Teaching Manual: Calculus I

Edwin Rutalebwa (PhD) & Narsis Mtega (PhD)

Department of General Studies
Dar es Salaam Institute of Technology

2012
Contents

Preface 1

1 Theory of Limits 1
1.1 Definitions of a Limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 An informal definition of limit . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 The formal definition of a limit . . . . . . . . . . . . . . . . . . . . 4
1.2 Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2 Laplace Transforms 14
2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Laplace Transforms of Some Standard Functions . . . . . . . . . . . . . . . 15
2.2.1 Laplace transform of a constant . . . . . . . . . . . . . . . . . . . . 15
2.2.2 Laplace transform of eat . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2.3 Laplace transform of sinh at . . . . . . . . . . . . . . . . . . . . . . 16
2.2.4 Laplace transform of cosh at . . . . . . . . . . . . . . . . . . . . . . 16
2.2.5 Laplace transform of sin at and cos at . . . . . . . . . . . . . . . . . 17
2.2.6 Laplace transform of tn . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 Laplace Transforms of the form eat f (t) . . . . . . . . . . . . . . . . . . . . 18
2.4 Laplace Transforms of the form tn f (t) where n is positive integer . . . . . . 20
2.5 Laplace Transforms of f (t)
t
. . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.6 Laplace Transforms of Unit Step Function and Unit Impulse Function . . . . 22
2.6.1 Properties Associated with the Unit Step Function . . . . . . . . . . 24
2.6.2 Laplace Transform of the Unit Impulse Function . . . . . . . . . . . 25
2.7 Laplace Transforms of Periodic Functions . . . . . . . . . . . . . . . . . . . 27

3 Inverse Laplace Transforms 28

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2 Inverse Laplace Transforms of Some Standard Functions . . . . . . . . . . . 28
3.3 Properties of inverse Laplace transform . . . . . . . . . . . . . . . . . . . . 29
3.4 Inverse Laplace Transforms using Partial Fractions . . . . . . . . . . . . . . 30
F (s)
3.5 Inverse Laplace Transforms of the Functions of the Form s
. . . . . . . . 32
3.6 Inverse Functions of Step Functions . . . . . . . . . . . . . . . . . . . . . . 33

4 Some Applications of Laplace Transforms 35

4.1 Laplace Transforms of the Derivatives . . . . . . . . . . . . . . . . . . . . . 35
4.2 Solution of Linear Differential Equations . . . . . . . . . . . . . . . . . . . 36

5 Fourier Series 41
5.1 Introduction to Fourier Series . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.2 Periodic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.3 Orthogonal functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.4 Odd and Even functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.5 The Fourier Series of a Function . . . . . . . . . . . . . . . . . . . . . . . . 45

1
 5.6 Dirichlet Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.7 Fourier series for periodic functions of period 2π . . . . . . . . . . . . . . . 48
5.8 Products of odd and even functions . . . . . . . . . . . . . . . . . . . . . . 50
5.9 Half Range Fourier Sine and Cosine Series . . . . . . . . . . . . . . . . . . 53
5.10 Complex Fourier Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.11 The Fourier Series of a Function . . . . . . . . . . . . . . . . . . . . . . . . 58
5.12 Dirichlet Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.13 Fourier series for periodic functions of period 2π . . . . . . . . . . . . . . . 61
5.14 Products of odd and even functions . . . . . . . . . . . . . . . . . . . . . . 63
5.15 Half Range Fourier Sine and Cosine Series . . . . . . . . . . . . . . . . . . 66
5.16 Complex Fourier Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

6 Functions of Several Variables 70

6.1 Functions of Two or More Variables . . . . . . . . . . . . . . . . . . . . . . 70
6.2 Partial Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.3 Higher-Oder Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.4 The Total Differential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.5 Chain Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
6.6 Applications of Partial Derivatives . . . . . . . . . . . . . . . . . . . . . . . 81

Exercise One 84

Exercise Two 85

Exercise Three 86

Exercise Four 88

References 89

2
 PREFACE

A Sound knowledge of engineering mathematics is a ’sine qua non’ for the modern engineer
to attain new heights in all aspects of engineering practice. Calculus I (with code GSU
07101) is the first among four courses in engineering mathematics offered at Dar es Salaam
Institute of Technology (DIT) a for a bachelor’s degree programme. The course aims at en-
abling students get knowledge and skills of limits and continuity, laplace transforms, fourier
series and functions of several variables and use it to solve engineering related problems.

This manual is written in a lucid, easy to understand language. Each topic has been well
coved in scope, content and several worked out examples, carefully selected to cover all
aspects of the topic are presented. There are more than 80 worked examples. At the end
of the manual a number of practice exercises are provided. Also selected reading texts for
further references are provided.

This manual has put into practice since 2005 here at DIT. We are hopeful that the manual
will be useful to both students and teachers not only at DIT but also to other engineering
colleges.

In spite of our best efforts, some errors might have crept in the manual. Any such errors and
all suggestions for improving future editions of the manual are welcome and will be greatly
acknowledged.

E. Rutalebwa
N. Mtega
Chapter 1

Theory of Limits

1.1 Definitions of a Limit

1.1.1 An informal definition of limit
If f (x) is defined for all x near a, except possibly at a itself, and if we can ensure that f (x)
is as close as we want to L by taking x close enough to a, but not equal to a, we say that
the function f approaches the limit L as x approaches a, and
lim f (x) = L
x→a

This is,
lim f (x) = L, for some number L, if and only if lim f (x) = lim+ f (x) = L
x→a x→a− x→a

Example 1.1 Evaluate

3x + 9
lim
x→−3 x2 − 9

Solution We examine some function values for x near −3.

x -2.80 -2.84 -2.89 -2.93 -2.98 -3.02 -3.07 -3.11 -3.16 -3.20
f(x) -0.52 -0.51 -0.51 -0.51 -0.50 -0.50 -0.49 -0.49 -0.49 -0.48

Based on this numerical evidence, it’s reasonable to conjecture that

3x + 9 3x + 9 1
lim − 2
= lim + 2 =− .
x→−3 x − 9 x→−3 x − 9 2
Further, note that
3x + 9 3(x + 3) 3 1
lim − 2
= lim − = lim − =− ,
x→−3 x − 9 x→−3 (x + 3)(x − 3) x→−3 x − 3 2
since (x − 3) → −6 as x → −3. Again, the cancellation of the factors of (x + 3) is valid
since in the limit as x → −3, x is close to −3, but x 6= −3, so that x + 3 6= 0.
Likewise,
3x + 9 1
lim + 2 =− .
x→−3 x − 9 2
Finally, since the function approaches the same value as x → −3 both from the right and
from the left (i.e., the one-sided limits are equal), we write
3x + 9 1
lim 2
=− .
x→−3 x −9 2

1
Example 1.2 Determine whether
3x + 9
lim exists
x→3 x2 − 9

Solution
We first compute some function values for x near to 3.

x 2.80 2.84 2.89 2.93 2.98 3.02 3.07 3.11 3.16 3.20
f(x) -15.00 -19.29 -27.00 -45.00 -135.00 135.00 45.00 27.00 19.29 15.00

Based on this numerical evidence, it appears that, as x → 3+ , 3x+9 x2 −9

is increasing without
bound. Thus
3x + 9
lim+ 2 does not exist
x→3 x − 9
Similarly, from the table of values for x < 3, we can say that
3x + 9
lim− 2 does not exist
x→3 x − 9

Since neither one-side limit exist, we say

3x + 9
lim does not exist
x→3 x2 − 9

Here, we considered both one-side limit for the sake of completeness. Of course, you should
keep in mind that if either one-side limit fails to exist, then the limit does not exist.
Example 1.3 Evaluate
x
lim
x→0 |x|

Solution
x x
lim+ = lim+ (since |x| = x, when x > 0)
x→0 |x| x→0 x
= lim+ 1
x→0
=1
and
x x
lim− = lim− (since |x| = −x, when x < 0)
x→0 |x| x→0 −x
= lim− −1
x→0
= −1
It now follows that
x
lim does not exist
x→0 |x|

Example 1.4 Evaluate limx→0 f (x), where f is defined by

 2
x + 2 cos x + 1, for x < 0
f (x) =
sec x − 4, for x ≥ 0
Solution Since f is defined by different expressions for x < 0 and for x ≥ 0, we must
consider one-sided limits. We have
lim− = lim− (x2 + 2 cos x + 1) = 2 cos 0 + 1 = 3.
x→0 x→0

Also, we have
lim f (x) = lim+ (sec x − 4) = sec 0 − 4 = 1 − 4 = −3.
x→0+ x→0

Since the one-sided limits are different, we have that limx→0 f (x) does not exist.

2
Limit Rules
If limx→a f (x) = L, limx→a g(x) = M , and k is a constant, then
1. Limit of a sum:
lim [f (x) + g(x)] = L + M
x→a

2. Limit of a difference:
lim [f (x) − g(x)] = L − M
x→a

3. Limit of a product:
lim f (x)g(x) = LM
x→a

4. Limit of a multiple:
lim kf (x) = kL
x→a

5. Limit of a quotient:
f (x) L
lim = , if M 6= 0.
x→a g(x) M
If m is an integer and n is a positive integer, then
6. Limit of a power:
lim [f (x)]m/n = Lm/n ,
x→a

provided L > 0 if n is even, and L 6= 0 if m < 0. If f (x) ≤ g(x) on an interval

containing a in its interior, then
7. Order is preserved:
L≤M
Rules 1-6 are also valid right limits and left limits. So is Rule 7, under the the assumption
that f (x) ≤ g(x) on the open interval extending in the appropriate direction from a.
Example 1.5 Evaluate
lim (x cot x)
x→0
Solution
 cos x   x 
lim (x cot x) = lim x = lim cos x
x→0 x→0 sin x x→0 sin x
 x  
= lim lim cos x
x→0 sin x x→0
limx→0 cos x 1
= sin x
= =1
limx→0 x 1

Limits at infinity and negative infinity (informal definition)

If the function f is defined on an interval (a, ∞) and if we can ensure that f (x) is as close
as we want to the number L by taking x large enough, then we say that f (x) approaches
the limit L as x approaches infinity, and we write
lim f (x) = L.
x→∞

If f is defined on an interval (−∞, b) and if we can ensure that f (x) is as close as we want
to the number M by taking x negative and large enough in absolute value, then we say that
f (x) approaches the limit M as x approaches negative infinity, and we write
lim f (x) = M.
x→−∞

3
Example 1.6 Find √ 
lim x2 + x − x .
x→∞

Solution We are trying to find the limit of the difference of two functions, each of which be-
comes arbitrarily large as x increases to infinity. We rationalize the expression by
√ multiplying
the numerator and the denominator ( which is 1) by the conjugate expression x2 + x + x:
√ √
√ ( x2 + x − x)( x2 + x + x)
2
lim ( x + x − x) = lim √
x→∞ x→∞ x2 + x + x
x + x − x2
2
= lim q
x→∞
x2 (1 + x1 ) + x
x 1 1
= lim q = lim q =
x→∞
x 1 + x1 + x
x→∞
1 + x1 + 1 2

√
(Here, x2 = x because x > 0 as x → ∞.)

