2x and y ′′ = 4ae−2x

Substituting we get

(4ae−2x ) + 3 (−2ae−2x ) + 2 (ae−2x ) = 4e−2x

or
0 = 4e−2x
an “impossible” situation since e−2x is never equal to zero! This is an “exceptional” case.
The problem is that the form we have assumed for yp is, in fact, part of yh
In fact, letting y ′′ + 3y ′ + 2y = 0, and hence solving r2 + 3r + 2 = 0, gives r = − 1,-2 and so

yh = c1 e−x + c2 e−2x

Since there is already a term with e−2x in the homogeneous part of the solution, we cannot
use this form in yp .
2.4. METHOD OF UNDETERMINED COEFFICIENTS 77

Note
The trick in such situations is to multiply the “normal” assumed term by a power of x which
is sufficiently high (but not higher) so that none of the “new” assumed terms for yp are in
yh .
Hence, here yp = x (ae−2x ) should work!
∴ y = axe−2x , y ′ = ae−2x − 2axe−2x
y ′′ = −2ae−2x − 2ae−2x + 4axe−2x = −4ae−2x + 4axe−2x
Now, on substituting in the given D.E. we get
(−4ae−2x + 4axe−2x ) + 3 (a−2x − 2axe−2x ) + 2axe−2x = 4e−2x
∴ − ae−2x = 4e−2x
Hence, a = −4 ∴ yp = −4xe−2x , and the general solution is given by
y = c1 e−x + c2 e−2x − 4xe−2x .

Notes

1. These exceptional cases can cause some problems, but these can usually be overcome
by some “experimentation”, using the given trick until no “contradiction” is reached.
It is important always to find yh first since its form, as noted above, can influence our
choice for yp .
2. You will not be required to carry out very “complicated” solutions using this method
- it is the “trial-and-error” approach that is worth noting as a technique! In any such
case, you just adjust the usual “yp ” by multiplying it by an appropriate power of x so
that no term in yh appears in the new yp , e.g. the yp for y ′′ − 2y ′ + y = ex needs to be
ax2 ex (Try it!).

Example 5

Find the general solution of

y ′′ + y ′ = x2 − 1
Solution
The characteristic equation is r2 + r = 0, with roots r = 0, −1. The homogeneous solution is
yh = A + Be−x .
For the particular solution we try yp = x (Ax2 + Bx + C) as the homogeneous solution
already contains an arbitrary constant.
Now yp′ = 3Ax2 + 2Bx + C
yp′′ = 6Ax + 2B
∴ 6Ax + 2B + 3Ax2 + 2Bx + C = x2 − 1
3Ax2 + (6A + 2B)x + (2B + C) = x2 − 1
On equating coefficients
3A = 1 6A + 2B = 0 2B + C = −1
A = 31 B = −1 C=1
78 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

Hence
1 1
yp = x ( x2 − x + 1) = x3 − x2 + x.
3 3
The general solution is
1
y = yh + yp = A + Be−x + x3 − x2 + x.
3

Example 6

Find the general solution of

y ′′ + 4y = x2 + 3ex y(0) = 0, y ′ (0) = 2
Solution
The characteristic equation is r2 + 4 = 0, with roots r = ±2i.
The homogeneous solution is yh = A cos 2x + B sin 2x.
For the particular solution we try
yp = Ax2 + bx + c + Dex
∴ yp′ = 2Ax + b + Dex
yp′′ = 2A + Dex
∴ 2A + Dex + 4 (Ax2 + bx + c + Dex ) = x2 + 3ex
On equating coefficients
4A = 1, 4b = 0, 2A + 4c = 0, 5D = 3
1 −1 3
A= , b = 0, c= , D= .
4 8 5
∴ The general solution is
1 1 3
y = A cos 2x + B sin 2x + x2 − + ex
4 8 5
Now, finding the solution given the initial conditions y(0) = 0, and y ′ (0) = 2, i.e. y = 0
when x = 0 and y ′ = 2 when x = 0.
1 1 3
y = A cos 2x + B sin 2x + x2 − + ex
4 8 5
3 1
∴ 0=A+ −
5 8
−19
∴ A=
40
′ 1 3
y = −2A sin 2x + 2B cos 2x + x + ex
2 5
3
∴ 2 = 2B +
5
7
∴ B=
10
−19 7 1 1 3
y= cos 2x + sin 2x + x2 − + ex
40 10 4 8 5

Tutorial 2 Question 2
2.5. METHOD OF VARIATION OF PARAMETERS 79

2.5 Method of variation of parameters

Read textbook section 4.6 (Variation of parameters)

Video (by Dr. R. Wood): Method of variations of parameters

Introduction
As noted above, the method of undetermined coefficients is effective only when the right-
hand side of the D.E. is of a special form. An alternative method, due to Lagrange, provides
a way of finding y p in all cases where the method of undetermined coefficients can be
applied, and also in additional cases. This method is rather ingenious, and is called the
method of variation of parameters. The method can, in fact, also be applied to linear D.E.s
which do not have constant coefficients.
We will consider only the case of a second order linear D.E. of the form

p(x)y ′′ + q(x)y ′ + r(x)y = s(x)

Notes

1. We will only consider examples where p, q, r are constant functions.

2. The disadvantage with this method of Lagrange is that it requires the finding of cer-
tain integrals which may, in practice, be quite difficult!

3. No justification for the method will be given or required: a proof will be found in any
standard text on D.E.s. You need to concentrate on the technique, and must learn the
required formulae, as detailed below.

Method of variation of parameters

Consider a linear D.E. of the form

a2 y ′′ + a1 y ′ + a0 y = R(x)

where a0 , a1 and a2 are constants and a2 ≠ 0.

Step 1: Find the solution of the reduced homogeneous D.E., i.e. find yh for the D.E.

a2 y ′′ + a1 y ′ + a0 y = 0.

Hence, yh = Ay1 (x) + By2 (x) as usual, where A and B are the two arbitrary
constants. For simplification we will use u = y1 (x) and v = y2 (x).

Step 2: Lagrange’s “trick” now lies in the method for finding yp . Following him we assume
that yp is of the form
yp = A(x)u(x) + B(x)v(x),
80 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

where u and v are the functions found in yh in Step 1, and A(x) and B(x) are
functions of x which must be found by solving simultaneously the two equations

A′ u + B ′ v = 0
{
A′ u′ + B ′ v ′ = aR2

Note We have written A′ , y1 , etc. for convenience, instead of A′ (x), y1 (x) etc.

Step 3: The above equations, when solved, actually yield A′ (x) and B ′ (x). A(x) and
B(x) must then be found by integration.
Note. Finding A′ (x) and B ′ (x) can be simplified using determinants.
For the two equations, using determinants we can write

0 v u 0
∣ ∣ ∣ ∣
R/a2 v ′ ′
u R/a2
A′ (x) = and B ′ (x) =
u v u v
∣ ∣ ∣ ∣
u′ v ′ u′ v ′

The determinant in the denominator need only be evaluated once and this also
simplifies the working.

Step 4: The required yp is then given at once by

yp = A(x)y1 + B(x)y2

and the general solution is, as always, given by

y = yh + yp

i.e
y = (Ay1 + By2 ) + (A1 (x)y1 + B1 (x)y2 )
where A and B are the 2 arbitrary constants for the given 2 nd order D.E and y1 , y2
A(x), B(x) are all functions which have been found.

Example 1

Solve
y ′′ − y = xex
Solution.
Here the homogeneous equation y ′′ − y = 0 yields yh = Aex + Be−x . Thus, if we were to
use the method of undetermined coefficients to find yp , we would have to make certain
“adjustments” as per the exceptions in the previous section. (Why?) To use the variation of
parameters method we let

yp = A(x)ex + B(x)e−x = Au + Bv

To find A(x) and B(x) we must solve

A′ ex + B ′ e−x = 0
{ xex
A′ ⋅ (ex ) + B ′ ⋅ (−e−x ) = 1
2.5. METHOD OF VARIATION OF PARAMETERS 81

We can easily eliminate B ′ , say, by adding the two equations:

∴ 2A′ ex = xex
1
∴ A′ = x
2
1
and so A(x) = x2
4
Note that no arbitrary constant is required here since we are finding a particular solution.
We can find B ′ similarly by eliminating A′ from the 2 equations or by substituting for A′
in the first equation, say. i.e.,
1
( x) ex + B ′ e−x = 0
2
1
∴ B ′ = − xe2x
2
and so
1
B(x) = ∫ (− xe2x ) dx
2
which can be found by integration by parts. i.e.
1 1
B(x) = − xe2x + e2x
4 8
1 1 1
∴ yp = ( x2 ) ex + (− xe2x + e2x ) e−x
4 4 8
or
1 1 1
yp = x2 ex − xex + ex .
4 4 8
Alternatively using determinants to find A(x) and B(x) we have u = ex , v = e−x , R = xex ,
and a2 = 1 giving

u v ex e−x
∣ ∣ = ∣ ∣ = ex ⋅ (−e−x ) − ex ⋅ (e−x ) = −ex−x − ex−x
u′ v ′ ex −e−x

= −e0 − e0 = −1 − 1 = −2
Then
0 v 0 e−x
∣ ∣ ∣ ∣
R/a2 v ′ xex /1 −e−x 0 − xex ⋅ e−x −x x
A′ (x) = = = = =
−2 −2 −2 −2 2
and
u 0 ex 0
∣ ′ ∣ ∣ ∣
u R/a2 ex xex ex ⋅ xex − 0 −xe2x
B ′ (x) = = == =
−2 −2 −2 2
as before.
The general solution to the given D.E. is
1 1 1
y = Aex + Be−x + x2 ex − xex + ex
4 4 8
or
1 1
y = Cex + Be−x + x2 ex − xex .
4 4
82 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

You should convince yourself that this method really works by checking this solution in the
given D.E. y ′′ − y = xex , and by finding yp by the method of undetermined coefficients.

Example 2

Solve
y ′′ + y = sec x.

(Note that this D.E. cannot be solved using the method of undetermined coefficients.)
Solution (Outline of steps)

1.
yh = A sin x + B cos x

∴ yp = A(x) sin x + B(x) cos x

2. Solve
′ ′
⎪ A sin x + B cos x = 0
⎧
⎪
⎨ ′ 1
⎪
⎪ A cos x − B ′ sin x = sec x =
⎩ cos x
∴ A′ = 1

and so A(x) = x.
(Hint: multiply the first equation by sin x and the second by cos x and add. This is
an important “trick” for such equations!)
Similarly B ′ = − tan x and so

B(x) = − ∫ tan xdx = ln(cos x).

3. The general solution is given at once by

y = A sin x + B cos x + x sin x + cos x(ln cos x).

For most practical purposes the method of undetermined coefficients supplemented by the
method of variation of parameters in cases where the former is inapplicable are enough to
solve any linear equation with constant coefficients.
However, there are a number of other methods, some of which are quite ingenious, or based
on tables of “formulas”, and so on. One particularly “elegant” method involves the use of
D-operator formulas and theorems.
We give a very brief introduction to this method in the next section. The textbook does not
cover this method in full detail, so the main source of information for this method is the
study guide.

Tutorial 2 Question 3
2.6. D-OPERATORS 83

2.6 D-operators
Video (by Dr. R. Wood): D-operators. Part 1
The concept of a differential operator is new for most of you. So, it is recommended to
spend more time on this section paying special attention to the examples.
D-operators can be used to obtain particular solutions of linear D.E.s with constant coeffi-
cients. A linear D.E. can be written in the form

ϕ(D)y = f (x),

where ϕ(D) is an operator of the form

ϕ(D) = an Dn + an−1 Dn−1 + . . . + a1 D + a0

where a0 , a1 , . . . , an are constants. Since a particular solution satisfies the given D.E., we
can write
ϕ(D)yp = f (x)
This equation can then be “solved” to give
1
yp = f (x),
ϕ(D)
1
where is an “operator” which operates on f (x). At this stage, it is not clear what
ϕ(D)
1
operator does, so we should treat carefully. However, if f (x) is a particular type of
ϕ(D)
1
function, some specific rules can be determined for the operation f (x).
ϕ(D)
1
The operator ϕ(D) is called inverse for ϕ(D). These operators “cancel” the effects of each
other, that is
1
ϕ(D) f (x) = f (x)
ϕ(D)
or
1
ϕ(D)f (x) = f (x).
ϕ(D)
Let us study some of the properties of differential operators. Suppose f (x) = emx , then

d mx
Df (x) = Demx = (e ) = memx
dx
d2 mx
D2 f (x) = D2 emx = (e ) = m2 emx
dx2
⋮
d n
Dn f (x) = Dn emx = n (emx ) = mn emx
dx
84 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

Example 1

Find
(2D3 − 3D + 4) e2x .
Here,
(2D3 − 3D + 4) e2x = 2D3 e2x − 3De2x + 4e2x
= 2 ⋅ (23 ) e2x − 3(2)e2x + 4e2x
= (2 ⋅ (23 ) − 3(2) + 4) e2x
= ϕ(2)e2x ,
where ϕ(m) = 2m3 − 3m + 4.
Note that ϕ(D)e2x = ϕ(2)e2x , where ϕ(D) is the operator 2D3 − 3D + 4 and ϕ(2) is the
polynomial 2 (23 ) − 3(2)(= 14). Hence, if ϕ(D) is a polynomial in D, then

ϕ(D)emx = ϕ(m)emx

Verify this result for (D2 + 4D + 3) e2x . It should be observed that ϕ(m) is a constant, and
not an operator.

Example 2

Determine
ϕ(D)e−2x if ϕ(D) = 3D2 + 2D + 5.

Here,
ϕ(D)e−2x = (3D2 + 2D + 5) e−2x
= 3D2 e−2x + 2De−2x + 5e−2x
= 3(−2)2 e−2x + 2(−2)e−2x + 5e−2x
= 13e−2x
= ϕ(−2)e−2x since ϕ(−2) = 13.
Now, since ϕ(D)emx = ϕ(m)emx
1 1
ϕ(D)emx = ϕ(m)emx
ϕ(D) ϕ(D)
1 mx
emx = ϕ(m) e
ϕ(D)
(since ϕ(m) is a constant ) or
1 mx 1 mx
e = e ,
ϕ(m) ϕ(D)
where we assumed ϕ(m) ≠ 0.
This result is used in obtaining a particular solution to D.E.s of the form ϕ(D)y = emx .

Example 3

Find a particular solution yp of the D.E.

y ′′ + 2y ′ + 3y = e−x .
2.6. D-OPERATORS 85

The D.E. is ϕ(D)yp = e−x , where ϕ(D) = D2 + 2D + 3

1 −x 1
∴ yp = e = e−x
ϕ(D) ϕ(−1)
by the result above
1
i.e. yp = e−x ( since ϕ(−1) = 2 ).
2
Confirm this result using undetermined coefficients, and check that yp = 12 e−x is a solution
of the given D.E. (This is a “powerful” method, is it not?! But, it is certainly not “all-
powerful”, see Example 5.)

Example 4

Find a particular solution yp of the D.E.

Here,
y ′′′ − 3y ′′ − 2y ′ − 5y = e2x
ϕ(D)yp = e2x ; where ϕ(D) = D3 − 3D2 − 2D − 5.
1 2x 1 2x −1 2x 1
∴ yp = e = e = e = − e2x
ϕ(D) ϕ(2) 13 13
1 2x
Hence, yp = − 13 e is a particular solution of the given D.E.

Example 5

Find a particular solution yp of the D.E.

y ′′ + 2y ′ − 8y = e2x

Now
1 2x
yp = e ; where ϕ(D) = D2 + 2D − 8
ϕ(D)
1 2x
∴ yp = e ; but ϕ(2) = 22 + 2 ⋅ 2 − 8 = 0
ϕ(2)
1
Hence is undefined! Hence, the result
ϕ(2)
1 mx 1 mx
e = e
ϕ(D) ϕ(m)
needs modification. The statement should be
1 mx 1 mx
e = e provided ϕ(m) ≠ 0.
ϕ(D) ϕ(m)
Example 5 is typical of those problems in which an alternative method of solution is needed,
as we have seen earlier. (“Adjustment” is needed here as e2x is part of the homogeneous
solution, yh i.e. Example 5 is an example of:
“Determine a particular solution yp of the D.E. ϕ(D)y = f (x), where f (x) is a part of the
solution of the homogeneous D.E. ϕ(D)y = 0.”
For example, if ϕ(D)y = emx and ϕ(D)emx = 0 (that is, emx is a solution of the homo-
geneous D.E.), then ϕ(m) = 0. This type of problem can only be solved using advanced
86 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

techniques based on the “Shift Theorem”, covered later. (We had the same “problem” of
adjustment previously, as in Example 4 of the last section.)
Note
The D -operator method can also be used to find a particular solution of the D.E.

ϕ(D)y = f (x),

where f (x) is sin mx or cos mx. Since eiθ = cos θ + i sin θ (known fact from Algebra!) then

eimx = cos mx + i sin mx

Hence,
cos mx = Re (eimx ) ; (the real part of eimx )
and
sin mx = Im (eimx ) ; ( the imaginary part of eimx ) .

Example 6

Find the particular solution of the D.E.

y ′′ + 2y ′ − 4y = cos 2x

Here,
1
yp = cos 2x
D2
+ 2D − 4
1
= 2 Re (e2ix )
D + 2D − 4
1
= Re { 2 e2ix }
D + 2D − 4
1
= Re {e2ix 2
} (by our earlier work)
(2i) + 2(2i) − 4
−8 − 4i
= Re {e2ix } (complex number arithmetic)
(−8 + 4i)(−8 − 4i)
−8 − 4i
= Re {(cos 2x + i sin 2x) ( )}
80
1 i
= Re { (−8 cos 2x + 4 sin 2x) + (−8 sin 2x − 4 cos 2x)}
80 80
and taking the real part,
1
yp = (−8 cos 2x + 4 sin 2x)
80
1 1
= sin 2x − cos 2x.
20 10
This has been found somewhat “more easily” than in finding yp by undetermined coeffi-
cients. Do you agree? Try it!
In this example, the identity
1 1 a + ib a + ib a + ib
= ⋅ = 2 2
=
a − ib a − ib a + ib a − (ib) a + b2
was used.
2.7. SHIFT THEOREM 87

Example 7

Find a particular solution of the D.E.

2y ′′ + 3y ′ + 15y = sin 3x

Now,
1
yp = Im {e3ix }
2(3i)2 + 3(3i) + 15
−1 − 3i
= Im {(cos 3x + i sin 3x) ( )}
30
1
= − (sin 3x + 3 cos 3x)
30
1 1
Hence a particular solution of the given D.E. is − 10 cos 3x − 30 sin 3x. (Not bad as a fast,
“recipe” method!)

2.7 Shift theorem

If a is any constant and y is any function, then
1 ax 1
e y = eax y
ϕ(D) ϕ(D + a)

with the extra rule:

1 1
a= a,
ϕ(D) ϕ(0)
where a is a constant, provided ϕ(0) ≠ 0.
For the case ϕ(0) = 0 simply factor out the desired multiple of D to reduce ϕ(D) and treat
separately.
Also,
1
a = ax
D
since 1/D is the inverse of the differential operator D, we simply integrate the constant
once and setting the constant of integration to zero.

Example 8

Find a particular solution of

y ′′ + 9y = cos 3x.
From before, trying
1
yp = Re { e3ix }
D2 +9
we try
1
Re {e3ix }
(3i)2 + 9
but (3i)2 + 9 = 0 and the method fails.
88 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

Now we can try the shifting theorem.

1 3ix 3ix 1 3ix 1 3ix 1
e = e 1 = e 1 = e 1
D2 + 9 (D + 3i)2 + 9 D2 + 6iD − 9 + 9 D2 + 6iD
We want to find
1
.
D2 + 6iD
Trying ϕ(0) gives 0 so we try factoring ϕ(D) = D2 + 6iD = D(D + 6i). Then
1 1 1 1 1 1 x
1= 1= 1= 1=
D2 + 6iD D(D + 6i) D 0 + 6i 6i D 6i
giving
1 3ix 3ix x e3ix
e = e = x
D2 + 9 6i 6i
Then
e3ix cos 3x + i sin 3x x
yp = Re { x} = x Re { } = sin 3x.
6i 6i 6
Check:
3x 1 x 1
yp′ = cos 3x + sin 3x = cos 3x + sin 3x
6 6 2 6
3x 1 3 3x
yp′′ = − sin 3x + cos 3x + cos 3x = − sin 3x + cos 3x
2 2 6 2
Then
3x 9x 3x 3x
yp′′ + 9yp = − sin 3x + cos 3x + sin 3x = − sin 3x + cos 3x + sin 3x = cos 3x.
2 6 2 2
So, yes, it works!

2.8 Finding yp when R is a polynomial

Video (by Dr. R. Wood): D-operators. Part 2
For any polynomial
R = an xn + . . . + a1 x + a0
we can always find
Dm R = 0.
Example 9

If R = x3 + 2x + 3 then D4 R = 0 (try it).

So if we can write
1 1 1
R= R = [1 − φ(D) + φ(D)2 + . . .] R
ϕ(D) a[1 + φ(D)] a
then the series (operating on R ) should terminate. Compare the formula above with the
geometric series
1
= 1 + x + x2 + x3 + ⋯
1−x
2.8. FINDING yp WHEN R IS A POLYNOMIAL 89

Example 10

dy
− y = x3 + 2x + 3
dx

We have ϕ(D) = D − 1 giving

1
yp = (x3 + 2x + 3)
D−1
−1
= (x3 + 2x + 3)
1−D
= − (1 + D + D2 + D3 + D4 + . . .) (x3 + 2x + 3)
= − [(x3 + 2x + 3) + (3x2 + 2) + (6x) + 6]
= −x3 − 3x2 − 8x − 11.

This is the same as the answer found using the method of undetermined coefficients and in
a lot less lines.

Example 11

Find a particular solution of

(D2 − 4D + 4) y = x

Solution
1 1 1
yp = x= x= 2
D2 − 4D + 4 4 − 4D + D 2
4 [1 − (D − D4 )]
1 D2 D2 2
= [1 + (D − ) + (D − ) + . . .] x
4 4 4
1 1 1
= (x + 1 + 0) = x + .
4 4 4

Sneaky trick
Knowing that the series expansion is generated using powers of D that occur in the de-
nominator on the LHS, if we see that Dm R = 0 then any terms of Dm or higher can be
eliminated at the beginning.
Since R = x then D2 x = 0 and we can simplify the calculation of yn as follows:

1 1 1
yp = x= x= x
D2
− 4D + 4 −4D + 4 4(1 − D)
1 1
= (1 + D + D2 + ⋯) x = (x + 1)
4 4

as before.

Tutorial 2 Question 4
90 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

2.9 Applications of linear differential equations

Read textbook section 3.5 (Electric circuits), section 5.7 (Electrical systems)

The differential equation

d2 x dx
a + b + cx = F (t),
dt dt
where a, b, c are given constants or functions of t and F (t) is a given function of t, occurs
frequently in applications of physics and engineering.
Second order linear differential equations for RLC circuits
In this section we will apply our standard methods for second order linear differential equa-
tions to the solution of circuits containing resistors, inductors and capacitors together - the
RLC circuit. You will be expected to be able to use the auxiliary equation method to find
complementary solutions of the homogeneous equation and the method of undetermined
coefficients to find particular solutions.
Consider now the RLC circuit shown below.

If the current flowing in the circuit is I and the charge on the capacitor is Q, then Kirchoff’s
Law gives
El + Er + Ec = E,
or
dI 1
L+ RI + ⋅ Q = E,
dt C
where E is the electromotive force supplied by the battery or generator.
dQ
Since I = dt , this equation becomes,

d2 Q dQ 1
L 2
+R + Q=E
dt dt C
a second order linear differential equation.
Example
2.9. APPLICATIONS OF LINEAR DIFFERENTIAL EQUATIONS 91

An RLC circuit is supplied by a generator having voltage given by E = 24 sin 10t. The
inductor is 1 henry, the resistor is 12 ohms and the capacitor is 0.01 farads. If at time t = 0
the switch is closed (i.e. at t = 0, Q = 0 and I = 0 ), find the charge Q and the current I at
time t.
Solution
The RLC circuit is as follows

and by Kirchoff’s Law we have

El + Er + Ec = E

so that
dI 1
L + RI + ⋅ Q = E, or
dt C
2
dQ dQ 1
L 2 +R + ⋅ Q = E,
dt dt C
which in this case becomes
d2 Q dQ
2
+ 12 + 100Q = 24 sin 10t
dt dt
To solve this equation we first find the corresponding solution Qc of the homogeneous
equation
d2 Q dQ
2
+ 12 + 100Q = 0
dt dt
The auxiliary equation is
m2 + 12m + 100 = 0
which has complex roots, m = −6 ± 8i.
Hence the complementary solution is

Qc = e−6t (A cos 8t + B sin 8t)

We next find a particular solution of the equation

d2 Q dQ
+ 12 + 100Q = 24 sin 10t
dt2 dt
by the method of undetermined coefficients. To do this we try a particular solution of the
equation of the form
Qp = a cos 10t + b sin 10t
92 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

(these are the only functions which can give the right-hand side when differentiated once
and twice).
Substituting Qp into the differential equation and solving for a and b gives a = − 15 and b = 0
(the solution of two simultaneous equations in a and b ). (Or try the D operator method.)
Hence
1
Qp = − cos 10t
5
and so the general solution is
1
Q = Qp + Qc = e−6t (A cos 8t + B sin 8t) − cos 10t
5
where the constants A and B can be found from the initial conditions (at t = 0, Q = 0 and
I = dQ
dt = 0). The condition Q = 0 at t = 0 gives

1
0 = e0 (A cos 0 + B sin 0) − cos 0
5
so that A = 51 .
Differentiating Q we get
dQ
I= = − 6e−6t (A cos 8t + B sin 8t)
dt
+ e−6t (−8A sin 8t + 8B cos 8t) + 2 sin 10t

so that, using the initial condition we have

1 8
0 = −6 ( cos 0 + B sin 0) + (− sin 0 + 8B cos 0) + 2 sin 0
5 5
3
which gives B = 20 .

Hence the charge Q is given by

e−6t 1
Q= (4 cos 8t + 3 sin 8t) − cos 10t
20 5
and the current I is
5e−6t
I =− sin 8t + 2 sin 10t.
2
The transient current, − 52 e−6t sin 8t dies away very quickly as t gets larger and the steady-
state current 2 sin 10t approximates the actual current when the transient current becomes
negligible. This is illustrated in the following sketch of the actual current I (steady-state
plus transient) and the steady-state current 2 sin 10t.
2.9. APPLICATIONS OF LINEAR DIFFERENTIAL EQUATIONS 93

The property of oscillation in the steady-state solution is very useful in certain applications.
For example, radio, television, radar and communication systems all exploit properties of
electrical oscillation (resonance). We tune our radios into certain frequencies and the oscil-
lating electronic circuits in the radio then amplify these frequencies to eventually produce
sound.
Another interesting application of electric circuits is in the analog computer. This device
simulates mechanical and physical systems by replacing the differential equation repre-
senting them by their electrical analogs, and then solving experimentally the resulting dif-
ferential equation for the electric circuits. The solution can be displayed on a cathode ray
oscilloscope in the form of a graph to those shown above for charge or current.

Tutorial 2 Question 5
94 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

2.10 Tutorial 2
Back to Chapter 2

Question 1

Find a general solution of each of the second-order linear differential equations below.

1. y ′′ − 3y ′ + 2y = 0
2. y ′′ + 5y ′ = 0
3. 3y ′′ − 10y ′ + 3y = 0
4. y ′′ + 25y = 0
5. y ′′ − 25y = 0
6. y ′′ − 4y ′ + 8y = 0
7. y ′′ + 10y ′ + 125y = 0
8. 9y ′′ − 6y ′ + 10y = 0
9. y ′′ = 7y
10. Textbook exercise set 4.2.
Problems 1, 5, 9, 13, 17, 19 (8,8e-th editions).
Problems 4, 13, 18, 19 (9th edition).

Answers

1. y(x) = C1 ex + C2 e2x
2. y(x) = C1 + C2 e−5x
x
3. y(x) = C1 e 3 + C2 e3x
4. y(x) = C1 cos 5x + C2 sin 5x
5. y(x) = C1 e5x + C2 e−5x
6. y(x) = e2x (C1 cos 2x + C2 sin 2x)
7. y(x) = e−5x (C1 cos 10x + C2 sin 10x)
x
8. y(x) = e 3 (C1 cos x + C2 sin x)
√ √
9. y(x) = C1 exp(x 7) + C2 exp(−x 7)
10. see textbook or detailed solutions

Detailed solutions to Question 1

Back to Chapter 2
2.10. TUTORIAL 2 95

Question 2

Use the method of undetermined coefficients to find a particular solution for each of the
given equations:

1. y ′′ + 2y ′ + 2y = 3 cos 2x

2. y ′′ + 9y = 5 cos 3x

3. Solve the initial problem y ′′ + y ′ − 2y = 2x, y(0) = 0, y ′ (0) = 1.

4. Textbook exercise set 4.4.

Problems 9, 13, 15, 17 (8,8e-th editions).
Problems 10, 13, 15, 19 (9th edition).

Answers

3 3
1. yp = − cos 2x + sin 2x
10 5
5
2. yp = x sin 3x
6
1 1
3. y = ex − e−2x − x −
2 2
4. see textbook or detailed solutions

Detailed solutions to Question 2

Back to Chapter 2

Question 3

Solve each of the following D.E.s using the method of variation of parameters:

1. y ′′ + y = tan x

2. y ′′ − y = ex

3. y ′′ + 3y ′ + 2y = 3e−2x + x

4. y ′′ − y = x2 ex

5. Textbook exercise set 4.6.

Problems 3, 9.(8,8e-th editions)
Problems 4, 9. (9th edition)

Answers

1. y = A cos x + B sin x − cos x ln(sec x + tan x).

96 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS

1
2. y = Aex + Be−x + xex
2
1 3
3. y = Ae−x + Be−2x + x − − 3xe−2x
2 4
1 x
4. y = Aex + Be−x + e (2x3 − 3x2 + 3x)
12
5. see textbook or detailed solutions

Detailed solutions to Question 3

Back to Chapter 2

Question 4

Use the inverse ϕ(D) operator method to find some of the particular solutions you have
previously found using other methods:

1. y ′′ − y = ex ;

2. y ′′ + 3y ′ + 2y = 3e−2x + x

3. (D2 − 1) y = x2 ex

4. y ′′ + 2y ′ + 2y = e−x cos x

Note: you only need find yp .

Answers
For parts 1-3, see exercise Tutorial 2 Question 3.
x
Answer to part 4: y = e−x sin x
2
Detailed solutions to Question 4
Back to Chapter 2

Question 5

1. It has been found experimentally that a 6 kg mass stretches a certain spring 0.5 m.
If the mass is pulled 0.25 m below the equilibrium position and released, the D.E.
which describes the motion is
d2 x
+ 20x = 0
dt2
where x is the displacement from the equilibrium position, t is the time in seconds
after the release of the spring, and g = 10 m/s2 .
Find the value of x at any time t after the release of the spring (i.e. the solution of
the D.E.).
Hint: The first sentence is irrelevant to the solution, while the first part of the second
sentence tells you that x(0) = 0.25 and ẋ(0) = 0, where ẋ ≡ dxdt .
2.10. TUTORIAL 2 97

2. The charge, Q, on a capacitor in an electric circuit which contains an inductance, a

resistance, a capacitance and a source of electromotive force, satisfies the following
D.E.
d2 Q dQ
2 2 + 16 + 50Q = 50
dt dt
dQ
At time t = 0, Q = 0 and dt = 0.
Find an expression for the charge on the capacitor.

3. Textbook exercise set 5.7. Problem 1

Answers:

1 √
1. x = cos(2 5t)
4
4
2. Q = −e−4t (cos 3t + sin 3t) + 1
3
3. see textbook

Detailed solutions to Question 5

Back to Chapter 2
98 CHAPTER 2. LINEAR DIFFERENTIAL EQUATIONS
Chapter 3

Infinite series

Chapter objectives

In this chapter you will learn to:

• determine series convergence and non-convergence using various methods;

• use power series to define certain functions; and

• determine the interval of convergence for a power series.

3.1 Introduction
Informally, an infinite series may be regarded as a sum of an infinite number of terms. But,
this raises many questions. What, if any, is the connection between the terms? Can we add
infinitely many numbers to get a sum? Indeed, can we ever add infinitely many numbers?
The ancient Greeks believed that no infinite set of numbers could possibly have a finite
sum. Because of this belief, they were caught in some logical paradoxes. We can avoid
their problems by considering first the idea of a sequence.

3.2 Sequences
Informally, a sequence is a set of numbers or terms arranged in a definite order according
to some rule, e.g.
2, 4, 8, . . . , 2n , . . .
This is a sequence whose first member is 2, second member is 4 = 22 , third member is 23
and so on. We call 2n , the general term or the n-th term.
Formally, a sequence is a function whose domain is a set of integers, usually the positive
integers, e.g. the sequence above should be denoted by

{(1, 2), (2, 4), (3, 8), (4, 16), . . . , (n, 2n ) , . . .}

99
100 CHAPTER 3. INFINITE SERIES

or by
{(n, 2n ) ∶ n = 1, 2, 3, . . .}
Note that the functional values are the important part of the ordered pairs. For this reason,
we just use these terms of the sequence. Thus, a sequence {(n, f (n)) ∶ n = 1, 2, 3, . . .} will
always be written simply as {f (n)}.
For instance, the sequence {n2 + 1} has its first term 2, its second term 5 and so on. Here we
are assuming that n takes on values n = 1, 2, . . . . We can also have the following sequence:
1, 2, 5, . . . , n2 + 1, . . . , where n = 0, 1, 2, . . . , if we so wish; or even
5, 10, 17, . . . , where n = 2, 3, . . . .
We will denote the general term of a sequence by an . Thus a sequence whose domain is
the positive integers will be given by a1 , a2 , a3 , . . .
For example, write down the first 4 terms of the sequence an ≡ (−1)n .
Here,
a1 = (−1)1 = −1
a2 = (−1)2 = 1
a3 = (−1)3 = −1
a4 = (−1)4 = 1

The sequence is −1, 1, −1, 1, −1, 1, . . . .

Note that the same sequence is given by bn = (−1)n+1 , n = 0, 1, 2, . . .
Now consider the sequences given by cn = n2 + 1 and by dn = n12 + 1. The first sequence is
2, 5, 10, 17, . . . and the terms are clearly getting bigger and bigger, i.e. cn → ∞ as n → ∞.
The second sequence is 2, 54 , 10 17 26
9 , 16 , 25 , . . .

It appears that the terms are approaching 1 as n takes on larger values i.e. dn → 1 as n → ∞
In general, we say that a sequence {an } converges if and only if lim an exists.
n→∞

If a sequence does not converge, it is said to diverge e.g. {n2 + 1} is divergent since
limn→∞ (n2 + 1) = ∞, i.e. does not exist (as a real number).
More examples:
{ n12 + 1} is convergent since limn→∞ { n12 + 1} = 1
{(−1)n } diverges since limn→∞ (−1)n does not exist. ((−1)n oscillates between − 1 and
1.)
Note. To test for convergence of a sequence {an } , we evaluate limn→∞ an by any appro-
priate method. (You may need to revise the evaluation of limits.)

Tutorial 3 Question 1

Note. There are a large number of results that can be proved about sequences but they are
beyond the scope of this subject. We will turn our attention instead to infinite series whose
3.3. SERIES 101

“existence”, i.e. whose sums, depend on the convergence or divergence of an appropriately

formed sequence.
Note. Be very careful not to be confused between sequences and series!

3.3 Series
Video (by Dr. R. Wood): Infinite series (sections 3.3-7)
The sigma notation for finite sums should be familiar to you as, for example, in
n
∑ f (x∗i ) ∆xi
i=1

which represents the sum:

f (x∗1 ) ∆x1 + f (x∗2 ) ∆x2 + . . . + f (x∗n ) ∆xn

An infinite series is written as

∞
∑ ai ≡ a1 + a2 + a3 . . . .
i=1

Of course, the index of summation, i, can be any letter.

We can also have series starting with a0 or a7 or, in general, aN as in
∞
∑ ai = aN + aN +1 + . . .
i=N

Note that a series is a sum, while a sequence is a function. However, a series is usually
formed by adding up the terms of a sequence, e.g.
∞
∑i = 1 + 2 + 3 + ...
i=1
∞
1 1 1 1
∑ 2
=1+ + + + ...
i=1 i 4 9 16
∞
1 1 1
∑ = 1 + + + ...
i=1 i 2 3
∞ ∞ ∞
1 1
We will show that ∑ i and ∑ do not “exist”, i.e. do not have a sum value but that ∑ 2
i=1 i=1 i i=1 i
∞
1 π2
does exist, i.e. adds up to a finite number. (In fact, it is known that the series ∑ 2 = ,
i=1 i 6
the proof being beyond the scope of this subject!) Let us consider the series
∞
∑ ai ≡ a1 + a2 + a3 + . . .
i=1

We form the partial sums, Sn , as follows:

S1 = a1
S2 = a1 + a2
S3 = a1 + a2 + a3
...
Sn = a1 + a2 + a3 + . . . an
102 CHAPTER 3. INFINITE SERIES

i.e. Sn , the n-th partial sum, is the sum of the first n terms of the series.
We now consider the sequence of these partial sums, {Sn } . If the sequence {Sn } converges
(i.e. limn→∞ Sn exists and equals S, say), then we say that the series is convergent and has
the value (sum) S. If {Sn } diverges, then ∑∞i=1 ai is called divergent.

Example 1

The series ∞
∑i = 1 + 2 + 3 + ...
i=1
n(n+1)
is divergent since the partial sum, Sn , = 1 + 2 + . . . + n which is known to be 2 , and
limn→∞ Sn = limn→∞ n(n+1)
2 = ∞.

Example 2

Does the series 1 − 1 + 1 − 1 + . . . converge or diverge?

Here, S1 = 1, S2 = 0, S3 = 1, . . . , and, in general,

1 if n is odd
Sn = {
0 if n is even

∴ ∑∞
i=1 Sn does not exist and the series is divergent.

Example 3

Consider ∞
3 3 3 3
∑ i
= + + + ...
i=1 10 10 100 1000
Here n
1
3 3 3 3 (1 − ( 10 ) )
Sn = + + ... + n = 1 .
10 102 10 10 1 − 10
Here we used the sum of a geometric progression formula.
1 n
3 ⎛ 1 − ( 10 ) ⎞
∴ lim Sn = lim
n→∞ n→∞ 10 ⎝ 1 − 1
10 ⎠

3 1 1 n
= ( ) , since ( ) → 0
10 9/10 10
1
=
3
Hence,
∞
3 1
∑ i
=
i=1 10 3
which is very “reasonable” since
∞
3 3 3
∑ i
= + + ...
i=1 10 10 100
1
can be written as 0.3333 . . . which we know is the decimal form of 3 (We can similarly
show that 1 = 0.9999 . . .).
3.3. SERIES 103

Note. The method used in Example 3 will work for any geometric series, i.e. any series
whose terms are in a geometric progression.
Hence ∞
∑ (ari−1 ) ≡ a + ar + ar2 . . .
i=1
converges to
a
1−r
if 0 < ∣r∣ < 1 and diverges if ∣r∣ > 1.
However, the general term in a sequence of partial sums may be difficult or impossible to
determine, e.g. for the series
∞
1
∑
i=1 i

no simple formula for Sn ≡ 1 + 21 + . . . + n1 in terms of n exists and so limn→∞ Sn cannot

be evaluated here. Hence, we will have to develop other methods and tests for convergence
and divergence. However, these tests will not give the sum of the series.
The necessary condition for ∑ an to be convergent is that limn→∞ an = 0, i.e. we have the
following test for divergence.
Consider ∞
∑ an
n=1

Evaluate limn→∞ an and if this does not equal 0, the series is divergent.
Be careful: if limn→∞ an = 0, the question is still open, and the series may be convergent
or divergent.
For example, the series
∞
1 1 1
1+ + ... + + ... ≡ ∑
2 n n=1 n

is divergent although
1
lim an = lim
= 0.
n→∞ n n→∞

There are also certain properties of series which may be useful:

a. If ∑ an and Σbn are convergent series and k is a constant, then Σ (an + bn ), Σ (an − bn )
and Σ (kan ) , Σ (kbn ) are also convergent. e.g. once we have shown that
∞
1
∑ 2
n=1 n

is convergent, we will immediately know that

∞
2 1 2 2
∑ 2
≡ 2 + + + . . . + + ...
n=1 n 2 9 n2

is a convergent series.
Also, the series
∞
∑ (2−n + 3−n )
n=1
104 CHAPTER 3. INFINITE SERIES

converges since
∞
∑ 1/2n
n=1

and
∞
∑ 1/3n
n=1

are both convergent geometric series.

b. If ∑ an is divergent and k is a constant, then ∑ kan is divergent.