1.1.2 The formal definition of a limit

When we say that f (x) has limit L as x approaches a, we are really saying that we can
ensure that the error |f (x) − L| will be less than any allowed tolerance, no matter how
small, by taking x close enough to a (but not equal to a). It is traditional to use , the
Greek letter ”epsilon”, for the size of the allowable error and δ, the Greek letter ”delta”
for the difference x − a that measures how close x must be to a to ensure that the error
is within that tolerance. These are the letters that Cauchy and Weierstrass used in their
pioneering work on limits and continuity in the nineteenth century.
If  is any positive, no matter how small, we must be able to ensure that |f (x) − L| < 
by restricting x to be close enough to (but not equal to) a. How close is close enough?
It is sufficient that the distance |x − a| from x to a be less than a positive number δ that
depends on . (See Figure 1.1.) If we can find such δ for any positive , we are entitled to
conclude that limx→a f (x) = L.

A formal definition of limit

We say that f (x) approaches the limit L as x approaches a, and we write

lim f (x) = L,
x→a

if the following condition is satisfied: for every number  > 0 there exists a number δ > 0,
possibly depending on , such that

if 0 < |x − a| < δ, then |f (x) − L| < .

Example 1.7 Testing the Definition

The formal definition of limit does not tell how to find the limit of a function, but it enables
us to verify that a suspected limit is correct. This following examples show how the definition
can be used to verify limit statements for specific functions. However, the real purpose of
the definition is not to do calculations like this, but rather to prove general theorems so that
the calculation of specific limits can be simplified.

Example 1.8 Show that

lim (5x − 3) = 2.
x→1

4
 Fig. 1.1: If x 6= a and |x − a| < δ, then |f (x) − L| < 

Fig. 1.2:

Fig. 1.3:

Solution Set c = 1, f (x) = 5x − 3, and L = 2 in the definition of limit. For any given
 > 0, we have to find a suitable δ > 0 so that if x 6= 1 and x is within distance δ of c = 1,
that is, whether
0 < |x − 1| < δ,
it is true that if f (x) is within distance  of L = 2, so

|f (x) − 2| < .

5
 Fig. 1.4:

We find δ by working backward from the -inequality:

|(5x − 3) − 2| =|5x − 5| < 
5|x − 1| < 
|x − 1| < /5.
Thus, we can take δ = /5 (see figure 1.5). If 0 < |x − 1| < δ = /5, then
|(5x − 3) − 2| = |5x − 5| = 5|x − 1| < 5(/5) = ,
which proves that limx→1 (5x − 3) = 2.
The value of δ = /5 is not the only value that will make 0 < |x − 1| < δ imply |5x − 5| < .
Any smaller positive δ will do as well. The definition does not ask for a ”best” positive δ,
just one that will work.

Fig. 1.5:

Example 1.9 Prove the following results

(a) limx→c x = c
(b) limx→c k = k (k constant)
Solution (a) Let  > 0 be given. We must find δ > 0 such that for all x
0 < |x − c| < δ implies |x − c| < 
The implication will hold if δ equals  or any smaller positive number (see figure 1.6 (a) ).
This proves that limx→c x = c.

6
(b) Let  > 0 be given. We must find δ > 0 such that for all x
0 < |x − c| < δ implies |k − k| < .
Since k − k = 0, we can use any positive number for δ and implication will hold (see figure
1.6 (b)). This proves that limx→c k = k.

(a) (b)

Fig. 1.6:

Finding Deltas Algebraically for Given Epsilons

In the above examples, the intervals for values about c for which |f (x) − L| was less than
 was symmetric about c and we could take δ to be half the length of that interval. When
such symmetry is absent, as it usually is, we take δ to be the distance from c to the interval’s
nearer endpoint.
√
Example 1.10 For the limit limx→5 x − 1 = 2, find a δ > 0 that works for  = 1. That
is, find a δ > 0 such that for all x
√
0 < |x − 5| < δ ⇒ | x − 1 − 2| < 1.
Solution We organize the search into two steps
√
1. Solving the inequality | x − 1 − 2| < 1 to find an interval containing x = 5 on which
the inequality holds for all x 6= 5.

√
| x − 1| < 1
√
−1< x−1−2<1
√
1< x−1<3
1<x−1<9
2 < x < 10
The inequality holds for all x in the open interval (2, 10), so it holds for all x 6= 5 in
this interval as well.
2. Find a value δ > 0 to place the centered interval 5 − δ < x < 5 + δ (centered at
x = 5) inside the interval (2, 10). The distance from 5 to the nearer endpoint of
(2, 10) is 3 (see figure 1.7 (a)). If we take δ = 3 or any smaller positive number, then
√ inequality 0 < |x − 5| < δ will automatically place x between 2 and 10 to make
the
| x − 1 − 2| < 1 (see figure 1.7 (b)):

7
 (a) (b)

Fig. 1.7:

How to Find Algebraically a δ for a Given f , L, c, and  > 0

The process of find a δ > 0 such that for all x

0 < |x − c| < δ ⇒ |f (x) − L| < 

can be accomplished in two steps

1. Solve the inequality |f (x) − L| <  to find an open interval (a, b) containing c on
which the inequality holds for all x 6= c.

2. Find a value of δ > 0 that places the open interval (c−δ, c+δ) centered at c inside the
interval (a, b). The inequality |f (x) − L| <  will hold for all x 6= c in this δ-interval.

Example 1.11 Prove that limx→2 f (x) = 4 if

 2
x , x 6= 2
f (x) =
1, x = 2

Solution Our task to show that given  > 0 there exists a δ > 0 such that for all x

0 < |x − 2| < δ ⇒ |f (x) − 4| < .

1. Solve the inequality |f (x) − 4| <  to find an open interval containing x = 2 on which
the inequality holds for all x 6= 2.
For c 6= 2, we have f (x) = x2 , and the inequality to solve is |x2 − 4| < :

|x2 − 4| < 
− < x2 − 4 < 
4 −  < x2 < 4 + 
√ √
4 −  < |x| < 4 + 
√ √
4−<x< 4+
√ √
The inequality |f (x) − 4| <  holds for all x 6= 2 in the open interval ( 4 − , 4 + )
(see figure below)

2. Find √ of δ > 0 that places the centered interval (2 − δ, 2 + δ) inside the interval
√ a value
( 4 − , 4 + ). √ √
Take δ to be the distance from x = 2 to the nearer endpoint of ( 4 − , 4 + ). In

8
 √ √
other worlds, take δ =√min{2 − 4√− , 4 +  − 2}, the minimum (the smaller) of
the two numbers 2 − 4 −  and 4 +  − 2. If δ has this or any smaller √ positive
value,
√ the inequality 0 < |x − 2| < δ will automatically place x between 4 −  and
4 +  to make |f (x) − 4| < . For all x,

0 < |x − 2| < δ ⇒ |f (x) − 4| < .

This completes the proof for  < 4.

If  ≥ 4, then
√ we take δ to the distance from x = 2√to the nearer endpoint of the
interval (0, 4 + ). In other words, take δ = min{2, 4 +  − 2}. (see figure 1.8)

Fig. 1.8:

1.2 Continuity
In ordinary language, to say that a certain process is ” continuous” is to say that it goes on
without interruption and without abrupt changes. In mathematics the word ”continuous”
has much the same meaning.
The concept of continuity is so important in calculus and its applications that we discuss it
with some care. First we treat continuity at a point c (a number c), and then we discuss
continuity on an interval.

Continuity at a Point
The basic idea is a follows: We are given a function f and a number c. We calculate (
if we can ) both limx→c f (x) and f (c). If these two numbers are equal, we say that f is
continuous at c. Here is the definition formally stated.

Definition 1.1 Let f be a function defined at least on an open interval (c − p, c + p). We

say that f is continuous at c if
lim f (x) = f (c)
x→c

If the domain of f contains an interval (c − p, c + p), then f can fail to be continuous at c

for only one of two reasons: either

(i) f has a limit as x tends to c, but limx→c 6= f (c) or

(ii) f has no limit as x tends to c.

9
In case (i) the number c is called aremovable discontinuity. The discontinuity can be
removed by redefining f at c. If the limit is L, redefine f at c to be L.
In case (ii) the number c is called an essential discontinuity. You can change the value of f
at a billion points in any way you like. The discontinuity will remain.
The function depicted in Figure 1.9 (a) has a removable discontinuity at c. The discontinuity
can be removed by lowering the dot into place (i.e., by redefining f at c to be L).
The function depicted in Figures 1.9 (b), 1.10, and 1.11 have essential discontinuity at c.
The discontinuity in Figure 1.9 (a) is, for obvious reasons, called jump discontinuity. The
functions of Figure 1.10 have infinite discontinuities.
In Figure 1.11, have tried to portray the Dirichlet function

1, x rational
f (x) =
−1 x irrational

(a) (b)

Fig. 1.9:

Fig. 1.10:

Fig. 1.11:

At no point c does f have a limit. Each point is an essential discontinuity. The function is
everywhere discontinuous.
Most of the functions that you have encountered so far are continuous at each point of their
domains. In particular, this is true for polynomials P ,

lim P (x) = P (c), (1.1)

x→c

10
for rational functions (quotients of polynomials) R = P/Q

P (x) P (c)
lim R(x) = lim = = R(c) provided Q(c) 6= 0, (1.2)
x→c x→c Q(x) Q(c)
and for the absolute value function,

lim |x| = |c|. (1.3)

x→c

As you were asked to show earlier,

√ √
lim x = c for each c > 0.
x→c

Theorem 1.1 If f and g are continuous at c, then

(i) f + g is continuous at c;

(ii) f − g is continuous at c;

(iii) αf is continuous at c for each real α;

(iv) f.g is continuous at c;

(v) f /g is continuous at c provided g(c) 6= 0

−x 3
Example 1.12 The function F (x) = 3|x| + x2x−5x+6 + 4 is continuous at all real numbers
other than 2 and 3. You can see this by noting that

F = 3f + g/h + k

where
f (x) = |x|, g(x) = x3 − x, h(x) = x2 − 5x + 6, k(x) = 4
Since f, g, h, k are everywhere continuous, F is continuous except at 2 and 3, the number
at which h takes on the value 0. (At those numbers F is not defined)

Definition 1.2 One-side Continuity

A function f is called

continuous from the left at c if lim f (x) = f (c).

x→c−1

It is called
continuous from the right at c if lim f (x) = f (c)
x→c+

The function of Figure 1.12 (a) is continuous from the right at 0; the function of Figure 1.12
(b) is continuous from left at 1. It follows from definition 1.1 that a function is continuous at
c iff it is continuous from both sides at c. Thus f is continuous at c iff f (c), limx→c− f (x),
limx→c+ f (x) all exist and are equal.