Note. From now on we will use n as the index of summation, as in ∑∞ n=1 an . Also, Σan
is an expression to indicate that it is immaterial whether n = 1, 2, . . . or n = 0, 1, 2, . . . or
indeed n = N, N + 1, . . .

Tutorial 3 Question 2

3.4 Series with positive terms

In order to develop tests for the convergence of series, we will first examine the case where
all of the terms of the series are positive, i.e. consider ∑∞
n=1 an and an ≥ 0 for all n

Integral test
If ∑∞
n=1 an is a series of positive terms, if f is a continuous, positive valued, decreasing
∞
function in [1, ∞) and if f (n) = an for all positive integers n, then ∑∞
n=1 an and ∫1 f (x)dx
both converge or both diverge.
This test requires us to form an appropriate improper integral, to evaluate that integral and
hence to infer at once whether the series converges or diverges.

Example 1

Examine the series

∞
1
∑ α
n=1 n

1 1
for convergence for various values of α. Here, we take f (x) = xα , as then f (n) ≡ nα .

For x ≥ 1, x1α is positive and continuous, and, in addition, for α > 0, it is decreasing.
Therefore we can apply the integral test to the case of α > 0 by considering the improper
∞
integral ∫1 x1α dx.(α ≤ 0 will have to be considered separately. )
3.4. SERIES WITH POSITIVE TERMS 105

Now,
∞ 1 N 1
∫1 dx = lim ∫ dx
xα N →∞ 1 xα
1 1 N
= lim [ ] , (α ≠ 1)
N →∞ −α + 1 xα−1 1
1 1 1
= lim { α−1
− }
N →∞ 1 − α N 1−α
− 1 for α > 1
= { 1−α
∞ for α < 1
Hence, the integral converges for α > 1 and diverges for 0 < α < 1. Thus the series ∑∞ 1
n=1 nα
converges for α > 1 and diverges for 0 < α < 1.
Hence, the integral converges for α > 1 and diverges for 0 < α < 1. Thus the series ∑∞ 1
n=1 nα
converges for α > 1 and diverges for 0 < α < 1.
For the case α = 1, the integral is
∞ 1
∫1 dx = lim [ln x]N
1 =∞
x N →∞

i.e. diverges.
Thus the harmonic series ∑∞ 1
n=1 n is again shown to be divergent. Also, for α < 0

∞ ∞
1
∑ α
≡ ∑ n−α ,
n=1 n n=1

where (−α) is positive. Hence, limN →∞ (n−α ) ≠ 0 and so, for α < 0, ∑∞ 1
n=1 nα diverges by
an earlier test. (What about the case α = 0?)

Example 2

Discuss the convergence or divergence of

∞
1
∑ .
n=2 n ln n

Note that here n starts its values at 2 since n = 0 and 1 would yield 0 in the denominator.
1
Check that x ln x is continuous, positive and decreasing for x ≥ 2. Hence we consider
∞ dx
∫2 x ln x
which can be shown to be a divergent integral. (Try it!)
Hence,
∞
1
∑
n=2 n ln n

diverges.

Tutorial 3 Question 3
106 CHAPTER 3. INFINITE SERIES

Comparison tests

Dominated comparison tests

1. If we are given a series ∑ an and can find a convergent series ∑ bn such that
0 ≤ an ≤ bn for all n, then ∑ an converges.

2. If we are given a series ∑ cn and can find a divergent series ∑ dn such that cn ≥ dn
for all n, then ∑ cn diverges.

Example 3

Does the series

∞
1
∑
n=1 2n +n
converge or diverge?
Here we spot that
1 1
< n
2n +n 2
for all n ≥ 1 and recall that ∑∞ 1
n=1 2n is convergent.

Hence ∑∞ 1
n=1 2n +n converges.

Example 4

Does the series

∞
1
∑ √
n=1 4 n − 1

converge or diverge?
Here we observe that
1 1
√ > √
4 n−1 4 n
for all n ≥ 1; that
1 1
∑ √ ≡ ∑ 1/2
n n
is divergent by Example 1; and that ∑ 4√1 n is thus also divergent. Hence, ∑ 4√1n−1 is diver-
gent.
A more useful comparison test is given by the following result.
The limit comparison test
an
Suppose that ∑ an and ∑ bn are two series with positive terms. If lim exists and equals
n→∞ bn
L > 0, then the two series both converge or both diverge.

Example 5
3.4. SERIES WITH POSITIVE TERMS 107

Does
∞
1
∑
n=1 3n2 − 2n + 1
converge or diverge? Here, the series itself suggests a comparison with
∞
1
∑ 2
.
n=1 n

We have
n2 1
lim = >0
n→∞ 3n2 − 2n + 1 3
Since we know that ∑ n12 converges, the series
∞
1
∑
n=1 3n2 − 2n + 1

converges too.

Example 6

The series
1
∑√ 2
3n + 1
is divergent since if we compare it with ∑ n1 in the limit comparison test, we get that

√ 1
3n2 +1 n 1 1
lim = lim √ = lim √ = √ >0
n→∞ 1/n n→∞ 3n2 + 1 n→∞ 3 + 1 3
n2

and ∑ n1 is known to be divergent.

Tutorial 3 Question 4

Ratio test

Given a series ∑ an , we consider

an+1
lim ≡ρ
n→∞ an

• If ρ < 1, the series is convergent.

• If ρ > 1, the series is divergent.

• If ρ = 1, the test fails and some other test will be necessary to decide convergence or
divergence.
108 CHAPTER 3. INFINITE SERIES

Example 7

Test for convergence:

∞
2n
∑
n=1 n

Here,
an+1 2n+1 n
lim = lim n
n→∞ an n→∞ 2 (n + 1)

2n
= lim
n→∞ n + 1
= 2 > 1.
∞
2n
∴ ∑ diverges by the ratio test.
n=1 n

Example 8

Apply the ratio test to determine whether

∞
1
∑ 3
n=1 n

is convergent or not.
Here,
an+1 1/(n + 1)3
lim = lim
n→∞ an n→∞ 1/n3
n 3
= lim ( )
n→∞ n + 1

=1.
Thus the test is inconclusive and some other test must be used (e.g. the integral test) to
show that this series is, in fact, convergent.
Note. There are a large number of other tests for convergence, most of which are highly
specialised. We will make particular use of the ratio test when we examine Power Series.

Tutorial 3 Question 5

3.5 Alternating series

We will now consider series where the terms are not all positive. First we examine alternat-
ing series where the terms are alternately positive and negative, e.g. a1 − a2 + a3 − a4 + . . . ,
where the an ’s are positive, i.e.
∞
∑ (−1)n+1 an .
n=1
3.6. ABSOLUTE CONVERGENCE 109

The alternating series test is as follows. Consider a series of the form

∞
∑ (−1)n+1 an .
n=1

If

a. lim an = 0 and
n→∞

b. an+1 ≤ an i.e. the terms are decreasing,

then the series is convergent.

Example

Test for convergence:

∞
1
∑ (−1)n+1
n=1 n2

1
a. lim an = lim = 0, and
n→∞ n→∞ n2

1 1
b. an+1 = 2
≤ 2 = an for all n.
(n + 1) n

∞
1
∴ ∑ (−1)n+1 is convergent.
n=1 n2

Tutorial 3 Question 6

3.6 Absolute convergence

We will now consider series in which the terms are positive or negative and not alternating,
e.g. √ √ √ √
∞ sin ( nπ )
3 3 3 3 3
∑ = + − − + ...
n=1 n2 2 8 32 50
In order to discuss the convergence of such series we have to consider the concept of abso-
lute convergence. Associated with any series ∑∞ n=1 an , where the an ’s are now positive or
negative, is the series
∞
∑ ∣an ∣ = ∣a1 ∣ + ∣a2 ∣ + ∣a3 ∣ + . . .
n=1

The series ∑ ∣an ∣ has positive terms and thus the tests for convergence discussed previously
can be applied to it. If ∑ ∣an ∣ can be shown to be convergent, then we say that the series
Σan is absolutely convergent.
110 CHAPTER 3. INFINITE SERIES

Once we have shown that a series is absolutely convergent (A.C.) we can use the theorem
(here accepted without proof) that if a series is absolutely convergent, then it is convergent.

Example 1

Show that
∞ sin ( nπ )
3
∑
n=1 n2
is absolutely convergent and hence is convergent. We must consider the series
∞ sin nπ
√ √ √ √
3 3 3 3 3
∑∣ 2 ∣= + + + + ...
n=1 n 2 8 32 50
i.e. √
3
2n2 n ≠ 3k
∣an ∣ = {
0 n = 3k
∞
√ √
3 3
If we compare this series with the series ∑ 2 , then ∣an ∣ ≤ 2 for all n.
n=1 2n 2n
∴ ∑ ∣an ∣ is convergent by the dominated comparison test, since
∞
√ √ ∞
3 3 1
∑ 2= ∑ 2
n=1 2n 2 n=1 n
∞
sin ( nπ )
3
is known to be convergent. Hence ∑ is absolutely convergent and, by the theo-
n=1 n2
rem, is thus also convergent.

Example 2

Test ∞
n+1
∑ (−1)n+1
n=1 n2
for absolute convergence and for convergence.
Here,
n+1 n+1
∣an ∣ = ∣(−1)n+1 ∣= 2
n2 n
∞
n+1
for all n, and so we consider the series ∑ 2
n=1 n
∞
1
If we compare this series with the divergent series ∑ using the limit comparison test,
n=1 n

n+1
2 n+1
lim n = lim ( )=1>0
n→∞ 1 n→∞ n
n

n+1
and so ∑ is also divergent.
n2
∞
n+1
Thus ∑ (−1)n+1 2 is not absolutely convergent, and the theorem cannot be applied to
n=1 n
this series.
3.7. POWER SERIES 111

∞
n+1
However, ∑ (−1)n+1 is an alternating series which meets the criteria for the alternat-
n=1 n2
∞
n+1
ing series test. (Show this!) Hence ∑ (−1)n+1 2 is convergent.
n=1 n
Note. Example 2 exhibits a series which is not absolutely convergent but is convergent.
Such a series is called conditionally convergent (C.C.).

Tutorial 3 Question 7

3.7 Power series

Read textbook section 8.2 (Power series and analytic functions)

Introduction
Thus far we have considered series whose terms are constants. Let us now consider series
whose terms are functions of a variable x.
We will not examine the convergence of the most general of such series i.e. ∑ fn (x) e.g.
∞
∑ (2 cos x)n but will consider only a special case, the power series.
n=1
∞
A power series has the form ∑ an (x−a)n , where the an ’s are constants which depend only
n=0
on n and a is a fixed constant, i.e.

a0 + a1 (x − a) + a2 (x − a)2 + . . .

When a value is substituted for x, this becomes an infinite series which may converge or
diverge, i.e. may or may not be said to have a “value” for that particular value of x.
Clearly, if we let x = a, then the power series becomes

a0 + a1 ⋅ 01 + a2 ⋅ 02 + . . .

i.e. the power series = a0 at x = a.

We want to ascertain for what other values of x a given power series converges.
Let us first consider a less general power series where a = 0 i.e. of the form
∞
∑ an x n = a0 + a1 x + a2 x 2 + . . .
n=0

e.g. consider the series

∞
∑ xn = 1 + x + x2 + . . . ,
n=0
112 CHAPTER 3. INFINITE SERIES

where an = 1 for all n. If we put x = 5, the power series becomes 1 + 5 + 52 + . . . + 5n + . . . ,

which we know is a divergent series since

lim 5n ≠ 0
n→∞

If we put x = 0, the power series equals 1 , and so converges trivially.

If we put x = 21 , the power series becomes 1 + 21 + 212 + . . . , which is a geometric series that
converges to 1−1 1 = 2.
2

In general, we note that 1 + x + x2 + . . . is a geometric series with common ratio x and so the
1
power series converges ( to 1−x ) for any x such that ∣x∣ < 1 i.e. −1 < x < 1, and diverges
for ∣x∣ ≥ 1, i.e. x ≥ 1 or x ≤ −1.
Radius and interval of convergence
Note. The series ∑∞ n
n=1 x converges for x in the interval (-1,1) and we call this the interval
of convergence of this power series. Further, the centre of this interval is at 0 and we say
that its ”radius of convergence” is 1 . If we can determine that a series is convergent for

∣x − a∣ < r

then the radius of convergence is r. The interval of convergence can then be determined by
simplifying
−r < x − a < r

and adding a gives

a−r <x<a+r

as the interval of convergence.

Also it may be necessary to evaluate the series at the endpoints in the case of equality to
∞
check for convergence. We will now show by examples that a power series ∑ an (x − a)n
n=0
always has an interval of convergence, centred at x = a and with “radius of convergence”
r, where 0 ≤ r ≤ ∞.

Example 1

Find all the values of x for which the power series

x
∑ n
n=0 (n + 1)2

converges. First we test the series for absolute convergence using the ratio test as follows.
Consider the series
∞
xn
∑∣ ∣,
n=0 (n + 1)2n
where the terms are positive and the general term is

xn ∣x∣n
∣ ∣ = .
(n + 1)2n (n + 1)2n
3.7. POWER SERIES 113

Now, for the ratio test, we must evaluate

∣x∣n+1 (n + 2)2n+1 ∣x∣ n + 1

ρ = lim n n
= lim ( )
n→∞ ∣x∣ (n + 1)2 n→∞ 2 n+2
∣x∣ n+1
= ⋅ ( lim )
2 n→∞ n + 2
∣x∣
=
2
The series
∞
xn
∑∣ ∣
n=1 (n + 1)2n
∣x∣ ∣x∣
converges if ρ < 1, i.e. if 2 < 1 i.e. if ∣x∣ < 2 and diverges if ρ > 1, i.e. if 2 > 1, i.e. ∣x∣ > 2.
Hence the series ∞
xn
∑ n
n=0 (n + 1)2

converges absolutely (and thus also converges) if ∣x∣ < 2, i.e. −2 < x < 2, and diverges if
∣x∣ > 2.
Hence the interval of convergence is (-2,2) (“centered” at 0 as we expected) and the radius
of convergence is 2.
We still do not know how the series behaves when ρ = 1, i.e. when ∣x∣ = 2, i.e. x = 2 or −2.
Accordingly, we must test the series
∞
xn
∑ n
n=0 (n + 1)2

for convergence at each of the endpoints of the interval in turn.

On substituting x = 2, the series becomes
∞ ∞
2n 1 1
∑ n
= ∑ = 1 + + ...
n=0 (n + 1)2 n=0 n + 1 2

which we know is the divergent harmonic series. On substituting x = −2, the series becomes
∞ ∞
(−2)n (−1)n 1 1
∑ n
= ∑ = 1 − + − ...
n=0 (n + 1)2 n=0 n + 1 2 3

which is an alternating series that we know (see the exercise in the section on Absolute
Convergence) to be convergent but not absolutely convergent.
To sum up: the series is absolutely convergent and thus convergent for x in (-2,2); it is
conditionally convergent for x = −2 (we say that the complete interval of convergence is
[-2,2) ); and it is divergent for x < −2 and x ≥ 2.

Example 2

Find the (complete) interval of convergence for

∞
xn x x2
∑ = 1 + + + ...
n=0 n! 1! 2!
114 CHAPTER 3. INFINITE SERIES

(Recall that n! = n(n − 1) . . . 2.1 and 0! ≡ 1.) First we consider

n+1
x
∣ (n+1)! ∣ ∣x∣
ρ = lim xn
= lim
n→∞ ∣ n! ∣ n→∞ (n + 1)

1
= ∣x∣ lim
n→∞ n + 1
= 0 for all x
Hence, for all x, ρ = 0 < 1 and so the series converges absolutely for all x. i.e. the interval
of convergence is (−∞, ∞). (No endpoints need be checked here!) (This is a case where
the “radius of convergence”, r = ∞. )

Example 3
∞
Find the interval of convergence for ∑ n!(x + 3)n .
n=0

Here,
∣(n + 1)!(x + 3)n+1 ∣
ρ = lim = lim (n + 1)∣x + 3∣
n→∞ ∣n!(x + 3)n ∣ n→∞

∞ for all x ≠ −3
={
0 for x = −3
Hence, the series converges only when x = −3 and diverges for all other x. The interval
thus consists of the single point, −3, and the “radius of convergence” is 0.

Example 4

Find the interval of convergence for

∞
n(n + 1)(x − 4)n
∑ (−1)n+1 .
n=1 5n
Now,
(−1)n−2 (n+1)(n+2)(x−4)n+1
5n+1
ρ= lim (−1)n+1 n(n+1)(x−4)n
n→∞
∣ 5n ∣
∣x − 4∣ n + 2
= lim ( )
n→∞ 5 n
∣x − 4∣
=
5
Hence the series converges absolutely for ∣x−4∣5 < 1 and diverges for ∣x−4∣
5 > 1, that is, the
series is A.C. for ∣x − 4∣ < 5 i.e. −5 < (x − 4) < 5 i.e. −1 < x < 9 and divergent for x < −1
and x > 9.
When x = −1, the series becomes
∞
n(n + 1)(−5)n ∞
∑ (−1)n+1 = ∑ (−1)2n+1 n(n + 1)
n=1 5n n=1
∞
= − ∑ n(n + 1)
n=1

which is a divergent series, since the n th term does not tend to 0 .

3.7. POWER SERIES 115

When x = 9, the series becomes ∑∞ n=1 (−1)

n+1 n(n + 1) which is similarly divergent. Hence

the (complete) interval of convergence is (-1,9) which is centered at x = 4 and has “radius”
5.
Functions defined by a power series
Since a power series has a unique sum at each point of its interval of convergence, it defines
a function given by
∞
F (x) = ∑ an (x − a)n , x ∈ I
n=0
the interval of convergence. For a particular power series, the function F may be hard or
impossible to find explicitly. However, the following properties of F, can be established
(the proofs are beyond the scope of this subject):

1. F (x) is continuous for x in I;

∞
′
2. F (x) is differentiable for x in I and F (x) = ∑ nan (x − a)n−1 , i.e. we can differen-
n=1
tiate the infinite series term-by-term;
∞
an
3. F (x) is integrable for x in I and ∫ F (x)dx = ∑ (x − a)n+1 i.e. we can
n=0 n + 1
integrate term-by-term.
4. The given an ’s of the power series are related to F (x) by
F (n) (a)
≡ an , for n = 0, 1, 2, . . .
n!
(F (0) (a) means F (a).)

Example 5

Consider the power series ∑ xn . This power series has the interval of convergence (-1,1)
n=0
since
∣xn+1 ∣
ρ = lim = lim ∣x∣ = ∣x∣
n→∞ ∣xn ∣ n→∞

and the series diverges at each end point. (Check all of this.) Now, it happens that this
1
power series 1 + x + x2 + . . . is a geometric series and so we know that its sum is 1−x for
∣x∣ < 1.
∞
1
∴ Here, F (x) = ≡ ∑ xn for all x in (-1,1)
1 − x n=0
1
Clearly is continuous, differentiable and integrable for all x in (-1,1).
1−x
1 2 n!
Also, F ′ (x) = 2
, F ′′ (x) = 3
and, in general, F (n) (x) =
(1 − x) (1 − x) (1 − x)n+1
∞
Now, for this power series ∑ xn , a = 0 and an ≡ 1 for all n.
n=0

F (n) (0) n!
Hence should equal 1 for all n, which is clearly so for F (n) (x) = .
n! (1 − x)n+1
116 CHAPTER 3. INFINITE SERIES

Taylor series
We now want to consider the converse problem: given a function f (x) “expand” it to give
a power series in powers of (x − a).
It can be shown that, if a power series exists for a particular series and if f has all order
derivatives in an interval containing a, then
∞
f (n) (a)
f (x) = ∑ (x − a)n
n=0 n!
f ′′ (a)
= f (a) + f ′ (a)(x − a) + (x − a)2 + . . .
2!
for all x in the interval.
∞
f (n) (a)
The series ∑ (x−a)n is called the Taylor series for f about a. If a = 0 in the Taylor
n=0 n!
series, we get
∞
f ′′ (0) 2 f (n) (0) n
f (0) + f ′ (0)x + x + ... = ∑ x
2! n=0 n!
which is often called the Maclaurin series for f .
Note
Examine carefully what the theorem above has as its suppositions. Only once we know
that a function actually has a power series expansion do we know that this power series is
in fact the Taylor series.
We can, of course, calculate the Taylor series corresponding to any given function f . Then
we can find the interval of convergence of the resulting power (Taylor) series using the
methods developed previously. But, we do not know that the Taylor series corresponding
to f actually converges to f (x) until we have checked that a certain “Remainder” tends to
zero.
This remainder is given by Rn (x) in Taylor’s formula, which states: If f has (n + 1)
derivatives in an interval I that contains the number a, then for x in I there is a number z
strictly between x and a such that

f ′ (a) f ′′ (a) f (n) (a)

f (x) = f (a) + (x − a) + (x − a)2 + . . . + (x − a)n + Rn (x),
1! 2! n!

f (n+1) (z)
where Rn (x) = (x − a)n+1 .
(n + 1)!
Fortunately, for all the elementary functions that we consider, it can be shown that the
Taylor series does converge to f (x) for all x in the interval of convergence, i.e. We will
write
∞
f (n) (a)
f (x) = ∑ (x − a)n
n=0 n!
and ignore the question of the remainder.

Example 6

Expand f (x) = cos x in powers of x and find the interval of convergence for the resulting
3.7. POWER SERIES 117

(Maclaurin) power series. Here a = 0 and we evaluate f (0), f ′ (0), f ′′ (0), . . . , in turn.
f (x) = cos x f (0) =1
f ′ (x) = − sin x f ′ (0) =0
f ′′ (x) = − cos x f ′′ (0) = −1
f ′′′ (x) = sin x f ′′′ (0) =0
f (4) (x) = cos x f (4) (0) =1
A pattern is now clearly established. Now the required series is given by
∞
f (n) (0) f ′ (0) f ′′ (0) 2
f (x) = ∑ (x)n = f (0) + x+ x + ...
n=0 n! 1! 2!
0 (−1) 2 0 3 1 4
i.e. cos x = 1 + x + x + x + x + ...
1! 2! 3! 4!
x2 x4
=1− + − ...
2! 4!
∞
x2n
= ∑ (−1)n
n=0 (2n)!
To find the interval of convergence of this power series, we consider
∣ (−1)n+1 x2(n+1) /(2(n + 1))!
ρ = lim
n→∞ ∣(−1)n x2n /(2n)!∣
1
= lim ∣x∣2
n→∞ (2n + 1)(2n + 2)
= 0 for all x.
Hence the power series converges absolutely for all x and, hence, (without examining the
appropriate remainder term) we have that, for all x
x2 x4 x6
cos x = 1 − + − + ...
2! 4! 6!
Example 7

Expand ln(1 + x) in a power series about x = 0 and find the interval of convergence of the
series.
Here a = 0 and we must find f (0), f ′ (0), . . .
f (x) = ln(1 + x)
1
f ′ (x) =
1+x
−1
f ′′ (x) =
(1 + x)2
2
f ′′′ (x) =
(1 + x)3
−2 ⋅ 3
f (4) (x) =
(1 + x)4
hence
f (0) = ln 1 = 0
f ′ (0) = 1
f ′′ (0) = −1
f ′′′ (0) = 2
f (4) (0) = −6 etc.
118 CHAPTER 3. INFINITE SERIES

In general, a pattern is developing and we note that

(−1)n+1 (n − 1)!
f (n) (x) = ;n ≥ 1
(1 + x)n

and thus f (n) (0) = (−1)n+1 (n − 1)!.

Hence, the (Maclaurin) power series corresponding to ln(1 + x) is
1 (−1) 2 2 3 (−1)n+1 (n − 1)! n
0+ x+ x + x + ... + x + ...
1! 2! 3! n!
x2 x 3 x4
=x− + − + ...
2 3 4
∞
xn
= ∑ (−1)n+1
n=1 n
Next, we consider
n+1
∣(−1)n+2 xn+1 ∣
ρ = lim n
n→∞ ∣(−1)n+1 xn ∣
n
= lim ∣x∣ ( )
n→∞ n+1
= ∣x∣.
∴ The series converges absolutely for ∣x∣ < 1, i.e. −1 < x < 1.
When x = −1, the series is
∞
(−1)n ∞ (−1)2n+1 ∞
1
∑ (−1)n+1 =∑ =−∑ .
n=1 n n=1 n n=1 n

which is divergent.
When x = 1, the series is ∑∞n=1 (−1)
n+1 1 which is convergent by the alternating series test,
n
but is not absolutely convergent. Hence, the complete interval of convergence for which
∞
xn
ln(1 + x) = ∑ (−1)n+1 is (−1, 1].
n=1 n

Example 8

Expand ex in a Taylor series in powers of x + 2.

The expansion is about x = −2 and so we must find f (−2), f ′ (−2), . . .
Hence,
f (x) = ex , f (−2) = e−2
f ′ (x), = ex f ′ (−2) = e−2
...
Clearly f (n) (−2) = e−2 hence
e−2 e−2
∴ ex = e−2 + (x + 2) + (x + 2)2 + . . .
1! 2!
∞
e−2
=∑ (x + 2)n
n=0 n!
3.7. POWER SERIES 119

(Show that this series converges for all x.)

Operations with power series
There are a large number of results that can be established using power series and some of
their properties as stated in the section on Functions defined by a Power Series.

Example 1

Prove that
sin x
lim =1
x→0 x
The Maclaurin series for sin x is easily found to be given by
x3 x5
x− + − ...,
3! 5!
which converges for all x
sin x x2 x4
=1− + ......
x 3! 5!
sin x x2 x5
∴ lim = lim (1 − + − . . .)
x→0 x x→0 3! 5!
= 1 ( all other terms → 0)

Example 2

In the section on Taylor series, Example 8 we sought the expansion of ex in powers of x+2.
It is easily shown that
∞
x xn
e =∑ for all x
n=0 n!
If we replace x by (x + 2) on the right-hand side of this result, we get a power series in
(x + 2). However, to preserve the equality, we must replace x by (x + 2) on the left-hand
side too ∞
(x + 2)n
∴ ex+2 = ∑
n=0 n!
∞
(x + 2)n
∴ ex ⋅ e2 = ∑
n=0 n!
∞
(x + 2)n
∴ ex = ∑ e−2 (as before!)
n=0 n!

Example 3

In the section on Taylor series, Example 7 we showed that

∞
xn
ln(1 + x) = ∑ (−1)n+1
n=1 n
with interval of convergence (-1,1]
Hence, for example, ln 1 = ln(1 + 0) = 0 as is well-known!
∞
1
ln 2 = ln(1 + 1) = ∑ (−1)n+1
n=1 n
1 1 1
= 1 − + − + ...
2 3 4
120 CHAPTER 3. INFINITE SERIES

Hence, we have found the sum of the alternating harmonic series. Also, we can determine
the value of ln 2, to as many decimal places as we wish, by taking a suitable number of
terms of the infinite series!
Further, since
d 1
ln(1 + x) = ,
dx 1+x
by differentiating the given power series term-by-term, we immediately get the Maclaurin
1
series for .
1+x
Also,
1 d x2 x3 x4
= (x − + − + . . .)
1 + x dx 2 3 4
= 1 − x + x2 − x3 + . . .
1
= 1 − (−x) + (−x)2 − (−x)3 + . . .
1−x
= 1 + x + x2 + x3 + . . .
which we have seen previously. (See Section on Functions defined by a power series,
Example 5.)
1
However, the series we have derived for 1−x can easily be shown to have interval of conver-
gence (-1,1) , and not (-1,1] as for ln(1 + x). (The endpoints of the interval of convergence
must thus still be examined separately for convergence.)
Note particularly the Binomial Series.
You would have already met the Binomial Theorem, which states that for a, b any real
numbers and n a positive integer,
n
n
(a + b)n = ∑ ( ) an−k bk
k=0
k

If we put a = 1, b = x, we get
n
n n n(n−1)...(n−k+1) n
(1 + x)n = ∑ ( ) xk , where ( ) = k! , ( )=1
k=0
k k 0

Newton extended the Binomial Theorem to the case where n is no longer a positive integer.
By listing the derivatives of f (x) = (1 + x)n , we can show that if n is any real number
and ∣x∣ < 1, then we get an infinite series:
∞
n
(1 + x)n = ∑ ( ) xk .
k=0
k

n
Note that if n is a given positive integer, then ( ) = 0 for k > n, whereas for other real n,
k

n − 21 (− 12 ) (− 32 ) (− 52 ) 1⋅3⋅5 15
( ) ≠ 0 for any k, e.g. ( )= =− =− .
k 3 3! 2⋅4⋅6 48

This concludes our examination of Series which have many uses and extensions. as we
will see in other subjects! In Chapter 5 we will examine the use of Series in solving certain
classes of Differential Equations.
3.7. POWER SERIES 121

This concludes our examination of Series which have many uses and extensions, as we will
see in other subjects! In Chapter 5 we will examine the use of Series in solving certain
classes of Differential Equations.

Tutorial 3 Question 8
122 CHAPTER 3. INFINITE SERIES

3.8 Tutorial 3
Back to Chapter 3

Question 1

In each of the following, indicate whether the sequence converges or diverges:

n2 + 1
1. { }
n3 + 2n + 1
n4 + 1
2. { }
n3 + 2n + 1
n3
3. { }
2n
n3
Hint: Use l’Hopital’s rule to evaluate lim
n→∞ 2n

4. {rn } for various positive values of r.

5. {1 + (−1)n }
1 + (−1)n
6. { }
n
1 k
7. a2k = 1, a2k+1 = 1 − ( )
2
n2 − n + 7
8. an =
2n3 + n2
sin2 n
9. an = √
n
√
n
10. an =
ln n
2n + 1
11. an =
en

Answers

1. Convergent

2. Divergent

3. Convergent

Convergent 0≤r≤1
4. {
Divergent r>1

5. Divergent
3.8. TUTORIAL 3 123

6. Convergent

7. Convergent
0
a1 = a2⋅0+1 = 1 − ( 12 ) = 0, a2 = a2⋅1 = 1
1
a3 = a2⋅1+1 = 1 − ( 21 ) = 12 , a4 = 1
2
a5 = 1 − ( 12 ) = 34 , etc.
limn→∞ an = 1.

8. Convergent: limn→∞ an = 0

9. Convergent: limn→∞ an = 0

10. Divergent

11. Convergent: limn→∞ an = 0

Detailed solutions to Question 1

Back to Chapter 3

Question 2

Determine whether a given series converges or diverges. If it converges, find its sum.

1 1 1
1. 1 + + +⋯+ n +⋯
3 9 3

2. 1 + 3 + 5 + 7 + ⋯ + (2n − 1) + ⋯

3. 1 − 2 + 4 − 8 + 16 − ⋯ + (−2)n + ⋯

4 4 4
4. 4 + + +⋯+ n +⋯
3 9 3

Answers

1. 3/2

2. Diverges (the k-th partial sum is k 2 )

3. Diverges (geometric series with ratio -2)

4. 6

Detailed solutions to Question 2

Back to Chapter 3
124 CHAPTER 3. INFINITE SERIES

Question 3

Use the integral test to test the given series for convergence

1. ∞
n
∑
n=1 n2
+1

2. ∞
1
∑√
n=1 n+1
3. ∞
1
∑
n=1 n2 +1

4. ∞
ln n
∑ 2
n=1 n

Answers

1. Diverges

2. Diverges

3. Converges

4. Converges

Detailed solutions to Question 3

Back to Chapter 3

Question 4

Use comparison tests to determine whether the series converge or diverge

∞
1
1. ∑
n=1 n2 +n+1
∞
1
2. ∑ √
n=1 n + n
∞
1
3. ∑ n
n=1 1 + 3

∞
√
n
4. ∑ 2
n=1 n + n

∞
sin2 n
5. ∑ 2
n=1 n + 1
3.8. TUTORIAL 3 125

Answers

1. Converges

2. Diverges

3. Converges

4. Converges

5. Converges

Detailed solutions to Question 4

Back to Chapter 3

Question 5

Use the ratio test to determine whether the series converge or diverge.

∞
10n
1. ∑ n
n=1 n

∞
n n
2. ∑ ( )
n=2 ln n
∞
3n
3. ∑
n=1 n!n

∞
n!n2
4. ∑
n=1 (2n)!

∞
3 n
5. ∑ n ( )
n=1 4

Detailed solutions to Question 5

Back to Chapter 3

Question 6

Determine whether the series converge or diverge.

∞
(−1)n+1
1. ∑
n=1 n2
∞
(−1)n+1 n
2. ∑
n=1 3n + 2

∞
(−1)n+1 n
3. ∑ √
n=1 n3 + 2
126 CHAPTER 3. INFINITE SERIES

∞
(−1)n+1 n
4. ∑
n=2 ln n
∞
(−1)n+1
5. ∑ √
n
n=1 2

Answers

1. Converges

2. Diverges

3. Converges

4. Diverges

5. Diverges

Detailed solutions to Question 6

Back to Chapter 3

Question 7
∞ ∞
(−1)n+1 (−1)n+1
a. Show that ∑ is A.C. and hence is convergent; that ∑ is not A.C. but
n=1 3n n=1 n
∞
is convergent, i.e. it is C.C.; and that ∑ 3n is not A.C., nor C.C., i.e. it is divergent.
n=1

b. Determine whether the series converge absolutely, conditionally, or diverge.

∞
(−10)n
1. ∑
n=1 n!
∞
(−1)n+1 nn
2. ∑ n
n=1 (n + 1)
∞ √ √
3. ∑ (−1)n ( n + 1 − n)
n=0

Answers

b.1 Converges absolutely

b.2 Diverges

b.3 Converges condiionally

Detailed solutions to Question 7

Back to Chapter 3
3.8. TUTORIAL 3 127

Question 8

Find the interval of convergence of the power series (do not forget to investigate conver-
gence at the endpoints)

∞
1. ∑ nxn
n=1

∞
nxn
2. ∑ n
n=1 2

∞
3. ∑ n!xn
n=1

∞
(−1)n nxn
4. ∑ n 3
n=1 2 (n + 1)

∞
5. ∑ (5x − 3)n
n=1

∞
2n (x − 3)n
6. ∑
n=1 n2
7. Textbook exercise set 8.2. Problems 1,3,5

Answers

1. (−1, 1)

2. (−2, 2)

3. [0, 0]

4. [−2, 2]

5. (2/5, 4/5)

6. [5/2, 7/2]

7. see textbook

Detailed solutions to Question 8

Back to Chapter 3
128 CHAPTER 3. INFINITE SERIES
Chapter 4

Series solution of differential equations

Chapter objectives

In this chapter you will learn to:

• use power series to find solutions of linear differential equations;

• classify singular points for certain differential equations; and

• obtain a power series solution given the point classification for a differential equation

4.1 Introduction
In this chapter, we return to the study of linear D.E.s, but now take up the question of the
solution of such D.E.s with non-constant (or variable) coefficients. We will confine our
attack to second order L.D.E.s of the form

a2 (x)y ′′ + a1 (x)y ′ + a0 (x)y = s(x)

but the method can be readily generalised. In fact, we will examine only homogeneous
L.D.E.s, i.e. we will consider D.E.s where s(x) = 0. Techniques for finding a particular
solution for the general case can be easily established, but are beyond the scope of this
subject. These study notes should be comprehensive enough for our purposes.
Notes.

1. To solve such D.E.s with non-constant coefficients, we will have to use series. Thus
some of the concepts from Chapter 4 will be important here. We will, in fact, not
have to worry too much about the convergence of our series solutions - the underly-
ing theory guarantees that such a power series solution converges in an interval of
convergence!
Further, we will only examine the “basic”’ case which yields a power series solution.
We will not examine the use of Taylor series to solve D.E.s at all, and will only
“briefly” examine more complicated cases which require a Frobenius type series for
their solution.
129
130 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

2. Until now we have been concerned with, and in fact have restricted ourselves to,
differential equations which could be solved exactly. There are certain differential
equations which are of extreme importance in higher mathematics and engineering
or other scientific applications but which cannot be solved exactly in terms of ele-
mentary functions by any methods. For example, the innocent looking differential
equation
xy ′′ + y ′ + xy = 0
cannot be solved exactly in terms of functions usually studied in elementary calculus,
such as the rational algebraic, trigonometric and inverse trigonometric, exponential,
and logarithmic functions. In such cases, short of recourse to higher transcendental
functions, series or other “approximate” methods provide the only alternatives.

4.2 Informal series approach to the solution of differential

equations

Read textbook section 8.3 (Power series solutions to linear differential equa-
tions)

Video (by Dr. R. Wood): Series solution of linear ODE (sections 4.2-3)

Example 1

Consider the linear DE

y ′′ − y = 0
with y(0) = 3 and y ′ (0) = 2.
Using previous methods the general solution can be shown to be

y = C1 e−x + C2 ex

Fitting the initial conditions gives

1 5
y = e−x + ex
2 2
We will now try an informal series approach to solve this differential equation. We will
assume (correctly) that the solution function has a power series expansion of the form

y = f (x) = a0 + a1 x + a2 x2 + a3 x3 + . . .

From before, we can differentiate the power series term by term to obtain
dy
= a1 + 2a2 x + 3a3 x2 + 4a4 x3 + . . .
dx
and
d2 y
= 2a2 + 3 ⋅ 2a3 + 4 ⋅ 3a4 x2 + 5 ⋅ 4a5 x3 + . . .
dx2
The DE then becomes
2a2 + 3 ⋅ 2a3 x + 4 ⋅ 3a4 x2 + 5 ⋅ 4a5 x3 + . . .
= a0 + a1 x + a2 x 2 + a3 x 3 + . . .
4.2. INFORMAL SERIES APPROACH TO THE SOLUTION OF DIFFERENTIAL EQUATIONS131

For this to be true for all x in the interval of convergence will require that the coefficients
of corresponding powers of x on both sides of the equality to be equal.
Equating coefficients give
2a2 = a0
3 ⋅ 2a3 = a1
4 ⋅ 3a4 = a2
5 ⋅ 4a5 = a3
and so on. Hence,
1
a2 = a0
2
1
a3 = a1
3⋅2
1 1 1 1
a4 = a2 = ⋅ a0 = a0
4⋅3 4⋅3 2 4⋅3⋅2
1 1 1 1
a5 = a3 = ⋅ a1 = a1
5⋅4 5⋅4 3⋅2 5⋅4⋅3⋅2
and looks like
1 1 1 1
a2 = a0 , a3 = a1 , a4 = a0 and a5 = a1
2! 3! 4! 5!
There is an obvious pattern and with a0 , a1 arbitrary the series solution appears to be

y = a0 + a1 x + a2 x 2 + a3 x 3 + a4 x 4 + a5 x 5 + . . .
1 1 1 1
= a0 + a1 x + a0 x 2 + a1 x 3 + a0 x 4 + a1 x 5 + . . .
2! 3! 4! 5!
x 2 x4
= a0 [1 + + + . . .]
2! 4!
= a0 y 1 + a1 y 2 .