Example 1.13 Determine the discontinuities, if any, of the following function:


 2x + 1, x≤0
f (x) = 1, 0<x≤1
 2
x + 1, x>1

Solution Clearly f is continuous at each point in the open intervals (−∞, 0), (0, 1), (1, ∞)
(On each of these intervals f is a polynomial, see Figure 1.13). Thus, we have to check

11
 (a) (b)

Fig. 1.12:

Fig. 1.13:

the behaviour of f at x = 0 and x = 1. The figure suggests that f is continuous at 0 and

discontinuous at 1. Indeed, that is the case:

f (0) = 1, lim f (x) = lim− f (2x + 1) = 1, lim = lim+ (1) = 1

x→0− x→0 x→x+ x→0

. This makes f continuous at 0. This situation is different at x = 1:

lim f (x) = lim− (1) = 1 and lim f (x) = lim+ (x2 + 1) = 2

x→1− x→1 x→1+ x→1

Thus f has an essential discontinuity at 1, a jump discontinuity.

Example 1.14 Determine the discontinuities, if any, of the following function:



 x3 , x ≤ −1
 2
 x − 2,
 −1 < x < 1
f (x) = 6 − x, 1≤x<4
6
, 4<x<7


 7−x


5x + 2, x ≥ 7.
Solution It should be clear that f is continuous at each point of the open intervals
(−∞, −1), (−1, 1), (1, 4), (4, 7), (7, ∞). All we have to check is the behaviour of f at
x = −1, 1, 4, 7. To do so, we apply definition 1.1
The function is continuous at x = −1 since f (−1) = (−1)3 = −1, limx→−1− f (x) =
limx→−1− (x3 ) = −1, and limx→−1+ f (x) = limx→−1+ (x2 − 2) = −1. Our findings at the
other three points are displayed in the following chart. Try to verify each entry.

c f(c) limx→c− f (x) limx→c+ f (x) Conclusion

1 5 -1 5 discontinuous
4 not defined 2 2 discontinuous
7 37 does not exist 37 discontinuous

12
The discontinuity at x = 4 is removable: if we redefine f at 4 to be 2, then f becomes
continuous at 4. The numbers 1 and 7 are essential discontinuities. The discontinuity at
1 is a jump discontinuity; the discontinuity at 7 is an infinite discontinuity: f (x) → ∞ as
x → 7−

13
Chapter 2

Laplace Transforms

In mathematics, the Laplace transform is a widely used integral transform. The Laplace
transform has many important applications throughout the sciences. It is named for Pierre-
Simon Laplace who introduced the transform in his work on probability theory. The Laplace
transform is related to the Fourier transform, but whereas the Fourier transform resolves a
function or signal into its modes of vibration, the Laplace transform resolves a function into
its moments. Like the Fourier transform, the Laplace transform is used for solving differential
and integral equations. In physics and engineering, it is used for analysis of linear systems
such as electrical circuits, harmonic oscillators, optical devices, and mechanical systems.
In this analysis, the Laplace transform is often interpreted as a transformation from the
time-domain, in which inputs and outputs are functions of time, to the frequency-domain,
where the same inputs and outputs are functions of complex angular frequency, in radians
per unit time. Also in engineering and physics; the output of a linear time invariant system
can be calculated by convolving its unit impulse response with the input signal. Given a
simple mathematical or functional description of an input or output to a system, the Laplace
transform provides an alternative functional description that often simplifies the process of
analyzing the behaviour of the system, or in synthesizing a new system based on a set of
specifications. Performing this calculation in Laplace space turns the convolution into a
multiplication; the latter being easier to solve because of its algebraic form. The Laplace
transform can also be used to solve differential equations and is used extensively in electrical
engineering. The Laplace transform reduces a linear differential equation to an algebraic
equation, which can then be solved by the formal rules of algebra. The original differential
equation can then be solved by applying the inverse Laplace transform.

2.1 Definition
Let f (t) be a real valued function defined for all t ≥ 0. Then the Laplace transform of f (t)
denoted by L{f (t)} is defined by
Z ∞
L{f (t)} = e−st f (t)dt (2.1)
0

where s is a real or a complex number.

If the integral on the right hand side (2.1) exists, it is a function of s and is usually denoted
by F (s). Here s is called the parameter.
Thus Z ∞
F (s) = L{f (t)} = e−st f (t)dt.
0
The Laplace transform is a linear transform, by which is meant that:

14
 1. The transform of a sum (or difference) of expression is the sum ( or difference) of the
individual transforms. That is

L{f (t) ± g(t)} = L{f (t)} ± L{g(x)}

2. The transform of an expression that is multiplied by a constant is the constant multi-

plied by the transform of the expression. That is

L{kf (t)} = kL{f (t)}.

Proof: (1) we have Z ∞

L{f (t)} = e−st f (t)dt
0
Therefore,
Z ∞
L{kf (t)} = e−st kf (t)dt
0
Z ∞
=k e−st f (t)dt
0
= kL{f (t)}

(2) Consider
Z ∞
L{f (t) + g(t)} = e−st [f (t) + g(t)]dt
Z0 ∞ Z ∞
−st
= e f (t)dt + e−st g(t)dt
0 0
= L{f (t)} + L{g(t)}

2.2 Laplace Transforms of Some Standard Functions

2.2.1 Laplace transform of a constant
Let f (x) = a. Where a is a constant. Then from the definition of Laplace transform, we
get
Z ∞
L(a) = e−st adt
0
Z ∞
=a e−st dt
0 −st ∞
e
=a
−s 0
−a −∞
= [e − e0 ]
s
a
= since e−∞ = 0 and e0 = 1
s
Hence
a
L(a) = . (2.2)
s
In particular cases, L(1) = 1s .

15
2.2.2 Laplace transform of eat
Substituting f (t) = eat in the definition of Laplace transform, we get
Z ∞
at
L(e ) = e−st eat dt
Z0 ∞
= e(−s+a)t dt
0 −(s−a)t ∞
e
=
−(s − a) 0
−1 −∞
= [e − e0 ]
s−a
1
= if s > a > 0
s−a
Therefore
1
L(eat ) = ,s > a > 0 (2.3)
s−a
Replacing a by −a, we get
1
L(e−at ) = , s > −a. (2.4)
s+a

2.2.3 Laplace transform of sinh at

We have
eat − e−at
sinh at =
2
Substituting
eat − e−at
f (t) = sinh at =
2
at −at
e −e 1
L(eat ) − L(e−at )
 
L(sinh at) = L{ }=
2 2
1 1 1
= { − }
2 s−a s+a
a
= 2 , if s > a
s − a2
Hence
a
L(sinh at) = ,s > a (2.5)
s 2 − a2

2.2.4 Laplace transform of cosh at

We have
eat + e−at
cosh at =
2
Substituting
eat + e−at
f (t) = cosh at =
2
at −at
e +e 1
L(eat ) + L(e−at )
 
L(cosh at) = L{ }=
2 2
1 1 1
= { + }
2 s−a s+a
s
= 2 , if s > a
s − a2
Hence
s
L(cosh at) = ,s > a (2.6)
s2 − a2

16
2.2.5 Laplace transform of sin at and cos at
We know by Euler’s formula that

eiat = cos at + i sin at

Therefore
L(cos at + i sin at) = L(eiat )
1
=
s − ia
s + ia
=
(s − ia)(s + ia)
s + ia
= 2
s + a2
s a
= 2 2
+i 2
s +a s + a2
On equating the real and imaginary parts, we obtain

s
L(cos at) =
+ a2 s2 (2.7)
a
L(sin at) = 2
s + a2

2.2.6 Laplace transform of tn

Let f (t) = tn , where n is a non-negative real number or n is a negative non-integers. Then
from the definition Z ∞
n
L(t ) = e−st tn dt
0
dx x
Substitute st = x, so that dt = s
and t = s
. When t = 0, x = 0 and when t = ∞,
x = ∞. Therefore
Z ∞  x n dt
n −x
L(t ) = e
0 s s
Z ∞
1
= e−x xn dx
sn+1 0
1
= n+1
s Γ(n + 1)

Thus
Γ(n + 1)
L(tn ) = (2.8)
sn+1
In particular if n is a non-negative integers, we have

Γ(n + 1) = n!

Hence,
n!
L(tn ) = (2.9)
sn+1
where n is a non-negative integer.

17
Example 2.1
 
2 1 − cos 2at
L(sin at) = L
2
1
= L(1 − cos 2at)
2
1
= [L(1) − L(cos 2at)]
2 
1 1 s
= −
2 s s2 + (2a)2
1 s2 + 4a2 − s2
 
=
2 s(s2 + 4a2 )
2a2
=
s(s2 + 4a2 )

Example 2.2
1
L(sin 5t cos 3t) = L{ [sin 8t + sin 2t]}
2
1
= {L(sin 8t) + L(sin 2t)}
2 
1 8 2
= +
2 s2 + 8 2 s2 + 2 2
5(s2 + 16)
= 2
(s + 64)(s2 + 4)

Example 2.3 
2, 0 < t < 3
If f (t) = , find L{f (t)}
t, t>3
Solution. Now,

Z ∞
L{f (t)} = e−st f (t)dt
Z0 3 Z ∞
−st
= e .2dt + e−st .tdt
0 3
 −st 3  −st ∞
e e e−st
=2 + t − 1.
−s 0 −s (−s)2 3
3e−2s e−3s
 
−2  −3s 
= e −1 +0− − − 2
s s s
2 s + 1 −3s
= + 2 .e
s s

2.3 Laplace Transforms of the form eatf (t)

If the Laplace transform of f (t) is known, then the Laplace transform of eat f (t) where a is
a constant can be determined by using the shifting property.

Shifting property:
If
L{f (t)} = F (s)

18
then
L{eat f (t)} = F (s − a)
Proof we have Z ∞
L{f (t)} = e−st f (t) = F (s)
0
Therefore
Z ∞
at
L{e f (t)} = e−st eat f (t)dt
Z0 ∞
= e−(s−a)t f (t)dt = F (s − a).
0

Replacing a by −a, we get,

L{e−at f (t)} = F (s + a)
Example 2.4 Find
L(e−t cos2 3t)
Solution. Consider
 
2 1 + cos 6t
L(cos 3t) = L
2
1
= [L(1) + L(cos 6t)]
2 
1 1 s
= +
2 s s2 + 6 2
s2 + 18
=
s(s2 + 36)
Therefore
(s + 1)2 + 18
L(e−t cos2 3t) = (s → s + 1)
(s + 1)((s + 1)2 + 36)
s2 + 2s + 19
=
(s + 1)(s2 + 2s + 37)
Example 2.5 Find
e3t sin3 2t
Solution. We have
1
sin3 A = (3 sin A − sin 3A)
4
Hence
1
sin3 2t = (3 sin 2t − sin 6t)
4
3 1
L(sin 2t) = [3L(sin 2t − L(sin 6t)]
4 
1 2 6
= 3. 2 −
4 s + 2 2 s2 + 6 2
48
= 2
(s + 4)(s2 + 36)
By using shifting Rule, we get, s → s − 3
48
L{e3t sin3 2t} =
[(s − 3)2 + 4][(s − 3)2 + 36]
48
= 2
(s − 6s + 13)(s2 − 6s + 45)

19
In view of the shifting property we can find the Laplace transform of the standard functions
discussed in the preceding section multiplied by eat or e−at
b
1. L(sin bt) = s2 +b 2 L(eat sin bt) = (s−a)b2 +b2
s s−a
2. L(cos bt) = s2 +b2 L(eat cos bt) = (s−a) 2 +b2
b
3. L(sinh bt) = s2 −b 2 L(e sinh bt) = (s−a)b2 −b2
at
s s−a
4. L(cosh bt) = s2 −b2 L(eat cosh bt) = (s−a) 2 −b2

5. L(tn ) = Γ(n+1)
sn+1
at n Γ(n+1)
L(e t ) = (s−a)n+1 , for n 6= 0

2.4 Laplace Transforms of the form tnf (t) where n is

positive integer
In this section we shall find the Laplace transforms of the function of the form tn f (t) where
n is a positive integer if the Laplace transform of f (t) is known.

Theorem 2.1 If L{f (t)} = F (s) then

dn
L{tn f (t)} = (−1)n {F (s)}
dsn
Proof: We shall prove the theorem for n = 1 i.e.
d
L{tf (t)} = − {F (s)}
ds
We have, Z ∞
F (s) = L{f (t)} = e−st f (t)dt
0
Differentiating w.r.t. ’s’, we have
Z ∞
d d −st
{F (s)} = {e f (t)}dt
ds 0 ds
In the R.H.S. we shall apply Leibnitz rule for differentiation under integral sign,
Z ∞
= e−st (−t)f (t)dt
0
Z ∞
=− e−st {tf (t)}dt
0
= −L{tf (t)}
−d −d
∴ L{tf (t)} = {F (s)} = [L{f (t)}]
ds ds
−d
Further, L{t2 f (t)} = L{t[tf (t)]} = L{tf (t)}
 ds
−d −d
= L{f (t)}
ds ds
d2
= (−)2 2 L{f (t)}
ds
2
2 d
= (−1) 2 {F (s)}
ds
By repeating this process of the above theorem, we get
dn
L{tn f (t)} = (−1)n . {F (s)}.
dsn

20
 f (t)
2.5 Laplace Transforms of t
f (t)
If Lf (t) is known then we can find the Laplace transform of t
by using the following.

Theorem 2.2 If L{f (t)} = F (s) and limt→0 f (t) t

exists then
  Z ∞
f (t)
L = F (s)ds (2.10)
t s
R∞
Proof: We have, F (s) = L{f (t)} = 0 e−st f (t)dt On integrating on both sides w.r.t. s
from s to ∞, we get
Z ∞
R ∞ R ∞ −st 
F (s)ds = s 0
e f (t)dt ds (2.11)
s
R ∞ R ∞ −st 
= 0 s
e ds f (t)dt, (2.12)
(By changing the order of integration.)
Now Z ∞  −st ∞
−st −e −1 −∞ +e−st
e ds = = [e − e−st ] = (e−∞ = 0)
s −t s t t
∴ Eqn (2.12) gives,
∞ ∞
e−st
Z Z  
f (t)
F (s)ds = f (t)dt = L
s 0 t t
Z ∞  
f (t)
∴ F (s) = L
s t
This complete the proof.

Example 2.6
−d
L{t sin at} = L(sin at)
ds  
−d a
=
ds s2 + a2
2as
= s
(s + a2 )2
Example 2.7 L(te−2t cos 2t). From theorem (2.1) above we have
s2 − 4
L(t cos 2t) =
(s2 + 4)2
Hence by using shifting Rule, we get s → s + 2
(s + 2)2 − 4
L{e−2t t cos 2t} =
[(s + 2)2 + 4]2
s(s + 4)
= 2
(s + 4s + 8)2
1
Example 2.8 L(t3 sin t). We have L(sin t) = s2 +1

d3
 
3 3 1
∴ L(t sin t) = (−1) . 3
ds s2 + 1
24s(s2 − 1)
=
(s2 + 1)4

21
 n o
1−eat 1−eat −aeat
Example 2.9 L t
. Now limt→0 t
= limt→0 1
= −a (By using L’Hospital
Rule).

Also L(1 − eat ) = L(1) − L(eat )

1 1
= −
 Zs ∞  s−a
at
 
1−e 1 1
L = − ds
t s s s−a
= [ln s − ln(s − a)]∞
s
  ∞
1
= ln
1 − as s
s
= − ln
 s − a
s−a
= ln
s
 1−cos at 1−cos at a sin at
Example 2.10 L t
. Consider limt→0 t
= limt→0 1
= 0 (By using L’Hospital
Rule).

We have L(1 − cos at) = L(1) − L(cos at)

1 s
= − 2
  Zs ∞ s + a2 
1 − cos at 1 s
L = − ds
t s s s 2 + a2
∞
1 2 2
= ln s − ln(s + a )
2 s
2
∞
1 s
= ln 2
2 s + a2 s
#∞
1 1
= ln
2 1 + as22
s
a2
  
1
= ln 1 + ln 1 + 2
2 s
1 s 2 + a2
= ln
2 s2

2.6 Laplace Transforms of Unit Step Function and

Unit Impulse Function
Unit Step Function (Heaviside function)
The unit step function or Heaviside function u(t − a) is defined as follows

0, when t ≤ a
u(t − a) = (2.13)
1, when t > a

where a ≥ 0. The graph of this function is as shown in Figure 2.1.

22
 Table 2.1: Laplace Transform Pairs

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

1. 1 1/s s>0
2. t 1/s2 s>0
3. tn (n = 1, 2, . . . n!/sn+1 s>0
4. ta (a > −1) Γ(a + 1)/sa+1 s>a
5. eat 1/(s − a) s>a
6. tn eat (n = 1, 2, . . . ) n!/(s − a)n+1 s>a
7. H(t − a) e−as /s s≥a
8. δ(t − a) e−as s > 0, a > 0
9. sin at a/(s2 + a2 ) s>0
10. cos at s/(s2 + a2 ) s>0
11. t sin at 2as/(s2 + a2 )2 s>0
12. t cos at (s2 − a2 )/(s2 + a2 )2 s>0
13. eat sin at b/[(s − a)2 + b2 ] s>a
14. eat cos at (s − a)/[(s − a)2 + b2 ] s>a
15. (1/2a3 ) sin at − (1/2a2 )t cos at 1/(a2 + a2 )2 s>0
16. (1/2a) sin at + (1/2a)t cos at s2 /(s2 + a2 )2 s>0
17. 1 − cos at a2 /[s(s2 + a2 )] s>0
18. at − sin at a2 /[s2 (s2 + a2 )] s>0
19. sinh at a/(s2 − a2 ) s > |a|
20. cosh at s/(s2 − a2 ) s > |a|
21. (1/2a3 ) sin at + (1/2a2 )t cosh at 1/(s2 − a2 )2 s > |a|
22. (1/2a)t sinh at s/(s2 − a2 )2 s > |a|
23. (1/2a) sinh at + (1/2)t cosh at s/ (s2 − a2 )2 s > |a|
24. sinh at − sin at 2a3 /(s4 − a4 ) s > |a|
25. cosh at − cos at 2a2 s/(s4 − a4 ) s > |a|

23
 Fig. 2.1: The unit step function

2.6.1 Properties Associated with the Unit Step Function

e−as
(i) L{u(t − a)} = s

(ii) L{f (t − a)u(t − a)} = e−as F (s) = e−as L{f (t)}.

Proof: (i) Using the definition of Laplace transform, we have
Z ∞
L{u(t − a)} = e−st u(t − a)dt
Z0 a Z ∞
−st
= e u(t − a)dt + e−st u(t − a)dt
Z0 a Z ∞ a
= e−st .0dt + e−st .1dt
0 a
∞
e−st
= 0+
−s a
−1 −∞ e−sa
= [e − e−sa ] =
s s
e−sa
Thus, L{u(t − a)} =
s
(ii) By definition, we have

Z ∞
L{f (t − a)u(t − a)} = e−st f (t − a)u(t − a)dt
Z0 a Z ∞
−st
= ef (t − a)u(t − a)dt + e−st f (t − a)u(t − a)dt
Z0 a Z ∞ a
= e−st f (t − a)(0)dt + e−st f (t − a).1dt
Z0 ∞ a

= e−st f (t − a)dt
a

Substituting t − a = x so that dt = dx, when t = a, x = 0, when t = ∞, x = ∞, t = a + x.