As expected for a second order linear DE the two solution functions y1 and y2 are linearly
independent. An alternative approach (we will use later) is to consider the general solution
of the second order linear DE as

y = C1 y 1 + C2 y 2

and generate y1 by setting a0 = 1 and a1 = 0. Similarly we should be able to determine y2

by setting a0 = 0 and a1 = 1. This ensures the two solution functions given by

y 1 = a0 + a1 x + a2 x 2 + . . .
= 1 + a2 x 2 + . . .

and
y 2 = a0 + a1 x + a2 x 2 + . . .
= 0 + x + a2 x 2 + . . .
= x + a2 x 2 + . . .
will be linearly independent. Basically if we can find two solution functions y1 and y2 for
the DE which are linearly independent then the linear combination

y = C1 y 1 + C2 y 2

for the second order DE should give the complete general solution.
132 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

Trying a0 = 1 and a1 = 0 gives

1 1 1
a2 = a0 = 1=
2! 2! 2!
1 1
a3 = a1 = ×0=0
3! 3!
1 1 1
a4 = a0 = 1=
4! 4! 4!
and we obtain
y 1 = a0 + a1 x + . . .
1 1
= 1 + 0 ⋅ x + x2 + 0x3 + x4 + . . .
2! 4!
x2 x4
=1+ + + ...
2! 4!
Trying a0 = 0 and a1 = 1 gives
1 1
a2 = a0 = ×0=0
2! 2!
1 1 1
a3 = a1 = ×1=
3! 3! 3!
1 1
a4 = a0 = ×0=0
4! 4!
1 1 1
a5 = a1 = ×1=
5! 5! 5!
and we obtain
y2 = a0 + a1 x + . . .
1 3 1
= 0 + x + 0x2 + x + 0x4 + x5 + . . .
3! 5!
x3 x5
=x+ + + ...
3! 5!
The full solution will then be
y = C1 y 1 + C2 y 2
x2 x4 x3 x5
= C1 [1 + + + . . .] + C2 [x + + + . . .]
2! 4! 3! 5!
as we found before.
Fitting the conditions we have
y(0) = C1 × (1 + 0) + C2 × 0 = C1 = 3
2x 4x3 3x2 5x5
y ′ = C1 (0 + + + . . .) + C2 (1 + + + . . .)
2! 4! 3! 5!
So
y ′ (0) = C1 × 0 + C2 × 1 = C2 = 2
and we obtain
x2 x4 x3 x 5
y = 3 {1 + + + . . .} + 2 {x + + + . . .} .
2! 4! 3! 5!
These series should look familiar. From previous studies on series you should recognise
x3 x5
x+ + + . . . = sinh x
3! 5!
4.2. INFORMAL SERIES APPROACH TO THE SOLUTION OF DIFFERENTIAL EQUATIONS133

and
x2 x4
1+ + + . . . = cosh x
2! 4!
We then have
y = C1 cosh x + C2 sinh x.
and with the initial conditions
y = 3 cosh x + 2 sinh x
Further, by reducing this solution to exponential form matches the previously found ana-
lytic solution:
ex + e−x ex − e−x
y =3⋅ +2⋅
2 2
3ex + 3e−x + 2ex − 2e−x
=
2
5ex + e−x 5 x 1 −x
= = e + e
2 2 2
as before.
More formal approach to series solution
We consider the more general form of a second order linear differential equation as
a2 (x)y ′′ + a1 (x)y ′ + a0 (x)y = 0
The form and existence of a series solution to this DE depends on whether the series is
taken about an ordinary point or a singular point of the DE.
Singular points for the DE are found by solving a2 (x) = 0. If any x satisfy this they are
called singular points and can be further classified. All other points are ordinary points for
the given DE. For a series solution about a singular point (depending on the classification
of the point) we will try the method of Frobenius. This will be covered in the next section.
For now we will only consider the series solution for a DE about an ordinary point. If x = a
is an ordinary point of the DE, then the substitution
y = a0 + a1 (x − a) + a2 (x − a)2 + a3 (x − a)3 + . . .
∞
= ∑ an (x − a)n
n=0

will lead to two series solutions involving two arbitrary constants and hence to the general
solution.
We can also construct a recurrence relation for the series solution by using the summation
form of the general series form as follows.
We write ∞
y = a0 + a1 x + a2 x 2 + . . . = ∑ an x n
n=0
Then ∞
y ′ = a1 + 2a2 x + 3a3 x2 + . . . = ∑ nan xn−1
n=1
and
y ′′ = 2a2 + 3 ⋅ 2a3 x + 4 ⋅ 3a4 x2 + . . .
∞
= ∑ n(n − 1)an xn−2
n=2
134 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

For Example 1 the DE becomes

∞ ∞
∑ n(n − 1)an xn−2 − ∑ an xn = 0
n=2 n=0

To simplify this under one summation we need the summation starting counter values to
be the same. To simplify our working we will also set the x indexes to be the same, in this
case xn . We do this step first, before adjusting the counter starting values. For the first sum
we simply add 2 to the counter and subtract 2 from the starting value giving:
∞ ∞
∑ (n + 2)(n + 2 − 1)an+2 xn+2−2 = ∑ (n + 2)(n + 1)an+2 xn .
n=0 n=0

If this is done correctly the summations are equivalent. This can be easily checked by
generating the first 2 terms for each sum and comparing as follows:
∞
∑ n(n − 1)an xn−2 = 2(2 − 1)a2 x2−2 + 3(3 − 1)a3 x3−2 + . . .
n=2
= 2 ⋅ 1a2 x0 + 3 ⋅ 2a3 x1 + . . .
= 2a2 + 6a3 x + . . .
∞
∑ (n + 2)(n + 1)an+2 xn =(0 + 2)(0 + 1)a0+2 x0 + (1 + 2)(1 + 1)a1+2 x1 + . . .
n=0
= 2 ⋅ 1a2 + 3 ⋅ 2a3 x + . . .
= 2a2 + 6a3 x + . . .
as before.
We then have ∞ ∞
∑ (n + 2)(n + 1)an+2 xn − ∑ an xn = 0
n=0 n=0
This adjustment has also resolved the first problem as the counter start values are now
equal and the terms can now be put in one summation. This will not always happen and
the counter start values will have to be adjusted. This is done by setting all the counter start
values to the highest start value. This is done by taking terms out of each summation till
they all start at the same highest value.
For example
∞ ∞ ∞
∑ an x n + ∑ b n x n + ∑ c n x n
n=2 n=1 n=0
The highest counter start value is 2 so we take terms out of the remaining summations until
their counter start values are 2 as follows:
∞ ∞ ∞
∑ an x n + b 1 x 1 + ∑ b n x n + c 0 x 0 + c 1 x 1 + ∑ c n x n
n=2 n=2 n=2
∞
= b1 x + c0 + c1 x + ∑ (an xn + bn xn + cn xn )
n=2

For our problem we now form:

∞
∑ [(n + 2)(n + 1)an+2 xn − an xn ]
n=0
∞
= ∑ [(n + 2)(n + 1)an+2 − an ] xn = 0
n=0
4.2. INFORMAL SERIES APPROACH TO THE SOLUTION OF DIFFERENTIAL EQUATIONS135

All the terms must be zero on the left hand side so we have
(n + 2)(n + 1)an+2 − an = 0
giving the recurrence relation
1
an+2 = an .
(n + 2)(n + 1)
For computer programming for a numerical solution using power series the general recur-
rence relation can be used.
Finding series terms for y1 and y2
For simple cases where the recurrence relation yields series which are odd and even, the
following approach can be taken to generate the general term for each series. For the even
series we want to find ∞
y1 = ∑ a2n x2n = a0 + a2 x2 + . . .
n=0
and for the odd series ∞
y2 = ∑ a2n+1 x2n+1 = a1 x + a3 x3 + . . .
n=0
The usual approach is to let the leading coefficient in each series be 1 and write the general
solution for the DE as
y = C1 y 1 + C2 y 2
This is equivalent to using the general recurrence relation and setting a0 = 1 with a1 = 0 to
find the even series and a0 = 0 with a1 = 1 to find the odd series.
This will work but a quicker approach for us to try will be to adjust the recurrence relation
for each case and use it to construct the general term as follows. For the even series we
want to find a2n so we adjust
1
an+2 = an
(n + 2)(n + 1)
to obtain a2n = . . .
The first step is to obtain an = . . . by subtracting 2 from the counter to give
1 1
an−2+2 = an = an−2 = an−2 .
(n − 2 + 2)(n − 2 + 1) n(n − 1)
Next replace n by 2n to obtain
1
a2n = a2n−2
2n(2n − 1)
and we have so far generated the corresponding recurrence relation for the even series from
the general recurrence relation. To find the general term a2n we simply expand as follows:
1
a2n−2 = a2n−2−2
(2n − 2)(2n − 2 − 1)
1
= a2n−4
(2n − 2)(2n − 3)
1
a2n−4 = a2n−4−2
(2n − 4)(2n − 4 − 1)
1
= a2n−6
(2n − 4)(2n − 5)
136 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

So we have
1 1 1
a2n = a2n−2 = ⋅ a2n−4
2n(2n − 1) 2n(2n − 1) (2n − 2)(2n − 3)
1 1
= ⋅ a2n−6
2n(2n − 1)(2n − 2)(2n − 3) (2n − 4)(2n − 5)

This process can be continued till we reach the last coefficient a0 .

This obviously can get a bit messy so we usually just have the beginning few factors and
the last few. The last few can be easily found by using the recurrence relation in reverse
from a0 . We want the last term to give some factor by a0 so we simply set 2n − 2 = 0 to
give n = 1 and we have
1 1
a2 = ⋅ a0 = a0
2(2 − 1) 2⋅1
As a check set n = 2 to obtain a2
1 1 1
a4 = a2 = ⋅ a0
4!(4 − 1) 4⋅3 2⋅1

1
So the last few factors will be . . . a0 and we have putting it all together
4⋅3⋅2⋅1
1
a2n = a0
2n(2n − 1)(2n − 2)(2n − 3) . . . 4 ⋅ 3 ⋅ 2 ⋅ 1

Now we have the general formula for a2n the next step is to simplify it. We look for factors
of some number (2, 3 etc. ) and any factorials (n!, 2n! etc. ) For this example we have
quickly
2n(2n − 1)(2n − 2)(2n − 3) . . . ⋅ 4 ⋅ 3 ⋅ 2 ⋅ 1 = (2n)!
and we obtain
1
a2n = a0
(2n)!
giving
∞ ∞
x2n
y1 = ∑ a2n x2n = ∑
n=0 n=0 (2n)!
x2 x4
=1+ + + ...
2! 4!
as we found before.
For the odd series we want to find a2n+1 so we adjust
1
an+2 = an
(n + 2)(n + 1)

similarly as before but now to obtain a2n+1 = . . .

First we subtract 1 (instead of 2 ) from the counter to give
1
an−1+2 = an+1 = an−1
(n − 1 + 2)(n − 1 + 1)
1
= an−1
(n + 1)n
4.2. INFORMAL SERIES APPROACH TO THE SOLUTION OF DIFFERENTIAL EQUATIONS137

Next replace n by 2n to obtain

1
a2n+1 = a2n−1
(2n + 1)2n
Expand to find the first few factors:
1
a2n−1 = a2n−3
(2n − 1)(2n − 2)
So
1 1
a2n+1 = ⋅ a2n−3
(2n + 1)2n (2n − 1)(2n − 2)
To find the last few factors - set n = 1 to get to a1 :
1 1
a3 = a1 = a1
(2 + 1) ⋅ 2 3⋅2
n = 2 gives
1 1 1
a5 = a3 = ⋅ a1
5⋅4 5⋅4 3⋅2
So the last few factors are
1
... a1
5⋅4⋅3⋅2
giving
1
a2n+1 = a1
(2n + 1)(2n)(2n − 1)(2n − 2) . . . 5 ⋅ 4 ⋅ 3 ⋅ 2
Trying to simplify we note that

(2n + 1)(2n)(2n − 1)(2n − 2) . . . 5 ⋅ 4 ⋅ 3 ⋅ 2 = (2n + 1)!

giving
1
a2n+1 = a1
(2n + 1)!
and if we set a1 = 1 we have
1
a2n+1 = a1
(2n + 1)!
giving
∞ ∞
x2n+1
y2 ∑ a2n+1 x2n+1 = ∑
n=0 n=0 (2n + 1)!
x x3 x5
= + + + ...
1! 3! 5!
x3 x5
=x+ + + ...
3! 5!
as we found before.
Series convergence
It is usually a good idea to check that the series solutions obtained converge (at least for
some region) especially if a numerical solution is wanted. For
∞
x2n
y1 = ∑
n=0 (2n)!
138 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

we try the ratio test and look at

an+1 x2n+2 (2n)!
lim ∣ ∣ = lim ∣ ⋅ 2n ∣
n→∞ an n→∞ (2n + 2)! x
x2 ⋅ (2n)!
= lim ∣ ∣
n→∞ (2n + 2)(2n + 1)(2n)!

∣x2 ∣
= lim =0
n→∞ (2n + 2)(2n + 1)

so converges for all x.

Similarly for
∞
x2n+1
y2 = ∑
n=0 (2n + 1)!
we look at
an+1 x2n+3 (2n + 1)!
lim ∣ ∣ = lim ∣ ⋅ ∣
n→∞ an n→∞ (2n + 3)! x2n+1
x2 ⋅ (2n + 1)!
= lim ∣ ∣
n→∞ (2n + 3)(2n + 2)(2n + 1)!

∣x2 ∣
= lim =0
n→∞ (2n + 3)(2n + 2)

and also converges for all x.

L.D.E.s with non-constant coefficients
We will now examine closely the underlying theory for the series solution of a L.D.E. with
non-constant coefficients. The method in practice requires a number of systematic steps:

Step 1 We identify and classify the singular points of the D.E. as shown in the next sec-
tion.
Step 2 If x = 0 is an ordinary point (i.e. x = 0 is not a singular point), then we assume
that ∞
y = ∑ an x n
n=0
is a solution (where a0 , a1 , a2 , . . . must be found), and substitute for y and its deriva-
tives in the given D.E.
Note: If x = 0 is a singular point, then we have to assume that
∞
y = x c ∑ an x n
n=0

is a solution (where c, a0 , a1 , a2 , . . . must be found.) This is the method of Frobe-

nius as discussed later.
Step 3 By equating coefficients, we find the values of the coefficients of the series in terms
of one or more of the “early” coefficients, a0 , a1 , . . ..
Step 4 We use the coefficients we have found to write down the series solution of the given
D.E. (We can, if we wish, check this solution!)
If possible, we try to write down the general term of the series. (Except for “obvi-
ous” cases, this will not be required in the examination.)
4.2. INFORMAL SERIES APPROACH TO THE SOLUTION OF DIFFERENTIAL EQUATIONS139

Step 5 If we have the correct number of arbitrary constants in our series solution, we have
at once the general solution of the D.E. (If we do not have the general solution,
then more “powerful”’ methods, beyond the scope of this subject, are required to
obtain the complete solution!)
Note that in the next section a general theorem is given which “controls” in advance
the situation that will occur in Step 5.)

The existence of series solutions

We will now consider the existence of series solutions of a second order D.E. of the form

a2 (x)y ′′ + a1 (x)y ′ + a0 (x)y = 0

Note
If the coefficient functions, a2 , a1 or a0 , contain a denominator involving x, then we must
first “clear” these denominators by multiplying each term in the D.E. by the appropriate
factor(s). (Thus, a2 (x), a1 (x) and a0 (x) will usually be polynomials.)
Difficulties can occur in the series solution of such L.D.E.s with variable coefficients if the
leading coefficient, here a2 (x) for y ′′ , is zero. Thus, having simplified the D.E. to include
no denominators, we must solve for a2 (x) = 0. The roots of this equation, if any, are called
the singular points of the D.E. All other values of x are called ordinary points i.e. x = a
is called an ordinary point of the D.E. if a2 (a) ≠ 0.
The next step is to write the DE in standard form. Providing a2 (x) ≠ 0, divide by a2 (x) to
obtain
y ′′ + p(x)y ′ + q(x)y = 0
If x = a is a singular point of the D.E. (i.e. if p(a) is not defined), then we classify the point
x = a further as follows.
If lim(x − a)p(x) and lim(x − a)2 q(x) both exist, then x = a is called a regular singular
x→a x→a
point.
If either of the limits does not exist, then x = a is called an irregular singular point.
Examples
y ′′ − 2xy ′ + 3y = 0 has no singular points. In particular, x = 0 is an ordinary point.
1 x
xy ′′ + y ′ + xy = 0 has a singular point at x = 0. Since lim x ⋅ = 1 and lim x2 = 0, hence
x→0 x x→0 x
x = 0 is a regular singular point.
2x2 y ′′ − xy ′ + (1 + x)y = 0 has a regular singular point at x = 0. (Why?)
2x2 (x + 2)y ′′ + (x + 2)y ′ + 3xy = 0 has a regular singular point at x = −2 and an irregular
singular point at x = 0. (Why?)

Tutorial 4 Question 1
140 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

General theorem about the existence of series solutions

We can now state the general theorem about the existence of series solutions of

a2 (x)y ′′ + a1 (x)y ′ + ao (x)y = 0

∞
i. If x = a is an ordinary point of the D.E., then the substitution y = ∑ an (x − a)n will
n=0
lead to two series solutions involving two arbitrary constants and hence to the general
solution of the D.E. (See Examples 1 and 2 in the next Section.)
∞
ii. If x = a is a regular singular point, then y = ∑ an (x − a)n may lead to one series
n=0
solution. (See Example.) In general, if x = a is a regular singular point, then the
∞
substitution y = (x − a)c ∑ an (x − a)n will lead (through the method of Frobenius),
n=0
to a particular solution or to the general solution. (See Example and method of
Frobenius in section on Series solution about a regular singular point.)

iii. If x = a is an irregular singular point, then Frobenius series solutions may or may not
exist.

Note
We will consider only power series solutions in powers of x, i.e. about x = 0. Thus for our
D.E. solutions we will be concerned only about the classification of x = 0 If x = 0 is an
∞
ordinary point of the D.E., then we let y = ∑ an xn , and proceed as in the next section.
n=0
∞
If x = 0 is a regular singular point of the D.E., then we let y = ∑ an xn+c , and proceed
n=0
as in the section on Series solution about a regular singular point. (If x = 0 is an irregular
singular point, then we try the Frobenius method, but we may or may not get any solution!)

4.3 Power series solution about an ordinary point

We will restrict our attention to second order linear D.E.s with variable coefficients, i.e. of
the form
a2 (x)y ′′ + a1 (x)y ′ + a0 (x)y = 0
The approach to the solution by series about x = 0 of such D.E.s can be summarised as
follows:

1. We assume that a power series solution of the form

∞
y = ∑ an x n = a0 + a1 x + a2 x 2 + . . .
n=0

is a solution; find the necessary derivatives of the power series and substitute in the
D.E. We then try to find the value of each an , usually by the technique of “equating
coefficients”.
4.3. POWER SERIES SOLUTION ABOUT AN ORDINARY POINT 141

2. If a series solution is obtained, we try to check that it is a solution.

3. If we find that it is a solution, we try to discover whether it is the only solution.

4. If it is not the only solution, we try to find the other solutions.

Notes

1. By the general theorem about the existence of series solutions stated previously, it is
known that when x = 0 is an ordinary point of the second order D.E., Step 1 (done
correctly!) will automatically yield the general solution, involving two linearly inde-
pendent series and two arbitrary constants. (See Examples 1 and 2 ) Hence Steps 2,3
and 4 are unnecessary in this case.

2. Nevertheless, it is often useful to carry out Step 2, i.e. to check each series solution in
the original D.E. (This should help you discover errors and also should improve your
confidence that the method works!)

3. The power series method may work if x = 0 is a regular singular point but all four
steps are likely to be needed. The methods for finding the other solution(s) are quite
“technical”.

4. More generally, if x = 0 is a regular singular point, it is better to use the same four
step approach but to assume that a Frobenius series of the form
∞
∑ an xn+c
n=0

is a solution, where the appropriate an and c values must be found. One such series
solution is always obtained in step one but all four steps may be necessary. Again,
quite complicated methods (beyond the scope of this subject) may be needed to find the
general solution.

Example 2

Find the solution of the D.E.

y ′′ − 2xy ′ + 3y = 0
with y(0) = 4 and y ′ (0) = −2
Solution
Here there are no singular points and hence x = 0, in particular, is an ordinary point and we
use the power series method. Assume that the solution is given by
∞
y = a0 + a1 x + a2 x 2 + . . . = ∑ an x n
n=0

then ∞
y ′ = a1 + 2a2 x + 3a3 x2 + . . . = ∑ nan xn−1
n=1

and ∞
y ′′ = 2a2 + 2 ⋅ 3a3 x + 3 ⋅ 4a4 x2 + . . . = ∑ n(n − 1)an xn−2 .
n=2
142 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

Substituting these series into the given D.E. gives

∞ ∞ ∞
n−2 n−1
∑ n(n − 1)an x − 2x ∑ nan x + 3 ∑ an x n = 0
n=2 n=1 n=0

or ∞ ∞ ∞
∑ n(n − 1)an xn−2 − ∑ 2nan xn + ∑ 3an xn = 0 .
n=2 n=1 n=0
In order to find the required values for the an ’s it is necessary to combine these infinite
series. To do this it is usually best to “adjust” each series, as necessary, to contain the same
general power of x.
Here, the last two series contain xn and so we re-write the first series as
∞ ∞
∑ n(n − 1)an xn−2 = ∑ (n + 2)(n + 1)an+2 xn
n=2 n=0

This is quite a “tricky” step and involves here replacing n by n + 2 to get the required xn
term. (Note that the new series has lower limit given by (n + 2) = 2, i.e. n = 0). A good
check is to examine the first term on each side. Here the first term on the left-hand side is
2(2 − 1)a2 x2−2 = 2a2 while on the r.h.s. it is (0 + 2) (0 + 1)a0+2 x0 = 2a2 , which checks.
Hence the D.E. can now be written as
∞ ∞ ∞
∑ (n + 2)(n + 1)an+2 xn − ∑ 2nan xn + ∑ 3an xn = 0
n=0 n=1 n=0

We now want to combine the infinite series by grouping powers of x. The series must all
“start” at the same value for n : here, n = 1 may be used.
∞ ∞ ∞
∴ {2a2 + ∑ (n + 2)(n + 1)an+2 xn } − ∑ 2nan xn + {3a0 + ∑ 3an xn } = 0
n=1 n=1 n=1
∞
∴ (2a2 + 3a0 ) + ∑ [(n + 2)(n + 1)an+2 − 2nan + 3an ] xn = 0
n=1

This equation is (finally!) in the form b0 + b1 x + b2 x2 + . . . = 0. By a theorem in a power

series analysis it is known that such a power series can only be equal to zero if b0 = b1 =
b2 = . . . = 0, i.e. all the coefficients must be equated to zero. Thus, in our equation, we
must have that
−3
(2a2 + 3a0 ) = 0, i.e. a2 = a0
2
and in general, for n ≥ 1,

[(n + 2)(n + 1)an+2 − 2nan + 3an ] = 0

giving the recurrence relation

2n − 3
an+2 = an , n ≥ 1
(n + 1)(n + 2)
Now, for n = 1,
2⋅1−3 −1
a3 = a1 = a1 ,
2⋅3 2⋅3
and for n = 2,
1 1 −3 −3 ⋅ 1
a4 = a2 = ( a0 ) = a0 ,
3⋅4 3⋅4 2 2⋅3⋅4
4.3. POWER SERIES SOLUTION ABOUT AN ORDINARY POINT 143

and so on.
All the an ’s are now known in value in terms of a0 (a2 , a4 , a6 , . . .) and a1 (a3 , a5 , a7 , . . .).
Thus a0 and a1 will be the required two arbitrary constants needed for the general solution.
The only problem remaining is to find a general expression for the coefficients in the power
series. This is often hard to do and can really only be done by “looking for patterns” in the
first coefficients. (You should thus be very careful not to destroy patterns by cancelling out
factors or by finding products too soon.)
Here we have
−3 −3
a2 = a0 = a0
2 2!
−1 −1
a3 = a1 = a1
2⋅3 3!
−3 ⋅ 1 −3 ⋅ 1
a4 = a0 = a0
2⋅3⋅4 4!
3 −1 ⋅ 3
a5 = a3 = a1
4⋅5 5!
5 −3 ⋅ 1 ⋅ 5
a6 = a4 = a0
5⋅6 6!
7 −1 ⋅ 3 ⋅ 7
a7 = a5 = a1
6⋅7 7!
⋮
The pattern in the coefficients should now be evident. (Try writing down a11 , a12 and then
calculating them to check your pattern). It is clear that the coefficients of the odd and even
powers of x are determined differently. We can find the general term using the recurrence
relation. If we want to find the coefficient of x2k , a2k , then we will have to let (n + 2) = 2k
in the recurrence relation.
4k − 7
∴ a2k = a2k−2 (n = 2k − 2)
(2k − 1)(2k)
To find the general form we construct the first few and last few factors as follows.
2(2k − 2) − 7
a2k−2 = a2k−2−2
(2k − 2 − 1)(2k − 2)
4k − 11
= a2k−4
(2k − 3)(2k − 2)
For a0 we set k = 1 giving
4−7 −3
a2 = a2−2 = a0
(2 − 1) ⋅ 2 1⋅2
and k = 2 gives
8−7 1
a4 = a4−2 = a2
(4 − 1) ⋅ 4 3⋅4
1 −3
= ⋅ a0
3⋅4 1⋅2
giving
4k − 7 4k − 11 1 ⋅ (−3)
a2k = ⋅ ... ⋅ ⋅ a0
(2k − 1)(2k) (2k − 3)(2k − 2) 3⋅4⋅1⋅2
(4k − 7)(4k − 11) . . . 1 ⋅ (−3)
= a0
(2k − 1)(2k)(2k − 3)(2k − 2) . . . 3 ⋅ 4 ⋅ 1 ⋅ 2
144 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

We can simplify this by rearranging the denominator to obtain

(2k)(2k − 1)(2k − 2)(2k − 3) . . . ⋅ 4 ⋅ 3 ⋅ 2 ⋅ 1 = (2k)!

So we have
(4k − 7)(4k − 11) . . . ⋅ 1 ⋅ (−3)
a2k = a0
(2k)!
which only makes sense if we start from k = 1 giving
3
a2 = − a0
2
The series is ∞
∑ a2k x2k = a0 + a2 x2 + . . .
k=0
So if a0 = 1 we have
∞
y1 = 1 + ∑ a2k x2k
k=1
∞
(4k − 7)(4k − 11) . . . 1 ⋅ (−3) 2k
=1+∑ x
k=1 (2k)!
which may be a little easier to follow if we rewrite it as
∞
(−3) ⋅ 1 . . . (4k − 7) 2k
y1 = 1 + ∑ x
k=1 (2k)!
Similarly, with n = 2k − 1,
4k − 5
a2k+1 = a2k−1
(2k)(2k + 1)
As before, we construct the first few and last few factors as follow:
2(2k − 2) − 5
a2k−1 = a2k−2−1
(2k − 2)(2k − 2 + 1)
4k − 9
= a2k−3
(2k − 2)(2k − 1)
For a1 we set k = 1 to obtain
4−5 −1
a3 = a2−1 = a1
2(2 + 1) 2⋅3
and k = 2 gives
8−5 3
a5 = a4−1 = a3
4(4 + 1) 4⋅5
3 −1 3 ⋅ (−1)
= ⋅ a1 = a1
4⋅5 2⋅3 5⋅4⋅3⋅2
giving
4k − 5 4k − 9 3 ⋅ −1
a2k+1 = ⋅ ... a1
(2k)(2k + 1) (2k − 2)(2k − 1) 5⋅4⋅3⋅2
−1 ⋅ 3 . . . (4k − 5)
= a1
(2k + 1)(2k)(2k − 1)(2k − 2) . . . 5 ⋅ 4 ⋅ 3 ⋅ 2
−1 ⋅ 3 . . . (4k − 5)
= a1
(2k + 1)!
4.3. POWER SERIES SOLUTION ABOUT AN ORDINARY POINT 145

Again, this only makes sense for k ≥ 1 so we obtain

∞
a1 x + ∑ a2k+1 x2k+1
k=1

and setting a1 = 1 gives

∞
−1 ⋅ 3 . . . (4k − 5) 2k+1
y2 = x + ∑ x
k=1 (2k + 1)!
The general solution is then
y = C1 y 1 + C2 y 2
∞ ∞
−3 ⋅ 1 . . . (4k − 7) 2k −3 ⋅ 1 . . . (4k − 5) 2k+1
= C1 {1 + ∑ x } + C2 {x + ∑ x }
k=1 (2k)! k=1 (2k + 1)!
Alternatively we could find the solution of the original D.E. by substituting the coefficients
we have calculated into the power series assumed at the start of the problem, i.e. into
∞
y = a0 + a1 x + a2 x2 + a3 x3 + . . . = ∑ an xn
n=0

−3 −1 −3.1
∴ y = a0 + a1 x + ( a0 ) x 2 + ( a1 ) x 3 + ( a0 ) x 4
2 3! 4!
−1.3 −3.1.5
+( a1 ) x 5 + ( a0 ) x6 + . . .
5! 6!
i.e.
−3 2 −3.1 4 −3.1.5 6
y = a0 {1 + x + x + x + . . . .}
2! 4! 6!
−1 −1.3 5 −1.3.7 7
+ a1 {x + x3 + x + x + . . .}
3! 5! 7!
or
∞
−3.1 . . . (4k − 7) 2k
y = a0 {1 + ∑ x }
k=1 (2k)!
∞
−3.1 . . . (4k − 5) 2k+1
+ a1 {x + ∑ x }
k=1 (2k + 1)!
This is the general solution of the D.E. y ′′ − 2xy ′ + 3y = 0 since there are two arbitrary
constants a0 and a1 . If initial conditions were specified, then corresponding values for a0
and a1 could then be calculated. Fitting the initial conditions:
y(0) = C1 y1 (0) + C2 y2 (0) = 4
y1 (0) = 1 + 0 = 1 y2 (0) = 0 + 0
So we have
y(0) = C1 ⋅ 1 + C2 ⋅ 0 = C1 = 4
∞
−3 ⋅ 1 . . . (4k − 7)
y1′ = 0 + ∑ ⋅ 2kx2k−1
k=1 (2k)!
∞
−3 −3 ⋅ 1 . . . (4k − 7) 2k−1
= ⋅ 2x1 + ∑ x
3! k=2 (2k − 1)!
So y1′ (0) = 0
∞
−1 ⋅ 3 . . . (4k − 5)
y2′ = 1 + ∑ (2k + 1)x2k
k=1 (2k + 1)!
∞
−1 ⋅ 3 . . . (4k − 5) 2k
=1+∑ x
k=1 (2k)!
146 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

So y2′ (0) = 1 giving

y ′ (0) = C1 y1′ (0) + C2 y2′ (0)
= 4 × 0 + C2 × 1 = C2 = −2
giving
y = 4y1 − 2y2
∞
−3 ⋅ 1 . . . (4k − 7)x2k
=4 {1 + ∑ }
k=1 (2k)!
∞
−1 ⋅ 3 . . . (4k − 5)x2k+1
− 2 {x + ∑ }
k=1 (2k + 1)!
Check for convergence of series
For y1 we look at (use k or n)

ak+1 −3 ⋅ 1 . . . (4k − 7)(4k − 3)x2k+2 (2k)!

lim ∣ ∣ = lim ∣
k→∞ ak k→∞ (2k + 2)!x2k ⋅ (−3) ⋅ 1 . . . (4k − 7)
(4k − 3) ⋅ x2 (2k)!
= lim ∣ ∣
k→∞ (2k + 2)(2k + 1)(2k)!

4k − 3
= lim ∣x∣2 = 0
k→∞ (2k + 2)(2k + 1)

So converges for all x.

Similarly for y2 we look at

ak+1 −1 ⋅ 3 . . . (4k − 5)(4k − 1)x2k+3 (2k + 1)!

lim ∣ ∣ = lim ∣ ∣
k→∞ ak k→∞ (2k + 3)!(−1) ⋅ 3 . . . (4k − 5) ⋅ x2k+1
(4k − 1)x2 (2k + 1)!
= lim ∣ ∣
k→∞ (2k + 3)(2k + 2)(2k + 1)!

4k − 1
= lim ∣x∣2 = 0
k→∞ (2k + 3)(2k + 2)

and also converges for all x.

So y = C1 y1 + C2 y2 converges for all x for C1 , C2 finite constants.

Example 3

Solve the D.E.

(x2 − 1) y ′′ + 2xy ′ − y = 0, subject to y(0) = 4, y ′ (0) = −3.

Solution
Let ∞
y = ∑ an x n ,
n=0

then ∞ ∞
y ′ = ∑ nan xn−1 and y ′′ = ∑ n(n − 1)an xn−2 .
n=1 n=2

The D.E. can be written as

x2 y ′′ − y ′′ + 2xy ′ − y = 0
4.3. POWER SERIES SOLUTION ABOUT AN ORDINARY POINT 147

Substituting the series into the D.E. gives

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

Hence,
∞ ∞ ∞ ∞
∑ n(n − 1)an xn − ∑ (n + 2)(n + 1)an+2 xn + ∑ 2nan xn − ∑ an xn = 0
n=2 n=0 n=1 n=0

and so
∞ ∞
∑ n(n − 1)an xn − 2 ⋅ 1a2 − 3 ⋅ 2a3 x − ∑ (n + 2)(n + 1)an+2 xn + 2a1 x
n=2 n=2
∞ ∞
+ ∑ 2nan xn − a0 − a1 x − ∑ an xn = 0
n=2 n=2

∴ (−2 ⋅ 1a2 − a0 ) + (−3 ⋅ 2a3 + 2a1 − a1 ) x

∞
+ ∑ [n(n − 1)an − (n + 2)(n + 1)an+2 + 2nan − an ] xn = 0
n=2
Hence, on equating the coefficients to zero
a0
−2 ⋅ 1a2 − a0 = 0 or a2 = −
2!
a1
−3 ⋅ 2 ⋅ a3 + 2a1 − a1 = 0 or a3 =
3!
For n ≥ 2, [n(n − 1)an − (n + 2)(n + 1)an+2 + 2nan − an ] = 0 i.e.,
n(n − 1) + 2n − 1 n2 + n − 1
an+2 = an = an
(n + 1)(n + 2) (n + 1)(n + 2)
5 −5
∴ a4 = a2 = a0
3⋅4 4!
19 −5 ⋅ 19
a6 = a4 = a0
5⋅6 6!
and so on.
In general (letting n = 2k − 2 ) in the recurrence relation,
4k 2 − 6k + 1 −5 ⋅ 19 . . . (4k 2 − 6k + 1)
a2k = a2k−2 = a0 ,
(2k − 1)(2k) (2k)!
11 11
a5 = a3 = a1
4⋅5 5!
29 11 ⋅ 29
a7 = a5 = a1
6⋅7 7!
⋮
4k 2 − 2k − 1 11 ⋅ 29 . . . (4k 2 − 2k − 1)
a2k+1 = a2k−1 = a1
2k(2k + 1) (2k + 1)!
Hence the general solution is given by y = a0 y1 + a1 y2 , where
1 2 5 4 5 ⋅ 19 6
y1 = 1 − x − x − x − . . . and
2! 4! 6!
1 11 11 ⋅ 29 7
y2 = x + x3 + x5 + x + ....
3! 5! 7!
148 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

Now when x = 0, y = 4 and so 4 = a0 {1} + a1 {0} therefore a0 = 4.

Also,
1 5 1 11
y ′ = a0 {− 2x − 4x3 − . . .} + a1 {1 + 3x2 + 5x4 + . . .}
2! 4! 3! 5!
′
and so, since y (0) = −3,
−3 = a0 {0} + a1 {1}
∴ a1 = −3.
Hence, the solution to the given initial value problem is

y = 4y1 − 3y2 .

Tutorial 4 Question 2

4.4 Series solution about a regular singular point

Read textbook section 8.6 (Method of Frobenius)

Video (by Dr. R. Wood): Series solution about a regular singular point
When x = 0 is a regular singular point of a given D.E., then the method of Frobenius
requires that we assume that the series solution is of the form

y = xc (a0 + a1 x + a2 x2 + . . .)

i.e. ∞
y = ∑ an xn+c
n=0

where c and the coefficients an must now be found.

Indicial equation
This method of Frobenius uses the indicial equation to find c, and uses the usual process
of equating coefficients to find the an ’s.
The approach is much as for the power series method, but the computations can become
more complicated! Careless mistakes can lead to horrendous problems, so be particularly
neat and careful!
The indicial equation does come out in the general working of the DE using the series
solution approach. However, there is also an alternative quick method which can be used to
derive the indicial equation. It is recommended this approach be taken as a check that your
series working is most likely correct if youl obtain the same equation. The quick method is
also very handy if you are more interested in investigating what form the general solution
may take instead of working fully through to the full solution.
4.4. SERIES SOLUTION ABOUT A REGULAR SINGULAR POINT 149

The quick method is as follows.

Suppose we are investigating a second order DE with a regular singular point at x = a. The
general form of the DE can be written as

a2 (x)y ′′ (x) + a1 (x)y ′ (x) + a0 (x)y(x) = 0

Given this form we rewrite the DE in standard form

y ′′ + p(x)y ′ + q(x)y = 0

We then evaluate
p0 = lim(x − a)p(x) and q0 = lim(x − a)2 q(x)
x→a x→a

Note. These limits would already have been determined for the classification of x = a as
being a regular singular point.
The indicial equation can be constructed as

c(c − 1) + p0 c + q0 = 0

Example 1

We will consider first an example where the method of Frobenius works without any ”com-
plications”: Solve the D.E.
2x2 y ′′ − xy ′ + (1 + x)y = 0
Solution
Writing the DE in standard form we obtain
x ′ 1+x
y ′′ − y + y=0
2x2 2x2
or
1 ′ 1+x
y ′′ − y + y=0
2x 2x2
giving
−1 1+x
p(x) = and q(x) = .
2x 2x2
Then
−1 1 1
p0 = lim(x − a)p(x) = lim x ⋅ = lim − = −
x→a x→0 2x x→0 2 2
and
1+x 1+x 1+0 1
q0 = lim(x − a)2 q(x) = lim x2 ⋅ 2
= lim = = .
x→a x→0 2x x→0 2 2 2
So the indicial equation should be

c(c − 1) + (− 12 ) c + 12 = 0 multiply by 2
2c(c − 1) − c + 1 = 2c2 − 2c − c + 1
= 2c2 − 3c + 1 = 0

with roots c = 1, 21 .
Let ∞
y = a0 xc + a1 xc+1 + a2 xc+2 + . . . = ∑ an xn+c .
n=0
150 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

Then ∞
y ′ = ca0 xc−1 + (c + 1)a1 xc + . . . = ∑ (n + c)an xn+c−1 ,
n=0

and
∞
′′ c−2 c−1
y = (c − 1)c ⋅ a0 x + c(c + 1)a1 x + . . . = ∑ (n + c − 1)(n + c)an xn+c−2
n=0

Substituting these series into the D.E. gives

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

Now ∞
∑ an xn+c+1
n=0

can be written as ∞
∑ an−1 xn+c .
n=1

So ∞ ∞
∑ {2(n + c − 1)(n + c) − (n + c) + 1}an xn+c + ∑ an−1 xn+c = 0,
n=0 n=1
or
∞
c
{2(c − 1)c − c + 1}a0 x + ∑ [{2(n + c − 1)(n + c) − (n + c) + 1}an + an−1 ] xn+c = 0.
n=1

Now, equating the coefficients of powers of x to zero, gives us

1. {2(c − 1)c − c + 1}a0 = 0

2. {2(n + c − 1)(n + c) − (n + c) + 1}an + an−1 = 0, for n ≥ 1.