Hence
Z ∞
L{f (t − a)u(t − a)} = e−s(a+x) f (x)dx
0
Z −sx
−sa
=e f (x)dx
Z ∞
−as
=e e−st f (t)dt change x to t
0
= e−as L{f (t)}
= e−as F (s). Hence proved.

24
Example 2.11 Find the Laplace transform of (2t − 1)u(t − 2)
Solution. Now 2t − 1 = 2(t − 2) + 3. Therefore, using Heaviside shift theorem, we get

L{(2t − 1)u(t − 2)} = L{[2(t − 2) + 3]u(t − 2)}

= e−2s L(2t + 3) Replacing t-2 by t
−2s
= e {2L(t) + L(3)}
 
−2s 2 3
=e +
s2 s

Example 2.12 Find the Laplace transform of t2 u(t − 3)

Solution. Now t2 = [(t − 3) + 3]2 = (t − 3)2 + 6(t − 3) + 9
Then
L{t2 u(t − 3)} = L{[(t − 3)2 + 6(t − 3) + 9]u(t − 3)}
= e−3s L(t2 + 6t + 9) Replacing t-3 by t
−3s 2
= e [L(t ) + 6L(t) + 9]
 
−3s 2 6 9
=e + + Using Heaviside shift theorem
s3 s2 s

Example 2.13 Find the Laplace transform of u(t − π/2). cos 2(t − π/2)
Solution.
L{u(t − π/2). cos 2(t − π/2)} = e−πs/2 .F (s) where F (s) = L(cos 2t)
s.e−πs/2
 
s
= e−πs/2 =
s2 + 4 s2 + 4

Example 2.14 Express the following function in terms of the Heaviside’s unit step function
and hence find its Laplace transform.
 −t
e , 0<t<3
f (t) =
0, t>3

Solution.
Now f (t) = e−t + [0 − e−t ]u(t − 3)
= e−t − e−t u(t − 3)
= e−t − e−(t−3) u(t − 3)e−3
∴ L{f (t)} = L(e−t − e−3 L{e−(t−3) u(t − 3)}
1
= − e−3 e−3s L(e−t )
s+1
1 1
= − e−3(s+1) .
s+1 s+1
−3(s+1)
1−e
=
s+1

2.6.2 Laplace Transform of the Unit Impulse Function

Graphically the Dirac delta or unit impulse δ(t − a) is represented by the horizontal axis with
a vertical line of infinite length at t = a (see figure 2.2.

Following are the results of Laplace Transform of Dirac Delta Functions:

L{(δ(t − a)} = e−as

25
 Fig. 2.2: Unit impulse function

L{δ(t)} = 1
Let us deal with the more general case of L{f (t).δ(t − a)}
We have Z ∞
L{f (t).δ(t − a)} = e−st .f (t).δ(t − a)dt.
0
Now the integrand e−st .f (t).δ(t − a) = 0 for all values of t except at t = a at which point
e−st = e−as , and f (t) = f (a).
Z ∞
−as
∴ L{f (t).δ(t − a)} = f (a).e δ(t − a)dt = f (a).e−as (1)
0

∴ L{f (t).δ(t − a)} = f (a)e−as

e.g.
L{6.δ(t − 4)}, a = 4, ∴ L{6.δ(t − 4)} = 6e−4s
L{t3 .δ(t − 2)} a = 2, ∴ L{t3 .δ(t − 2)} = 8e−2s
Example 2.15
Impulses of 1, 4, 7 units occurs at t = 1, t = 3 and t = 4 respectively, in directions shown
in figure below. Write down an expression for f (t) and determine its Laplace transform.
Solution We have f (t) = 1.δ(t − 1) − 4.δ(t − 3) + 7.δ(t − 4). Then
L{f (t)} = e−s − 4e−3s + 7e−4s

26
2.7 Laplace Transforms of Periodic Functions
A function f (t) is said to be periodic function with period α > 0, if f (t + α = f (t).

Theorem 2.3 If f (t) is a periodic function of period α > 0, then

Z α
1
L{f (t)} = e−st f (t)dt (2.14)
1 − e−sα 0

Example 2.16

3t, 0 < t < 2
If f (x) = and f (t) = f (t + 4), find L{f (t)}.
6, 2 < t < 4
Solution Since f (t) is a periodic function with period α = 4 from 2.14, we get
Z 4
1
L{f (t)} = e−st f (t)dt
1 − e−4s 0
Now,

Z 4 Z 2 Z 4
−st −st
e f (t)dt = e .3tdt +e−st f (t)dt
0 0 2
 −st −st
2  −st 4
e e e
=3 t. − 1. 2
+ 6
−s (−s) 0 −s 2
−2s −2s
   
e e 1 6 −4s
e − e−2s


=3 −2 − 2 −3 0− 2 −
s s s s
−2s −4s
3 3e 6e
= 2− 2 −
s s s
3
= 2 1 − e−2s − 2se−4s


s

3 (1 − e−2s − 2se−4s )
∴ L{f (t)} = .
s2 (1 − e−4s )

27
Chapter 3

Inverse Laplace Transforms

3.1 Introduction
If L{f (t)} = F (s), then f (t) is called the Inverse Laplace Transform of F (s) and sym-
bolically, we write f (t) = L−1 {F (s)}. Here L−1 is called the inverse Laplace transform
operator. For example
 
at 1 −1 1
(i) L(e ) = , s > a, L = eat
s−a s−a
 
1 −1 1
(ii) L(t) = 2 , L =t
s s2

3.2 Inverse Laplace Transforms of Some Standard

Functions
L(1) = 1s , L−1  1s = 1

1. Since
2. L(eat ) = s−a1
, s > a, L−1 s−a 1
= eat
Replacing a by -a, we get
L−1 s+a = e−at
 1

L(sin at) = s2a+a , L−1 s2 +a = sin at, L−1 s2 +a

 a  1
3. 2 2 = sinaat
s − s
4. L(cos at) = s2 +a 2, L s2 +a2 = cos at.
a
L−1  s2 −a
a
= sinh at, L−1 s2 −a
 1
5. L(sinh at) = s2 −a 2, 2 2 = sinha at
6. s
L(cosh at) = s2 −a2 , L−1 s2 −as
2 = cosh at
n!
7. L(tn) = sn+1 where n is a positive integer, we get
−1 n!
L  sn+1 = tn
n
L−1 sn+11
= tn!
Replacing n by n − 1, we get

tn−1
 
−1 1
L =
sn (n − 1)!
In particular
 
−1 1
L =1
s
t2 − 1
 
−1 1
L = =t
s2 (2 − 1)!
t3−1 t2
 
−1 1
L = =
s3 (3 − 1)! 2

28
Since,
tn
 
n Γ(n + 1) −1 1
L(t ) = , L = , n > −1
sn+1 sn+1 Γ(n + 1)

Table 3.1: The following table gives list of the Inverse Laplace Transform of some
standard functions.

3.3 Properties of inverse Laplace transform

1. Linearity Property
If a and b are two constants then
L−1 {aF (a) + bG(s)} = aL−1 {F (s)} + bL−1 {G(s)}
This result can be extended to more than two functions. This shows that like L, L−1
is also a linear operator.

Example 3.1
       
−1 2 3s 4 −1 1 −1 s −1 1
L − + = 2L − 3L + 4L
s − 3 s2 + 16 s2 − 9 s−3 s2 + 4 2 s2 − 32
sinh 3t
= 2e3t − 3 cos 4t + 4
3
2. Shifting Property
If
L−1 {F (s)} = f (t)
then
{L−1 F (s − a)} = eat f (t) = eat L−1 {F (s)}
This follows immediately from the result.
If
L{f (t)} = F (s)
then
L{eat f (t)} = F (s − a)

Replacing a by −a, we get

L−1 {F (s + a)} = e−at f (t) = e−at L−1 {F (s)}

29
Example 3.2
   
−1 1 −1 1
L 2
=L
s − 2s + 5 (s − 1)2 + 22
 
t −1 1
=eL
s2 + 2 2
sin 2t
= et
2
1 t
= e sin 2t.
2
Example 3.3
   
−1 s−3 −1 s−3
L =L
s2 − 6s + 13 (s − 3)2 + 22
 
3t −1 s
=e L
s2 + 2 2
= e3t cos 2t.

3.4 Inverse Laplace Transforms using Partial Frac-

tions
In this method we first resolve the given rational function of s into partial fractions and then
find the inverse Laplace transform of each fraction.

Example 3.4 Find the inverse Laplace transform of

2s − 1
s2 − 5s + 6
Solution

2s − 1 2s − 1 A B
Let = = +
s2 − 5s + 6 (s − 2)(s − 3) s−2 s−3
2s − 1 = A(s − 3) + B(s − 2)
Put s = 3, B=5
Put s = 2, A = −3
2s − 1 3 5
∴ 2
=− +
s − 5s + 6 s−2 s−3
     
−1 2s − 1 −1+ 1 −1 1
Then L = −3L + 5L
s2 − 5s + 6 s−1 s−3
2t 3t
= −3e + 5e

Example 3.5 Find the inverse Laplace transform of

s
(2s − 1)(3s − 1)
s A B
Let = +
(2s − 1)(3s − 1) 2s − 1 3s − 1
S = A(3s − 1) + B(2s − 1)

30
 1 1
Put s = , A = 1, and s = , B = −1
2 3
s 1 1
= −
(2s − 1)(3s − 1) 2s − 1 3s − 1
     
−1 s −1 1 −1 1
∴ L =L −L
(2s − 1)(3s − 1) 2s − 1 3s − 1
( ) ( )
1 1
= L−1
− L−1
2 s − 12 3 s − 13


   
1 −1 1 1 −1 1
= L − L
2 s − 12 3 s − 13
1 1 1 1
= e 2 t − e 3 t.
2 3
Example 3.6 Find the inverse Laplace transform of
2s − 3
(s − 1)(s − 2)(s − 3)

Solution.
2s − 3 A B C
Let = + +
(s − 1)(s − 2)(s − 3) s−1 s−2 s−3
2s − 3 = A(s − 2)(s − 3) + B(s − 1)(s − 3) + C(s − 1)(s − 2)
−1
Put s = 1, ⇒A=
2
3
s = 3, ⇒C=
2
s = 2, ⇒ B = −1

2s − 3 − 12 1 3
Thus, = − + 2
(s − 1)(s − 2)(s − 3) s−1 s−2 s−3

   
−1 2s − 3 −1 −1 1 1 3 1
L = L − L−1 + L−1
(s − 1)(s − 2)(s − 3) 2 s−1 s−2 2 s−3
−1 t 3
= e − e2t + e3t .
2 2
Example 3.7 Find the inverse Laplace transform of
4s + 5
(s − 1)2 (s + 2)

Solution
4s + 5 A B C
Let 2
= + 2
+
(s − 1) (s + 2) s − 1 (s − 1) s+2
4s + 5 = A(s − 1)(s + 2) + B(s + 2) + C(s − 1)2
Put s = 1, ⇒B=3
−1
s = −2, ⇒C=
3
To find A, put s = 0,
Then 5 = −2A + 2B + C

31
 1
This gives A=
3
Thus the partial fraction is
1 1
4s + 5 3 3 3
= + −
(s − 1)2 (s + 2) s − 1 (s − 1)2 s + 2

       
−1 4s + 5 1 −1 1 −1 1 1 −1 1
∴ L = L + 3L − L
(s + 1)2 (s + 2) 3 s−1 (s − 1)2 3 s+2
  
1 t 1 1 −2t
= e + 3et L−1 − e
3 s2 3
1 t 1
= e + 3et .t − e−2t
3 3
1 t 1
= e + 3tet − e−2t .
3 3

3.5 Inverse Laplace Transforms of the Functions of

the Form F s(s)
We have proved that if
L{f (t)} = F (s)
Z t 
F (s)
then L f (t)dt =
0 s
  Z t
F (s)
Hence L−1 = f (t)dt (3.1)
s 0

Example 3.8 Evaluate  

−1 1
L
s(s + a)
Solution

 
−1 1
Consider L = e−at
(s + a)
Using equation (3.1), we get
  Z t  −at t
−1 1 −at e 1
1 − e−at


L = e dt = =
s(s + a) 0 −a 0 a
Example 3.9 Evaluate  
−1 1
L
s(s2 + a2 )
 
−1 1 1
We have L = sin at
s 2 + a2 a
  Z t
−1 1 1
∴ L 2 2
= sin atdt
s(s + a ) 0 a
 t
1 (− cos at)
=
a a 0
1
= 2 (1 − cos at)
a

32
3.6 Inverse Functions of Step Functions
The main points are

(a) 
0, 0 < t < c
u(t − c) =
1, t≥c

(b)
e−cs
L{u(t − c)} =
s
1
L{u(t)} =
s
(c)
L{u(t − c).f (t − c)} = e−cs .F (s) where F (s) = L{f (t)}

(d) If
F (s) = L{f (t)},
then
e−cs .F (s) = L{u(t − c).f (t − c)}
−4s
Example 3.10 Find the function whose transform is e s2 .
The numerator corresponds to e−cs where c = 4 and therefore indicates u(t − 4).
Then
1
= F (s) = L(t) ∴ f (t) = t
s2
 −4s 
−1 e
∴ L = u(t − 4).(t − 4)
s2
Remember that in writing the final result, f (t) is replaced by f (t − c).

Example 3.11 Determine

s.e−s
 
−1
L
s2 + 9
Solution
The numerator contains e−s which indicates u(t − 1). Also
s
= F (s) = L(cos 3t)
s2 +9
∴ f (t) = cos 3t ∴ f (t − 1) = cos 3(t − 1).
s.e−s
 
−1
∴ L = u(t − 1). cos 3(t − 1)
s2 + 9
Remember that, having obtained f (t), the result contains f (t − c).

33
 Table of Laplace Transforms
f ( t ) = L -1 {F ( s )} F ( s ) = L { f ( t )} f ( t ) = L -1 {F ( s )} F ( s ) = L { f ( t )}
1 1
1. 1 2. e at
s s-a
n! G ( p + 1)
3. t n , n = 1, 2,3,K 4. t p , p > -1
s n +1 s p +1
p 1 × 3 × 5L ( 2n - 1) p
, n = 1, 2, 3,K
n - 12
5. t 3 6. t n+ 1
2s 2
2n s 2
a s
7. sin ( at ) 8. cos ( at )
s2 + a2 s2 + a2
2as s2 - a2
9. t sin ( at ) 10. t cos ( at )
(s + a2 ) (s + a2 )
2 2 2
2

2a 3 2as 2
11. sin ( at ) - at cos ( at ) 12. sin ( at ) + at cos ( at )
(s + a )
2 2 2
(s + a )
2 2 2

s(s - a ) 2 2
s ( s + 3a )
2 2

13. cos ( at ) - at sin ( at ) 14. cos ( at ) + at sin ( at )

(s + a )
2 2 2
(s + a )
2 2 2

s sin ( b ) + a cos ( b ) s cos ( b ) - a sin ( b )

15. sin ( at + b ) 16. cos ( at + b )
s2 + a2 s2 + a2
a s
17. sinh ( at ) 18. cosh ( at )
s - a2
2
s - a2
2

b s-a
19. e at sin ( bt ) 20. e at cos ( bt )
(s - a) (s - a)
2 2
+ b2 + b2
b s-a
21. e at sinh ( bt ) 22. e at cosh ( bt )
( s - a ) - b2 ( s - a ) - b2
2 2

n! 1 æsö
23. t ne at , n = 1, 2,3,K 24. f ( ct ) Fç ÷
(s - a)
n +1
c ècø
uc ( t ) = u ( t - c ) e - cs d (t - c )
25. 26. e - cs
Heaviside Function s Dirac Delta Function
27. uc ( t ) f ( t - c ) e - cs F ( s ) 28. uc ( t ) g ( t ) e - cs L { g ( t + c )}
ect f ( t ) F ( s - c) t n f ( t ) , n = 1, 2,3,K ( -1) F ( n) ( s )
n
29. 30.
1 ¥ t F (s)
31. f (t ) ò F ( u ) du 32. ò f ( v ) dv
t s 0
s
T
t
ò e - st f ( t ) dt
33. ò f ( t - t ) g (t ) dt F (s)G (s) 34. f (t + T ) = f (t ) 0
0
1 - e - sT
35. f ¢ (t ) sF ( s ) - f ( 0 ) 36. f ¢¢ ( t ) s F ( s ) - sf ( 0 ) - f ¢ ( 0 )
2

37. f ( n)
(t ) s F (s) - s
n n -1
f ( 0 ) - s n - 2 f ¢ ( 0 )L - sf (
n - 2)
( 0 ) - f ( n-1) ( 0 )

34
Chapter 4

Some Applications of Laplace

Transforms

4.1 Laplace Transforms of the Derivatives

If the Laplace transform of f (t) is known then by the following R t results we can find the
Laplace transforms of the derivatives f 0 (t), f 00 (t), · · · , f n (t) and 0 f (t)dt.

Laplace transforms of the derivatives: Functions of exponential or-

der.
A continuous function f (t), t > 0 is said to be exponential order if

lim e−st f (t) = 0

t→∞

Theorem 4.1 If f (t) is of exponential order and f 0 (t) is continuous then

L{f 0 (t)} = sL{f (t)} − f (0) (4.1)

Proof: By the definition of Laplace transform, we have

Z ∞
0
L{f }(t) = e−st f 0 (t)dt
0

Applying integration by parts

Z ∞
 −st ∞
= e f (t) 0 − f (t)e−st (−s)dt
Z ∞0
= 0 − f (0) + s e−st f (t)dt
Z ∞0
= −f (0) + s e−st f (t)dt
0
= −f (0) + sL{f (t)}
0
L{f (t)} = sL{f (t)} − f (0)

Laplace transform of f 00 (t)

L{f 00 (t)} = s2 L{f (t)} − sf (0) − f 0 (0) (4.2)

Let
f 0 (t) = g(t) so that f 00 (t) = g 0 (t)

35
Consider
Lf 00 (t) = L{g 0 (t)}
= sL{g(t)} − g(0), using (4.1)
= sL{f 0 (t)} − f 0 (0)
= s[sL{f (t)} − f (0)] − f 0 (0)
L{f 00 (t)} = s2 L{f (t)} − sf (0) − f 0 (0)

Similarly,
L{f 000 (t)} = s3 L{f (t)} − s2 f (0) − sf 0 (0) − f 00 (0) (4.3)
L{f n (t)} = sn L{f (t)} − sn−1 f (0) − sn−2 f 0 (0) · · · f n−1 (4.4)
If f (t) = y then (4.4) can be written in the form

L(y n ) = sn L(y) − sn−1 y(0) − sn−2 y 0 (0) − · · · − y n−1 (0)

where y 0 , y 00 , · · · y (n) denotes the successive derivatives.

4.2 Solution of Linear Differential Equations

One of the important application of Laplace transform is to solve linear differential equations
with constant coefficients with initial conditions. For example, consider a second order linear
differential equation
d2 y dy
2
+ a0 + a1 y = f (t)
dx dx
i.e. y 00 + a0 y 0 + a1 y = f (t)
where a0 , a1 are constants with initial conditions y(0) = A and y 0 (0) = B.
Taking Laplace transforms on both sides of the above equation and using the formulae on
Laplace transforms of the derivatives y 0 and y 00 .
We recall the formulae for immediate reference.

L(y 0 ) = sL(y) − y(0)

L(y 00 ) = s2 L(y) − sy(0) − y 0 (0)
L(y 000 ) = s3 L(y) − s2 y(0) − sy 0 (0) − y 00 (0)

and so on.
To solve a differential equation by Laplace transforms, we go through four distinct stages
(a) Rewrite the equation in terms of Laplace transforms.

(b) Insert the given initial conditions.

(c) Rearrange the equation algebraically to give the transform of the solution.

(d) Determine the inverse transform to obtain the particular solutions.

The procedure is summarized in Figure (4.1).

Example 4.1 Solve using Laplace transforms.

d2 y dy
− 3 + 2y = e3t
dt2 dt
given that y(0) = 0 and y 0 (0) = 0

36
Solution. Given equation is y 00 − 3y 0 + 2y = e3t .
Taking Laplace transforms on both sides, we get

L(y 00 ) − 3L(y 0 ) + 2L(y) = L(e3t )

1
i.e. s2 L(y) − sy(0) − y 0 (0) − 3[sL(y) − y(0)] + 2L(y) =
s−3
where y(0) = 0 and y 0 (0) = 0
1
i.e., (s2 − 3s + 2)L(y) = using the initial conditions
s−3
1
i.e., L(y) =
(s2 − 3s + 2)(s − 3)
1
=
(s − 1)(s − 2)(s − 3)
 
−1 1
∴ y=L
(s − 1)(s − 2)(s − 3)
 1 1 
−1 2 1 2
=L − + using partial fractions
s−1 s−2 s−3
     
1 −1 1 −1 1 1 −1 1
= L −L + L
2 s−1 s−2 2 s−3
1 1
y = et − e2t + e3t
2 2
This is the required solution.