The first coefficient, a0 , is always taken to be non-zero and so 2(c − 1)c − c + 1 = 0. This is
called the indicial equation (as we found before) and can now be solved for the appropriate
values of c, i.e.
2c2 − 3c + 1 = 0
(2c − 1)(c − 1) = 0
Hence the indicial roots are c1 = 1 and c2 = 21 .
Next, from (2), we obtain the recurrence relation for the coefficients, here
−1
an = an−1 , n ≥ 1,
2(n + c − 1)(n + c) − (n + c) + 1
and then substitute c1 and c2 in turn for c.
Note
The relationship between the values of c1 and c2 is crucial in determining whether only
one or two distinct series will be obtained by this substitution - see the Theorem following.
This example, in fact, demonstrates the “best” case.
Case I
4.4. SERIES SOLUTION ABOUT A REGULAR SINGULAR POINT 151

Put c = c1 = 1 in the recurrence relation.

−1
∴ an = an−1 , n ≥ 1. (Do this substitution carefully.)
n(2n + 1)
Hence
−1 −1
a1 = a0 = a0
1⋅3 1!3
−1 1 1
a2 = a1 = a0 = a0
2⋅5 1⋅3⋅2⋅5 2!3 ⋅ 5
−1 −1 −1
a3 = a2 = a0 = a0
3⋅7 1⋅3⋅2⋅5⋅3⋅7 3!3 ⋅ 5 ⋅ 7
...
(−1)n
an = a0 .
n!3 ⋅ 5 . . . (2n + 1)
Hence, a series solution is given by
1 1 1
y1 = x {1 − x+ x2 − x3 + . . .}
1!3 2!3 ⋅ 5 3!3 ⋅ 5 ⋅ 7
Note that it is convenient to take a0 ≡ 1 as we are just after a solution now and will put in
the arbitrary constant later.
Case II
1
Put c = c2 = 2 in the recurrence relation. After careful calculation we get that
−1
∴ an = an−1 for n ≥ 1.
n(2n − 1)
Hence a second series solution is given by

1 1 1 (−1)n
y2 = x 2 {1 − x+ x2 − . . . + xn + . . .}
1!1 2!1 ⋅ 3 n!1 ⋅ 3 . . . (2n − 1)
Thus the general solution of the given D.E. has yielded two different series solutions.
(These two series are, in fact, linearly independent (l.i.).
Thus the general solution is given by

y = Ay1 + By2

i.e.,
1 1
y =Ax {1 −
x+ x2 − . . .}
1!3 2!3 ⋅ 5
1 1 1
+Bx 2 {1 − x+ x2 − . . .}
1!1 2!1 ⋅ 3
where A and B are arbitrary constants.
Theorem A
As mentioned in the above example, the results we obtain on using the indicial roots in the
recurrence relation are determined by a Theorem which is as follows:

i. If the indicial roots differ by a constant which is not an integer or zero, the general
solution of the D.E. is always obtained.
152 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

ii. If the indicial roots differ by a non-zero integer, there are two possibilities:
a. The general solution is obtained by use of the smaller indicial root.
b. No solution is obtained by use of the smaller indicial root. However, in all cases
where the roots differ by a non-zero integer, one solution is always obtained by
using the larger root.
iii. If the indicial roots differ by zero, i.e. are equal, only one solution is obtained.

Notes

1. You will be required only to carry out “simple” solutions based on case (i).
2. To find the general solution in cases (ii)(b) and (iii) quite complicated methods must
be used, which are beyond the scope of this subject! Those methods are studied in
MTH418 (Topics in Calculus)

For completeness, a more general theorem is given below for dealing with D.E.s where
x = 0 is a regular singular point.
Theorem B
If x = 0 is a regular singular point of the D.E. p(x)y ′′ + q(x)y ′ + r(x)y = 0, then the D.E.
has two linearly independent series solutions y1 and y2 (and hence the general solution is
given by y = Ay1 + By2 ).
If c1 ≥ c2 are the indicial roots, then the l.i. solutions y1 and y2 are obtained as follows:

∞ ∞
a. If c1 − c2 is not an integer or zero, then y1 = xc1 ∑ an xn and y2 = xc2 ∑ bn xn where the
n=0 n=0
an ’s and bn ’s are obtained from the recurrence relation using c1 and c2 in turn.
∞
b. If c1 = c2 , then y1 = xc1 ∑ an xn where the an ’s are obtained from the recurrence relation
n=0
∞
and y2 = y1 ln x + xc1 ∑ bn xn where the bn ’s are obtained by substituting this form for
n=0
y2 in the given D.E. and again equating coefficients.
∞
c. If c1 − c2 = N, a positive integer, then y1 = xc1 ∑ an xn where the an ’s are obtained from
n=0
∞
the recurrence relation using c1 and y2 = ay1 ln x + xc2 ∑ bn xn where a and the bn ’s are
n=0
obtained by substituting this form for y2 in the given D.E.

Example

Try to solve xy ′′ + y ′ + xy = 0, using the method of Frobenius.

Solution
x = 0 is a regular singular point. Let
∞
y = ∑ an xn+c
n=0
4.4. SERIES SOLUTION ABOUT A REGULAR SINGULAR POINT 153

and so ∞
y ′ = ∑ (n + c)an xn+c−1
n=0
∞
y ′′ = ∑ (n + c)(n + c − 1)an xn+c−2
n=0

Now, on substituting into the D.E. and “adjusting” as before, we get that

{c(c − 1)a0 + ca0 } xc−1 + {c(c + 1)a1 + (c + 1)a1 } xc

∞
+ ∑ {(n + c)(n + c − 1)an + (n + c)an + an−2 } xn+c−1 = 0
n=2

Hence, equating coefficients as usual, we get:

Coefficient of xc−1 :
{c(c − 1) + c}a0 = 0
But, as always, we take a0 ≠ 0

∴ c(c − 1) + c = 0 i.e., c2 = 0 ∴ c = 0.

Trouble ahead: Case (ii) in the general existence theorem, and Case (b) in Theorem B!!
Coefficient of xc ∶
{(c + 1)c + (c + 1)}a1 = 0
∴ (c + 1)2 a1 = 0
But c ≠ −1 (c equals 0), and so a1 must be equal to zero here.
Recurrence relation
−1
an = an−2 , n ≥ 2
(n + c)2
Hence,
1
a2 = − a0
(c + 2)2
1
a3 = − a1 = 0
(c + 3)2
1 1
a4 = − 2
a2 = a0 ,
(c + 4) (c + 2) (c + 4)2
2

and so on.
Now, if we put c = 0, we get
1 1
a2 = − 2
a0 , a 4 = 2 2 a0 , . . . ,
2 2 ⋅4
and
a1 = a3 = a5 = . . . = 0
Thus, as expected, we do not obtain the general series solution but only the one series
solution
1 1
y = a0 (1 − 2 x2 + 2 2 x4 − . . .)
2 24
The series in brackets is known as J0 (x), the Bessel function of order zero. i.e. J0 (x) is a
“part” of the series solution of the given D.E.
154 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

Note
After considerable effort (not required by you here!) it can be shown that th second linearly
independent series solution has the “appalling” appearance

x2 x4 1 x6 1 1
y = J0 (x) ln x + { 2
− 2 2
(1 + ) + 2 2 2
(1 + + ) + . . .}
2 2 ⋅4 2 2 ⋅4 ⋅6 2 3

4.5 Summary of approach to the solution of D.E.s by use

of series
1. Given the D.E., determine the nature of the point x = 0.

2. If x = 0 is an ordinary point, then determine the general solution of the D.E. using
∞
the power series, y = ∑ an xn
n=0

∞
3. If x = 0 is a regular singular point, use the Frobenius series y = xc ∑ an xn to deter-
n=0
mine the indicial roots, c1 ≥ c2

a. If c1 − c2 is not an integer or zero, determine the two Frobenius series solutions

and hence the general solution.
b. If c1 = c2 , determine the one Frobenius series solution.
c. If c1 − c2 = N, a positive integer, then use c2 (the smaller root) to try to determine
the general solution of the D.E. involving two Frobenius series. If this fails, use
c1 (the larger root) to determine one Frobenius series solution.
(When there is only one Frobenius series solution then the general solution can be
obtained using more powerful methods.)

4. If x = 0 is an irregular singular point, then a Frobenius series solution of the form

∞
∑ an xn+c may or may not exist. In general, the solution of such a D.E. is very
n=0
difficult and beyond the scope of this subject.

Tutorial 4 Question 3
4.6. TUTORIAL 4 155

4.6 Tutorial 4
Back to Chapter 4

Question 1

Textbook exercise set 8.3. Problems 11, 13, 15

Detailed solutions to Question 1
Back to Chapter 4

Question 2

1. Find series solutions for each of the following:

a. y ′ = y − x
b. y ′′ + y = 0
c. y ′′ − y ′ − 2y = 0

(These can also be solved by the methods of Chapters 1 and/or 2 and 3. )

2. Solve
(1 − x2 ) y ′′ − 2xy ′ + 6y = 0, y(1) = 1, y ′ (0) = 0

3. Show that the solution of the initial value problem

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

is y = 3y1 − y2 , where
1 1
y1 (x) = 1 − x2 + x4 − x6 + . . .
6 30
and
1 7 5 19 7
y2 (x) = x − x3 + x − x + ....
4 160 1920
4. Textbook exercise set 8.3 Problems 19, 23, 25

Solutions

1. a.
c 2 c 3
y = (c + 1) + (c + 1)x + x + x + ...
2! 3!
= cex + x + 1
b.
x2 x4 x6 x3 x5
y = c1 (1 − + − + . . .) + c2 (x − + − . . .)
2! 4! 6! 3! 5!
= c1 cos x + c2 sin x
156 CHAPTER 4. SERIES SOLUTION OF DIFFERENTIAL EQUATIONS

c.
1 1
y = a0 (1 + x2 + x3 + x4 + . . .)
3 4
1 2 1 3 5 4
+ a1 (x + x + x + x + . . .)
2 2 24
This series solution must be capable of being transformed into the Taylor series for
c1 e2x + c2 e−x . (Why?) (Try to transform your solution thus!)

2. The general solution is

2 1 4
y = a0 (1 − 3x2 ) + a1 (x − x3 − x5 − x7 − . . .)
3 5 35
−1
and the particular solution is y = 2 + 23 x2 . (A finite series)

Detailed solutions to Question 2

Back to Chapter 4

Question 3

Solve each of the following as far as “possible”,

a. 4xy ′′ + 2y ′ + y = 0

b. y ′′ + xy ′ + y = 0

c. xy ′′ + y = 0.

d. Textbook exercise set 8.6. Problems 1,3,5,7,9,11,17 (8th ed),19,21,23

Answers
3 5
x x2 x3 1 x2 x2
a. y = A (1 − + − + . . .) + B (x 2 − + − . . .) which actually is represented
2! 4! 6! 3! 5!
√ √
by y = A cos x + B sin x
x2 x4 x6 x3 x5 x7
b. y = A (1 − + − + . . .) + B (x − + − + . . .)
2 2⋅4 2⋅4⋅6 3 3⋅5 3⋅5⋅7
(A power series solution, as expected. Why?)
x2 x3 x4
c. y = A (x − + − + . . .)
1!2! 2!3! 3!4!
(not the general solution).

d. See textbook.

Detailed solutions to Question 3

Back to Chapter 4
Chapter 5

Gamma function (special functions)

Chapter objectives

In this chapter you will learn:

• the definition and the simplest properties of the Gamma function

• evaluate Gamma function

5.1 Introduction
Recall. The concept of a function. The elementary functions studied in earlier subjects.
In Computer Aided Mathematics 1 with Applications (MTH101), the elementary mathe-
matical functions were studied and their main properties noted. In particular, algebraic
functions, the exponential and logarithmic and trigonometric functions were considered.
In addition to these, there are a great many special functions. Some of these are specialized
in the sense of having a very specific use. However, others have a wide use in engineering,
applied mathematics, statistics and the further general theory of mathematics.
This part of this subject is concerned almost entirely with just one of the “non-elementary”
special functions - the Gamma function.

5.2 Definition of the Gamma function

Video (by Dr. R. Wood): Gamma function
Recall. The definition of an improper integral, and its evaluation as a limit.
There are several equivalent definitions of the Gamma function. Here, the starting point,
the basic definition, is taken to be the Euler integral. The Gamma function (denoted by the
capital Greek letter, Γ, is defined by
∞
Γ(x) = ∫ e−t tx−1 dt for x > 0 (5.2.1)
0
157
158 CHAPTER 5. GAMMA FUNCTION (SPECIAL FUNCTIONS)

This is one of the more important of the non-elementary functions, with uses in applied
mathematics and in mathematical statistics.
It is also called the generalized factorial function, for reasons seen in the next section.
First, consider the definition as a mathematical statement. Note that the integral is improper
because of the infinite upper limit of integration.
Recall. In evaluating, an improper integral
∞ N
∫0 f (t)dt = lim ∫ f (t)dt
N →∞ 0

The integral is, however, definite. The symbol t is an integration parameter which does not
appear in the result of the definite evaluation from zero to infinity. That is to say,
∞
∫0 e−t tx−1 dt

should be seen as a function of x [which is exactly what equation (5.2.1) says; lefthand-side
equals right-hand-side and, by notation, LHS is a function of x and nothing else].
Equation (5.2.1), then, defines a function of the real variable x. (The Gamma function of a
complex variable z can be similarly defined by the same form of equation.) It is important
to note that the definition (5.2.1) is only for x > 0. At this stage, therefore, the Gamma
function is defined only for positive values although a later extension removes this (see
section on Extended definition).

The relation between the Gamma function and the ordinary factorial
function

Recall.

• The ordinary factorial function of elementary algebra,

n! = n(n − 1)(n − 2) . . . 3 ⋅ 2 ⋅ 1,

where n is a positive integer.

• The factorial property, n! = n(n − 1)!

• The technique of integration by parts. ∫ udv = uv − ∫ vdu

• L’Hôpital’s (second) rule: If

lim ∣f (x)∣ = lim ∣g(x)∣ = ∞,

x→c x→c

and if
f ′ (x)
lim = L,
x→c g ′ (x)

then
f (x)
lim =L
x→c g(x)

• Iterated use of L’Hôpital’s rule.

5.2. DEFINITION OF THE GAMMA FUNCTION 159

Integrate the Euler integral, equation (5.2.1) by parts:

∞ tx ∞ 1 ∞ −t x 1 ∞
∫0 e−t tx−1 dt = [e−t ] + ∫ e t dt = 0 + ∫ e−t tx dt for x > 0
x 0 x 0 x 0
For x > 0, the part ’cleared of integration’ obviously evaluates to zero at t = 0. For the limit
at t → ∞, apply L’Hôpital’s rule:

tx xtx−1
lim = lim = ... = 0
t→∞ et t→∞ et

The remaining integral is the Gamma function of x + 1. Hence the important recurrence
relation
Γ(x + 1)
Γ(x) =
x
or
xΓ(x) = Γ(x + 1). (5.2.2)
This is the functional relation of the Gamma function.

Example 1

Evaluate Γ(1) by substitution of x = 1 into the defining equation (5.2.1):

∞
∞
Γ(1) = ∫ e−t dt = [−e−t ]0 = 1
0

Now by use of the recurrence relation (5.2.2) we simply find:

Γ(2) = 1 ⋅ Γ(1) = 1
Γ(3) = 2 ⋅ Γ(2) = 2 ⋅ 1 = 2!
Γ(4) = 3 ⋅ Γ(3) = 3 ⋅ 2 ⋅ 1 = 3!
Γ(5) = 4 ⋅ Γ(4) = 4 ⋅ 3 ⋅ 2 ⋅ 1 = 4!
Γ(6) = 5 ⋅ Γ(5) = 5 ⋅ 4 ⋅ 3 ⋅ 2 ⋅ 1 = 5!

and so on. . .
In general, then, for integer n, we find that

Γ(n + 1) = n! n = 1, 2, 3, . . . (5.2.3)

Hence the connection between the Gamma function and the ordinary factorial function.
Note the change in the “argument” - the n plus one.

Extended definition. The graph of the Gamma function

The ideas of the previous section can be extended in two ways: firstly to define a general-
ized factorial; secondly (and of a more immediate importance) to extend the definition of
the Gamma function.
As to the first, equation (5.2.3) can be generalized to define x! as

x! = Γ(x + 1)
160 CHAPTER 5. GAMMA FUNCTION (SPECIAL FUNCTIONS)

This can be used as a definition for fractional or even negative numbers [a concept which
obviously does not fit in with the particular case (positive integer n) of the elementary
factorial function]. Thus, 12 ! = Γ ( 23 ) and (− 12 )! = Γ ( 12 ) .
What is of interest here is that now

0! = Γ(1) = 1

a result which in previous subjects may have been stated without any further justification.
Recall, for example, the evaluation of combinations and of certain discrete probabilities in
statistics.
Now, to extend the present (x > 0) definition of the Gamma function, consider the repeated
application of the recurrence:

Γ(x + 1) Γ(x + 2) Γ(x + k + 1)

Γ(x) = = = ... =
x x(x + 1) x(x + 1)⋯(x + k)

for k an integer, where, note, x ≠ 0, −1, −2, . . .

This relation is used to define the Gamma function for negative x. For k, the smallest
integer can be chosen such that x + k + 1 > 0 . Γ(x + k + 1) is known from the integral
definition (5.2.1).
So there are two parts to the complete definition. The Gamma function is defined:
⎧ ∞ −t x−1
⎪
⎪
⎪ ∫0 e t dt for x > 0
⎪
⎪
Γ(x) = ⎨ Γ(x + k + 1) x<0
⎪
⎪
⎪ k integer, for {
⎪
⎩ x(x + 1)⋯(x + k)
⎪ x ≠ −1, −2, . . .

The Gamma function remains undefined for zero and all negative integers.
Tables of values of the Gamma function have, of course, been compiled. These cover only
a unit interval of x, for example 1 ≤ x ≤ 2. Other values can be found simply by using the
recurrence formula. A short table follows.
5.2. DEFINITION OF THE GAMMA FUNCTION 161

Gamma Function Table for 1 ≤ x ≤ 2

x Γ(x) x Γ(x) x Γ(x) x Γ(x)
1.00 1.00000 1.25 0.90640 1.50 0.88623 1.75 0.91906
1.01 0.99433 1.26 0.90440 1.51 0.88659 1.76 0.92137
1.02 0.98884 1.27 0.90250 1.52 0.88704 1.77 0.92376
1.03 0.98355 1.28 0.90072 1.53 0.88757 1.78 0.92623
1.04 0.97844 1.29 0.89904 1.54 0.88818 1.79 0.92877
1.05 0.97350 1.30 0.89747 1.55 0.88887 1.80 0.93138
1.06 0.96874 1.31 0.89600 1.56 0.88964 1.81 0.93408
1.07 0.96415 1.32 0.89464 1.57 0.89049 1.82 0.93685
1.08 0.95973 1.33 0.89338 1.58 0.89142 1.83 0.93969
1.09 0.95546 1.34 0.89222 1.59 0.89243 1.84 0.94261
1.10 0.95135 1.35 0.89115 1.60 0.89352 1.85 0.94561
1.11 0.94740 1.36 0.89018 1.61 0.89468 1.86 0.94869
1.12 0.94359 1.37 0.88931 1.62 0.89592 1.87 0.95184
1.13 0.93993 1.38 0.88854 1.63 0.89724 1.88 0.95507
1.14 0.93642 1.39 0.88785 1.64 0.89864 1.89 0.95838
1.15 0.93304 1.40 0.88726 1.65 0.90012 1.90 0.96177
1.16 0.92980 1.41 0.88676 1.66 0.90167 1.91 0.96523
1.17 0.92670 1.42 0.88636 1.67 0.90330 1.92 0.96877
1.18 0.92373 1.43 0.88604 1.68 0.90500 1.93 0.97240
1.19 0.92089 1.44 0.88581 1.69 0.90678 1.94 0.97610
1.20 0.91817 1.45 0.88566 1.70 0.90864 1.95 0.97988
1.21 0.91558 1.46 0.88560 1.71 0.91057 1.96 0.98374
1.22 0.91311 1.47 0.88563 1.72 0.91258 1.97 0.98768
1.23 0.91075 1.48 0.88575 1.73 0.91467 1.98 0.99171
1.24 0.90852 1.49 0.88595 1.74 0.91683 1.99 0.99581
1.25 0.90640 1.50 0.88623 1.75 0.91906 2.00 1.00000

Example 2

Evaluate Γ(2.3) and Γ(−0.4)

a. Γ(2.3) = 1.3Γ(1.3) = 1.3 × 0.89747 (from the table) = 1.166711 ≈ 1.167.

Γ(0.6) Γ(1.6) 0.89352
b. Γ(−0.4) = −0.4 = (−0.4)⋅(+0.6) = −0.24 = −3.723.
162 CHAPTER 5. GAMMA FUNCTION (SPECIAL FUNCTIONS)

The graph of the Gamma Function

By careful consideration of its definition, properties and values, the graph of the Gamma
function can be drawn as follows:

Figure 5.1: Euler’s Gamma function Γ(x)

The general features of the graph of the Gamma function above should be noted. Especially
note the asymptotes at x = 0, −1, −2, −3, . . ., the value of the function increasing without
limit as these arguments are approached. For negative x, the function assumes alternately
positive and negative values in successive intervals of definition, due, of course, to the sign
of the denominator in the recurrence relation (8).
As x → −∞, the local maxima and minima tend to zero, with a flattening of the peaks. The
curve segments are asymmetric. With increasing positive x, Γ(x) increases very rapidly
indeed.
Note. The values of the Gamma function of half-integral arguments, Γ (k + 21 ) can be
calculated from

a. the recurrence relation, xΓ(x) = Γ(x + 1), and

√
b. Γ ( 12 ) = π

There is also a multiplication formula for Gamma functions, which often proves useful. We
state it here, without proof:
π
Γ(x) ⋅ Γ(1 − x) = , x ≠ 0, ±1, ±2, . . .
sin xπ
√
Using this, with x = 12 , confirms that Γ ( 12 ) = π. Another example would be:
1 2 π 2π
Γ( ) ⋅ Γ( ) = π = √ .
3 3 sin 3 3
5.2. DEFINITION OF THE GAMMA FUNCTION 163

Tutorial 5 Question 1
164 CHAPTER 5. GAMMA FUNCTION (SPECIAL FUNCTIONS)

5.3 Tutorial 5
Back to Chapter 5

Question 1

Evaluate the following Γ functions, using the table of Γ function values:

1. Γ(1.7)

2. Γ(3.3)

3. Γ(5.7)

4. Γ(0.6)

5. Γ(−1.4)

6. Γ(−2.3)

Answers

1. 0.90864

2. 2.6834

3. 72.528

4. 1.4892

5. 2.6593

6. –1.4471

Back to Chapter 5
Chapter 6

Laplace transforms

Chapter objectives

In this chapter you will learn to:

• use the integral formula to find the Laplace transform of a function;

• find functions given their Laplace transform;
• use tables to find functions and Laplace transforms; and
• solve initial value problems using Laplace transforms

6.1 Introduction
In this chapter, we will examine a very important concept, the Laplace Transform, which
is an example of a useful family called Integral Transforms. Such transforms are very sig-
nificant in a number of theoretical and applied areas of mathematics, numerical analysis
and computing. We will consider just a brief introduction to the topic here, and then ex-
amine how the Laplace Transform can be used to solve some of the linear D.E.s that we
examined in Chapter 2.

6.2 Definition of Laplace transform

Read textbook section 7.2 (Definition of the Laplace transform)

Video (by Dr. D. Demskoy): Laplace transform: definition and simplest properties
An integral transform is a relation of the form
β
F (s) = ∫ K(s, t)f (t)dt
α

where a given function f is transformed into another function F by means of an integral.

F is called the transform of f. Function K is called the kernel of the transformation.
165
166 CHAPTER 6. LAPLACE TRANSFORMS

Transforms are useful in making the solution of a problem, especially linear differential
equations, simpler.
We shall see that the Laplace transform reduces a differential equation to an algebraic
equation, and so simplifies its solution. If f (t) is defined for t ≥ 0, then the Laplace
transform of f is a function F defined by
∞
F (s) = ∫ e−st f (t)dt
0

(Here, e−st is the kernel.) The domain of the transform function F is the set of real values
of s for which the (improper) integral exists. Since e−st is a decreasing function of s, this
domain is s > s0 for some s0 determined by f .
The Laplace transform is usually represented by the operator L. Thus
∞
L{f (t)} = F (s) = ∫ e−st f (t)dt .
0

Example 1

Find L{f (t)}, where f (t) = 1, t ≥ 0.

∞
L{f (t)} = ∫ e−st dt
0
N
1
= lim [− e−st ]
N →∞ S 0
1 −sT 1
= lim [− e + ]
N →∞ s s
Now providing s > 0, e−sT → 0 as N → ∞.
1
∴ L{f (t)} = ,s > 0
s
In this example, s0 = 0.
Before stating the theorem which gives the conditions under which the Laplace transform
exists, we need to define what is meant by piecewise continuity.
A function f is piecewise continuous on an interval α ≤ t ≤ β if

i. f is continuous on each open sub-interval (ti−1 , ti );

ii. f approaches a finite limit as the end-points of each sub-interval are approached from
within the sub-interval.

An example of a piecewise continuous function is shown in the diagram below:

6.2. DEFINITION OF LAPLACE TRANSFORM 167

We now state the Theorem for the existence of Laplace transforms. Theorem 1 Suppose:

a. f is piecewise continuous on 0 ≤ t ≤ T for any positive T .

b. ∣f (t)∣ ≤ M es0 t when t ≥ 0. Then the Laplace transform exists for s > s0 .

We have already seen what is meant by piecewise continuous. To understand the second
condition, we look at some examples of functions satisfying and not satisfying (b).

i. f (t) = sin t
−1 ≤ sin t ≤ 1 i.e. ∣ sin t∣ ≤ 1 = 1e0⋅t
∴ M = 1, s0 = 0.
∴ f (t) = sin t satisfies condition b.

ii. f (t) = cosh t

cosh t = 12 (et + e−t )
∴ ∣ cosh t∣ < 1e1⋅t for t > 0
∴ M = 1, s0 = 1
∴ f (t) = cosh t satisfies condition b.

iii. f (t) = tn (n a non-negative integer)

It can be proved that ∣tn ∣ < n!et for t > 0
∴ M = n!, s0 = 1
∴ f (t) = tn satisfies condition b.
2
iv. f (t) = et
2
et is always greater than M es0 t , no matter how large M, s0 are.
2
∴ f (t) = et does not satisfy condition b.
168 CHAPTER 6. LAPLACE TRANSFORMS

Further examples of Laplace transforms

Example 2

Find L{f (t)} where f (t) = eat , t ≥ 0

∞
L (eat ) = ∫ e−st eat dt
0
T
= lim ∫ e−(s−a)t dt
T →∞ 0
1 −(s−a)t T
= lim [− e ]
T →∞ s−a 0
1 −(s−a)T 1
= lim [− e + ]
T →∞ s−a s−a
1
= , provided s > a.
s−a
Note: s > a means (s − a)T > 0, and hence e−(s−a)T → 0 as T → ∞

Example 3

Find L{f (t)} where f (t) = sin at t ≥ 0.

∞
L(sin at) =F (s) = ∫ e−st sin atdt
0
T
= lim ∫ e−st sin atdt
T →∞ 0
T T
1 1
= lim {[− e−st cos at] − ∫ (−se−st ) (− cos at) dt} (integrating by parts)
T →∞ a 0 0 a
1 s T [Note: e−sT → 0 as T → ∞
= − lim ∫ e−st cos atdt, s > 0
a a T →∞ 0 provided s > 0.]
T
1 s 1 s T
= − lim {[ e−st sin at] + ∫ e−st sin atdt} (integrating by parts again)
a a T →∞ a 0 a 0
1 s s ∞
= − ⋅ ∫ e−st sin atdt
a a a 0
i.e.
1 s2
F (s) = − 2 F (s)
a a
s2 1
F (s) (1 + 2 ) =
a a
2
a +s 2 1
∴ F (s) ( 2
)=
a a
1 a2
∴ F (s) = ⋅ 2
a s + a2
a
= 2 , s>0
s + a2
We do not, however, always have to determine Laplace transforms from the definition. One
property that helps is the linearity of the Laplace transform.
6.2. DEFINITION OF LAPLACE TRANSFORM 169

Linearity of L
The Laplace transform is a linear operator.
Theorem 2
Suppose L{f (t)} = F (s) for s > s1 , and L{g(t)} = G(s) for s > s2 . Then,

L {c1 f (t) + c2 g(t)} = c1 L{f (t)} + c2 L{g(t)}

= c1 F (s) + c2 G(s) for s > max (s1 , s2 ) .

Proof ∞
L {c1 f (t) + c2 g(t)} = ∫ e−st [c1 f (t) + c2 g(t)] dt
0
∞ ∞
= c1 ∫ e−st f (t)dt + c2 ∫ e−st g(t)dt
0 0
= c1 L{f (t)} + c2 L{g(t)}
= c1 F (s) + c2 G(s) for s > max (s1 , s2 ) .
We apply the linearity property in the following example.

Example 4

Find L{f (t)}, where f (t) = cosh at.

Now
1 1
cosh at = eat + e−at
2 2
1 at 1 −at
∴ L{cosh at} = L { e + e }
2 2
1 1
= L (eat ) + L (e−at ) (by the linearity property)
2 2
1 1 1 1
= + , s > a (from Example 2),
2s−a 2s+a
s+a+s−a
=
2 (s2 − a2 )
s
= 2
s − a2
The next example makes use of the definition of the Gamma function, covered previously
in Chapter 5 .

Example 5

Find L{f (t)}, where f (t) = ta (a positive)

∞
L (ta ) = F (s) = ∫ e−st ta dt
0

Let x = st ∴ dx = sdt. Then

∞ x a1
∴ F (s) = ∫ e−x ( ) dx
0 s s
1 ∞
= e−x xa dx
sa+1 ∫0
Γ(a + 1)
= , s > 0. (definition of Gamma function)
sa+1
170 CHAPTER 6. LAPLACE TRANSFORMS

Tutorial 6 Question 1

6.3 Inversion of transforms

Read textbook sections 7.3 and 7.4 (Properties of the Laplace transform and
Inverse Laplace transform)

In order to use Laplace transforms, it is usually necessary to be able to recover the function
f (t) when L{f (t)} = F (s) is known.
This is called inverting the transform, and can be regarded as the action on F (s) of an
operator L−1 which is the inverse of L.
A general method of inversion involves a complex integral which is beyond the scope of
this subject. We will therefore use known transforms as a basis for inversion. This is
illustrated in the following example.

Example 6

If

1
F (s) = , a ≠ b,
(s − a)(s − b)

find L−1 {F }.
We use partial fractions on

1
.
(s − a)(s − b)

Let

1 A B
= +
(s − a)(s − b) s − a s − b
A(s − b) + B(s − a)
=
(s − a)(s − b)
∴ 1 ≡ A(s − b) + B(s − a)
1
s = a gives A =
a−b
−1
s = b gives B =
a−b
6.4. TRANSFORMS OF DERIVATIVES 171

1 1 1 1
∴ = ( − )
(s − a)(s − b) a − b s − a s − b
1 1 1
∴ L−1 {F (s)} = L−1 { ( − )}
a−b s−a s−b
1 −1 1 1 −1 1
= L ( )− L ( )
a−b s−a a−b s−b
1 at 1 bt
= e − e (from Example 2)
a−b a−b
1
= (eat − ebt ) .
a−b

6.4 Transforms of derivatives

We intend to use Laplace transforms to solve differential equations. Hence we need to con-
sider what happens when we find L {f ′ (t)} . Roughly speaking, differentiation of functions
corresponds to the multiplication of transforms by s Hence the Laplace transform reduces
a calculus problem to an algebraic problem.
To be more precise, the theorem that describes the transform of derivatives is as follows.
Theorem 3
Suppose the following conditions are satisfied

• f (t) is continuous, and f ′ (t) is piecewise continuous on any interval 0 ≤ t ≤ A.

• there exist constants, M, s0 such that ∣f (t)∣ ≤ M es0 t for t ≥ 0.

Then L {f ′ (t)} exists for s > s0 and L {f ′ (t)} = sL{f (t)} − f (0).
We will prove this theorem for the case of f ′ (t) continuous. (The proof for f ′ (t) piecewise
continuous is similar, but requires the range of integration to be broken up into parts.)
Proof
∞
L {f ′ (t)} = ∫ e−st f ′ (t)dt (definition)
0
T
= lim ∫ e−st f ′ (t)dt
T →∞ 0
∞
= lim [e−st f (t)]0 + s ∫ e−st f (t)dt (using integration by parts)
T
T →∞ 0
= lim [e−sT f (T ) − f (0)] + sL{f (t)}
T →∞
= sL{f (t)} − f (0) for s > s0

as required.
We can go on to find the transforms of higher derivatives.

L {f ′′ (t)} = sL {f ′ (t)} − f ′ (0),

= s[sL{f (t)} − f (0)] − f ′ (0) (6.4.1)
∴ L {f ′′ (t)} = s2 L{f (t)} − sf (0) − f ′ (0)
172 CHAPTER 6. LAPLACE TRANSFORMS

and
L {f ′′′ (t)} = sL {f ′′ (t)} − f ′′ (0)
= s [s2 L{f (t)} − sf (0) − f ′ (0)] − f ′′ (0)
= s3 L{f (t)} − s2 f (0) − sf ′ (0) − f ′′ (0).
Using mathematical induction, we can show, that, for n a positive integer,

L {f (n) (t)} = sn L{f (t)} − sn−1 f (0) . . . − sf (n−2) (0) − f (n−1) (0)

Recall: f (n) (t) means the nth derivative of f (t).

Some applications of the above results are as follows.

Example 7

Find L{f (t)}, where f (t) = t2 .

We could use ∞
L{f (t)} = ∫ e−st t2 dt
0
and integrate by parts.
Instead, it is more convenient to use transform of the derivative property:

L {f ′′ (t)} = s2 L{f (t)} − sf (0) − f ′ (0)

s2 L{f (t)} = L {f ′′ (t)} + sf (0) + f ′ (0)
L {f ′′ (t)} + sf (0) + f ′ (0)
L{f (t)} =
s2
L(2) + s × 02 + 2 × 0
=
s2
2
+0+0 1
= s 2 since L(1) = .
s s
Thus
2
L{f (t)} = .
s3

Example 8

Find L{f (t)} where f (t) = sin t.

From (6.4.1),

L {f ′′ (t)} + sf (0) + f ′ (0)

L{f (t)} =
s2
−L{f (t)} + 0 + 1
= since f ′′ (t) = − sin t = −f (t)
s2
(s2 + 1) L{f (t)} = 1
1
L{f (t)} = 2 .
s +1
6.4. TRANSFORMS OF DERIVATIVES 173

Example 9

Find L{f (t)} where f (t) = cos nt.

Here
f ′ (t) = −n sin nt, f ′′ (t) = −n2 cos nt

therefore From (6.4.1)

L {f ′′ (t)} + sf (0) + f ′ (0)

L{f (t)} =
s2
2
−n L{f (t)} + s
=
s2
∴ (s2 + n2 ) L{f (t)} = s
s
∴ L{f (t)} = 2
s + n2
Tables of Laplace transforms
Using different techniques, we are now building up a list of Laplace transforms of common
functions. Such lists are published in tables, and we make use of these tables in solving
problems, particularly differential equations. Rather than perform Laplace transforms from
first principles every time, we make use of the tables wherever possible.
A restricted table is given below, and this table will appear on the examination paper. You
are free to quote results from the table in your solutions.

Some functions f (t) and their Laplace transforms L(f )

Table 1

f (t) L(f ) f (t) L(f )

1
1 1 1/s 6 eat
s−a
s
2 t 1/s2 7 cos ωt
s + ω2
2

ω
3 t2 2!/s3 8 sin ωt
s + ω2
2

n! s
4 tn n+1
9 cosh at 2
s s − a2
Γ(a + 1) a
5 ta 10 sinh at
sa+1 s 2 − a2

Here n = 0, 1, 2, . . . and α is any positive number.

Tutorial 6 Question 2
174 CHAPTER 6. LAPLACE TRANSFORMS

6.5 Solving initial value differential equations

Read textbook section 7.5 (Solving initial value problems)

Video (by Dr. R. Wood): Laplace Transform and solution of ODEs

We are now in a position to see how Laplace transforms can help us to solve differential
equations, in particular, differential equations with initial conditions.
We begin with an example, which could be solved by finding the auxiliary equation (as in
Chapter 2 - Second order L.D.E.). We will, however, show how it can be solved by taking
Laplace transforms.

Example 10

Solve the differential equation

y ′′ − y ′ − 2y = 0, where y(0) = 1, y ′ (0) = 0

We assume the solution y = y(t), and its first two derivatives satisfy the conditions for
applying the Laplace transform. Then

L (y ′′ − y ′ − 2y) = L(0)
∴ L (y ′′ ) − L (y ′ ) − 2L(y) = 0 (linearity property)
∴ s2 L(y) − sy(0) − y ′ (0) − sL(y) + y(0) − 2L(y) = 0 (using transform of derivatives)
∴ (s2 − s − 2) L(y) + (1 − s)y(0) − y ′ (0) = 0

Recall that we used the notation L{f (t)} = F (s).

Here, we use L(y) = Y (s).

∴ (s2 − s − 2) Y (s) + (1 − s)y(0) − y ′ (0) = 0

But y(0) = 1, y ′ (0) = 0.

∴ (s2 − s − 2) Y (s) = s − 1

s−1
∴ Y (s) =
s2 −s−2

We use partial fractions for

s−1
.
s2 − s − 2
6.5. SOLVING INITIAL VALUE DIFFERENTIAL EQUATIONS 175

Let
s−1 A B A(s + 1) + B(s − 2)
= + =
s2 − s − 2 s − 2 s + 1 (s − 2)(s + 1)
∴ s − 1 ≡ A(s + 1) + B(s − 2)
1
s = 2 gives A =
3
2
s = −1 gives B = .
3
1 2
∴ Y (s) = +
3(s − 2) 3(s + 1)
1 2 1
∴ y(t) = e2t + e−t since L (eat ) =
3 3 s−a
(Table 1, No. 6)
We now look at a second order non-homogeneous differential equation.

Example 11

Solve the differential equation

y ′′ + y = sin 2t, where y(0) = 2, y ′ (0) = 1

With the same assumptions as before,

L (y ′′ ) + L(y) = L(sin 2t)

2
∴ s2 Y (s) − sy(0) − y ′ (0) + Y (s) = 2
s +4
2
∴ s2 Y (s) − 2s − 1 + Y (s) = 2
s +4
2
∴ (s2 + 1) Y (s) = 2 + 2s + 1
s +4
2s3 + s2 + 8s + 6
∴ Y (s) = 2
(s + 4) (s2 + 1)
5 2
2s 3 3
= 2 + −
s + 1 s2 + 1 s2 + 4
using partial fractions
5 1
∴ y(t) = 2 cos t + sin t − sin 2t (using Table 1, Nos. 7, 8 ).
3 3

Example 11

Solve the differential equation

y (4) − y = 0,
where y(0) = 0, y ′ (0) = 1, y ′′ (0) = 0, y ′′′ (0) = 0.
We could solve this homogeneous equation by writing it as (D4 − 1) y = 0.
Auxiliary equation is r4 − 1 = 0.

∴ r = ±1, ±i
∴ y(t) = Aet + Be−t + C cos t + D sin t
176 CHAPTER 6. LAPLACE TRANSFORMS

We then use the initial conditions to find the values of A, B, C and D. Alternatively, we
can take Laplace transforms.
L (y (4) − y) = 0
∴ L (y (4) ) − L(y) = 0
∴ s4 Y (s) − s3 y(0) − s2 y ′ (0) − sy ′′ (0) − y ′′′ (0) − Y (s) = 0
∴ s4 Y (s) − s2 − Y (s) = 0
∴ (s4 − 1) Y (s) = s2
s2 s2
∴ Y (s) = 4 = 2
s − 1 (s + 1) (s2 − 1)
1 1
2 2
= +
s2 + 1 s2 − 1
and using partial fractions we obtain
1
y(t) = (sin t + sinh t) (Table 1, Nos. 8, 10).
2
From these examples, we see the advantages of using the Laplace transform in initial value
problems. There is no need to determine a general solution, and subsequently find values
for arbitrary constants. The given initial conditions are included automatically when the
transform is taken.

Tutorial 6 Question 3

6.6 Transforms of integrals

We know that differentiation and integration are inverse processes, and we have already
seen that roughly speaking, differentiation of a function corresponds to the multiplication
of its transform by s. As we might expect, then, integration of a function corresponds to
division of its transform by s.
The precise statement of the relevant theorem is as follows.
Theorem 4
If f (t) is piecewise continuous and satisfies ∣f (t)∣ ≤ M es0 t t ≥ 0, then
t 1
L {∫ f (τ )dτ } = L{f (t)}, s > 0, s > s0
0 s
From the last formula, we can form a companion statement for L−1 .
1 t
L{f (t)} = L {∫ f (τ )dτ }
s 0
1 t
i.e. F (s) = L {∫ f (τ )dτ }
s 0
1 t
∴ L−1 { F (s)} = ∫ f (τ )dτ.
s 0
6.7. SHIFTING THEOREMS 177

Example 13

Find f (t) if
3
L{f (t)} = .
s2 +s

Now
3 1 3
= ( )
s2 + s s s + 1
and
3
L−1 ( ) = 3e−t
s+1
therefore
1 3 t
L−1 { ( )} = ∫ 3e−τ dτ
s s+1 0
= −3 [e−τ ]0
t

= −3e−t + 3.

Tutorial 6 Question 4

6.7 Shifting theorems

Read textbook section 7.3 (Properties of the Laplace transform)

To obtain a better idea of the power of the Laplace transform, we look at properties which
concern shifting on the s-axis. We can also have shifting on the t-axis, but we leave treat-
ment of that to the subject MTH418 (Topics in Calculus).
Theorem 5 (s-shifting)
If
L{f (t)} = F (s) where s > s0
then
L {eat f (t)} = F (s − a)
and
L−1 {F (s − a)} = eat f (t) (inverse).

Proof ∞
F (s) = ∫ e−st f (t)dt (definition)
0
∞
∴ F (s − a) = ∫ e−(s−a)t f (t)dt
0
∞
=∫ e−st eat f (t)dt
0
= L {eat f (t)}
178 CHAPTER 6. LAPLACE TRANSFORMS

Example 14

Find L (eat tn ) .
We know that
n!
L (tn ) =
sn+1
n!
∴ L (eat tn ) = F (s − a) =
(s − a)n+1
Example 15

Find L (eat sin ωt)

We know that
ω
L(sin ωt) =
s2
+ ω2
ω
∴ L (eat sin ωt) =
(s − a)2 + ω 2
Example 16

Solve the differential equation

y ′′ + 2y ′ + 2y = 0
with the initial condition y(0) = 0, y ′ (0) = 1.
We have
L (y ′′ + 2y ′ + 2y) = 0
∴ L (y ′′ ) + 2L (y ′ ) + 2L(y) = 0
∴ s2 Y (s) − sy(0) − y ′ (0) + 2sY (s) − 2y(0) + 2Y (s) = 0
∴ s2 Y (s) − 1 + 2sY (s) + 2Y (s) = 0
∴ (s2 + 2s + 2) Y (s) = 1
1 1
∴ Y (s) = 2 = (on completing the square)
s + 2s + 2 (s + 1)2 + 1

1
∴ y(t) = L−1 { }
(s + 1)2 + 1
1
= L−1 {F [s − (−1)]} where F (s) =
s2 +1
= e−t sin t

Tutorial 6 Question 5
6.8. TUTORIAL 6 179

6.8 Tutorial 6
Back to Chapter 6

Question 1

1. Sketch the graphs of the following functions, and determine whether f is continuous,
piecewise continuous, or neither on the interval 0 ≤ t ≤ 3 :
⎧
⎪
⎪ t2 , 0 ≤ t ≤ 1
⎪
⎪
i. f (t) = ⎨2 + t,1 < t ≤ 2
⎪
⎪
⎪
⎩ 6 − t, 2 < t ≤ 3
⎪
⎧
⎪
⎪ t2 , 0 ≤ t ≤ 1
⎪
ii. f (t) = ⎨ 1 , 1 < t ≤ 2
⎪
⎪
⎩ 3−t , 2<t≤3
⎪

2. Use the definition ∞

L{f (t)} = ∫ e−st f (t)dt
0
to find the Laplace transform of:

i. 3t + 4
ii. eat+b (a, b constants)
iii. cos(at) (a constant)

3. Textbook exercise set 7.2. Problems 1,3,5,9,11,21,23,25,27

180 CHAPTER 6. LAPLACE TRANSFORMS

Answers

Piecewise continuous

1. i.

Continuous

ii.
3 4
2. i. +
s2 s
eb
ii.
s−a
s
iii. 2
s + a2
Detailed solutions to Question 1.2

Back to Chapter 6

Question 2

1. Use partial fractions to find f (t) if F (s) = L{f } is:

1
i.
s(s + 1)
6.8. TUTORIAL 6 181

1
ii.
(s − 4)(s − 1)
3s
iii. 2
s + 2s − 8
s2 − 6s + 4
iv. 3
s − 3s2 + 2s
2. Use the table of Laplace transforms to find the transforms of the following functions:

i. t2 + at + b (a, b constants )
ii. cos(ωt + θ) (ω, θ constants)
iii. cos2 ωt (ω constant )
iv. 3t5
v. cosh 4t

3. Use L (f ′ ) = sL(f ) − f (0) to find the transform of sin2 ωt.

4. Use L (f ′′ ) = s2 L(f ) − sf (0) − f ′ (0) to find the transform of t cos t.

5. Textbook exercise set 7.6. Problem 51 (8th ed).

Answers

1. i. 1 − e−t
1
ii. (e4t − et )
3
iii. 2e−4t + e2t
iv. 2 + et − 2e2t .
2 a b
2. i. 3
+ 2+
s s s
s cos θ − ω sin θ
ii.
s2 + ω 2
1 s
iii. Use: cos2 ωt = 21 + 12 cos 2ωt. Ans: + 2
2s 2s + 8ω 2
3 ⋅ 5!
iv.
s6
s
v. 2 .
s − 16
2ω 2
3.
s (s2 + 4ω 2 )
s2 − 1
4. 2
(s2 + 1)
5. see textbook

Detailed solutions to Question 2

Back to Chapter 6
182 CHAPTER 6. LAPLACE TRANSFORMS

Question 3

Using Laplace transforms, solve the following differential equations:

1. 4y ′′ + π 2 y = 0, y(0) = 0, y ′ (0) = 1.

2. y ′′ + 5y ′ + 6y = 0, y(0) = 0, y ′ (0) = 1.

3. y ′′ + 25y = t, y(0) = 1, y ′ (0) = 0.04.

4. Textbook exercise set 7.5. Problems 5, 7, 11 (see Example 3), 13

Answers

2
1. y = π sin π2 t

2. y = e−2t − e−3t
1
3. y = cos 5t + 25 t.

4. see textbook

Detailed solutions to Question 3

Back to Chapter 6

Question 4

Use
1 t
L−1 { F (s)} = ∫ f (τ )dτ
s 0

to find f (t) if L(f ) equals:

4
1.
s3 − 4s
1
2. here a is a constant
s2 + as
8
3. Hint: Apply the formula twice.
s4 − 4s2

Answers

1. cosh 2t − 1

2. 1
a (1 − e−at )

3. sinh 2t − 2t.

Detailed solutions to Question 4

Back to Chapter 6
6.8. TUTORIAL 6 183

Question 5

1. Using the s-shifting theorem, find the Laplace transforms of:

i. t2 e−2t
ii. e−t cos t
iii. e−2t sin 3t
iv. 2t3 e−t/2

2. Use the inverse form of the s -shifting theorem to find f (t) if L(f ) equals:
1
i. 3
(s + 12 )
1
ii.
s2 + 2s + 5
6
iii. 2
s − 4s − 5
3. Using Laplace transforms, and the s -shifting theorem, solve the differential equa-
tions:

i. y ′′ − 4y ′ + 5y = 0, y(0) = 1, y ′ (0) = 2
ii. 9y ′′ − 6y ′ + y = 0, y(0) = 3, y ′ (0) = 1
iii. y ′′ + 2y ′ + 5y = 9 cosh 2t + 4 sinh 2t, y(0) = 1, y ′ (0) = 2

4. Textbook exercise set 7.3. Problems 1, 3, 5;

5. Textbook exercise set 7.4. Problems 1, 3, 5, 7, 9;

Answers

2
1. i.
(s + 2)3
s+1
ii. 2
s + 2s + 2
3
iii.
(s + 2)2 + 9
12
iv. 4
(s + 12 )
1 2 −t/2
2. i. 2t e
1 −t
ii. 2 e sin 2t
iii. 2e2t sinh 3t

3. i. y = e2t cos t
ii. y = 3et/3
iii. y = e−t sin 2t + cosh 2t.

4. see textbook
184 CHAPTER 6. LAPLACE TRANSFORMS

5. see textbook

Detailed solutions to Question 5

Back to Chapter 6
Chapter 7

Solutions to selected tutorial questions

7.1 Tutorial 1

Question 2

Part 1.i

The equation is
3xdx + 4ydy = 0.
We denote
M = 3x, N = 4y.
Calculating the derivatives, we obtain
∂y M = 0, ∂x N = 0.
We see that they are equal, so the equation is exact. Further, we calculate

F = ∫ M dx = ∫ 3xdx = 23 x2 + ϕ(y),

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y
we obtain
ϕ′ = 4y
which implies that
ϕ = 2y 2 + C.
Therefore the general solution is 23 x2 +2y 2 +C = 0 or after multiplication by 2 and redefining
the arbitrary constant:
3x2 + 4y 2 = C.
The solution can be written explicitly:
√
y = ± 12 C − 3x2 .

Back to Tutorial 1 Question 2

185
186 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Part 1.ii

The equation can be rewritten as

(y − x) dx + (x + y) dy = 0.

So, we denote
M = y − x, N = x + y.
Then, calculating the derivatives, we obtain

∂M ∂N
= 1, = 1.
∂y ∂x

We see that they are equal, so the equation is exact. Further, we calculate

F = ∫ M dx = ∫ (y − x) dx = − 21 x2 + yx + ϕ(y),

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y

we obtain
x + ϕ′ = x + y
which implies that
ϕ = 21 y 2 + C.
Therefore the general solution is − 21 x2 + yx + 12 y 2 + C = 0 which can be re-written as

x2 − 2xy − y 2 = C

after multiplication by 2 and re-defining the arbitrary constant.

Back to Tutorial 1 Question 2

Part 1.iii

The equation can be re-written as

−xdx + (x + y) dy = 0.

So, we denote
M = −x, N = x + y.
Calculating the derivatives, we obtain

∂M ∂N
= 0, = 1.
∂y ∂x

We see that they are not equal, so the equation is not exact!
Back to Tutorial 1 Question 2
7.1. TUTORIAL 1 187

Part 1.iv

The equation can be re-written slightly:

(ye−x − sin x) dx + (−e−x − 2y) dy = 0.

Then, we denote
M = ye−x − sin x, N = −e−x − 2y.
Calculating the derivatives, we obtain

∂M ∂N
= e−x , = e−x .
∂y ∂x

We see that the derivatives are equal, so the equation is exact. Further, we calculate

F = ∫ M dx = ∫ (ye−x − sin x) dx = −ye−x + cos x + ϕ(y),

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y

we obtain
−e−x + ϕ′ = −e−x − 2y
which implies that
ϕ = −y 2 + C.
Therefore the general solution is

−ye−x + cos x − y 2 + C = 0

or
ye−x − cos x + y 2 = C

Back to Tutorial 1 Question 2

Part 1.v

The equation is
(x2 + x) dy + (2yx + 1 + 2 cos x) dx = 0.
So, we denote
M = 2yx + 1 + 2 cos x, N = x2 + x.
Calculating the derivatives, we obtain

∂M ∂N
= 2x, = 2x + 1.
∂y ∂x

We see that they are not equal, so the equation is not exact!
Back to Tutorial 1 Question 2
188 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Part 2.i

The equation is
2xydx + (1 + x2 ) dy = 0.
We denote:
M = 2yx, N = x2 + 1.
Calculating the derivatives, we obtain
∂M ∂N
= 2x, = 2x.
∂y ∂x
Hence the derivatives are equal and the equation is exact. Further, we calculate

F = ∫ M dx = ∫ 2yxdx = yx2 + ϕ(y),

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y
we obtain
x2 + ϕ′ = x2 + 1
which implies that
ϕ = y + C.
Therefore the general solution is

yx2 + y + C = 0.

The solution can be written explicitly:

C
y=− .
x2+1
Substituting x = 1, y = 3 into this solution we find C = 6.
Back to Tutorial 1 Question 2

Part 2.ii

The equation is
(2yx + 2y 2 e2x ) dx + (x2 + 2ye2x ) dy = 0.
So, we denote
M = 2yx + 2y 2 e2x , N = x2 + 2ye2x .
Calculating the derivatives, we obtain
∂M ∂N
= 2x + 4ye2x , = 2x + 4ye2x .
∂y ∂x
We see that they are equal, so the equation is exact. Further, we calculate

F = ∫ M dx = ∫ (2yx + 2y 2 e2x ) dx = yx2 + y 2 e2x + ϕ(y),

7.1. TUTORIAL 1 189

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y
we obtain
x2 + 2ye2x + ϕ′ = x2 + 2ye2x
which implies that
ϕ = C.
Therefore the general solution is

yx2 + y 2 e2x + C = 0.

Substituting x = 0, y = 1 into this solution we find C = −1.

Back to Tutorial 1 Question 2

Part 3

The equation is
ydx + (4x − y 2 ) dy = 0
So, we find
M = y, N = 4x − y 2 .
We note that
∂x N − ∂y M 3
Q= =
M y
is a function of y only, so the integrating factor can be calculated:
3dy
IF = exp (∫ Qdy) = exp (∫ ) = y3.
y
Multiplying the initial equation by the integrating factor we calculate an exact equation!
The equation takes the form

y 4 dx + (4y 3 x − y 5 ) dy = 0.

Again, we denote
M = y 4 , N = 4y 3 x − y 5 .
Calculating the derivatives, we obtain
∂M ∂N
= 4y 3 , = 4y 3 .
∂y ∂x
Now we see that the equation is exact. Further, we calculate

F = ∫ M dx = ∫ y 4 dx = y 4 x + ϕ(y),

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y
190 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

we obtain
4y 3 x + ϕ′ = 4y 3 x − y 5
which implies that
ϕ = − 61 y 6 + C.
Therefore the general solution is y 4 x − 61 y 6 + C = 0 which is equivalent to
6y 4 x − y 6 = C.

Back to Tutorial 1 Question 2

Question 3

Part 3.i

The equation can be rewritten as

ydy 3xdx
= − .
y2 + 1 x2 + 2
Now the variables x and y are separated. Hence we can integrate:
ydy 3xdx
∫ y 2 + 1 = ∫ − x2 + 2 .
Calculating the latter we get
1
2 ln (y 2 + 1) = − 23 ln (x2 + 2) + C.
Then, after multiplying by 2 we get
ln (y 2 + 1) = −3 ln (x2 + 2) + 2C
or
ln (y 2 + 1) + 3 ln (x2 + 2) = 2C.
Using the properties of ln function we get
ln ((y 2 + 1)(x2 + 2)3 ) = 2C
and finally
(y 2 + 1)(x2 + 2)3 = e2C
or
(y 2 + 1)(x2 + 2)3 = A
Here A = e2C is an arbitrary constant.
Below is a detailed calculation of the integral on the right:
3x
∫ − x2 + 2 dx = Constant multiple rule: ∫ cf (x)dx = c ∫ f (x)dx

x Substitution: u = x2 + 2
−3 ∫ 2 dx = du = 2xdx
x +2
du
−3 ∫ 12 = Constant multiple rule: ∫ cf (x)dx = c ∫ f (x)dx
u
du
− 32 ∫ =
u
− 23 ln u + C
= − 23 ln (x2 + 2) + C
7.1. TUTORIAL 1 191

Back to Tutorial 1 Question 3

Part 3.ii

This is a separable equation. Indeed, we can bring it to the form

dy
= −2e3x dx
y
hence
dy
∫ = −2e3x dx.
y ∫
Integrating the latter we obtain
ln y = − 23 e3x + C
or basically the same:
3 ln y + 2e3x = C,
where C is a constant of integration. If required we can state the solution in explicit form:

y(x) = exp (− 23 e3x + C) .

Back to Tutorial 1 Question 3

Part 3.iii

This is a separable equation. Indeed, we can cast it to the form

ydy
= 1 xdx
y2 + 1 4
hence
ydy 1
∫ y2 + 1 = ∫ 4 xdx.

Integrating the latter we find

1
2 ln (y 2 + 1) = 18 x2 + C,

where C is a constant of integration. If we multiply the solution by 8 we rewrite as follows

4 ln (y 2 + 1) − x2 = C,

After the substitution of x = 1, y = 0, we have

C = −1

hence
4 ln (y 2 + 1) − x2 = −1
or
x2 − 4 ln (y 2 + 1) = 1.

Back to Tutorial 1 Question 3

192 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Question 3.iv

This is a separable equation. Indeed, it can be written as

dy dx
=
y2 +1 x
hence
dy dx
∫ y2 + 1 = ∫ x .
Integrating the latter we obtain

arc tan (y) = ln ∣x∣ + C,

where C is a constant of integration. Further, solving for y we get the explicit solution

y(x) = tan (ln ∣x∣ + C) .

Back to Tutorial 1 Question 3

Part 3.v

This is a separable equation. Indeed, we can cast it to the form

(y 2 + 1) dy dx
=− 3
y x
hence
(y 2 + 1) dy dx
∫ =∫ − 3.
y x
Integrating the latter we get
1 2 1 1
2y + ln ∣y∣ =
+C 2
x2
or equivalent solution (after multiplication by -2 and renaming C)
1
− y 2 − 2 ln ∣y∣ = C
x2
where C is a constant of integration.
Back to Tutorial 1 Question 3

Part 3.vi

Textbook exercise set 2.2. Problem 11

This is a separable equation. Indeed, it can be written in the form

dy dx
=
sec y 1 + x2
2

hence
dy dx
∫ sec2 y = ∫ 1 + x2 .
7.1. TUTORIAL 1 193

Integrating the latter we obtain

1
4 sin 2y + 12 y = arc tan x + C,
where C is a constant of integration.
Here is a detailed calculation of the integral on the left

dy
Use formula: 1
= cos2 y
∫ sec2 y = sec2 y

2 cos y 2 = 1
cos (2y) + 1
∫ cos ydy = Use formula: 2 2

1 1
∫ ( 2 cos (2y) + 2 )dy =
1 1
∫ 2 cos (2y) dy + ∫ 2 dy

1
=∫ 2 cos (2y) dy + 21 y =
1 1 Substitution: u = 2y
2 ∫ cos (2y) dy + 2 y = du = 2dy
1
4 sin (2y) + 12 y + C
Back to Tutorial 1 Question 3

Textbook exercise set 2.2. Problem 15

Separable equation:
dy
= − exp (cos x) sin xdx
y2
hence
dy
∫ y 2 = ∫ − exp (cos x) sin xdx.
Integrating the latter (using substitution u = cos x) we get
1
− = exp (cos x) + C,
y
where C is a constant of integration. Further, solving for y we get the explicit solution
1 1
y(x) = − = ,
exp (cos x) + C A − exp (cos x)
where A = −C.
Back to Tutorial 1 Question 3

Textbook exercise set 2.2. Problem 19

This is a separable equation. It can be written as follows

dy
√ = 2 cos xdx
y+1
194 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

hence
dy
∫ √ = ∫ 2 cos xdx.
y+1
Integrating the latter we have
√
2 y + 1 = 2 sin x + C,

where C is a constant of integration. Further, solving for y we find the explicit solution
C 2
y(x) = (sin x + ) − 1.
2

After the substitution of x = π, y = 0, we get

C 2 C2
0 = (sin π + ) −1= −1
2 4
hence C = 2 and
y (x) = (sin x + 1)2 − 1 = sin2 x + 2 sin x.
Back to Tutorial 1 Question 3

Textbook exercise set 2.2. Problem 25

This is a separable equation:

dy
= x2 dx
y+1
hence
dy 2
∫ y + 1 = ∫ x dx.
Integrating the latter we have
ln (y + 1) = 31 x3 + C,
where C is a constant of integration. Further, solving for y we get the explicit solution
1 3 +C
y(x) = e 3 x − 1.

On substitution of x = 0, y = 3, we obtain eC − 1 = 0 hence C = 0 and

1 3
y (x) = 4e 3 x − 1.

Back to Tutorial 1 Question 3

Question 4

Textbook exercise set 2.5. Problem 7

The equation is
2xydx + (y 2 − 3x2 ) dy = 0
7.1. TUTORIAL 1 195

Then, we denote
M = 2xy, N = y 2 − 3x2 .

Note that
∂x N − ∂y M 4
Q= =−
M y
is a function of y only, so the integrating factor can be calculated:

4dy 1
IF = exp (∫ Qdy) = exp (∫ − ) = 4.
y y

Multiplying the initial equation by the integrating factor we get an exact equation! The
equation is
2xdx (−y 2 + 3x2 ) dy
− =0
y3 y4
So, we have
2x −y 2 + 3x2
M= , N = − .
y3 y4
Calculating the derivatives, we obtain

∂M 6x ∂N 6x
=− 4, =− 4.
∂y y ∂x y

We see that they are equal, so the equation is exact. Further, we calculate

2xdx x2
F = ∫ M dx = ∫ = 3 + ϕ(y),
y3 y

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y

we obtain
3x2 ′ −y 2 + 3x2
− + ϕ = −
y4 y4
which implies that
1
ϕ = − + C.
y
Therefore the general solution is

x2 1
− + C = 0.
y3 y

Back to Tutorial 1 Question 4

196 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Textbook exercise set 2.5. Problem 9

The equation1 is
(x4 − x + y) dx − xdy = 0
So, we get
M = x4 − x + y, N = −x.
We note that
∂y M − ∂x N 2
P= =−
N x
is a function of x only, so the integrating factor is
2 1
IF = exp (∫ P dx) = exp (∫ − dx) = 2 .
x x
Multiplying the initial equation by the integrating factor we obtain an exact equation! The
equation is
(x4 − x + y) dx dy
− =0
x2 x
So, we have
x4 − x + y 1
M= 2
, N =− .
x x
Calculating the derivatives, we obtain
∂M 1 ∂N 1
= 2, = 2.
∂y x ∂x x
We see that they are equal, so the equation is exact. Further, we calculate
(x4 − x + y) dx 1 3 y
F = ∫ M dx = ∫ = 3 x − − ln ∣x∣ + ϕ(y),
x2 x
where ϕ(y) is a function of y only. Substituting F into
∂F
= N,
∂y
we obtain
1 1
− + ϕ′ = −
x x
which implies that
ϕ = C.
Therefore the general solution is
1 3 y
3x − − ln ∣x∣ + C = 0
x
or after multiplication by x it takes the form
1 4
3x − y − x ln ∣x∣ + Cx = 0.

Back to Tutorial 1 Question 4

1
It is slightly different in 9th ed
7.1. TUTORIAL 1 197

Textbook exercise set 2.5. Problem 11

The equation is
(y 2 + 2xy)dx − x2 dy = 0
hence
M = y 2 + 2xy, N = −x2 .
We note that
∂x N − ∂y M 2
Q= =−
M y
is a function of y only, so the integrating factor can be calculated:

2dy 1
IF = exp (∫ Qdy) = exp (∫ − ) = 2.
y y

Multiplying the equation by the integrating factor we obtain the equation

(y + 2x) dx x2 dy
− 2 =0
y y

Then, we have
y + 2x x2
M= , N = − 2.
y y
Calculating the derivatives, we obtain

∂M 2x ∂N 2x
=− 2, =− 2.
∂y y ∂x y

We see that they are equal, so the equation is exact. Further, we calculate

(y + 2x) dx x2
F = ∫ M dx = ∫ =x+ + ϕ(y),
y y

where ϕ(y) is a function of y only. Substituting F into

∂F
= N,
∂y

we obtain
x2 ′ x2
− + ϕ = −
y2 y2
which implies that
ϕ = C.
Therefore the general solution is

x2
x+ + C = 0.
y

Back to Tutorial 1 Question 4

198 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Question 5

Part 5.i

What we’re given is a linear first order ODE whose standard form is
5y 7
y′ + = 2x − .
x x
The integrating factor for this equation is

IF = exp (∫ pdx)
5dx
= exp (∫ )
x
= x5 .
After multiplication by the integrating factor we can re-write the equation as
d
(x5 y) = (2x2 − 7) x4 ,
dx
and therefore the solution is
1 1 C
y = 5 ∫ (2x2 − 7) x4 dx = 5 ( 27 x7 − 75 x5 + C) = 27 x2 − 75 + 5 ,
x x x
where C is a constant of integration.
Back to Tutorial 1 Question 5

Part 5.ii

What we’re given is a linear first order ODE whose standard form is
y ′ + 2xy = 2x.
The integrating factor for this equation is

IF = exp (∫ pdx)

= exp (∫ 2xdx)
2
= ex .
After multiplication by the integrating factor we can re-write the equation as
d x2 2
(e y) = 2xex ,
dx
and therefore the solution is
y = e−x ∫ 2xex dx = e−x (ex + C) = 1 + Ce−x ,
2 2 2 2 2

where C is a constant of integration.

Back to Tutorial 1 Question 5
7.1. TUTORIAL 1 199

Part 5.iii

This is a linear first order ODE whose standard form is

3y
y′ + = x.
x
The integrating factor for this equation is

IF = exp (∫ pdx)
3dx
= exp (∫ )
x
= x3 .
After multiplication by the integrating factor we can bring the equation to the form
d
(x3 y) = x4 ,
dx
and therefore the solution is
1 1 C
y = 3 ∫ x4 dx = 3 ( 15 x5 + C) = 15 x2 + 3 ,
x x x
where C is a constant of integration.
Back to Tutorial 1 Question 5

Part 5.iv

What we have is a linear first order ODE whose standard form is

y ′ + 3y = e−2x .
The integrating factor for this equation is

IF = exp (∫ pdx)

= exp (∫ 3dx)
= e3x .
After multiplication by the integrating factor the equation can be written as
d 3x
(e y) = ex ,
dx
and therefore the solution is
y = e−3x ∫ ex dx = e−3x (ex + C) = e−2x + Ce−3x ,

where C is a constant of integration.

On substitution of x = 0, y = 5 we get
y (x) = (ex + 4) e−3x .
Back to Tutorial 1 Question 5
200 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Part 5.v

We have a linear first order ODE whose standard form is

y
y ′ + 12 = x.
x
The integrating factor for this equation is

IF = exp (∫ pdx)

1 dx
= exp (∫ 2 )
√ x
= x.
After multiplication by the integrating factor the equation can be written as
d √ √
( xy) = x3 ,
dx
and therefore the solution is
√ √ √ √ √
y = 1/ x ∫ x3 dx = 1/ x ( 25 x5 + C) = 25 x2 + C/ x,

where C is a constant of integration.

After the substitution x = 1, y = 1 we obtain
√
y (x) = 25 x2 + 35 / x.
Back to Tutorial 1 Question 5

Part 5.vi

This is a linear first order ODE whose standard form is

y ′ − 2y = 0.
The integrating factor for this equation is

IF = exp (∫ pdx)

= exp (∫ −2dx)
= e−2x .
After multiplication by the integrating factor we can re-write the equation as
d −2x
(e y) = 0,
dx
and therefore the solution is
y = e2x ∫ 0 = e2x (C) = Ce2x ,

where C is a constant of integration.

Back to Tutorial 1 Question 5
7.1. TUTORIAL 1 201

Part 5.vii

Textbook exercise set 2.3. Problem 7 (8th ed); Problem 12 (9th ed)
Taking terms with y on the left we get the equation
y
y ′ − = 2x + 1.
x
The integrating factor for this equation is

IF = exp (∫ pdx)
dx
= exp (∫ − )
x
1
= .
x
After multiplication by the integrating factor we can bring the equation to the form
d 1 2x + 1
( y) = ,
dx x x
and therefore the solution is
(2x + 1) dx
y = x∫ = x (2x + ln ∣x∣ + C) = 2x2 + x ln ∣x∣ + Cx,
x
where C is a constant of integration.
Back to Tutorial 1 Question 5

Textbook exercise set 2.3. Problem 9

After dividing by x the equation takes the form

2y 1
y′ + = .
x x4
The integrating factor for this equation is

IF = exp (∫ pdx)
2dx
= exp (∫ )
x
= x2 .
After multiplication by the integrating factor we can cast the equation to the form
d 1
(x2 y) = 2 ,
dx x
and therefore the solution is
1 dx 1 −1 + Cx 1 C
y= 2 ∫ 2
= 2( ) = − 3 + 2,
x x x x x x
where C is a constant of integration.
Back to Tutorial 1 Question 5
202 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Textbook exercise set 2.3. Problem 11

After dividing by −dt the equation takes the form

y ′ − y = t + 1.

The integrating factor for this equation is

IF = exp (∫ pdt)

= exp (∫ −dt)
= e−t .
After multiplication by the integrating factor we can re-write the equation as
d −t
(e y) = e−t (t + 1) ,
dt
and therefore the solution is

y = et ∫ e−t (t + 1) dt = et (−2e−t − e−t t + C) = −2 − t + Cet ,

where C is a constant of integration.The last step involved integration by parts

−t
RRR u = t + 1
RRR u′ = 1 RRR
RRR = −e−t t − e−t + −t
∫ e (t + 1) dt = RRR RRR ∫ e dt
RR dv = e−t dt v = −e−t RR
= −2e−t − e−t t.
Back to Tutorial 1 Question 5

Question 6

Part 6.i

The equation is
z ′ − z = xz 2
This is a Bernoulli equation. Let us introduce the new variable y = z −1 or equivalently
z = y −1 . Then using the chain rule we calculate the derivative
dy 1 dz
=− 2
dx z dx
which implies
dz 1 dy
=− 2 .
dx y dx
Substituting this and z in the equation, we find

−y ′ − y = x.

This is a linear first order ODE whose standard form is

y ′ + y = −x,
7.1. TUTORIAL 1 203

where the coefficient in front of y is

p = 1.
The integrating factor for this equation is

IF = exp (∫ pdx)

= exp (∫ dx)
= ex .
After multiplication by the integrating factor we can cast the equation to the form
d x
(e y) = −ex x,
dx
and therefore the solution is

y = e−x ∫ −ex xdx = e−x (ex − ex x + C) = 1 − x + Ce−x ,

where C is a constant of integration. Finally, substituting y = z −1 into the obtained solution

and solving for z, we find the solution of the Bernoulli equation
1
z(x) = .
1 − x + Ce−x
The solution above involved integration by parts

−x
RRR u = −x
RRR u′ = −1 RRR
RRR = xe−x − −x
∫ −xe dx = RRR RRR ∫ e dx
RR dv = e−x dx v = −e−x RR
= xe−x + e−x .

Back to Tutorial 1 Question 6

Part 6.ii

Remark. The solutions below use slightly different notation for the unknown function and
independent variables. It is assumed everywhere that the unknown z is a function of x:
z(x). This should be taken into account when comparing solutions given here with the
ones in the textbook.

Textbook exercise set 2.6. Problem 21

The equation is
z
z′ + = x2 z 2
x
We introduce the new variable y = z −1 or equivalently z = y −1 . Then we calculate the
derivative
dy 1 dz
=− 2
dx z dx
204 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

which implies
dz 1 dy
=− 2 .
dx y dx
Substituting this and z in the equation, we have
y
−y ′ + = x2 .
x
We got a linear first order ODE whose standard form is
y
y′ − = −x2 ,
x
where the coefficient in front of y is
1
p=− .
x
The integrating factor for this equation is

IF = exp (∫ pdx)
dx
= exp (∫ − )
x
1
= .
x
After multiplication by the integrating factor we can re-write the equation as
d y
( ) = −x,
dx x
and therefore the solution is

y = x ∫ −xdx = x (− 21 x2 + C) = − 12 x3 + Cx,

where C is a constant of integration. Finally, substituting y = z −1 into the obtained solution

and solving for z, we get the solution of the Bernoulli equation
2
z(x) = .
x (2C − x2 )
Back to Tutorial 1 Question 6

Textbook exercise set 2.6. Problem 23

The equation is
2z
z′ − = −x2 z 2
x
Let us introduce the new variable y = z −1 or equivalently z = y −1 , then the derivative is
dy 1 dz
=− 2
dx z dx
or
dz 1 dy
=− 2 .
dx y dx
7.1. TUTORIAL 1 205

Substituting the latter and z in the equation, we obtain

2y
−y ′ − = −x2
x
or
2y
y′ + = x2 .
x
Note the coefficient in front of y:
2
p= .
x
The integrating factor is
IF = exp (∫ pdx)
2dx
= exp (∫ )
x
= x2 .
After multiplication by the integrating factor we can bring the equation to the form

d
(x2 y) = x4 ,
dx
and therefore the solution is
1 4 1 1 5 1 3 C
y= ∫ x dx = ( 5 x + C) = 5 x + ,
x2 x2 x2
where C is a constant of integration. Finally, substituting y = z −1 into the obtained solution
and solving for z, we find the solution of the Bernoulli equation

5x2
z(x) = .
x5 + 5C
Back to Tutorial 1 Question 6

Textbook exercise set 2.6. Problem 25

The equation is
z
z′ + = −xz 3
x
1
The new variable is y = z −2 hence z = y − 2 . The derivative is given by

dy 2 dz
=− 3
dx z dx
hence
dz dy
= − 2√1y3 .
dx dx
Substituting this and z in the equation, we get
y
− 12 y ′ + = −x.
x
206 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

This is a linear first order ODE whose standard form is

2y
y′ − = 2x,
x
where the coefficient in front of y is
2
p=− .
x
The integrating factor for this equation is

IF = exp (∫ pdx)
2dx
= exp (∫ − )
x
1
= .
x2
After multiplication by the integrating factor the equation can be written in the form
d y 2
( 2) = ,
dx x x
and therefore the solution is
2dx
y = x2 ∫ = x2 (2 ln ∣x∣ + C) = 2x2 ln ∣x∣ + Cx2 ,
x
where C is a constant of integration. Finally, substituting y = z −2 into the obtained solution
and solving for z, we obtain the solution (implicit) of the Bernoulli equation
1
− x2 (2 ln ∣x∣ + C) = 0.
z2
Back to Tutorial 1 Question 6

Textbook exercise set 2.6. Problem 27

The equation is
2z z 2
z′ − =
x x2
We introduce the new variable y = z −1 or equivalently z = y −1 . Then we calculate the
derivative
dy 1 dz
=− 2
dx z dx
hence
dz 1 dy
=− 2 .
dx y dx
Substituting those in the equation, we get the linear equation
2y 1
−y ′ − =
x x2
whose standard form is
2y 1
y′ + = − 2,
x x
7.1. TUTORIAL 1 207

where the coefficient in front of y is

2
p= .
x
The integrating factor for this equation is

IF = exp (∫ pdx)
2dx
= exp (∫ )
x
= x2 .

After multiplication by the integrating factor the equation can be written as

d
(x2 y) = −1,
dx
and therefore the solution is
1 1 1 C
y= 2 ∫ −dx = 2 (−x + C) = − + 2 ,
x x x x
where C is a constant of integration. Finally, substituting y = z −1 into the obtained solution
and solving for z, we obtain the solution of the Bernoulli equation

x2
z(x) = .
C −x
Back to Tutorial 1 Question 6

Question 7

Part 7.i

The equation can be written in the form

dz z z 2
= + .
dx x x2
We see that it has the form dz/dx = f ( xz ) . Introducing the new variable y = z
x (or z = yx)
we calculate
dz dy
=y+x .
dx dx
Substituting these into the equation we obtain

dy
y+x = y (1 + y)
dx
or
dy y 2
=
dx x
This is a separable equation. Indeed, it can be re-written as

dy dx
=
y2 x
208 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

hence
dy dx
∫ y2 = ∫ x .
Integrating the latter we get
1
− = ln (x) + C,
y
where C is a constant of integration. Further, solving for y we get the explicit solution
1
y(x) = − .
ln (x) + C

Finally on substituting y = z/x in the previous, we obtain

x
z(x) = − .
ln (x) + C

After the substitution x = 1, z = 1 we obtain C = 1 hence

x x
z (x) = − = .
ln (x) − 1 1 − ln (x)
Back to Tutorial 1 Question 7

Part 7.ii

The equation can be written in the form

dz 3z
=2+ .
dx x
We see that it has the form dz/dx = f ( xz ) . Introducing the new variable y = z
x (or z = yx)
we calculate
dz dy
=y+x .
dx dx
Substituting these into the equation we obtain
dy
y+x = 2 + 3y
dx
or
dy 2 + 2y
=
dx x
This is a separable equation. Indeed, it can be written as
dy 2dx
=
y+1 x
hence
dy 2dx
∫ y+1 =∫ x .
Integrating the latter we find

ln (y + 1) = 2 ln (x) + C,
7.1. TUTORIAL 1 209

where C is a constant of integration. Further, solving for y we obtain the explicit solution

y(x) = eC x2 − 1 = Ax2 − 1,

where eC = A = const. Finally on substituting y = z/x in the previous, we obtain

z(x) = Ax3 − x.

Back to Tutorial 1 Question 7

Part 7.iii

The equation can be written in the form

dz z z 2
= + cos ( ) .
dx x x

We see that it has the form z ′ = f ( xz ) . Introducing the new variable y = z

x (or z = yx) we
calculate
dz dy
=y+x .
dx dx
Substituting these into the equation we obtain
dy
y+x = y + cos2 y
dx
or
dy cos2 y
=
dx x
This is a separable equation. Indeed, it can be written as
dy dx
2
=
cos y x
hence
dy dx
∫ cos2 y = ∫ x .
Integrating the latter we get
tan y = ln x + C,
where C is a constant of integration. Further, solving for y we get the explicit solution

y(x) = arc tan (ln (x) + C) .

Finally on substituting y = z/x in the previous, we obtain

z(x) = xarc tan (ln (x) + C) .

After the substitution x = 1, z = π/4 we get C = 1 hence

z (x) = xarc tan (ln (x) + 1) .