Fig. 4.1:

2
Example 4.2 Solve using Laplace transforms ddt2y − 5 dy
dt
+ 6y = sin t, given y(0) = 1
10
and
y 0 (0) = 21
10
.
Solution. Given equation is
y 00 − 5y 0 + 6y = sin t
Taking Laplace transforms on both sides, we get

L(y 00 ) − 5L(y 0 ) + 6L(y) = L(sin t)

1
i.e., s2 L(y) − sy(0) − y 0 (0) − 5[sL(y) − y(0)] + 6L(y) =
s2 +1

37
where y(0) = 1
10
and y 0 (0) = 21
10
i.e.,
 
2 s 21 1 1
s L(y) − − − 5 sL(y) − + 6L(y) = 2
10 10 10 s +1

1 1
(s2 − 5s + 6)L(y) = + (s + 16)
s2 + 1 10
1 1 s + 16
L(y) = 2 2
+ 2
(s + 1)(s − 5s + 6) 10 (s − 5s + 6)
1 1 s + 16
i.e., L(y) = +
(s2 + 1)(s − 3)(s − 2) 10 (s − 2)(s − 3)
1 1 −1 1  
10
s + 10 5 10 1 −18 19
= + + + + By using partial fractions
s2 + 1 s − 2 s − 3 10 s − 2 s − 3
 
−2 2 1 s 1
= + + +
s − 2 s − 3 10 s2 + 1 s2 + 1

Therefore,
  
−1 −2 2 1 s 1
y=L + + +
s − 2 s − 3 10 s2 + 1 s2 + 1
        
−1 1 −1 1 1 −1 s −1 1
y = −2L + 2L + L +L
s−2 s−3 10 s2 + 1 s2 + 1
1
= −2e2t + 2e3t + (cos t + sin t)
10

Simultaneous differential equations

Convert the simultaneous differential equations into simultaneous algebraic equations by
taking the Laplace transform of each equation in turn. Insert the initial values. Solve the
simultaneous algebraic equations in the usual manner and take the inverse Laplace transform
of the algebraic solutions to find the solutions to the simultaneous differential equations.

Example 4.3 Solve the pair of simultaneous equations

y 0 − x = et
x0 + y = e−t

given that at t = 0, x = 0, and y = 0.

(a) We first express both equations in Laplace transforms.

1
(sL(y) − y(0)) − L(x) =
s−1
1
(sL(x) − x(0)) − L(y) =
s+1

(b) Then we insert the initial conditions, x(0) = 0 and y(0) = 0

1 
∴ sL(y) − L(x) = 
s−1

1 
sL(x) − L(y) = 
s+1

38
(c) We now solve these for L(y) and L(x) by normal algebraic method. Eliminating L(y)
we have
1
sL(y) − L(x) =
s−1
2 s
sL(y) + s L(x) =
s+1
2 1 s2 − 2s − 1
∴ (s2 + 1)L(x) = − =
s+1 s−1 (s + 1)(s − 1)
2
s − 2s − 1
∴ L(x) =
(s − 1)(s + 1)(s2 + 1)

s2 − 2s − 1 A B Cs + D
L(x) = 2
≡ + + 2
(s − 1)(s + 1)(s + 1) s−1 s+1 s +1
∴ s2 − 2s − 1 =A(s + 1)(s2 + 1) + B(s − 1)(s2 + 1)
+ (s − 1)(s + 1)(Cs + D)
Putting s = 1 and s = −1 gives A = − 21 and B = − 12 .
Comparing coefficients of s3 and the constant terms gives C = 1 and D = 1.
1 1 1 1 ss + 1
∴ L(x) = − + 2
2s−1 2s+1 s +1
1 1
x = − et − e−t + cos t + sin t
2 2
and
s 
s2 L(y) − sL(x) = 
s−1

1 
L(y) + sL(x) = 
s+1
s 1 s2 + 2s − 1
∴ (s2 + 1)L(y) = + =
s−1 s+1 (s − 1)(s + 1)
s2 + 2s − 1 A B Cs + D
∴ L(y) = ≡ + +
(s − 1)(s + 1)(s2 + 1) s−1 s+1 s2 + 1
∴ s2 + 2s − 1 =A(s + 1)(s2 + 1) + B(s − 1)(s2 + 1)
+ (s − 1)(s + 1)(Cs + D)
Putting s = 1 and s = −1 gives A = 21 and B = 12 . Equating coefficients of s3 and the
constant terms gives C = − and D = 1.
1 1 1 1 s 1
∴ L(y) = + − 2 + 2
2s−1 2s+1 s +1 s +1
1 1
∴ y = et + e−t − cos t + sin t
2 2
Example 4.4 Solve the equations

2y 0 − 6y + 3x = 0

3x0 − 3x − 2y = 0
given that x(0) = 1 and y(0) = 3.
Expressing these in Laplace transforms, we have

2(sL(y) − y(0)) − 6L(y) + 3L(x) = 0

39
 3(sL(x) − x(0)) − 3L(x) − 2L(y) = 0
Then we insert the initial conditions and simplify, obtaining

3L(x) + (2s − 6)L(y) = 6

(3s − 3)L(x) − 2L(y) = 3

(a) To find L(x), multiply the second equation by s − 3 and add to the first equation i.e.
)
3L(x) + (2s − 6)L(y) = 6
(s − 3)(3s − 3)L(x) − (2s − 6)L(y) = 3(s − 3)

Add to obtain
[(s − 3)(3s − 3) + 3]L(x) = 3s − 9 + 1
∴ (3s2 − 12s + 12)L(x) = 3s − 3
(s2 − 4s + 4)L(x) = s − 1
s−1 A B A(s − 2) + B
∴ L(x) = 2
≡ + 2
=
(s − 2) s − 2 (s − 2) (s − 2)2
∴ s − 1 = A(s − 2) + B
giving A = 1 and B = 1
1 1
∴ L(x) = + ∴ x = e2t + te2t
s − 2 (s − 2)2

(b) Similarly, we can find L(y)

   
6s − 9 1 A B 1 A(s − 2) + B
L(y) = 2
≡ + =
2(s − 2) 2 s − 2 (s − 2)2 2 (s − 2)2

∴ 6s − 9 = A(s − 2) + B ∴ A = 6; B = 3
 
1 6 3 1  2t
∴ L(y) = + 2
∴ y= 6e + 3te2t
2 s − 2 (s − 2) 2

40
Chapter 5

Fourier Series

5.1 Introduction to Fourier Series

We have seen earlier that many functions can be expressed in the form of infinite series.
Problems involving various forms of oscillations are common in fields of modern technology
and Fourier series, with which we shall now be concerned, enable us to represent a peri-
odic function as an infinite trigonometrical series in sine and cosine terms. One important
advantage of a Fourier series is that it can represent a function containing discontinuities,
whereas Maclaurin’s and Taylor’s series require the function to be continuous throughout.

Fourier series provides a method of analyzing periodic functions into their constituent com-
ponents. Alternating current and voltages, displacement, velocity and acceleration of slider-
crank mechanisms and acoustic waves are typical practical examples in engineering and
science where periodic functions are involved and often requiring analysis.

5.2 Periodic functions

A function f (x) is said to be periodic if f (x + T ) = f (x) for all values of x, where T is
some positive number. T is the interval between two successive repetitions and is called the
period of the function f (x). For example, y = sin x is periodic in x with period 2π since
sin x = sin(x + 2π) = sin(x + 4π), and so on. In general, if y = sin ωt then the period of
the waveform is 2π/ω.
Example 5.1 The function sin x has periods 2π, 4π, 6π, . . . , since sin(x + 2π), sin(x +
4π), sin(x + 6π), . . . all are equal sin x. However, 2π is the least period or the period of
sin x.

Example 5.2 The period of sin nx or cos nx, where n is a positive integer, is 2π/n.
e.g. y = 5 sin 2x, the amplitude is 5. The period is π (or 180o ) and there are thus 2 complete
cycles in 360o
Example 5.3 The period of tan x is π.
Other examples of periodic functions are

41
Integrals of periodic functions
1. Z π
dx = [x]π−π = 2π
−π

2. Z π
cos nxdx = 0
−π

3. Z π
sin nxdx = 0
−π

4. Z π
cos mx cos nxdx = πδmn
−π

1 ifm = n
where δmn =
6 n
0 ifm =
δmn is called the Kronecker delta.

5. Z π
sin mx sin nxdx = πδmn
−π

6. Z π
cos mx sin nxdx = 0
−π

Note that the same results are obtained no matter what the end points of the integrals are,
provided that the interval between them is one period. So, for example

Z k+2π  k+2π
sin nx
cos nxdx = (n 6= 0)
k n k
sin(nk + 2nπ) sin nk
= −
n n
=0 because sin(x + 2nπ) = sin x

42
5.3 Orthogonal functions
If two different functions f (x) and g(x) are defined on the interval a ≤ x ≤ b and
Z b
f (x)g(x)dx = 0
a

then we say that the two functions are orthogonal to each other on the interval a ≤ x ≤ b. In
the previous section we have seen that the trigonometric functions sin nx and cos nx where
n = 0, 1, 2, . . . form an infinite collection of periodic functions that are mutually orthogonal
on the interval −π ≤ x ≤ π, indeed on any interval of width 2π. That is
Z π
cos mx cos nxdx = 0 for m 6= n
−π
Z π
sin mx sin nxdx = 0 for m 6= n
−π

and Z π
cos mx sin nxdx = 0 for m 6= n
−π

5.4 Odd and Even functions

Before taking up our study of Fourier series, in the next chapter, we need to define even,
odd and periodic functions (period functions already defined).
Let f be defined on an x interval, finite or infinite, that is centered at x = 0. We say that
f is an even function if

f (−x) = f (x) (5.1)

and an odd function if

f (−x) = −f (x) (5.2)

for all x in that interval. That is, the graph of f is symmetric about x = 0 if f is even,
and antisymmetric about x = 0 if f is odd. Examples are shown in 5.1. For example,
2
5, x2 , 3x4 , cos x, sin |x|, and e−x are even and x, 3x3 , 2x5 , sin x and x cos x are odd.