Back to Tutorial 1 Question 7

210 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Question 8

Part 8.1

Textbook exercise set 3.2. Problem 5

The equation that describes the phenomenon is given by

x′ = a − 10ax,
where a = 5 × 10−5 . We also have the initial condition x(0) = 1. This is a linear first order
ODE whose standard form is
x′ + 10ax = a.
The integrating factor is then

IF = exp (10a ∫ dt)

= e10at .
After multiplication by the integrating factor we can re-write the equation as
d 10at
(e x) = e10at a,
dt
and therefore the solution is
x = e−10at ∫ e10at adt = e−10at ( 10
1 10at
e + C) = 1
10 + Ce−10at ,

where C is a constant of integration. Solving the initial problem x(0) = 1 and substituting
a = 5 × 10−5 we obtain the solution
1
1 9 − 2000 t
x (t) = 10 + 10 e .
Further, we calculate
1
1 9 − 2000
x (1) = 10 + 10 e = 0.97.
Now, if x(t) = 0.2 then solving
1
1 9 − 2000 t
0.2 = 10 + 10 e
for t we find
t = 73.24.
Back to Tutorial 1 Question 8

Textbook exercise set 3.2. Problem 9

The Malthusian law for population growth is given by the formula

p (t) = p0 ekt .
We also have two conditions
p(0) = p0 = 1000 and p (7) = 3000.
Using these conditions we find k = 71 ln (3) = 0.1569. The answer to the question is then
p (30) = 1000 exp ( 30
7 ln (3)) = 110868.
7.1. TUTORIAL 1 211

Textbook exercise set 3.2. Problem 13

On substituting the conditions

p(0) = p0 = 1000, pa = p (7) = 3000 and pb = p (14) = 5000

in the equations

− (pa pb − 2p0 pb + p0 pa ) pa ln ( ppb0(−p a +p0 )

(−pb +pa ) )
p1 = , A= ,
−p2a + p0 pb p1 ta

we obtain
1
p1 = 6000, A = 42000 ln (5) .
Hence the predicted limiting population is p1 = 6000. The population in 2020 is estimated
by substituting t = 30 in the formula
−p0 p1
p (t) =
−p0 − exp (−Ap1 t) p1 + exp (−Ap1 t) p0

i.e.
p (30) = 6000000/ (1000 + 5000 exp (− 30
7 ln (5))) = 5969.8458.

Back to Tutorial 1 Question 8

Part 8.2

We are given a linear first order ODE:

I ′ + 2I = 10e−2t .

The integrating factor for this equation is

IF = exp (2 ∫ dt) = e2t .

After multiplication by the integrating factor we can rewrite the equation as

d 2t
(e I) = 10,
dt
and thus the solution is

I = e−2t ∫ 10dt = e−2t (10t + C) = 10te−2t + Ce−2t ,

where C is a constant of integration. Using the condition I(0) = 0 we find C = 0 hence the
solution is
I (t) = 10te−2t .

Back to Tutorial 1 Question 8

212 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

7.2 Tutorial 2
Question 1

Part 1
y ′′ − 3y ′ + 2y = 0.
The latter implies the quadratic equation

k 2 − 3k + 2 = 0

whose roots are

k1 = 1, k2 = 2.
The general solution, corresponding to this, is given by

y = Aex + Be2x .

Here and below A and B are arbitrary constants.

Back to Tutorial 2 Question 1

Part 2
y ′′ + 5y ′ = 0.
Quadratic equation
k (k + 5) = 0
has roots:
k1 = −5, k2 = 0.
The general solution, corresponding to this, reads

y = Ae−5x + B.

Back to Tutorial 2 Question 1

Part 3
3y ′′ − 10y ′ + 3y = 0.
Quadratic equation:
k 2 − 10
3 k+1=0

The roots are

k1 = 3, k2 = 31 .
The general solution, corresponding to this, is
1
y = Ae3x + Be 3 x .

Back to Tutorial 2 Question 1

Part 4
y ′′ + 25y = 0.
The quadratic equation
k 2 + 25 = 0
7.2. TUTORIAL 2 213

has the roots

k1 = −5i, k2 = 5i.
The general solution, corresponding to this, is given by

y = A cos (5x) + B sin (5x) .

Back to Tutorial 2 Question 1

Part 5
y ′′ − 25y = 0.
The quadratic equation
k 2 − 25 = 0
has the roots
k1 = −5, k2 = 5.
The general solution, corresponding to this, has the form

y = Ae−5x + Be5x .

Back to Tutorial 2 Question 1

Part 6
y ′′ − 4y ′ + 8y = 0.
The quadratic equation
k 2 − 4k + 8 = 0
has the roots
k1 = 2 − 2i, k2 = 2 + 2i.
The general solution, corresponding to this, reads

y = e2x (A cos (2x) + B sin (2x)) .

Back to Tutorial 2 Question 1

Part 7
y ′′ + 10y ′ + 125y = 0.
The quadratic equation
k 2 + 10k + 125 = 0
has the roots
k1 = −5 − 10i, k2 = −5 + 10i.
The general solution, corresponding to this, is

y = e−5x (A cos (10x) + B sin (10x)) .

Back to Tutorial 2 Question 1

Part 8
9y ′′ − 6y ′ + 10y = 0.
The quadratic equation
k 2 − 32 k + 10
9 =0
214 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

has the roots

k1 = 13 − i, k2 = 31 + i.
The general solution, corresponding to this, reads
1
y = e 3 x (A cos x + B sin x) .
Back to Tutorial 2 Question 1

Part 9
y ′′ = 7y.
The latter implies the quadratic equation
k2 − 7 = 0
whose roots are √ √
k1 = 7, k2 = − 7.
The general solution, corresponding to this, is given by
√ √
y = Ae 7x
+ Be− 7x
.
Back to Tutorial 2 Question 1

Part 10

Textbook exercise set 4.2. Problem 1

y ′′ + 6y ′ + 9y = 0.
The quadratic equation
k 2 + 6k + 9 = 0
has only one root k = −3. The general solution, corresponding to this, reads
y = Ae−3x + Bxe−3x .
Note that the second term in the solution is the first one multiplied by x. Here A and B are
arbitrary constants.
Back to Tutorial 2 Question 1

Textbook exercise set 4.2. Problem 5

y ′′ − 5y ′ + 6y = 0.
The quadratic equation
k 2 − 5k + 6 = 0
has the roots
k1 = 2, k2 = 3.
The general solution, corresponding to this, reads
y = Ae2x + Be3x .
Back to Tutorial 2 Question 1
7.2. TUTORIAL 2 215

Textbook exercise set 4.2. Problem 9

4y ′′ − 4y ′ + y = 0.
The quadratic equation
k 2 − k + 41 = 0
has only one root k = 1/2. The general solution, corresponding to this, has the form
1 1
y = Ae 2 x + Bxe 2 x .

Note that the second term in the solution is the first one multiplied by x.
Back to Tutorial 2 Question 1

Textbook exercise set 4.2. Problem 13

y ′′ + 2y ′ − 8y = 0.
The latter implies the quadratic equation

k 2 + 2k − 8 = 0

whose roots are

k1 = −4, k2 = 2.
The general solution, corresponding to this, is

y = Ae−4x + Be2x .

Here A and B are arbitrary constants.Substituting the initial values y(0) = 3, y ′ (0) = −12
into the obtained solution we obtain the algebraic system

−12 = −4A + 2B, 3 = A + B.

Solving this system, we get

A = 3, B = 0.
So, the solution reads
y (x) = 3e−4x .
Back to Tutorial 2 Question 1

Textbook exercise set 4.2. Problem 17

y ′′ − 2y ′ − 2y = 0.
The quadratic equation
k 2 − 2k − 2 = 0
has the roots √ √
k1 = 1 − 3, k2 = 1 + 3.
The general solution, corresponding to this, reads
√ √
y = A exp ((1 − 3) x) + B exp ((1 + 3) x) .
216 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Here A and B are arbitrary constants. Substituting the initial values y(0) = 0, y ′ (0) = 3
into the obtained solution we get the algebraic system
√ √
0 = A + B, 3 = A + A 3 − B 3 + B.

Solving this system, we get √ √

1
A= 2 3, B = − 12 3.
So, the solution reads
1
√ √ √ √
y (x) = 2 3 exp ((1 + 3) x) − 21 3 exp (− ( 3 − 1) x) .

Back to Tutorial 2 Question 1

Textbook exercise set 4.2. Problem 19

y ′′ + 2y ′ + y = 0.
The quadratic equation
k 2 + 2k + 1 = 0
has only one root k = −1. The general solution, corresponding to this, is

y = Ae−x + Bxe−x .

Substituting the initial values y(0) = 1, y ′ (0) = −3 into the obtained solution we have the
algebraic system
−3 = −A + B, 1 = A.
Solving this system, we get
A = 1, B = −2.
So, the solution is
y (x) = e−x − 2xe−x .
Back to Tutorial 2 Question 1

Question 2

Part 1

y ′′ + 2y ′ + 2y = 3 cos (2x) .
We have an inhomogeneous second order equation with constant coefficients. We will be
looking for a particular solution in the form

yp = K cos (2x) + L sin (2x) ,

where K and L some constants to be found. , we calculate the derivatives:

yp′ = −2K sin (2x) + 2L cos (2x) ,

yp′′ = −4K cos (2x) − 4L sin (2x) .
7.2. TUTORIAL 2 217

Substituting these expressions in the original equation we calculate

(−4K − 2L) sin (2x) + (4L − 2K) cos (2x) = 3 cos (2x) .

Equating the coefficients at sin and cos, we get

−4K − 2L = 0, −2K + 4L − 3 = 0

which gives
3
K = − 10 , L = 35 .
Therefore a particular solution y = yp is given by

y = 35 sin (2x) − 10
3
cos (2x) .

Back to Tutorial 2 Question 2

Part 2

y ′′ + 9y = 5 cos (3x) .
We are looking for a particular solution in the form

yp = xK cos (3x) + xL sin (3x)

Calculate the derivatives:

yp′ = (L − 3xK) sin (3x) + (K + 3xL) cos (3x) ,
yp′′ = (−6K − 9xL) sin (3x) + (6L − 9xK) cos (3x) .
Substituting these expressions in the equation we get

−6K sin (3x) + 6L cos (3x) = 5 cos (3x) .

Equating the coefficients at sin and cos, we get

−6K = 0, 6L − 5 = 0

which means
K = 0, L = 65 .
Therefore a particular solution y = yp has the form

y = 56 sin (3x) x.

Back to Tutorial 2 Question 2

Part 3
y ′′ + y ′ − 2y = 2x.
First consider the homogeneous equation:

y ′′ + y ′ − 2y = 0.

The latter implies the quadratic equation

k2 + k − 2 = 0
218 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

whose roots are

k1 = −2, k2 = 1.
The general solution, corresponding to this, reads
yh = Ae−2x + Bex ,
where A and B are arbitrary constants.
Now we’re looking for a particular solution in the form
yp = c0 + c1 x,
where ci are some constants. We have:
yp′ = c1 , yp′′ = 0.
Substituting this in the original equation we obtain
c1 − 2c0 − 2c1 x = 2x.
Equating the coefficients at different powers of x, we obtain
−2c1 − 2 = 0, c1 − 2c0 = 0.
Solving this, we find
c1 = −1, c0 = − 21 .
Therefore the general solution y = yh + yp reads
y = Ae−2x + Bex − 21 − x.
Substituting the initial values y(0) = 0, y ′ (0) = 1 into the obtained solution we obtain the
algebraic system
0 = A + B − 21 , 1 = −2A + B − 1.
Solving this system, we get
A = − 21 , B = 1.
Finally, the solution is
y (x) = − 12 e−2x + ex − 12 − x.
Back to Tutorial 2 Question 2

Part 4

Textbook exercise set 4.4. Problem 9

y ′′ + 2y ′ − y = 10.
We’re looking for a particular solution in the form
y p = c0 ,
where c0 is a constant. We have:
yp′ = 0, yp′′ = 0
hence
−c0 = 10
or
c0 = −10.
Back to Tutorial 2 Question 2
7.2. TUTORIAL 2 219

Textbook exercise set 4.4. Problem 13

y ′′ − y ′ + 9y = 3 sin (3x) .
We are looking for a particular solution in the form
yp = K sin (3x) + L cos (3x) ,
where K and L are some constants to be found. The derivatives are
yp′ = 3K cos (3x) − 3L sin (3x) ,
yp′′ = −9K sin (3x) − 9L cos (3x) .
Substituting these expressions in the equation we calculate
−3K cos (3x) + 3L sin (3x) = 3 sin (3x) .
Equating the coefficients at sin and cos, we obtain
3L − 3 = 0, −3K = 0
which gives
K = 0, L = 1.
Therefore a particular solution y = yp has the form
y = cos (3x) .
Back to Tutorial 2 Question 2

Textbook exercise set 4.4. Problem 15

y ′′ − 5y ′ + 6y = xex .
We are looking for a particular solution in the form
yp = (c0 + c1 x)ex ,
where ci are some constants. The derivatives of yp are
yp′ = (c0 + c1 x + c1 ) ex
yp′′ = (c0 + c1 x + 2c1 ) ex .
Substituting these expressions in the original equation we calculate
ex (2c0 + 2c1 x − 3c1 ) = xex .
Factoring out the exponential function, we obtain
2c0 − 3c1 = 0, 2c1 − 1 = 0.
Solving this we find
c0 = 43 , c1 = 21 .
Therefore a partial solution reads
yp = ( 34 + 12 x)ex .
Back to Tutorial 2 Question 2
220 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Textbook exercise set 4.4. Problem 17

y ′′ − 2y ′ + y = 8ex .
We’re looking for a particular solution in the form

yp = x2 ex c0 ,

where c0 is a constant. The derivatives of yp are

yp′ = (2xc0 + x2 c0 ) ex
yp′′ = (2c0 + 4xc0 + x2 c0 ) ex .

Substituting these expressions in the original equation we obtain

2ex c0 = 8ex .

Factoring out the exponential function, we get

2c0 − 8 = 0

hence
c0 = 4.
Thus a partial solution has the form

yp = 4x2 ex .

Back to Tutorial 2 Question 2

Question 3

Part 1

y ′′ + y = tan (x) .
First we restrict our attention to the homogeneous equation:

y ′′ + y = 0.

The latter implies the quadratic equation

k2 + 1 = 0

whose roots are

k1 = −i, k2 = i.
The general solution, corresponding to this, is

yh = A cos x + B sin x.

Here A and B are arbitrary constants. Now we’re looking for a particular solution of the
form
y = M (x) cos x + N (x) sin x.
7.2. TUTORIAL 2 221

First, we calculate the derivative:

y ′ = M ′ cos x − M sin x + N ′ sin x + N cos x

and assume that

M ′ cos x + N ′ sin x = 0.
Next, we calculate the second derivative:

y ′′ = −M ′ sin x − M cos x + N ′ cos x − N sin x.

Substituting the derivatives in the equation, we calculate :

−M ′ sin x + N ′ cos x = tan (x) .

This needs to be solved along with

M ′ cos x + N ′ sin x = 0.

From the last two equations we get

sin2 x
M′ = − , N ′ = sin x.
cos x
Integrating this we get

sin2 xdx
M =∫ − = sin x − ln (sec (x) + tan (x)) ,
cos x
N = ∫ sin xdx = − cos x.

Therefore the general solution has the form

1 + sin x
y = yh + yp = A cos x + B sin x − cos x ln ( ).
cos x
Back to Tutorial 2 Question 3

Part 2
y ′′ − y = ex .
First solve the homogeneous equation:

y ′′ − y = 0.

The latter implies the quadratic equation

k2 − 1 = 0

whose roots are

k1 = −1, k2 = 1.
The general solution, corresponding to this, has the form

yh = Ae−x + Bex

where A and B are arbitrary constants.

222 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Now we’re looking for a particular solution of the form

y = M (x) e−x + N (x) ex .

First, we calculate the derivative:

y ′ = M ′ e−x − M e−x + N ′ ex + N ex

and assume that

M ′ e−x + N ′ ex = 0.
Next, we calculate the second derivative:

y ′′ = −M ′ e−x + M e−x + N ′ ex + N ex .

Substituting the derivatives in the equation, we calculate :

−M ′ e−x + N ′ ex = ex .

This needs to be solved along with

M ′ e−x + N ′ ex = 0.

From the last two equations we obtain

M ′ = − 21 e2x , N ′ = 21 .

Integrating this we calculate

M = ∫ − 12 e2x dx = − 14 e2x ,
1
N =∫ 2 dx = 21 x.

Therefore the general solution is

y = yh + yp = Ae−x + Bex − 41 ex + 12 xex

or
y = Ae−x + Cex + 12 xex
Back to Tutorial 2 Question 3

Part 3
y ′′ + 3y ′ + 2y = 3e−2x + x.
First solve the homogeneous equation:

y ′′ + 3y ′ + 2y = 0.

The latter implies the quadratic equation

k 2 + 3k + 2 = 0

whose roots are

k1 = −2, k2 = −1.
7.2. TUTORIAL 2 223

The general solution, corresponding to this, is given by

yh = Ae−2x + Be−x ,

where A and B are arbitrary constants.

Now we’re looking for a particular solution of the form

y = M (x) e−2x + N (x) e−x .

First we calculate the derivative:

y ′ = M ′ e−2x − 2M e−2x + N ′ e−x − N e−x

and assume that

M ′ e−2x + N ′ e−x = 0.
Next, we calculate the second derivative:

y ′′ = −2M ′ e−2x + 4M e−2x − N ′ e−x + N e−x .

Substituting the derivatives in the equation, we get :

−2M ′ e−2x − N ′ e−x = 3e−2x + x.

This needs to be solved along with

M ′ e−2x + N ′ e−x = 0.

From the last two equations we calculate

M ′ = −3 − e2x x, N ′ = (3 + e2x x) e−x .

Integrating this we calculate

M = ∫ (−3 − e2x x) dx = −3x + 41 e2x − 12 e2x x,

N = ∫ (3 + e2x x) e−x dx = −3e−x − ex + xex .

Therefore the general solution has the form

y = yh + yp = Ae−2x + Be−x − 3xe−2x − 43 + 12 x − 3e−2x

or
y = Ce−2x + Be−x − 3xe−2x − 43 + 12 x
Back to Tutorial 2 Question 3

Part 4
y ′′ − y = x2 ex .
First we restrict our attention to the homogeneous equation:

y ′′ − y = 0.

The characteristic equation

k2 − 1 = 0
224 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

has the roots

k1 = −1, k2 = 1
Hence the yh solution is given by

yh = Ae−x + Bex .

Here A and B are arbitrary constants.

Now we’re looking for a particular solution of the form

y = M (x) e−x + N (x) ex .

Firsly, we calculate the derivative:

y ′ = M ′ e−x − M e−x + N ′ ex + N ex

and assume that

M ′ e−x + N ′ ex = 0.
Next, we calculate the second derivative:

y ′′ = −M ′ e−x + M e−x + N ′ ex + N ex .

Substituting the derivatives in the equation, we obtain :

−M ′ e−x + N ′ ex = x2 ex .

This needs to be solved along with

M ′ e−x + N ′ ex = 0.

From the last two equations we calculate

M ′ = − 21 x2 e2x , N ′ = 21 x2 .

Integrating this we obtain

M = ∫ − 12 x2 e2x dx = − 18 e2x + 14 e2x x − 14 x2 e2x ,

1 2
N =∫ 2 x dx = 61 x3 .

Therefore the general solution is given by

y = yh + yp = Ae−x + Bex − 81 ex + 14 xex − 41 x2 ex + 16 x3 ex

or
y = Ae−x + Cex + 14 xex − 14 x2 ex + 61 x3 ex
Back to Tutorial 2 Question 3

Part 5
7.2. TUTORIAL 2 225

Textbook exercise set 4.6. Problem 3

y ′′ + 2y ′ + y = e−x .
First solve the homogeneous equation:

y ′′ + 2y ′ + y = 0.

From the quadratic equation

k 2 + 2k + 1 = 0
we find that it has only one root k = −1. The general solution, corresponding to this, has
the form
yh = Ae−x + Bxe−x .
Note that the second term in the solution is the first one multiplied by x. Here A and B are
arbitrary constants.
Now we’re looking for a particular solution of the form

y = M (x) e−x + N (x) xe−x .

Firsly, we calculate the derivative:

y ′ = M ′ e−x − M e−x + N ′ xe−x + N e−x − N xe−x

and assume that

M ′ e−x + N ′ xe−x = 0.
Next, we calculate the second derivative:

y ′′ = −M ′ e−x + M e−x + N ′ e−x − 2N e−x − N ′ xe−x + N xe−x .

Substituting the derivatives in the equation, we calculate :

−M ′ e−x + N ′ e−x − N ′ xe−x = e−x .

This needs to be solved along with

M ′ e−x + N ′ xe−x = 0.

From the last two equations we get

M ′ = −x, N ′ = 1.

Integrating this we get

M = ∫ −xdx = − 21 x2 ,

N = ∫ dx = x.
Therefore the general solution reads

y = yh + yp = Ae−x + Bxe−x + 21 x2 e−x .

Back to Tutorial 2 Question 3

226 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Textbook exercise set 4.6. Problem 9

y ′′ − y = 4 + 2x.
First we solve the homogeneous equation:
y ′′ − y = 0
which implies the quadratic equation
k 2 − 1 = 0.
The roots are
k1 = −1, k2 = 1.
The general solution, corresponding to this, has the form
yh = Ae−x + Bex .
Here A and B are arbitrary constants.
Now we’re looking for a particular solution of the form
y = M (x) e−x + N (x) ex .
Firsly, we calculate the derivative:
y ′ = M ′ e−x − M e−x + N ′ ex + N ex
and assume that
M ′ e−x + N ′ ex = 0.
Next, we calculate the second derivative:
y ′′ = −M ′ e−x + M e−x + N ′ ex + N ex .
Substituting the derivatives in the equation, we calculate :
−M ′ e−x + N ′ ex = 4 + 2x.
This needs to be solved along with
M ′ e−x + N ′ ex = 0.
From the last two equations we obtain
M ′ = −ex (2 + x) , N ′ = e−x (2 + x) .
Integrating this we get

M = ∫ −ex (2 + x) dx = −ex − xex ,

N = ∫ e−x (2 + x) dx = −3e−x − e−x x.

Therefore the general solution is

y = yh + yp = Ae−x + Bex − 4 − 2x.
Back to Tutorial 2 Question 3

Remark. In the solutions below we use the geometric series expansion which is
1
= 1 + x + x2 + . . . .
1−x
7.2. TUTORIAL 2 227

Question 4

Part 1
We re-write the equation in the form

(D2 − 1)y = ex ,

and then as
1
y= ex .
(D − 1) (D + 1)
We denote φ(D) = (D − 1) (D + 1) and check that φ(1) = 0. Using the shifting theorem
(we have to make the substitution D → 1+D when moving the exponential function through
the D−operator), we obtain
1
y = ex 2 1,
D + 2D
which can then be re-written as
1 1 1
y = ex 1 = ex x.
D+2 D D+2
We can re-write the latter as
1
y = 12 ex x.
1 + 12 D
Using the geometric series expansion, we obtain

y = 12 ex (1 − 12 D + . . . )x
= 12 ex (x − 12 ) .

Thus a particular solution is

yp = 12 ex (x − 21 ) .
Back to Tutorial 2 Question 4

Part 2
We solve the problem in two steps: solve separately for equations

y ′′ + 3y ′ + 2y = 3e−2x , y ′′ + 3y ′ + 2y = x

and then combine the answers. We re-write the equation in the form

(D2 + 3D + 2)y = 3e−2x ,

and then as
1
y= 3e−2x .
(D + 2) (D + 1)
We denote φ(D) = (D + 2) (D + 1) and check that φ(−2) = 0. Using the shifting theorem
(we have to make the substitution D → −2 + D when moving the exponential function
through the D−operator), we obtain

1
y = 3e−2x 1,
D2 − D
228 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

which can then be re-written as

1 1 1
y = 3e−2x 1 = 3e−2x x.
D−1 D D−1
We can re-write the latter as
1
y = −3e−2x x.
1−D
Using the geometric series expansion, we find
yp = −3e−2x (D + 1 + . . . )x
= −3e−2x (x + 1) .
Now, let’s turn to the second equation. We re-write the equation in the form

(D2 + 3D + 2)y = x,

and then as
1
y= x.
(D + 2) (D + 1)
Then we find
1
1 2
y= x = x.
D2 + 3D + 2 1 + 12 D2 + 23 D
Using the geometric series expansion, we get
yp = 21 (1 − 21 D2 − 32 D + . . . )x
= 21 x − 43 .
So, combining the solutions we get

y (x) = −e−2x (3 + 3x) + 21 x − 34 .

Back to Tutorial 2 Question 4

Part 3
We re-write the equation in the form

(D2 − 1)y = x2 ex ,

and then as
1
y= x2 ex .
(D − 1) (D + 1)
We denote φ(D) = (D − 1) (D + 1) and check that φ(1) = 0. Using the shifting theorem
(we have to make the substitution D → 1+D when moving the exponential function through
the D−operator), we obtain
1
y = ex 2 x2 ,
D + 2D
which can then be re-written as
1 1 2 1 1 3
y = ex x = ex x.
D+2 D D+23
We can re-write the latter as
1
y = 61 ex x3 .
1 + 21 D
7.2. TUTORIAL 2 229

Using the geometric series expansion, we get

y = 16 ex (1 − 12 D + 41 D2 − 18 D3 + . . . )x3
= 61 ex (x3 − 23 x2 + 32 x − 43 ) .

Note that the last term can be absorbed into the general solution thus a particular solution
is
yp = 16 ex (x3 − 32 x2 + 32 x) .
Back to Tutorial 2 Question 4

Part 4
We re-write the equation in the form

(D2 + 2D + 2)y = e−x cos x ,

and then as
1
y= e−x cos x .
D2 + 2D + 2
This is equivalent to
1
y = Re ( exp ((−1 + i) x))
D2
+ 2D + 2
We denote φ(D) = D2 + 2D + 2 and check that φ(−1 + i) = 0. Using the shifting theorem
(we have to make the substitution D → −1 + i + D when moving the exponential function
through the D−operator), we obtain
1
y = Re (exp ((−1 + i) x) 1) ,
D2 + 2Di
which can then be re-written as
1 1 1 1
y = Re (exp ((−1 + i) x) 1) = Re (exp ((−1 + i) x) )
D D + 2i D 2i
= Re (− 21 i exp ((−1 + i) x) x) .

Finally, taking the real part, we obtain the particular solution

yp = 12 xe−x sin x.

Back to Tutorial 2 Question 4

Question 5

Part 1

x′′ + 20x = 0.
The quadratic equation
k 2 + 20 = 0
has the roots √ √
k1 = −2i 5, k2 = 2i 5.
230 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Hence the general solution is given by

√ √
x = A cos (2 5t) + B sin (2 5t) ,

where A and B are arbitrary constants.

For the derivative we get
√ √ √ √
x′ = 2 5B cos (2 5t) − 2A 5 sin (2 5t) .

Substituting the initial conditions t = 0, x (0) = 41 and x′ (0) = 0 into the solution and its
derivative we find √
0 = 2B 5, 14 = A
and thus
B = 0, A = 41 .
Hence the solution reads √
x (t) = 14 cos (2 5t) .
Back to Tutorial 2 Question 5

Part 2

2Q′′ + 16Q′ + 50Q = 50.

First solve the homogeneous equation:

Q′′ + 8Q′ + 25Q = 0

whose characteristic equation is

k 2 + 8k + 25 = 0.

The roots are

k1 = −4 − 3i, k2 = −4 + 3i.
Hence the solution Qh reads

Qh = e−4t (A cos (3t) + B sin (3t)) .

Here A and B are arbitrary constants. Now we’re looking for a particular solution in the
form
Qp = c0 ,
where c0 is a constant. We have:

Q′p = 0, Q′′p = 0.

Substituting this in the original equation we obtain

25c0 = 25.

Thus
c0 = 1.
7.2. TUTORIAL 2 231

The general solution y = yh + yp is given by

Q = e−4t (A cos (3t) + B sin (3t)) + 1.

For the derivative we find

Q′ = A (−4 cos (3t) − 3 sin (3t)) e−4t + B (−4 sin (3t) + 3 cos (3t)) e−4t .

Substituting the initial conditions t = 0, Q (0) = 0 and Q′ (0) = 0 into the solution and its
derivative we get
0 = 1 + A, 0 = 3B − 4A.
Solving for A and B we get
A = −1, B = − 34 .
Hence the solution reads

Q (t) = − 43 e−4t sin (3t) − e−4t cos (3t) + 1.

Back to Tutorial 2 Question 5

Part 3

Textbook exercise set 5.7 Problem 1

The circuit is described by the equation

4Q′′ + 100Q′ + 100Q = 20

which is an inhomogeneous second order equation with constant coefficients. We consider

the homogeneous equation:
Q′′ + 25Q′ + 25Q = 0
whose characteristic equation is

k 2 + 25k + 25 = 0.

The roots are √ √

k1 = − 25 5 25 5
2 − 2 21, k2 = − 2 + 2 21.

Solution Qh , corresponding to this, is given by

√ √
Qh = A exp ((− 252 − 5
2 21) t) + B exp ((− 25
2 + 5
2 21) t) ,

where A and B are arbitrary constants.

Now we’re looking for a particular solution in the form

Q p = c0 ,

where c0 is some constant. We have: Q′p = 0, Q′′p = 0. Hence we get 25c0 = 5 or c0 = 51 .

Therefore the general solution y = yh + yp is
√ √
Q = A exp ((− 25
2 − 5
2 21) t) + B exp ((− 25
2 + 5 1
2 21) t) + 5 .
 232 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

For the derivative we get

√ √ √ √
Q′ = (− 25 5 5
2 + 2 21) exp ( 2 (−5 + 21) t) B + (− 25 5 5
2 − 2 21) exp (− 2 (5 + 21) t) A.

Substituting the initial conditions t = 0, Q (0) = 4 and Q′ (0) = 0 into the solution and its
derivative we get
√ √
0 = (− 25 5 25 5
2 + 2 21) B + (− 2 − 2 21) A, 4 = B + A + 5 .
1

Solving for A and B we find

19 19
√ 19
√
A= 10 − 42 21, B = 42 21 + 19
10 .

Hence the solution has the form

√ √ √ √
Q (t) = exp ( 52 (−5 + 21) t) ( 19 19 5
42 21 + 10 ) + exp (− 2 (5 + 21) t) ( 19 19 1
10 − 42 21) + 5 .

Differentiating this solution we get

√ √ √ √
I(t) = Q′ (t) = − 19 5
21 21 exp ( 2 (−5 + 21) t) + 19 5
21 21 exp (− 2 (5 + 21) t) .

Back to Tutorial 2 Question 5

7.3 Tutorial 3
Question 1

Part 1
1 1
n2 + 1 n + n3 0+0
= lim 2 1 = = 0 ( convergent )
h3 + 2n + 1 n→∞ 1 + n2 + n3 1 + 0 + 0
Part 2
n4 + 1 n + n13
lim = lim
n→∞ n3 + 2n + 1 n→∞ 1 + 22 + 13
n n
= lim n = ∞ ( divergent )
n→∞

Part 3
n3 ∞
lim n
= ( , l’Hopital )
n→∞ 2 ∞
d(n3 )
dn 3n2
= lim d2 n = lim
n→∞ n→∞ 2n ln 2
dn
d(3n2 )
dn 6n
= lim = lim n
n→∞ d(2n ln 2) n→∞ 2 (ln 2)2
dn
d(6n)
dn 6
= lim = lim n
n→∞ d(2n (ln 2)2 ) n→∞ 2 (ln 2)3
dn
= 0, as lim 2n = ∞
n→∞
(convergent)
Back to Tutorial 3 Question 1
 7.3. TUTORIAL 3 233

Part 4 The rigorous proof is quite complicated, but obviously rn = rn−1 r < r < 1 (r < 1).
So the sequence rn stays less than 1 and decreasing. We accept without proof that it
decreasing all the way to zero.
1
If r > 1 then r < 1 hence

1
lim = 0 → lim rn = ∞
n→∞ rn n→∞

Part 5 an = 1 + (−1)4 hence {an } = {0, 2, 0, 2, . . . , } The sequence consists of alternating

numbers: 0 and 2 hence it has ho limit by definition (divergent).

Part 6
1 + (−1)n 0, n − odd
an = ={ 2
n n , n − even

Since
2
lim =0
n→∞ n
then
1 + (−1)n
lim an = lim = 0 (convergent)
n→∞ n→∞ n
Part 7
⎧
⎪ 1, n - even
⎪
an = ⎨ 1
n−1
2
⎪
⎪
⎩ 1 − ( 2
) n - odd
n−1
Since → ∞ as n → ∞, then
2
n−1
1 2
lim ( ) = 0.
n→∞ 2

Hence
lim an = 1 (convergent)
n→∞

Part 8
1 1 7
n2 − n + 7 n − n2 + n3
lim = lim
n→∞ 2n3 + n2 n→∞ 2 + n1
0−0+0
= = 0 (convergent )
2+0
Part 9
sin2 n
an = √
n
Note that 0 < sin2 n < 1, thus 0 < an < √1n .
√
Since limn→∞ n = ∞, then limn→∞ √1n = 0 hence

sin2 n
lim √ = 0.
n→∞ n

Back to Tutorial 3 Question 1

 234 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Part 10 √
n ∞
lim = ( , l’Hopital )
n→∞ ln n ∞
√ 1
√
d n 2 n
= lim d ln n = lim 1
n→∞ n→x
n
√dn
n
= lim =∞ (divergent)
n→∞ 2

Part 11
2n + 1 ∞
lim = ( , l’Hopital )
n→∞ en ∞
d(2n +1)
dn 2n ln 2
= lim de n = lim
n→∞
dn
n→∞ en
2 n
= ln 2 lim ( ) = 0
n→∞ e

2
Since e < 1 then
2 n
lim ( ) = 0 (converges)
n→∞ e

Back to Tutorial 3 Question 1

Question 2

Part 1
∞
1 1 1 1 n−1
1+ + + ... + n + ... = ∑ ( )
3 9 3 n=1 3
1 3 1
= 1 = (Geometric sevies with q = )
1− 3 2 3

Part 2
1 + 3 + 5 + 7 + . . . + (2n − 1) + . . .
Series diverges because
lim an = lim (2n − 1) ≠ 0.
n→∞ n→∞

Part 3
1 − 2 + 4 − 8 + 16 − . . . + (−2)n + . . .
Series diverges because

lim qn = lim (−2)n ≠ 0 (doesn’t exist )

n→∞ n→∞

Part 4
∞
4 4 4 1 n−1 4
4+ + + ... + n + ... = ∑ 4( ) = =6
3 9 3 n=1 3 1 − 31
Geometric series with a0 = 4, q = 13 .

Back to Tutorial 3 Question 2

 7.3. TUTORIAL 3 235

Question 3

Part 1 We have
n 1 2
∫ n2 + 1 dn = 2 ln (n + 1)
hence ∞ n
∫1 dn = ∞ − 12 ln (2) = ∞.
n2
+1
The series diverges.
Part 2
1 √
∫ √ dn = 2 n + 1
n+1
∞ 1 √
∫1 √ dn = ∞ − 2 2 = ∞.
n+1
The series diverges.
Part 3
1
∫ n2 + 1 dn = arc tan (n)
∞ 1 1
∫1 n2 + 1 dn = 4 π.
The series converges.
Part 4
ln (n) ln (n) 1
∫ 2
dn = − −
n n n
∞ ln (n)
∫1 dn = 1.
n2
The series converges. Here is the detailed integration (by parts):
RRR u = ln (n) u′ = 1 RRR
ln (n) n RRR = − 1 ln (n) + 1
∫ dn = RRRRR RRR ∫ dn
n2 RRR dv = n12 dn v = − n1 RR n n2
1 1
= − ln (n) − .
n n

Back to Tutorial 3 Question 3

Question 4

Part 1
1 1
< 2, n ⩾ 1
n2 +n+1 n
The series ∞
1
∑ 2
n=1 n
is known to converge. So
∞
1
∑
n=1 n2 +n+1
also converges.
 236 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Part 2
1 1 1
√ ⩾ = , n⩾1
n + n n + n 2n
The series
∞
1 1 ∞ 1
∑ = ∑
n=1 2n 2 n=1 n
diverges. Thus the series
∞
1
∑ √
n=1 n + n
is divergent.

Part 3
1 1
< , n⩾1
1 + 3n 3n
The series
∞
1
∑ n
n=1 3
1
converges as a geometric series with q = 3 . Hence
∞
1
∑ n
n=1 1 + 3

is convergent.

Part 4 √ √
n n 1
2
⩽ 2 = 3, n⩾1
n +n n n2
The series
∞
1
∑ 3
n=1 n2
3
converges as an α -series with α = 2 > 1 (see Example 1). Hence
∞
√
n
∑ 2
n=1 n + n

converges.

Part 5
sin2 n 1 1
2
⩽ 2 ⩽ 2, n ⩾ 1
n +1 n +1 n
The series
∞
1
∑ 2
n=1 n

converges. Thus
∞
sin2 n
∑ 2
n=1 n + 1

converges as well.

Back to Tutorial 3 Question 4

7.3. TUTORIAL 3 237

Question 5

Part 1 We have
10n
an = .
nn
Consider the expression

an+1 10n+1 nn 10 n n 10 1
= n+1 = ( ) = n.
an (n + 1) 10 n n+1 n+1 n + 1 (1 + n1 )

Taking into account that

1 n
lim (1 + ) = e
n→∞ n
we obtain
an+1 10
lim = lim =0<1
n→∞ an n→∞ (n + 1)e

hence the series converges.

Part 2 We have
n n
an = ( ) .
ln n
n
Note that > 1, so the series diverges. Let us use the ratio test to show the same.
ln n
Consider the expression
n+1 n
an+1 n+1 ln n n n+1 n n+1 ln n
=( ) ( ) =( ) ( ) .
an ln(n + 1) n n ln(n + 1) ln(n + 1)

Since
n+1 n n+1
lim ( ) = e, lim = ∞ (L’Hospital’s rule)
n→∞ n n→∞ ln(n + 1)
n
ln n
lim ( ) =1
n→∞ ln(n + 1)

then
an+1
lim =∞
n→∞ an

and the series diverges.

Part 3 We have
3n
an = .
n!n
Consider the expression

an+1 3n+1 n!n 3n

= n
= 2.
an (n + 1)! (n + 1) 3 (n + 1)

Since
3n
lim 2 =0<1
n→∞ (n + 1)
the series converges.
Back to Tutorial 3 Question 5
 238 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Part 4 We have
n!n2
an = .
(2n)!
Consider the expression
2 2
an+1 (n + 1)! (n + 1) (2n)! 1 (n + 1)
= =2 .
an (2n + 2)!n!n2 (2n + 1) n2

Since
2
1 (n + 1)
lim =0
n→∞ 2 (2n + 1) n2

we conclude that the series converges.

Part 5 We have
n
an = n ( 43 ) .
Consider the expression

3 (n+1)
an+1 (n + 1) ( 4 ) 3n+1
= n = 4 .
an n ( 34 ) n

Since
3 n+1 3
lim =
n→∞ 4 n 4

we conclude that the series converges.

Back to Tutorial 3 Question 5

Question 6

Part 1 We have
1
an = .
n2
Since
1
lim =0
n→∞ n2
and
1 1
an+1 = 2 < = an , for any n > 1
(n + 1) n2
we conclude that the series converges.

Part 2
n
an = .
3n + 2
We have
n 1
lim = 3 ≠0
n→∞ 3n + 2
hence the series diverges.
Back to Tutorial 3 Question 6
 7.3. TUTORIAL 3 239

Part 3
n
an = √ .
n3 + 2
Since
n
lim √ =0
n→∞ n3 + 2
and
n+1 n
an+1 = √ <√ = an , for any n > 1
3
(n + 1) + 2 n3 + 2

we conclude that the series converges.

Part 4
n
an = .
ln (n)
Since
n
lim =∞≠0
n→∞ ln (n)

we conclude that the series diverges.

Part 5
1
an = 1 .
2n
Since
1
lim 1 =1≠0
n→∞ 2n
we conclude that the series diverges.
Back to Tutorial 3 Question 6

Question 7
∞ ∞
(−1)n+1 1
a. Consider ∑ n
. The respective series with positive terms is ∑ n
which is
n=1 3 n=1 3
∞
(−1)n+1
a geometric series with q = 31 < 1 hence it converges. Thus ∑ converges
n=1 3n
absolutely.
∞
(1)n+1
Further, cousider ∑ . The respective series with positive terms is the harmonic
n=1 n
∞
1
series ∑ which is divergent.
n=1 n
On the other hand
1 1
a) > and
n n+1
1
b) lim = 0
n→∞ n

∞
(−1)n+1
thus ∑ converges conditionally.
n=1 3n
Back to Tutorial 3 Question 7
240 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

b.1 The series can be written as

∞ n
(−1) 10n
∑ .
n=1 n!
The series with positive terms is
∞
∑ an ,
n=1
where
10n
an = .
n!
We will use the ratio test to test the convergence of the latter series. Consider the
expression
an+1 10n+1 n! 10
= n
= .
an (n + 1)!10 n+1
Since
10
lim =0
n→∞ n + 1
we conclude that the series converges hence the series in question converges abso-
lutely.
b.2 We are given an alternating series
∞
∑ (−1)n+1 an
n=1

with
nn
an = n.
(n + 1)
Since
nn −1
lim n =e ≠0
n→∞ (n + 1)

we conclude that the series diverges.

b.3 We are given an alternating series
∞
1 − ∑ (−1)n+1 an
n=1

where
√ √ 1
an = n+1− n= √ √ .
n+1+ n
Since
1
lim an = lim √ √ =0
n→∞ n→∞ n+1+ n
and
1 1
an+1 = √ √ <√ √ = an , for any n > 1
n+2+ n+1 n+1+ n
we see that series converges. Now consider the series with positive terms
∞
1 1 ∞ 1
∑√ √ > ∑√ .
n=1 n + 1 + n 2 n=1 n + 1
The latter series is known to be divergent hence the former series diverges as well.
Thus the series in question converges conditionally.
Back to Tutorial 3 Question 7
7.3. TUTORIAL 3 241

Question 8

1. ∞
∑ nxn
n=1

We denote
an = nxn
and consider the expression
an+1 (n + 1) xn+1 (n + 1) x
∣ ∣=∣ n
∣=∣ ∣.
an nx n
Then we calculate
(n + 1) x
lim ∣ ∣ = ∣x∣ .
n→∞ n
According to the ratio test the series converges if

∣x∣ < 1 i.e.

−1 < x < 1

This gives us the interval of convergence (without endpoints):

x ∈ (−1, 1).

The radius of convergence is R = 1.

Now consider the end points. Firstly we set x = −1 to get the series
∞
n
∑ n (−1)
n=1

which is an alternating series of the form

∞
∑ (−1)n an
n=1

with
an = n.
Since
lim n = ∞ ≠ 0
n→∞

we conclude that the series diverges at x = −1.

Secondly we set x = 1 to obtain the series
∞
∑n
n=1

Since
lim n = ∞ ≠ 0
n→∞

we conclude that the series diverges at x = 1. The interval of convergence is therefore

(−1, 1).
Back to Tutorial 3 Question 8
242 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

2. ∞
nxn
∑ n
n=1 2

We denote
nxn
an =
2n
and consider the expression
an+1 (n + 1) xn+1 2n (n + 1) x
∣ ∣=∣ n+1 n
∣ = ∣ 21 ∣.
an 2 nx n
Then we calculate
(n + 1) x
lim ∣ 21 ∣ = ∣ 21 x∣ .
n→∞ n
According to the ratio test the series converges if

∣ 12 x∣ < 1, i.e.
−1 < 21 x < 1
−2 < x < 2

This gives us the interval of convergence (without endpoints):

x ∈ (−2, 2).

The radius of convergence is R = 2.

Now consider the end points. Firstly we set x = −2 to get the series
∞ ∞ n
n (−2) n
∑ n
= ∑ n (−1)
n=1 2 n=1

which is an alternating series of the form

∞
∑ (−1)n an
n=1

with
an = n.
Since
lim n = ∞ ≠ 0
n→∞

we conclude that the series diverges at x = −2.

Secondly we set x = 2 to obtain the series
∞
∑n
n=1

Since
lim n = ∞ ≠ 0
n→∞

we conclude that the series diverges at x = 2. The interval of convergence is therefore

(−2, 2).
Back to Tutorial 3 Question 8
7.3. TUTORIAL 3 243

3. ∞
∑ n!xn
n=1

We denote
an = n!xn
and consider the expression
an+1 (n + 1)!xn+1
∣ ∣=∣ ∣ = ∣(n + 1) x∣ .
an n!xn
Then we calculate
lim ∣(n + 1) x∣ = ∞ for x ≠ 0.
n→∞

Thus the interval of convergence consists of one point x = 0.

Back to Tutorial 3 Question 8

4.
∞ ∞
(−1)n nxn n∣x∣n
∑∣ 3 ∣ = ∑ 3
n=1 2n (n + 1) n
n=1 2 (n + 1)

We denote
n∣x∣n
an = 3
2n (n + 1)
and consider the expression
4 4
an+1 (n + 1) xn+1 2n (n + 1) x
∣ ∣=∣ 3 ∣ = ∣ 12 3 ∣.
an n+1
2 (n + 2) nx n (n + 2) n
Then we calculate
4
(n + 1) x
lim ∣ 1 ∣ = ∣ 12 x∣ .
n→∞ 2 (n + 2) n
3

According to the ratio test the series converges if

∣ 12 x∣ < 1
−1 < 12 x < 1
−2 < x < 2

This gives us the interval of convergence (without endpoints):

x ∈ (−2, 2).

The radius of convergence is R = 2.

Now consider the end points. Firstly we set x = −2 to get the series
∞ n ∞
(−1)n (−2) n
∑ 3 =∑ 3.
n=1 2n (n + 1) n=1 (n + 1)
Since
∞ ndn 3
∫1 3 = ≠∞
(n + 1) 8
we conclude that the series converges at x = −2.
244 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Secondly we set x = 2 to obtain the series

∞
(−1)n n
∑ 3
n=1 (n + 1)
which is an alternating series with
n
an = 3.
(n + 1)
Since
n
lim 3 =0
n→∞ (n + 1)
and
n+1 n
an+1 = 3 < 3 = an , for any n > 1
(n + 2) (n + 1)
we conclude that the series converges. The interval of convergence is thus [−2, 2].
Back to Tutorial 3 Question 8
5. ∞
n
∑ (5x − 3)
n=1
We denote
n
an = (5x − 3)
and consider the expression
n+1
an+1 (5x − 3)
∣ ∣=∣ n ∣ = ∣5x − 3∣ .
an (5x − 3)
Then we calculate
lim ∣5x − 3∣ = ∣5x − 3∣ .
n→∞
According to the ratio test the series converges if
∣5x − 3∣ < 1
−1 < 5x − 3 < 1
2 < 5x < 4
2 4
5 <x< 5

This gives us the interval of convergence (without endpoints):

x ∈ ( 25 , 45 ).

The radius of convergence is R = 51 .

2
Now consider the end points. Firstly we set x = 5 to get the series
∞
n
∑ (−1)
n=1

which is an alternating series of the form

∞
∑ (−1)n an .
n=1
7.3. TUTORIAL 3 245

We have
an = 1.
Since
lim 1 = 1 ≠ 0
n→∞

we conclude that the series diverges.

4
Secondly we set x = 5 to obtain the series
∞
∑1
n=1

Since
lim 1 = 1 ≠ 0
n→∞

we conclude that the series diverges at x = 54 . The interval of convergence is therefore

( 52 , 45 ).
Back to Tutorial 3 Question 8

6.
∞ n
2n (x − 3)
∑
n=1 n2
We denote n
2n (x − 3)
an =
n2
and consider the expression
n+1
an+1 2n+1 (x − 3) n2 2 (x − 3) n2
∣ ∣=∣ 2 n ∣ = ∣ 2 ∣.
an (n + 1) 2n (x − 3) (n + 1)

Then we calculate
2 (x − 3) n2
lim ∣ 2 ∣ = ∣2x − 6∣ .
n→∞ (n + 1)
According to the ratio test the series converges if

∣2x − 6∣ < 1
−1 < 2x − 6 < 1
5 < 2x < 7
5 7
2 <x< 2

This gives us the interval of convergence (without endpoints):

x ∈ ( 52 , 72 ).

The radius of convergence is R = 21 .

5
Now consider the end points. Firstly we set x = 2 to get the series
n
∞ 2n (− 12 ) ∞ n
(−1)
∑ = ∑
n=1 n2 n=1 n
2
246 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

which is an alternating series of the form

∞
∑ (−1)n an .
n=1

We have
1
an = .
n2
Since
1
lim =0
n→∞ n2

and
1 1
an+1 = 2 < = an , for any n > 1
(n + 1) n2
we conclude that the series converges at x = 52 .
7
Secondly we set x = 2 to obtain the series with positive terms
∞
1
∑ 2
.
n=1 n

This series is know to converge (α-series with α > 1). We also use the integral test:
∞ 1
∫1 dn = 1.
n2
Since the integral is finite, then according to the integral test the series converges at
x = 72 . The interval of convergence is thus [ 52 , 72 ].
Back to Tutorial 3 Question 8

7. Textbook exercise set 8.2 Q 1

∞ n
(x − 1)
∑ n
n=0 2 (n + 1)

We denote n
(x − 1)
an = n
2 (n + 1)
and consider the expression
n+1
an+1 (x − 1) 2n (n + 1) 1 (x − 1) (n + 1)
∣ ∣ = ∣ n+1 n ∣ = ∣2 ∣.
an 2 (n + 2) (x − 1) n+2

Then we calculate
(x − 1) (n + 1)
lim ∣ 1 ∣ = ∣ 21 x − 12 ∣ .
n→∞ 2 n+2
According to the ratio test the series converges if

∣ 21 x − 12 ∣ < 1
−1 < 21 x − 12 < 1
− 12 < 21 x < 3
2

−1 < x < 3
7.3. TUTORIAL 3 247

This gives us the interval of convergence (without endpoints):

x ∈ (−1, 3).

The radius of convergence is R = 2.

Now consider the end points. Firstly we set x = −1 to get the series
∞ ∞ n n
(−2) (−1)
∑ n =∑
n=0 2 (n + 1) n=0 n + 1

which is an alternating series of the form

∞
∑ (−1)n an .
n=0

We have
1
an = .
n+1
Since
1
lim =0
n→∞ n + 1

and
1 1
an+1 = < = an , for any n > 1
n+2 n+1
we conclude that the series converges.
Secondly we set x = 3 to obtain the series
∞
1
∑
n=0 n + 1

Then we calculate the integral

∞ 1
∫0 dn = ∞.
n+1
Since the integral is infinite the series diverges at x = 3. The interval of convergence
is thus [−1, 3).
Back to Tutorial 3 Question 8

8. Textbook exercise set 8.2 Q 3

∞ n
n2 (x + 2)
∑
n=0 2n
We denote n
n2 (x + 2)
an =
2n
and consider the expression
2 n+1 2
an+1 (n + 1) (x + 2) 2n (n + 1) (x + 2)
∣ ∣=∣ n ∣ = ∣ 21 ∣.
an n+1 2
2 n (x + 2) n2

Then we calculate
2
(n + 1) (x + 2)
lim ∣ 21 ∣ = ∣ 21 x + 1∣ .
n→∞ n2
248 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

According to the ratio test the series converges if

∣ 12 x + 1∣ < 1
−1 < 21 x + 1 < 1
−2 < 12 x < 0
−4 < x < 0

This gives us the interval of convergence (without endpoints):

x ∈ (−4, 0).

The radius of convergence is R = 2.

Now consider the end points. Firstly we set x = −4 to get the series
∞ n∞
n2 (−2) n
∑ n
= ∑ (−1) n2
n=0 2 n=0

which is an alternating series of the form

∞
∑ (−1)n an .
n=0

We have
an = n 2 .
Since
lim n2 = ∞ ≠ 0
n→∞

we conclude that the series diverges.

Secondly we set x = 0 to obtain the series
∞
∑ n2
n=1

Since
lim n2 = ∞ ≠ 0
n→∞

we conclude that the series diverges at x = 0. The interval of convergence is therefore

(−4, 0).
Back to Tutorial 3 Question 8

9. Textbook exercise set 8.2 Q 5

∞ n
3 (x − 2)
∑
n=1 n3
We denote n
3 (x − 2)
an =
n3
and consider the expression
n+1
an+1 (x − 2) n3 (x − 2) n3
∣ ∣=∣ 3 n ∣ = ∣ 3 ∣.
an (n + 1) (x − 2) (n + 1)
7.3. TUTORIAL 3 249

Then we calculate
(x − 2) n3
lim ∣ 3 ∣ = ∣x − 2∣ .
n→∞ (n + 1)
According to the ratio test the series converges if

∣x − 2∣ < 1
−1 < x − 2 < 1
1<x<3

This gives us the interval of convergence (without endpoints):

x ∈ (1, 3).

The radius of convergence is R = 1.

Now consider the end points. Firstly we set x = 1 to get the series
∞ n
3 (−1)
∑
n=1 n3

which is an alternating series of the form

∞
∑ (−1)n an .
n=1

We have
3
an = .
n3
Since
3
lim =0
n→∞ n3
and
3 3
an+1 = 3 < = an , for any n > 1
(n + 1) n3
we conclude that the series converges.
Secondly we set x = 3 to obtain the series
∞
3
∑ 3
n=1 n

Then we calculate the integral

∞ 3
∫1 dn = 32 .
n3
Since the integral is finite the series converges at x = 3. The interval of convergence
is thus [1, 3].
Back to Tutorial 3 Question 8
250 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

7.4 Tutorial 4
Question 1

1. Textbook exercise set 8.3 Q 11

y ′ + (x + 2) y = 0
We are looking for a solution in the form
∞
y = ∑ an x n .
n=0

Then the derivative of the solution is

∞
y ′ = ∑ nan xn−1 .
n=1

After substituting these formulae in the equation we obtain

∞ ∞
∑ nan xn−1 + (x + 2) ∑ an xn = 0
n=1 n=0

or ∞ ∞ ∞
∑ nan xn−1 + 2 ∑ an xn + ∑ an xn+1 = 0.
n=1 n=0 n=0

On adjusting the indices we obtain

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

Setting n = 0 we obtain the relation

a1 + 2a0 = 0

which implies
a1 = −2a0 .
Equating the coefficients at different powers of x we find the recurrent formula

(n + 1) an+1 + 2an + an−1 = 0

or
2an + an−1
an+1 = − .
n+1
Setting n = 1, 2, . . . in the recurrent formula we calculate

a1 = −2a0 , a2 = 23 a0 , a3 = − 13 a0 , a4 = − 24
5
a0 .

The solution takes the form

y = a0 − 2a0 x + 23 a0 x2 − 13 a0 x3 − 24
5
a0 x 4 + . . .

Back to Tutorial 4 Question 1

7.4. TUTORIAL 4 251

2. Textbook exercise set 8.3 Q 13.

Here we slightly change notation by setting z = y, then the equation becomes

y ′′ − x2 y = 0

Then, we are looking for a solution in the form

∞
y = ∑ an x n .
n=0

Hence ∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2

After substituting these formulae in the equation we obtain

∞ ∞
∑ (n − 1) nan xn−2 − x2 ∑ an xn = 0
n=2 n=0

or ∞ ∞
∑ (n − 1) nan xn−2 − ∑ an xn+2 = 0.
n=2 n=0

On adjusting the indices we obtain

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

Setting n = 0, 1 we obtain
n=0∶ 2a2 = 0
n=1∶ 6a3 = 0
which imply
a2 = 0, a3 = 0.
Equating the coefficients at different powers of x we get the recurrent formula

(n + 1) (n + 2) an+2 − an−2 = 0

or
an−2
an+2 = .
(n + 1) (n + 2)
Setting n = 2, 3, . . . in the recurrent formula we find
1 1
a2 = 0, a3 = 0, a4 = 12 a0 , a5 = 20 a1 .

The solution takes the form

1 1
y = a0 + a1 x + 12 a0 x4 + 20 a1 x 5 + . . .

Back to Tutorial 4 Question 1

3. Textbook exercise set 8.3 Q 15

y ′′ + (x − 1) y ′ + y = 0
252 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

We are looking for a solution in the form

∞
y = ∑ an x n .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2

After substituting these formulae in the equation we obtain

∞ ∞ ∞
∑ (n − 1) nan xn−2 + (x − 1) ∑ nan xn−1 + ∑ an xn = 0
n=2 n=1 n=0

or
∞ ∞ ∞ ∞
∑ (n − 1) nan xn−2 − ∑ nan xn−1 + ∑ an nxn + ∑ an xn = 0.
n=2 n=1 n=1 n=0

On adjusting the indices we find

∞ ∞ ∞ ∞
∑ (n + 1) (n + 2) an+2 xn − ∑ (n + 1) an+1 xn + ∑ an nxn + ∑ an xn = 0.
n=0 n=0 n=1 n=0

Setting n = 0 we obtain the relation

2a2 − a1 + a0 = 0 .

On solving the above we obtain

a2 = 21 a1 − 12 a0 .

Equating the coefficients at different powers of x we find the recurrent formula

(n + 1) (n + 2) an+2 − (n + 1) an+1 + (n + 1)an = 0

or
an+1 − an
an+2 = .
n+2
Setting n = 0, 1, . . . in the recurrent formula we calculate

a2 = 21 a1 − 12 a0 , a3 = − 61 a1 − 16 a0 , a4 = − 61 a1 + 12
1
a0 , a 5 = 1 1 1
20 a0 , 2 a1 − 2 a0 = 12 a1 − 12 a0 .

The solution takes the form

y = a0 + a1 x + ( 21 a1 − 21 a0 ) x2 + (− 16 a1 − 16 a0 ) x3 + (− 16 a1 + 12
1 1
a0 ) x4 + 20 a0 x 5 + . . .

Back to Tutorial 4 Question 1

7.4. TUTORIAL 4 253

Question 2

Part 1 a)
y′ − y + x = 0
We are looking for a solution in the form
∞
y = ∑ an x n .
n=0

Then the derivative of the solution is

∞
y ′ = ∑ nan xn−1 .
n=1

After substituting these formulae in the equation we obtain

∞ ∞
∑ nan xn−1 − ∑ an xn + x = 0
n=1 n=0

or ∞ ∞
∑ nan xn−1 − ∑ an xn = −x.
n=1 n=0

On adjusting the indices we obtain

∞ ∞
∑ (n + 1) an+1 xn − ∑ an xn = −x.
n=0 n=0

Equating the coefficients at different powers of x we find the recurrent formula

2a2 − a1 = −1, n = 1
(n + 1) an+1 − an = 0, n ≠ 1
or
an
an+1 = , n ≠ 1.
n+1
Setting n = 0, 1, . . . in the recurrent formula we find

a1 = a0 , a2 = 12 (a0 − 1), a3 = 61 (a0 − 1), a4 = 1

24 (a0 − 1).

On setting c = a0 − 1 the solution takes the form

y = (c + 1) + (c + 1)x + 21 cx2 + 16 cx3 + 24

1
cx4 + . . .

Back to Tutorial 4 Question 2

b)
y ′′ + y = 0
We are looking for a solution in the form
∞
y = ∑ an x n .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2
254 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

After substituting these formulae in the equation we obtain

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

or ∞ ∞
∑ (n − 1) nan xn−2 + ∑ an xn = 0.
n=2 n=0

On adjusting the indices we obtain

∞ ∞
∑ (n + 1) (n + 2) an+2 xn + ∑ an xn = 0.
n=0 n=0

Equating the coefficients at different powers of x we get the recurrent formula

(n + 1) (n + 2) an+2 + an = 0

or
an
an+2 = − .
(n + 1) (n + 2)
Setting n = 0, 1, . . . in the recurrent formula we get

a2 = − 21 a0 , a3 = − 61 a1 , a4 = 1
24 a0 , a5 = 1
120 a1 .

The solution takes the form

y = a0 + a1 x − 21 a0 x2 − 16 a1 x3 + 24
1 1
a0 x4 + 120 a1 x 5 + . . .
1 4
= a0 (1 − 21 x2 + 24 x + . . . ) + a1 (x − 16 x3 + 120
1
x5 + . . . ).

Back to Tutorial 4 Question 2

c)
y ′′ − y ′ − 2y = 0
We are looking for a solution in the form
∞
y = ∑ an x n .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2

After substituting these formulae in the equation we obtain

∞ ∞ ∞
∑ (n − 1) nan xn−2 − ∑ nan xn−1 − 2 ∑ an xn = 0
n=2 n=1 n=0

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

On adjusting the indices we find

∞ ∞ ∞
∑ (n + 1) (n + 2) an+2 xn − ∑ (n + 1) an+1 xn − 2 ∑ an xn = 0.
n=0 n=0 n=0
7.4. TUTORIAL 4 255

Equating the coefficients at different powers of x we find the recurrent formula

(n + 1) (n + 2) an+2 − (n + 1) an+1 − 2an = 0

or
an+1 n + an+1 + 2an
an+2 = .
(n + 1) (n + 2)
Setting n = 0, 1, . . . in the recurrent formula we get

a2 = 21 a1 + a0 , a3 = 12 a1 + 13 a0 , a4 = 5
24 a1 + 14 a0 , a5 = 11
120 a1
1
+ 12 a0 .

The solution takes the form

y = a0 + a1 x + ( 21 a1 + a0 ) x2 + ( 12 a1 + 13 a0 ) x3
5
+ ( 24 a1 + 14 a0 ) x4 + ( 120
11 1
a1 + 12 a0 ) x 5 + . . .

Back to Tutorial 4 Question 2

Part 2
(1 − x2 ) y ′′ − 2xy ′ + 6y = 0
We are looking for a solution in the form
∞
y = ∑ an x n .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2

After substituting these formulae in the equation we obtain

∞ ∞ ∞
(1 − x2 ) ∑ (n − 1) nan xn−2 − 2x ∑ nan xn−1 + 6 ∑ an xn = 0
n=2 n=1 n=0

or
∞ ∞ ∞ ∞
∑ (n − 1) nan xn−2 − ∑ an n (n − 1) xn − 2 ∑ an nxn + 6 ∑ an xn = 0.
n=2 n=2 n=1 n=0

On adjusting the indices we find

∞ ∞ ∞ ∞
∑ (n + 1) (n + 2) an+2 xn − ∑ an n (n − 1) xn − 2 ∑ an nxn + 6 ∑ an xn = 0.
n=0 n=2 n=1 n=0

Setting n = 0 . . . 1 we obtain the relations

n=0∶ 2a2 + 6a0 = 0

.
n=1∶ 6a3 x + 4a1 x = 0

On solving the above we calculate

a2 = −3a0 , a3 = − 32 a1 .

Equating the coefficients at different powers of x we calculate the recurrent formula

(n + 1) (n + 2) an+2 − an n (n − 1) − 2an n + 6an = 0

256 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

or
an (n2 + n − 6)
an+2 = .
(n + 1) (n + 2)
Setting n = 0, 1, . . . in the recurrent formula we calculate
a2 = −3a0 , a3 = − 32 a1 , a4 = 0, a5 = − 15 a1 , a6 = 0, . . .
The solution takes the form
y = a0 + a1 x − 3a0 x2 − 32 a1 x3 − 15 a1 x5 + . . .
Back to Tutorial 4 Question 2
Part 3
(2 + x2 ) y ′′ − xy ′ + 4y = 0
We are looking for a solution in the form
∞
y = ∑ an x n .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2

After substituting these formulae in the equation we obtain

∞ ∞ ∞
(2 + x2 ) ∑ (n − 1) nan xn−2 − x ∑ nan xn−1 + 4 ∑ an xn = 0
n=2 n=1 n=0
or
∞ ∞ ∞ ∞
2 ∑ (n − 1) nan xn−2 + ∑ an n (n − 1) xn − ∑ an nxn + 4 ∑ an xn = 0.
n=2 n=2 n=1 n=0

On adjusting the indices we find

∞ ∞ ∞ ∞
2 ∑ (n + 1) (n + 2) an+2 xn + ∑ an n (n − 1) xn − ∑ an nxn + 4 ∑ an xn = 0.
n=0 n=2 n=1 n=0

Setting n = 0, 1 we obtain the relations

n=0∶ 4a2 + 4a0 = 0
.
n=1∶ 12a3 x + 3a1 x = 0
On solving the above we get
a2 = −a0 , a3 = − 41 a1 .
Equating the coefficients at different powers of x we obtain the recurrent formula
2 (n + 1) (n + 2) an+2 + an n (n − 1) − an n + 4an = 0
or
an (n2 − 2n + 4)
an+2 = − 21 .
(n + 1) (n + 2)
Setting n = 0, 1, . . . in the recurrent formula we get
a2 = −a0 , a3 = − 14 a1 , a4 = 61 a0 , a5 = 7
160 a1 , . . . .

The solution takes the form

y = a0 + a1 x − a0 x2 − 41 a1 x3 + 61 a0 x4 + 160
7
a1 x 5 + . . .
Back to Tutorial 4 Question 2
7.4. TUTORIAL 4 257

Part 4 Textbook exercise set 8.3 Q 19

y ′ − 2xy = 0

We are looking for a solution in the form

∞
y = ∑ an x n .
n=0

Then the derivative of the solution is

∞
y ′ = ∑ nan xn−1 .
n=1

After substituting these formulae in the equation we obtain

∞ ∞
∑ nan xn−1 − 2x ∑ an xn = 0
n=1 n=0

or ∞ ∞
∑ nan xn−1 − 2 ∑ an xn+1 = 0.
n=1 n=0
On adjusting the indices we obtain
∞ ∞
∑ (n + 1) an+1 xn − 2 ∑ an−1 xn = 0.
n=0 n=1

Setting n = 0 we obtain the relation

n=0∶ a1 = 0 .

On solving the above we find

a1 = 0.
Equating the coefficients at different powers of x we obtain the recurrent formula

(n + 1) an+1 − 2an−1 = 0

or
2an−1
an+1 = .
n+1
Setting n = 1, 2, . . . in the recurrent formula we obtain

a1 = 0, a2 = a0 , a3 = 0, a4 = 21 a0 .

The solution takes the form

y = a0 + a0 x2 + 21 a0 x4 + . . .

Back to Tutorial 4 Question 2

Textbook exercise set 8.3 Q 23

y ′′ − x2 y ′ − xy = 0

We are looking for a solution in the form

∞
y = ∑ an xn .
n=0
258 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2

After substituting these formulae in the equation we obtain

∞ ∞ ∞
∑ (n − 1) nan xn−2 − x2 ∑ nan xn−1 − x ∑ an xn = 0
n=2 n=1 n=0

or ∞ ∞ ∞
∑ (n − 1) nan xn−2 − ∑ an nxn+1 − ∑ an xn+1 = 0.
n=2 n=1 n=0

On adjusting the indices we obtain

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

Setting n = 0 . . . 1 we obtain the relations

n=0∶ 2a2 = 0
.
n=1∶ 6a3 x − a0 x = 0

On solving the above we obtain

a2 = 0, a3 = 61 a0 .

Equating the coefficients at different powers of x we obtain the recurrent formula

(n + 1) (n + 2) an+2 − an−1 (n − 1) − an−1 = 0

or
an−1 n
an+2 = .
(n + 1) (n + 2)
Setting n = 1, 2, . . . in the recurrent formula we get

a2 = 0, a3 = 61 a0 , a4 = 16 a1 , a5 = 0, 16 a0 = 16 a0 .

The solution takes the form

y = a0 + a1 x + 16 a0 x3 + 16 a1 x4 + . . .

Back to Tutorial 4 Question 2

Textbook exercise set 8.3 Q 25

y ′′ + 3xy ′ − y = 0

We are looking for a solution in the form

∞
y = ∑ an x n .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2
7.4. TUTORIAL 4 259

After substituting these formulae in the equation we obtain

∞ ∞ ∞
∑ (n − 1) nan xn−2 + 3x ∑ nan xn−1 − ∑ an xn = 0
n=2 n=1 n=0

or
∞ ∞ ∞
∑ (n − 1) nan xn−2 + 3 ∑ an nxn − ∑ an xn = 0.
n=2 n=1 n=0

On adjusting the indices we obtain

∞ ∞ ∞
∑ (n + 1) (n + 2) an+2 xn + 3 ∑ an nxn − ∑ an xn = 0.
n=0 n=1 n=0

Setting n = 0 we obtain the relation

n=0∶ 2a2 − a0 = 0 .

On solving the above we calculate

a2 = 12 a0 .

Equating the coefficients at different powers of x we calculate the recurrent formula

(n + 1) (n + 2) an+2 + 3an n − an = 0

or
an (3n − 1)
an+2 = − .
(n + 1) (n + 2)
Setting n = 0, 1, . . . in the recurrent formula we find

a2 = 21 a0 , a3 = − 13 a1 , a4 = − 24
5
a0 , a 5 = 2 1
15 a1 , 2 a0 = 12 a0 .

The solution takes the form

y = a0 + a1 x + 21 a0 x2 − 13 a1 x3 − 24
5 2
a0 x4 + 15 a1 x 5 + . . .

Back to Tutorial 4 Question 2

Question 3

a)
4xy ′′ + 2y ′ + y = 0
We are looking for a solution in the form
∞
y = ∑ an xn+c .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ (n + c) an xn+c−1 , y ′′ = ∑ (n + c − 1) (n + c)an xn−2+c .
n=0 n=0
260 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

After substituting these formulae in the equation we obtain

∞ ∞ ∞
4x ∑ (n + c − 1) (n + c)an xn−2+c + 2 ∑ (n + c) an xn+c−1 + ∑ an xn+c = 0
n=0 n=0 n=0

or
∞ ∞ ∞
n+c−1 n+c−1
4 ∑ an (n + c) (n + c − 1) x + 2 ∑ (n + c) an x + ∑ an xn+c = 0.
n=0 n=0 n=0

On adjusting the indices we get

∞ ∞ ∞
4 ∑ an+1 (n + 1 + c) (n + c) xn+c + 2 ∑ (n + 1 + c) an+1 xn+c + ∑ an xn+c = 0.
n=−1 n=−1 n=0

Setting n = −1 and assuming a0 ≠ 0 we obtain the indicial equation

4c (c − 1) + 2c = 0.

On solving the above we get

c = 0, c = 21 .
Equating the coefficients at different powers of x for n > −1 we calculate the recurrent
formula
4an+1 (n + 1 + c) (n + c) + 2 (n + 1 + c) an+1 + an = 0
or
an
an+1 = − 21 .
(n + 1 + c) (2c + 2n + 1)
Firstly, substituting c = 0 in the recurrent formula we obtain
an
an+1 = − 21
(n + 1) (1 + 2n)

The coefficients of the series are then obtained by setting n = 0, 1, . . . in the previous
recurrent formula we obtain

a1 = − 21 a0 , a2 = 1
24 a0 ,
1
a3 = − 720 a0 , a 4 = 1
40320 a0 .

1
Secondly, substituting c = 2 in the recurrent formula we find
an
an+1 = − 21
(2n + 3) (n + 1)

By setting n = 0, 1, . . . in the previous recurrent formula we calculate

a1 = − 61 a0 , a2 = 1
120 a0 ,
1
a3 = − 5040 a0 , a4 = 1
362880 a0 .

The solution takes the form

√
y = A x (1 − 16 x + 120
1 1
x2 − 5040 1
x3 + 362880 x4 − . . . )

+B (1 − 12 x + 24
1 2 1
x − 720 1
x3 + 40320 x4 − . . . )

Back to Tutorial 4 Question 3

7.4. TUTORIAL 4 261

b)
y ′′ + xy ′ + y = 0.
The point x = 0 is an ordinary point of the equation. So, we are looking for a solution
in the form ∞
y = ∑ an x n .
n=0
Then the derivatives of the solution are
∞ ∞
y ′ = ∑ nan xn−1 , y ′′ = ∑ (n − 1) nan xn−2 .
n=1 n=2

After substituting these formulae in the equation we obtain

∞ ∞ ∞
∑ (n − 1) nan xn−2 + x ∑ nan xn−1 + ∑ an xn = 0
n=2 n=1 n=0

or ∞ ∞ ∞
∑ (n − 1) nan xn−2 + ∑ an nxn + ∑ an xn = 0.
n=2 n=1 n=0
On adjusting the indices we calculate
∞ ∞ ∞
∑ (n + 1) (n + 2) an+2 xn + ∑ an nxn + ∑ an xn = 0.
n=0 n=1 n=0

Setting n = 0 we obtain t 2a2 + a0 = 0 implying

a2 = − 21 a0 .

Equating the coefficients at different powers of x we obtain the recurrent formula

(n + 1) (n + 2) an+2 + an n + an = 0

or
an
an+2 = − .
n+2
Setting n = 0, 1, . . . in the recurrent formula we find

a2 = − 21 a0 , a3 = − 13 a1 , a4 = 18 a0 , a5 = 1
15 a1 , . . .

The solution takes the form

y = a0 + a1 x − 21 a0 x2 − 13 a1 x3 + 18 a0 x4 + 15
1
a1 x 5 + . . .

Back to Tutorial 4 Question 3

c)
xy ′′ + y = 0
We are looking for a solution in the form
∞
y = ∑ an xn+c .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ (n + c) an xn+c−1 , y ′′ = ∑ (n + c − 1) (n + c)an xn−2+c .
n=0 n=0
262 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

After substituting these formulae in the equation we obtain

∞ ∞
x ∑ (n + c − 1) (n + c)an xn−2+c + ∑ an xn+c = 0
n=0 n=0

or ∞ ∞
∑ an (n + c − 1) (n + c) xn+c−1 + ∑ an xn+c = 0.
n=0 n=0
On adjusting the indices we obtain
∞ ∞
∑ an+1 (n + 1 + c) (n + c) xn+c + ∑ an xn+c = 0.
n=−1 n=0

Setting n = −1 and assuming a0 ≠ 0 we obtain the indicial equation

c (c − 1) = 0

hence
c = 0, c = 1.
Equating the coefficients at different powers of x for n > −1 we obtain the recurrent
formula
an+1 (n + 1 + c) (n + c) + an = 0
or
an
an+1 = − .
(n + 1 + c) (n + c)
Firstly, substituting c = 0 in the recurrent formula we obtain

(n + 1) nan+1 = −an

which, however, gives a0 = 0 for n = 0 and then an = 0 for all n. Thus c = 0 does not
yield any non-trivial solutions.
Secondly, substituting c = 1 in the recurrent formula we find
an
an+1 = −
(n + 1) (n + 2)
By setting n = 0, 1, . . . in the previous recurrent formula we obtain

a1 = − 21 a0 , a2 = 1
12 a0 ,
1
a3 = − 144 a0 , . . .

Thus we have only one solution

y = Ax (1 − 12 x + 12
1 2 1
x − 144 x3 + . . . )

Back to Tutorial 4 Question 3

d) Solutions to textbook exercises
Set 8.6. Problem 1.

(x2 − 1) y ′′ + xy ′ + 3y = 0
Dividing through by the coefficient at the second derivative: x2 − 1 and then factorising,
we obtain
x 3y
y ′′ + y′ + = 0.
(x − 1) (x + 1) (x − 1) (x + 1)
7.4. TUTORIAL 4 263

Thus we have
x 3
p(x) = , q(x) = .
(x − 1) (x + 1) (x − 1) (x + 1)
The singular points are
x = −1, x = 1.
For x = −1 we need to calculate the limits:
x 3x + 3
lim (x + 1)p(x) = lim = 21 , lim (x + 1)2 q(x) = lim =0
x→−1 x→−1 x − 1 x→−1 x→−1 x − 1

and for x = 1 they are

x 3x − 3
lim(x − 1)p(x) = lim = 21 , lim(x − 1)2 q(x) = lim = 0.
x→1 x→1 x+1 x→1 x→1 x+1
Since both limits for the point x = −1 exist we conclude that this point is a regular
singular point. The point x = 1 is also a regular singular point.
Back to Tutorial 4 Question 3

Set 8.6. Problem 3.

(x2 + 1) y ′′ + 7x2 y ′ − 3xy = 0

Dividing through by the coefficient at the second derivative: x2 + 1 we obtain
7x2 ′ −3xy
y ′′ + y + 2 = 0.
x2 + 1 x +1
Thus we have
7x2 3x
p(x) = 2
, q(x) = − 2 .
x +1 x +1
The singular points are
x = −i, x = i.
For x = −i we need to calculate the limits:
7x2 3x (x + i)
lim (x + i)p(x) = lim = − 72 i, lim (x + i)2 q(x) = lim =0
x→−i x→−i x − i x→−i x→−i −x + i

and for x = i they are

7x2 7 3x (−x + i)
lim(x − i)p(x) = lim = 2 i, lim(x − i)2 q(x) = lim = 0.
x→i x→i x+i x→i x→i x+i
Since both limits for the point x = −i exist we conclude that this point is a regular
singular point. The point x = i is a regular singular point too.
Back to Tutorial 4 Question 3

Set 8.6. Problem 5.

2
(x2 − 1) y ′′ − (x − 1) y ′ + 3y = 0
Dividing through by the coefficient at the second derivative: (x2 − 1)2 and then factoris-
ing, we obtain
−(x − 1) 3y
y ′′ + 2
′
2y + 2 2 =0
(x − 1) (x + 1) (x − 1) (x + 1)
264 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

or
y′ 3y
y ′′ − 2 + 2 2 = 0.
(x − 1) (x + 1) (x − 1) (x + 1)
Thus we have
1 3
p(x) = − 2, q(x) = 2 2.
(x − 1) (x + 1) (x − 1) (x + 1)
The singular points are
x = −1, x = 1.
For x = −1 we need to calculate the limits:
1
lim (x + 1)p(x) = lim − = undef ined,
x→−1 x→−1 (x − 1) (x + 1)

3
lim (x + 1)2 q(x) = lim = 43
x→−1 x→−1 (x − 1)2

and for x = 1 they are

1 3
lim(x − 1)p(x) = lim − 2 = − 41 , lim(x − 1)2 q(x) = lim 2 = 43 .
x→1 x→1 (x + 1) x→1 x→1 (x + 1)
Since one of the limits for the point x = −1 does not exist, we conclude that this point is
an irregular singular point. Since both limits for the point x = 1 exist we conclude that
this point is a regular singular point.
Back to Tutorial 4 Question 3

Set 8.6. Problem 7.

2
(x2 − x − 2) y ′′ + (x2 − 4) y ′ − xy = 0
Dividing through by the coefficient at the second derivative: (x2 − x − 2)2 and then
factorising, we obtain
(x − 2) (x + 2) −xy
y ′′ + 2 2y
′
+ 2 2 =0
(x + 1) (x − 2) (x + 1) (x − 2)
or
(x + 2) y ′ xy
y ′′ + 2 − 2 2 = 0.
(x − 2) (x + 1) (x + 1) (x − 2)
Thus we have
x+2 x
p(x) = 2, q(x) = − 2 2.
(x − 2) (x + 1) (x + 1) (x − 2)
The singular points are
x = −1, x = 2.
For x = −1 we need to calculate the limits:
x+2
lim (x + 1)p(x) = lim = undef ined,
x→−1 x→−1 (x + 1) (x − 2)

x
lim (x + 1)2 q(x) = lim − = 91
x→−1 x→−1 (x − 2)2
7.4. TUTORIAL 4 265

and for x = 2 they are

2+x x
lim(x − 2)p(x) = lim 2 = 94 , lim(x − 2)2 q(x) = lim − 2 = − 92 .
x→2 x→2 (x + 1) x→2 x→2 (x + 1)

Since one of the limits for the point x = −1 does not exist, then we conclude that this
point is an irregular singular point. Since both limits for the point x = 2 exist we
conclude that this point is a regular singular point.
Back to Tutorial 4 Question 3

Set 8.6. Problem 9.

2
(x2 + 2x − 8) y ′′ + (3x + 12) y ′ − x2 y = 0
Dividing through by the coefficient at the second derivative: (x2 + 2x − 8)2 and then
factorising, we obtain

3x + 12 −x2 y
y ′′ + 2 2 y ′
+ 2 2 =0
(4 + x) (x − 2) (4 + x) (x − 2)
or
3y ′ x2 y
y ′′ + 2 − 2 2 = 0.
(4 + x) (x − 2) (4 + x) (x − 2)
Thus we have
3 x2
p(x) = 2 , q(x) = − 2 2.
(4 + x) (x − 2) (4 + x) (x − 2)

The singular points are

x = −4, x = 2.
For x = −4 we need to calculate the limits:
3 1 x2
lim (4 + x)p(x) = lim 2 = lim (4 + x)2 q(x)
12 , x→−4 = lim − 2 = − 94
x→−4 x→−4 (x − 2) x→−4 (x − 2)

and for x = 2 they are

3
lim(x − 2)p(x) = lim = undef ined,
x→2 x→2 (4 + x) (x − 2)

x2
lim(x − 2)2 q(x) = lim − 2 = − 91 .
x→2 x→2 (4 + x)

Since both limits for the point x = −4 exist we conclude that this point is a regular
singular point. Since one of the limits for the point x = 2 does not exist, we conclude
that this point is an irregular singular point.
Back to Tutorial 4 Question 3

Set 8.6. Problem 11.

x2 y ′′ − 2xy ′ − 10y = 0
266 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Dividing through by x2 we obtain

2y ′ 10y
y ′′ − − 2 = 0.
x x
Thus we have
2 10
p(x) = − , q(x) = − 2 .
x x
We need to calculate the limits:

lim xp(x) = lim(−2) = −2, lim x2 q(x) = lim(−10) = −10.

x→0 x→0 x→0 x→0

The indicial equation is c (c − 1) − 2c − 10 = 0 or

c2 − 3c − 10 = 0.

Solving it for c we obtain

c = −2, c = 5.
This implies that a solution is to be sought in either of the forms

1 ∞
y = 2 ∑ an xn
x n=0
or ∞
y = x ∑ an x n .
5
n=0

Back to Tutorial 4 Question 3

Set 8.6. Problem 17.

(x − 1) y ′′ + (x2 − 1) y ′ − 12y = 0
2

Dividing through by the coefficient at the second derivative: (x − 1)2 and then factoris-
ing, we obtain
(x − 1) (x + 1) ′ −12y
y ′′ + 2 y + 2 =0
(x − 1) (x − 1)
or
(x + 1) y ′ 12y
y ′′ + − 2 = 0.
x−1 (x − 1)
Thus we have
x+1 12
p(x) = , q(x) = − 2.
x−1 (x − 1)
We need to calculate the limits:

lim(x − 1)p(x) = lim x + 1 = 2, lim(x − 1)2 q(x) = lim −12 = −12.

x→1 x→1 x→1 x→1

The indicial equation is c (c − 1) + 2c − 12 = 0 or

c2 + c − 12 = 0.

Solving it for c we obtain

c = −4, c = 3.
7.4. TUTORIAL 4 267

Thus a solution is to be sought in either of the forms

∞
1
y= 4 ∑ an (x − 1)n
(x − 1) n=0
or ∞
y = (x − 1)3 ∑ an (x − 1)n .
n=0

Back to Tutorial 4 Question 3

Set 8.6. Problem 19.

9x2 y ′′ + 9x2 y ′ + 2y = 0
We are looking for a solution in the form
∞
y = ∑ an xn+c .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ (n + c) an xn+c−1 , y ′′ = ∑ an (n + c) (n + c − 1) xn−2+c .
n=0 n=0

After substituting these formulae in the equation we obtain

∞ ∞ ∞
9x2 ∑ an (n + c) (n + c − 1) xn−2+c + 9x2 ∑ (n + c) an xn+c−1 + 2 ∑ an xn+c = 0
n=0 n=0 n=0

or
∞ ∞ ∞
9 ∑ an (n + c) (n + c − 1) xn+c + 9 ∑ an (n + c) xn+1+c + 2 ∑ an xn+c = 0.
n=0 n=0 n=0

On adjusting the indices we get

∞ ∞ ∞
n+c n+c
9 ∑ an (n + c) (n + c − 1) x + 9 ∑ an−1 (n + c − 1) x + 2 ∑ an xn+c = 0.
n=0 n=1 n=0

Setting n = 0 and assuming a0 ≠ 0 we obtain the indicial equation 9c (c − 1) + 2 = 0 or

9c2 − 9c + 2 = 0.

On solving the above we find

c = 31 , c = 23 .
Equating the coefficients at different powers of x for n > 0 we find the recurrent formula

9an (n + c) (n + c − 1) + 9an−1 (n + c − 1) + 2an = 0

or
9an−1 (n + c − 1)
an = − .
(3c − 1 + 3n) (3c − 2 + 3n)
1
Firstly, substituting c = 3 in the recurrent formula we calculate

an−1 (3n − 2)
an = −
n (−1 + 3n)
268 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

The coefficients of the series are then obtained by setting n = 1, 2, . . . in the previous
recurrent formula we get

a1 = − 12 a0 , a2 = 51 a0 , a3 = − 120
7
a0 , a4 = 7
528 a0 , . . .

2
Secondly, substituting c = 3 in the recurrent formula we get

an−1 (−1 + 3n)

an = −
(1 + 3n) n
By setting n = 1, 2, . . . in the previous recurrent formula we calculate

a1 = − 21 a0 , a2 = 5
28 a0 ,
1
a3 = − 21 a0 , a 4 = 11
1092 a0 , . . .

The solution takes the form

1
y = Ax 3 (1 − 12 x + 15 x2 − 120
7 7
x3 + 528 x4 − . . . )
2
+Bx 3 (1 − 21 x + 28
5 2 1 3
x − 21 11
x + 1092 x4 − . . . )

Back to Tutorial 4 Question 3

Set 8.6. Problem 21.

x2 y ′′ + xy ′ + x2 y = 0
We are looking for a solution in the form
∞
y = ∑ an xn+c .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ (n + c) an xn+c−1 , y ′′ = ∑ an (n + c) (n + c − 1) xn−2+c .
n=0 n=0

After substituting these formulae in the equation we obtain

∞ ∞ ∞
x2 ∑ an (n + c) (n + c − 1) xn−2+c + x ∑ (n + c) an xn+c−1 + x2 ∑ an xn+c = 0
n=0 n=0 n=0

or ∞ ∞ ∞
∑ an (n + c) (n + c − 1) xn+c + ∑ an (n + c) xn+c + ∑ an xn+2+c = 0.
n=0 n=0 n=0

On adjusting the indices we find

∞ ∞ ∞
∑ an (n + c) (n + c − 1) xn+c + ∑ an (n + c) xn+c + ∑ an−2 xn+c = 0.
n=0 n=0 n=2

Setting n = 0 and assuming a0 ≠ 0 we obtain the indicial equation c (c − 1) + c = 0 or

c2 = 0.

Hence
c = 0.
7.4. TUTORIAL 4 269

Equating the coefficients at different powers of x for n > 0 we get the recurrent formula

an n (n − 1) + an n + an−2 = 0

or
an−2
an = − .
n2
The coefficients of the series are then obtained by setting n = 1, 2, 3, . . . in the previous
recurrent formula we find

a1 = 0, a2 = − 41 a0 , a3 = 0, a4 = 1
64 a0 ,
1
a5 = 0, a6 = − 2304 ,....

The solution takes the form

y = a0 (1 − 14 x2 + 64
1 4 1
x − 2304 x6 + . . . ) .

Back to Tutorial 4 Question 3

Set 8.6. Problem 23.

x2 y ′′ + (x2 + x) y ′ − y = 0
We are looking for a solution in the form
∞
y = ∑ an xn+c .
n=0

Then the derivatives of the solution are

∞ ∞
y ′ = ∑ (n + c) an xn+c−1 , y ′′ = ∑ an (n + c) (n + c − 1) xn−2+c .
n=0 n=0

After substituting these formulae in the equation we obtain

∞ ∞ ∞
x2 ∑ an (n + c) (n + c − 1) xn−2+c + (x2 + x) ∑ (n + c) an xn+c−1 − ∑ an xn+c = 0
n=0 n=0 n=0

or
∞ ∞ ∞ ∞
∑ an (n + c) (n + c − 1) xn+c + ∑ an (n + c) xn+c + ∑ an (n + c) xn+1+c − ∑ an xn+c = 0.
n=0 n=0 n=0 n=0

On adjusting the indices we obtain

∞ ∞ ∞ ∞
∑ an (n + c) (n + c − 1) xn+c + ∑ an (n + c) xn+c + ∑ an−1 (n + c − 1) xn+c − ∑ an xn+c = 0.
n=0 n=0 n=1 n=0

Setting n = 0 and assuming a0 ≠ 0 we obtain the indicial equation c (c − 1) + c − 1 = 0 or

c2 − 1 = 0.

On solving the above we calculate

c = −1, c = 1.
270 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Equating the coefficients at different powers of x for n > 0 we obtain the recurrent
formula
an (n + c) (n + c − 1) + an (n + c) + an−1 (n + c − 1) − an = 0
or
an−1
.
an = −
n+1+c
Firstly, substituting c = −1 in the recurrent formula we find
an−1
an = −
n
The coefficients of the series are then obtained by setting n = 1, 2, . . . in the previous
recurrent formula we get
(−1)n
a1 = −a0 , a2 = 21 a0 , a3 = − 16 a0 , a4 = 1
24 a0 , . . . , an = n! .

Secondly, substituting c = 1 in the recurrent formula we obtain

an−1
an = −
n+2
By setting n = 1, 2, . . . in the previous recurrent formula we obtain

a1 = − 13 a0 , a2 = 1
12 a0 ,
1
a3 = − 60 a0 , a 4 = 1
360 a0 .

The solution takes the form

y = Ax (1 − 13 x + 12
1 2 1 3
x − 60 1
x + 360 x4 − . . . )

+Bx−1 (1 − x + 21 x2 − 16 x3 + 24
1 4
x − ...)

Back to Tutorial 4 Question 3

7.5 Tutorial 6
Question 1

Problem 2.i.
We calculate the indefinite integral (by parts):
RRR u = 3t + 4 u′ = 3 RRR
−ts
dt = RRRRR RRR = − (3t + 4)e−ts + 3 −ts
∫ (3t + 4) e RRR s∫ e dt
RRR dv = e−ts dt v = − e s s
−ts
RR
(3t + 4)e−ts 3e−ts
=− − 2 .
s s
Note that limt→∞ te−ts = 0. Thus we have
∞
∞ 4 3 3t 4 3 4 3
−ts −ts
∫0 (3t + 4) e dt = e (− − 2 − ) ∣ = 0 − (− − 2 ) = + 2 .
s s s 0 s s s s
7.5. TUTORIAL 6 271

Problem 2.ii
We have
at+b −ts b (a−s)t eb (a−s)t
∫ e e dt = e ∫ e dt =
a−s
e

hence ∞
∞ eb (a−s)t eb eb
at+b −ts
∫0 e e dt = e ∣ =0− =
a−s 0
a−s s−a
provided a − s < 0.

Problem 2.iii
This question involves two integrations by parts:
RRR u = cos (at) u′ = − sin (at) a RRR
cos −ts
e dt RRR RRR = − cos (at) e−ts − a −ts
∫ (at) = RRR −ts e−ts RRR s s ∫ e sin (at) dt
RR dv = e dt v = − s RR
RRR u = sin (at) u′ = cos (at) a RRR
= RRRRR RRR = − cos (at) e−ts + sin (at) ae−ts − a2 cos (at) e−ts dt.
e−ts a e−ts a R
R s s 2 s 2 ∫
RRR dv = − s dt v = s2 RRR

Thus we have

−ts a2 −ts cos (at) e−ts sin (at) ae−ts

∫ cos (at) e dt + s2 ∫ cos (at) e dt = − s
+
s2
or
s 2 + a2 −ts cos (at) e−ts sin (at) ae−ts
cos (at) e dt = − + .
s2 ∫ s s2
Note that
lim e−ts cos(at) = lim e−ts sin(at) = 0
t→∞ t→∞

hence
∞
s 2 + a2 ∞ −ts cos (at) e−ts sin (at) ae−ts 1
∫ cos (at) e dt = (− + ) ∣ = .
s2 0 s s2 0
s

Finally we obtain
∞ s
∫0 cos (at) e−ts dt = .
s2 + a2

Back to Tutorial 6 Question 1

Question 2

1. i.
1
.
s (s + 1)
For this fraction the decomposition has the form

1 A B
= + .
s (s + 1) s s + 1
272 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Multiplying by the common denominator and collecting the coefficients at s, we

obtain
1 = (A + B) s + A
Comparing the coefficients we find the system A = 1, A + B = 0 hence

A = 1, B = −1.

The required partial fractions decomposition is given by

1 1 1
= − .
s (s + 1) s s + 1

Hence we have
f (t) = −e−t + 1.
Back to Tutorial 6 Question 2
ii.
1
.
(s − 4) (s − 1)
We’re looking for a decomposition of the form
1 A B
= + .
(s − 4) (s − 1) s − 1 s − 4

Multiplying by the common denominator and collecting the coefficients at s, we

obtain
1 = (A + B) s − 4A − B
which implies that
−4A − B = 1, A + B = 0.
Solving this system we get
A = − 31 , B = 31 .
Thus the required partial fractions decomposition is given by
1 1 1
= − 31 + 31 .
(s − 4) (s − 1) s−1 s−4

Hence we have
f (t) = 31 e4t − 13 et .
Back to Tutorial 6 Question 2
iii.
3s
.
s2 + 2s − 8
For this fraction the decomposition has the form
3s A B
= + .
(s + 4) (s − 2) s + 4 s − 2

Multiplying by the common denominator and collecting the coefficients at s, we

find
3s = (A + B) s − 2A + 4B
7.5. TUTORIAL 6 273

which gives the system

−2A + 4B = 0, A + B = 3

whose solution is
A = 2, B = 1.
The partial fractions decomposition is given by
3s 2 1
= + .
(s + 4) (s − 2) s + 4 s − 2
Hence we have
f (t) = e2t + 2e−4t .
Back to Tutorial 6 Question 2
iv.
s2 − 6s + 4
.
s (s2 − 3s + 2)
For this fraction the decomposition has the form
s2 − 6s + 4 A B C
= + + .
s (s − 1) (s − 2) s − 1 s s − 2
Multiplying by the common denominator and collecting the coefficients at s, we
obtain
s2 − 6s + 4 = (A + C + B) s2 + (−2A − C − 3B) s + 2B
Comparing the coefficients we find the system

2B = 4, −2A − C − 3B = −6, A + C + B = 1.

Solving this system we calculate

A = 1, B = 2, C = −2.

Thus the required partial fractions decomposition is

s2 − 6s + 4 1 2 2
= + − .
s (s − 1) (s − 2) s − 1 s s − 2
Hence we have
f (t) = −2e2t + 2 + et .
Back to Tutorial 6 Question 2

2. i. Use the property of the Laplace transform: L(t2 +at+b) = L(t2 )+aL(t)+bL(1).
ii. Use the identity

cos(ωt + θ) = cos (ωt) cos (θ) − sin (ωt) sin (θ) .

iii. Use the identity

1 + cos(2ωt)
cos2 (ωt) = .
2
Back to Tutorial 6 Question 2
274 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

3. Denote f = sin2 (ωt). We have f (0) = sin2 (0) = 0 hence

L(f ′ ) = sL(f )

or
L(f ′ )
L(f ) = .
s
Further, we calculate the derivative

f ′ = 2ω sin(ωt) cos(ωt) = ω sin(2ωt)

and then apply the Laplace transform to get

2ω 2
L(f ′ ) = .
s2 + 4ω 2
Hence
2ω 2
L(f ) = .
s(s2 + 4ω 2 )
Back to Tutorial 6 Question 2

4. Denote f = t cos t. We have

f ′ = cos (t) − t sin (t) , f ′′ = −2 sin (t) − t cos (t)

which implies f (0) = 0, f ′ (0) = 1. The formula L(f ′′ ) = s2 L(f ) − sf (0) − f ′ (0)
then becomes
L(f ′′ ) = s2 L(f ) − 1.
On the other hand
2
L(f ′′ ) = −2L(sin t) − L(t cos t) = − − L(f ).
s2 +1
From the previous we get
2
s2 L(f ) − 1 = − − L(f ).
s2 + 1
Solving for L(f ) we obtain
s2 − 1
L(f ) = .
(s2 + 1)2
Back to Tutorial 6 Question 2
7.5. TUTORIAL 6 275

5. Textbook exercise set 7.6. Problem 51

(a) We use the formula
Γ(n + 1)
L(tn ) = .
sn+1
We have
1 Γ( 1 )
L(t− 2 ) = √2 .
s
It is given that √
Γ ( 12 ) = π
hence √
1
−2 π
L(t )= .
s
Back to Tutorial 6 Question 2
(b) We have
7 Γ( 9 )
L(t 2 ) = √ 2 .
s9
We will use the multiplication formula:
Γ(x + 1) = xΓ(x)
to calculate Γ( 29 ). We have
√
Γ ( 29 ) = 72 Γ ( 72 ) = 35
(5)
4 Γ 2 = 105
(3)
8 Γ 2 = 105
(1)
16 Γ 2 = 105
16 π.
The answer is therefore √
7
105 π
L(t 2 ) = 16 .
s9
Back to Tutorial 6 Question 2

Question 3

1.
4y ′′ + π 2 y = 0, y(0) = 0, y ′ (0) = 1.
Applying the Laplace transform to the equation, we obtain
4s2 L (y) − 4y ′ (0) − 4sy (0) + π 2 L (y) = 0.
On substitution of the initial conditions, we get
(4s2 + π 2 ) L (y) − 4 = 0.
Solving for L(y) gives
π
4 2 2
L (y) = 2 2
= π 2.
4s + π s + ( π2 )
2

Applying the inverse Laplace transform to the right hand side, we obtain the solution
2
y= sin ( 12 πx) .
π
Back to Tutorial 6 Question 3
276 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

2.
y ′′ + 5y ′ + 6y = 0, y(0) = 0, y ′ (0) = 1.
Applying the Laplace transform to the equation, we obtain

s2 L (y) − y ′ (0) − sy (0) + 5sL (y) − 5y (0) + 6L (y) = 0.

On substitution of the initial conditions, we get

(5s + s2 + 6) L (y) − 1 = 0.

Solving for L(y) gives

1
L (y) =
(s + 3) (s + 2)
For this fraction the decomposition has the form
1 A B
= + .
(s + 3) (s + 2) s + 2 s + 3
Multiplying by the common denominator and collecting the coefficients at s, we
calculate
1 = (A + B) s + 3A + 2B
Comparing the coefficients we get the system

3A + 2B = 1, A + B = 0.

Solving this system we obtain

A = 1, B = −1.

Thus the required partial fractions decomposition has the form

1 1 1
= − .
(s + 3) (s + 2) s + 2 s + 3
Applying the inverse Laplace transform to the right hand side, we obtain the solution

y = e−2t − e−3t .

Back to Tutorial 6 Question 3

3.
y ′′ + 25y = x, y(0) = 1, y ′ (0) = 1
25 .

Applying the Laplace transform to the equation, we obtain

1
s2 L (y) − y ′ (0) − sy (0) + 25L (y) = .
s2
On substitution of the initial conditions, we get

1 1
(s2 + 25) L (y) − 25 −s= .
s2
Solving for L(y) gives
2
+ 25s3
1 25 + s
L (y) = 25 2 2
s (s + 25)
7.5. TUTORIAL 6 277

For this fraction the decomposition has the form

25 + s2 + 25s3 A + Bs C
= 2 + .
s2 (s2 + 25) s + 25 s2

Multiplying by the common denominator and collecting the coefficients at s, we

calculate
25 + s2 + 25s3 = s3 B + (A + C) s2 + 25C
Comparing the coefficients we obtain the system

25C = 25, A + C = 1, B = 25

which gives
A = 0, B = 25, C = 1.
Thus the required partial fractions decomposition is
2
+ 25s3
1 25 + s s 1 1
25 2 2
= 2 + 25 .
s (s + 25) s + 25 s2

Applying the inverse Laplace transform to the right hand side, we obtain the solution
1
y = cos (5t) + 25 t.

Back to Tutorial 6 Question 3

4. Textbook Exercise Set 7.5 Problem 5

y ′′ + y = t2 + 2, y(0) = 1, y ′ (0) = −1.

Applying the Laplace transform to the equation, we obtain

2 2
s2 L (y) − y ′ (0) − sy (0) + L (y) = + .
s3 s
On substitution of the initial conditions, we get
2 2
(s2 + 1) L (y) + 1 − s = + .
s3 s
Solving for L(y) gives
2 + 2s2 − s3 + s4
L (y) =
s3 (s2 + 1)
For this fraction the decomposition has the form

2 + 2s2 − s3 + s4 A + Bs C
= 2 + .
s3 (s2 + 1) s + 1 s3

After multiplying by the common denominator we get

2 + 2s2 − s3 + s4 = s3 A + s4 B + Cs2 + C

Comparing the coefficients we find

A = −1, B = 1, C = 2.
278 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

Thus the required partial fractions decomposition is given by

2 + 2s2 − s3 + s4 s − 1 2
3 2
= 2 + 3.
s (s + 1) s +1 s

Applying the inverse Laplace transform to the right hand side, we obtain the solution

y = cos (t) − sin (t) + t2 .

Back to Tutorial 6 Question 3

5. Textbook Exercise Set 7.5 Problem 7

y ′′ − 7y ′ + 10y = 9 cos (t) + 7 sin (t) , y(0) = 5, y ′ (0) = −4.

Applying the Laplace transform to the equation, we obtain

9s + 7
s2 L (y) − y ′ (0) − sy (0) − 7sL (y) + 7y (0) + 10L (y) = .
s2 + 1

On substitution of the initial conditions, we get

9s + 7
(s2 − 7s + 10) L (y) + 39 − 5s = .
s2 + 1

Solving for L(y) gives

14s − 32 − 39s2 + 5s3

L (y) =
(s2 + 1) (s − 2) (s − 5)

For this fraction the decomposition has the form

14s − 32 − 39s2 + 5s3 A B C + sK

2
= + + 2 .
(s + 1) (s − 2) (s − 5) s − 5 s − 2 s +1

Multiplying by the common denominator and collecting the coefficients at s, we

calculate

14s − 32 − 39s2 + 5s3 = (K + B + A) s3 + (C − 2A − 7K − 5B) s2

+ (10K − 7C + A + B) s + 10C − 2A − 5B.

Comparing the coefficients we calculate the system

A+B+K = 5, −2A−5B+10C = −32, A+B−7C+10K = 14, −2A−5B+C−7K = −39.

7.5. TUTORIAL 6 279

We will solve this system using Gaussian elimination. We have

⎛ 1 1 0 1 5 ⎞ R2′ =R2 +2R1 ⎛ 1 1 0 1 5 ⎞

⎜ −2 −5 10 0 −32 ⎟ R3 =R3 −R1 ⎜ 0 −3 10 2 −22 ⎟
′
⎜ ⎟ ⎜ ⎟
⎜ 1 1 −7 10 14 ⎟ R4′ =R4 +2R1 ⎜ 0 0 −7 9 9 ⎟
⎝ −2 −5 1 −7 −39 ⎠ Ð→ ⎝ 0 −3 1 −5 −29 ⎠
⎛ 1 1 0 1 5 ⎞ ⎛ 1 1 0 1 5 ⎞
R4′ =R4 −R2 ⎜ 0 −3 10 2 −22 ⎟ R4′ =R4 −R3 ⎜ 0 −3 10 2 −22 ⎟
⎜ ⎟ ⎜ ⎟
Ð→ ⎜ 0 0 −7 9 9 ⎟ Ð→ ⎜ 0 0 −7 9 9 ⎟
⎝ 0 0 −9 −7 −7 ⎠ ⎝ 0 0 −2 −16 −16 ⎠
⎛ 1 1 0 1 5 ⎞ ⎛ 1 1 0 1 5 ⎞
1
R4 =− 2 R4 ⎜ 0 −3 10 2 −22 ⎟ R3 ↔R4 ⎜ 0 −3 10 2 −22 ⎟
′
⎜ ⎟ ⎜ ⎟
Ð→ ⎜ 0 0 −7 9 9 ⎟ Ð→ ⎜ 0 0 1 8 8 ⎟
⎝ 0 0 1 8 8 ⎠ ⎝ 0 0 −7 9 9 ⎠
⎛ 1 1 0 1 5 ⎞ ⎛ 1 1 0 1 5 ⎞
1
R4′ =R4 +7R3 ⎜ 0 −3 10 2 −22 ⎟ R4′ = 65 R4 ⎜ 0 −3 10 2 −22 ⎟
⎜ ⎟ ⎜ ⎟
Ð→ ⎜ 0 0 1 8 8 ⎟ Ð→ ⎜ 0 0 1 8 8 ⎟
⎝ 0 0 0 65 65 ⎠ ⎝ 0 0 0 1 1 ⎠
R3 =R3 −8R4 ⎛ 1
′ 1 0 0 4 ⎞ ⎛ 1 1 0 0 4 ⎞
R2 =R2 −2R4 ⎜ 0 −3 10 0 −24 ⎟ R2 =R2 −10R3 ⎜ 0 −3 0 0 −24 ⎟
′ ′
⎜ ⎟ ⎜ ⎟
R1′ =R1 −R4 ⎜ 0 0 1 0 0 ⎟ Ð→ ⎜ 0 0 1 0 0 ⎟
Ð→ ⎝ 0 0 0 1 1 ⎠ ⎝ 0 0 0 1 1 ⎠
⎛ 1 1 0 0 4 ⎞ ⎛ 1 0 0 0 −4 ⎞
1
R2′ =− 3 R2 ⎜ 0 1 0 0 8 ⎟ R1′ =R1 −R2 ⎜ 0 1 0 0 8 ⎟
⎜ ⎟ ⎜ ⎟
Ð→ ⎜ 0 0 1 0 0 ⎟ Ð→ ⎜ 0 0 1 0 0 ⎟
⎝ 0 0 0 1 1 ⎠ ⎝ 0 0 0 1 1 ⎠

Thus the solutions is

A = −4, B = 8, C = 0, K = 1.
Thus the required partial fractions decomposition is given by

14s − 32 − 39s2 + 5s3 4 8 s

= − + + .
(s2 + 1) (s − 2) (s − 5) s − 5 s − 2 s2 + 1

Applying the inverse Laplace transform to the right hand side, we obtain the solution

y = −4e5t + 8e2t + cos (t) .

Back to Tutorial 6 Question 3

6. Textbook Exercise Set 7.5 Problem 11

y ′′ − y = t − 2, y(2) = 3, y ′ (2) = 0.

First we make the substitution t − 2 = x. This brings the problem to the form

y ′′ − y = x, y(0) = 3, y ′ (0) = 0.

Applying the Laplace transform to the equation, we obtain

1
s2 L (y) − y ′ (0) − sy (0) − L (y) = .
s2
280 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

On substitution of the initial conditions, we get

1
(s2 − 1) L (y) − 3s = .
s2
Solving for L(y) gives
1 + 3s3
L (y) =
s2 (s − 1) (s + 1)
For this fraction the decomposition has the form

1 + 3s3 A B C
2
= + 2+ .
s (s − 1) (s + 1) s − 1 s s+1

Multiplying by the common denominator and collecting the coefficients at s, we

calculate
1 + 3s3 = (A + C) s3 + (A + B − C) s2 − B
Comparing the coefficients we obtain

B = −1, A + B − C = 0, A + C = 3

which yields
A = 2, B = −1, C = 1.
Thus the required partial fractions decomposition has the form

1 + 3s3 2 1 1
2
= − 2+ .
s (s − 1) (s + 1) s − 1 s s+1

Applying the inverse Laplace transform to the right hand side, we obtain the solution

y = 3 cosh (x) + sinh (x) − x.

The solution for the original equation is obtained by substituting x = t − 2:

y = 3 cosh (t − 2) + sinh (t − 2) − t + 2.

Back to Tutorial 6 Question 3

7. Textbook Exercise Set 7.5 Problem 13

y ′′ − y ′ − 2y = −8 cos (t) − 2 sin (t) , y(π/2) = 1, y ′ (π/2) = 0.

On making the substitution t − π2 = x we bring the problem to the form

y ′′ − y ′ − 2y = 8 sin (x) − 2 cos (x) , y(0) = 1, y ′ (0) = 0.

Applying the Laplace transform to the equation, we obtain

−8 + 2s
s2 L (y) − y ′ (0) − sy (0) − sL (y) + y (0) − 2L (y) = − .
s2 + 1
On substitution of the initial conditions, we get
−8 + 2s
(−s + s2 − 2) L (y) + 1 − s = − .
s2 + 1
7.5. TUTORIAL 6 281

Solving for L(y) gives

7 − s − s2 + s3
L (y) =
(s2 + 1) (s + 1) (s − 2)

For this fraction the decomposition has the form

7 − s − s2 + s3 A B + sC K
2
= + 2 + .
(s + 1) (s + 1) (s − 2) s − 2 s +1 s+1

Multiplying by the common denominator and collecting the coefficients at s, we

calculate

7−s−s2 +s3 = (A + K + C) s3 +(−2K + A + B − C) s2 +(K + A − B − 2C) s+A−2K−2B

Comparing the coefficients we calculate the system

A − 2K − 2B = 7, K + A − B − 2C = −1, −2K + A + B − C = −1, A + K + C = 1.

Solving this system we find

K = −1, A = 53 , B = − 11 7
5 , C = 5.

Thus the required partial fractions decomposition has the form

7 − s − s2 + s3 1 7s − 11 1
2
= 35 + 15 2 − .
(s + 1) (s + 1) (s − 2) s−2 s +1 s+1

Applying the inverse Laplace transform to the right hand side, we obtain the solution
−x
y = 53 e2x + 75 cos (x) − 11
5 sin (x) − e .

Recalling that x = t − π2 we find the solution to the original problem

π π
y = 53 e2(t− 2 ) + 75 cos (t − π2 ) − 11
5 sin
(t − π2 ) − e−(t− 2 )
π
−t+ 2
= 35 e2t−π + 75 sin t + 11
5 cos t − e .

Back to Tutorial 6 Question 3

Question 4

1. We have
4 2 2
= 2 .
s3 − 4s s s − 4
We know that
2
L−1 ( ) = sinh 2t
s2 −4
hence
t
2 2 t
−1
L ( 2 ) = 2 ∫ sinh 2τ dτ = cosh 2τ ∣ = cosh 2t − 1.
ss −4 0
0
282 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

2. We have
1 1 1
= .
s2 + as s s + a
We know that
1
L−1 ( ) = e−at
s+a
hence
t
1 1 t 1 1
−1
L ( ) = ∫ e−aτ dτ = − e−aτ ∣ = − (e−at − 1).
ss+a 0 a 0
a

3. We have
8 2 4
= 3 .
s4 − 4s 2 s s − 4s
Previously we’ve calculated that

4
L−1 ( ) = cosh 2t − 1
s3 − 4s
hence
t
2 4 t
−1
L ( 3 ) = ∫ (2 cosh 2τ − 2)dτ = (sinh 2τ − 2τ )∣ = sinh 2t − 2t.
s s − 4s 0
0

Back to Tutorial 6 Question 4

Question 5

1. i. We have
2
L(t2 ) =
.
s3
Applying the s-shifting theorem we obtain

2
L(t2 e−2t ) =
(s + 2)3

ii. We have
s
L(cos t) = .
s2 +1
Applying the s-shifting theorem we obtain

s+1
L(e−t cos t) = .
(s + 1)2 + 1

iii. We have
3
L(sin 3t) = .
s2 +9
Applying the s-shifting theorem we obtain

3
L(e−2 sin 3t) = .
(s + 2)2 + 9
7.5. TUTORIAL 6 283

iv. We have
6
L(t3 ) =
.
s4
Applying the s-shifting theorem we obtain
12
L(2t3 e−t/2 ) =
(s + 21 )4
Back to Tutorial 6 Question 5
2. i. We know that
2
L−1 ( 21 3
) = 12 t2
s
hence
2
L−1 ( 12 ) = 21 t2 e−t/2 .
(s + 12 )3
ii. We first complete the square
1 1
= .
s2 + 2s + 5 (s + 1)2 + 4
We know that
1
L−1 ( ) = 12 sin 2t
s2 +4
hence
1
L−1 ( ) = 21 e−t sin 2t.
(s + 1)2 + 4
iii. We first complete the square
6 3×2
= .
s2 − 4s − 5 (s − 2)2 − 9
We know that
3
L−1 ( ) = sinh 3t
s2 −9
hence
3×2
L−1 ( ) = 2e2t sinh 3t.
(s − 2)2 − 9
Back to Tutorial 6 Question 5
3. i.
y ′′ − 4y ′ + 5y = 0, y(0) = 1, y ′ (0) = 2.
Applying the Laplace transform to the equation, we obtain
s2 L (y) − y ′ (0) − sy (0) − 4sL (y) + 4y (0) + 5L (y) = 0.
On substitution of the initial conditions, we get
(−4s + s2 + 5) L (y) + 2 − s = 0.
Solving for L(y) gives
s−2 s−2
L (y) = =
s2
− 4s + 5 (s − 2)2 + 1
Applying the inverse Laplace transform to the right hand side and using the s-
shifting theorem, we obtain the solution
y = e2t cos (t) .
284 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS

ii.
9y ′′ − 6y ′ + y = 0, y(0) = 3, y ′ (0) = 1.

Applying the Laplace transform to the equation, we obtain

9s2 L (y) − 9y ′ (0) − 9sy (0) − 6sL (y) + 6y (0) + L (y) = 0.

On substitution of the initial conditions, we get

(−6s + 9s2 + 1) L (y) + 9 − 27s = 0.

Solving for L(y) gives

9 3
L (y) = =
3s − 1 s − 31

Applying the inverse Laplace transform to the right hand side, we obtain the
solution
1
y = 3e 3 t .

iii.
y ′′ + 2y ′ + 5y = 9 cosh (2t) + 4 sinh (2t) , y(0) = 1, y ′ (0) = 2.

Applying the Laplace transform to the equation, we obtain

9s + 8
s2 L (y) − y ′ (0) − sy (0) + 2sL (y) − 2y (0) + 5L (y) = .
s2 − 4

On substitution of the initial conditions, we get

9s + 8
(2s + s2 + 5) L (y) − 4 − s = .
s2 − 4

Solving for L(y) gives

5s − 8 + 4s2 + s3 5s − 8 + 4s2 + s3
L (y) = =
(s − 2) (s + 2) (2s + s2 + 5) (s − 2) (s + 2) ((s + 1)2 + 4)

For this fraction the decomposition has the form

5s − 8 + 4s2 + s3 A B + sC K
2
= + 2
+ .
(s − 2) (s + 2) ((s + 1) + 4) s − 2 (s + 1) + 4 s + 2

Multiplying by the common denominator and collecting the coefficients at s, we

calculate

s3 +4s2 +5s−8 = (C + A + K) s3 +(B + 4A) s2 +(9A + K − 4C) s−4B−10K+10A

Comparing the coefficients we find the system

A + C + K = 1, 4A + B = 4, 9A − 4C + K = 5, 10A − 4B − 10K = −8.

7.5. TUTORIAL 6 285

We solve this system using Gaussian elimination:

⎛ 1 0 1 1 1 ⎞ R2′ =R2 −4R1 ⎛ 1 0 1 1 1 ⎞

⎜ 4 1 0 0 4 ⎟ R3 =R3 −9R1 ⎜ 0 1 −4
′ −4 0 ⎟
⎜ ⎟ ⎜ ⎟
⎜ 9 0 −4 1 5 ⎟ R4′ =R4 −10R1 ⎜ 0 0 −13 −8 −4 ⎟
⎝ 10 −4 0 −10 −8 ⎠ Ð→ ⎝ 0 −4 −10 −20 −18 ⎠
⎛ 1 0 1 1 1 ⎞ ⎛ 1 0 1 1 1 ⎞
R4′ =R4 +4R2 ⎜ 0 1 −4 −4 0 ⎟ R4′ =R4 −2R3 ⎜ 0 1 −4 −4 0 ⎟
⎜ ⎟ ⎜ ⎟
Ð→ ⎜ 0 0 −13 −8 −4 ⎟ Ð→ ⎜ 0 0 −13 −8 −4 ⎟
⎝ 0 0 −26 −36 −18 ⎠ ⎝ 0 0 0 −20 −10 ⎠
⎛ 1 0 1 1 1 ⎞ R3 =R3 +8R4 ⎛ 1 0
′ 1 0 12 ⎞
1
R4 =− 20 R4 ⎜ 0 1
′ −4 −4 0 ⎟ R2′ =R2 +4R4 ⎜ 0 1 −4 0 2 ⎟
⎜ ⎟ ⎜ ⎟
Ð→ ⎜ 0 0 −13 −8 −4 ⎟ R1′ =R1 −R4 ⎜ 0 0 −13 0 0 ⎟
⎝ 0 0 0 1 12 ⎠ Ð→ ⎝ 0 0 0 1 12 ⎠
1
⎛ 1 0 1 0 2 ⎞ R′ =R2 +4R3 ⎛ 1 0 0 0 12 ⎞
1 2
R3′ =− 13 R3 ⎜ 0 1 −4 0 2 ⎟ ⎜ 0 1 0 0 2 ⎟
⎜ ⎟ R1′ =R1 −R3 ⎜ ⎟
Ð→ ⎜ 0 0 1 0 0 ⎟ ⎜ 0 0 1 0 0 ⎟
⎝ 0 0 0 1 1 ⎠ Ð→
2
⎝ 0 0 0 1 1 ⎠
2

Thus the solution is

A = 12 , B = 2, C = 0, K = 1
2

hence the required partial fractions decomposition is given by

5s − 8 + 4s2 + s3 1 1 2
2
= 21 + 21 + .
(s − 2) (s + 2) ((s + 1) + 4) s+2 s − 2 (s + 1)2 + 4

Applying the inverse Laplace transform to the right hand side, we obtain the
solution
y = cosh (2t) + e−t sin (2t) .

Back to Tutorial 6 Question 5

286 CHAPTER 7. SOLUTIONS TO SELECTED TUTORIAL QUESTIONS
Chapter 8

Conclusion

Where do we go from here?

In this subject we have considered the solution of Differential Equations. Of course, there
are many other applications of integration (and of differentiation), some of which will be
considered in other subjects. Series are also used as a theoretical and an applied tool in
Complex Analysis (MTH307) and in Numerical Methods (MTH203).
In this subject we have also introduced the vast and important topic of Differential Equa-
tions. It is important theoretically because of the powerful tools that have been developed
to solve general problems and because of the as yet unsolved problems that still challenge
mathematicians. It is important practically because so many real-life situations in a diverse
range of areas have as their mathematical model, a Differential Equation which must be
solved and then have its solution interpreted properly. We have focussed on linear D.E.s
in this subject because they occur so frequently as models and because they require such
intriguing and varied methods of solution.
There are, of course, many problems that remain:
What other types of D.E.s can we classify and how do we solve them? How can Numer-
ical Analysis help us solve a whole range of problems by numerical means? What other
applications of D.E.s are there, both known and unknown? What are Partial Differential
Equations? How do we classify them? How do we solve them? What applications do they
have? Etcetera, etcetera, etcetera! In other words, like most subjects in mathematics, no
matter what their level, this has been only a beginning.
In brief, many exciting, elegant and challenging mathematical moments lie ahead!

287