(a) (b)

Fig. 5.1: (a) Even and (b) odd function

There are several useful algebraic properties of even and odd functions, such as the following:

43
 even + even = even (5.3a)
even × even = even (5.3b)
odd + odd = odd (5.3c)
odd × odd = even (5.3d)
even × odd = odd (5.3e)
To prove (5.3e), for example, let F (x) be even and let G(x) be odd. Then F (−x)G(−x) =
F (x)[−G(x)] = −F (x)G(x), in accord with 5.2. In addition, two useful integral properties
are as follows. If f is even, then

Z A Z A
f (x) = 2 f (x)dx, (f even) and if f is odd, then (5.4a)
−A 0
Z A
f (x)dx = 0, (f odd) (5.4b)
−A

for if we interpret the integral in (5.4a) as areas ( positive above the x axis, negative below
R0 RA
it) then the area −A f (x)dx is equal to the 0 f (d)dx due the symmetry of the graph of f .
R0 RA
And in the case of (5.4b)the area −A f (x)dx and 0 f (x)dx are negatives of each other,
due to the antisymmetry of the graph of f , and hence cancel.
Alternatively, (5.3a) and (5.3b) follow directly from (5.1) and (5.2), respectively. For exam-
ple, if f is odd, then

Z A Z 0 Z A
f (x)dx = f (x)dx +f (x)dx
−A −A 0
Z A Z A
= f (−t)(−dt) + f (x)dx (x = −t)
0 0
Z A Z A
= f (−t)(dt) + f (x)dx
0 0
Z A Z A
= −f (t)dt + f (x)dx (oddness of f)
0 0
Z A Z A
=− f (x)(dt) + f (x)dx (x = t)
0 0
=0
as stated in (5.4b)

Note carefully that a given function is not necessary even or odd: it may be both even and
odd, or it may be either. Every function can be uniquely decomposed into sum of an even
function, say fe and an odd function, say fo , as demonstrated by the simple identity

f (x) + f (−x) f (x) − f (−x)

f (x) = +
2 2 (5.5)
≡ fe (x) + fo (x)
for observe that
f (−x) − f (x) f (x) − f (−x)
fo (−x) = =− = −fo (x)
2 2
and, similarly, that fe (−x) = fe (x).

44
Example 5.4 Surely f (x) = ex is neither even nor odd. Since (Fig. 5.2) it is neither
symmetry nor antisymmetry about x = 0. Putting f (x) = ex and f (−x) = e−x into (5.5)
gives
ex + e−x ex − e−x
fe (x) = and fo =
2 2
x
as the even and odd part of e respectively. In fact, we recognize these functions as cosh x
and sinh x. So it is interesting that we can think of cosh x and sinh x as the even and odd
parts of ex respectively.

Fig. 5.2: Even and odd parts of ex

5.5 The Fourier Series of a Function

Let f (x) be defined on [−L, L]. We want to choose numbers a0 , a2 , . . . and b1 , b2 , . . . such
that ∞
1 X
f (x) = a0 + [ak cos(kπx/L) + bk sin(kπx/L)] (5.6)
2 k=1

This decomposition of the function into a sum of terms , each representing the influence of
a different fundamental frequency on the behavior of the function.
To determine a0 , integrate equation (5.24) term by term to get

Z L Z L
1
f (x)dx = a0 dx
−L 2 −L
∞ 
X Z L Z L 
+ ak cos(kπx/L)dx + bk sin(kπx/L)dx
k=1 −L −L

1
= a0 (2L) = πa0
2
because all of the integrals in the summation are zero. Then

1 L
Z
a0 = f (x)dx (5.7)
L −L

To solve for the other coefficients in the proposed equation (5.24), we will use the following
three facts, which follow by routine integrations. Let m and n be integers. Then
Z L
cos(nπx/L) sin(mπx/L)dx = 0. (5.8)
−L

45
Furthermore, if n 6= m then
Z L Z L
cos(nπx/L) cos(mπx/L)dx = sin(nπx/L) sin(mπx/L)dx = 0. (5.9)
−L −L

And, if n 6= 0, then
Z L Z L
2
cos (nπx/L)dx = sin2 (nπx/L)dx = L. (5.10)
−L −L

Now let n be any positive integer. To solve for an , multiply equation (5.24) by cos(nπx/L)
and integrate the resulting equation to get

Z L Z L
1
f (x) cos(nπx/L)dx = a0 cos(nπx/L)dx
−L 2 −L
X∞  Z L Z L 
+ ak cos(kπx/L) cos(nπx/L)dx + bk sin(kπx/L) cos(nπx/L)dx
k=1 −L −L

Because equations (5.26) and (5.27), all of the terms on the right are zero except the
coefficient of an , which occurs in the summation when k = n. The last equation reduces to
Z L Z L
f (x) cos(nπx/L)dx = an cos2 (nπx/L)dx = an L
L −L

by equation (5.28). Therefore

Z L
1
an = f (x) cos(nπx/L)dx. (5.11)
L −L

This expression contains a0 if we let n = 0.

Similarly, if we multiply equation (5.24) by sin(nπx/L) and integrate, we obtain

1 L
Z
bn = f (x) sin(nπx/L)dx. (5.12)
L −L
The numbers Z L
1
an = f (x) cos(nπx/L)dx for n = 0, 1, 2, . . . (5.13)
L −L
Z L
1
bn = f (x) sin(nπx/L)dx for n = 0, 1, 2, . . . (5.14)
L −L

are called the Fourier coefficients of f on [-L,L]. When these numbers are used, the series
in (5.24) is called the Fourier series of f on [-L,L].

Example 5.5 (a) Find the Fourier coefficients corresponding to the function

0 −5 < x < 0
f (x) =
3 0 < x < 5 Period = 10

(b) Write the corresponding Fourier series

Solution
The graph of f (x) is

46
 Fig. 5.3:

(a) Period = 2L = 10 and L = 5. Choose the interval c to c + 2L as −5 to 5, so that

c = −5. Then
1 c+2L 1 5
Z Z
an = f (x) cos(nπx/L)dx = f (x) cos(nπx/5)dx
L c 5 −5
Z 0 Z 5
3 5
 Z
1
= (0) cos(nπx/5)dx + (3) cos(nπx/5)dx = cos(nπx/5)
5 −5 0 5 0
 5
3 5
= sin(nπx/5) = 0 if n 6= 0
5 nπ 0

If n = 0 Z 5 Z 5
3 3
a0 = cos(0πx/5)dx = dx = 3.
5 0 5 0

1 c+2L 1 5
Z Z
bn = f (x) sin(nπx/L)dx = f (x) sin(nπx/5)dx
L c 5 −5
Z 0 Z 5
3 5
 Z
1
= (0) sin(nπx/5)dx + (3) sin(nπx/5)dx = sin(nπx/5)
5 −5 0 5 0
 5
3 5 3(1 − cos nπ)
= − cos(nπx/5) =
5 nπ 0 nπ

(b) The corresponding Fourier series is

∞ ∞
a0 X 3 X 3(1 − cos nπ)
+ (an cos(nπx/L) + bn sin(nπ/L)) = + sin(nπx/5)
2 n=1
2 n=1 nπ

3 6 1 1
= + sin(xπ/5) + sin(3xπ/5) + sin(5πx/5) + . . .
2 π 3 5

5.6 Dirichlet Conditions

Suppose that
1. f (x) is defined except possibly at a finite number of points in (−L, L)

2. f (x) is periodic outside (−L, L) with period 2L

3. f (x) and f 0 (x) are piecewise continuous in −L, L.

Then the series (5.24) with Fourier coefficients converges to
(a) f (x) if x is a point of continuity
f (x+0)+f (x−0)
(b) 2
if x is a point of discontinuity

47
Here f (x + 0) and f (x − 0) are the right and left- hand limits of f (x) at x and represent
lim→+0 f (x + ) and lim→0+ f (x − ), respectively.

0 −5 < x < 0
Example 5.6 How should f (x) be defined at x = −5, x = 0, and x = 5
3 0<x<5
in order that the Fourier series will converge to f (x) for −5 ≤ x ≤ 5?

Solution
Since f (x) satisfies the Dirichlet conditions, we can say that the series converges to f (x) at
all points of continuity and to f (x+0)+f
2
(x−0)
at points of discontinuity. At x = −5, 0, and 5,
which are points of discontinuity, the series converge to (3 + 0)/2 = 3/2 as seen from the
graph. If we redefine f (x) as follows,


 3/2 x = −5
0 −5 <x<0



f (x) = 3/2 x=0 Period = 10
3 0<x<5




3/2 x=5


Then the series will converge to f (x) for −5 ≤ x ≤ 5.

5.7 Fourier series for periodic functions of period 2π

Given that certain conditions are satisfied then it is possible to write a periodic function of
period 2π as a series expansion of the orthogonal periodic functions just discussed. That is,
if f (x) is defined on the interval −π ≤ x ≤ π where f (x + 2πn) = f (x) then
∞
a0 X
f (x) = + (an cos nx + bn sin nx)
2 n=1

The coefficients of an
Z π
f (x) cos mxdx
−π
∞
!
Z π
a0 X
= + (an cos nx + bn sin nx) cos mxdx
−π 2 n=1
∞
a0 π
Z X Z π
= cos mxdx + an cos nx cos mxdx
2 −π n=1 −π

X∞ Z π
+ bn sin nx cos mxdx
n=1 −π
∞ ∞
a0 X X
= ×0+ an πδnm + bn × 0
2 n=1 n=1
=am π
and so
1 π
Z
am = f (x) cos mxdx
π −π
1 π
Z
an = f (x) cos nxdx, n = 0, 1, 2, . . .
π −π

48
The coefficient of bn
Z π
f (x) sin mxdx
−π
∞
!
Z π
a0 X
= (an cos nx + bn sin nx) sin mxdx
−π 2 n=1
∞
a0 π
Z X Z π
= sin mxdx + an cos nx sin mxdx
2 −π n=1 −π

X∞ Z π
+ bn sin nx sin mxdx
n=1 −π
∞ ∞
a0 X X
= ×0+ an × 0 + bn πδnm
2 n=1 n=1
=bm π
1 π
Z
∴ bn = f (x) sin nxdx, n = 0, 1, 2, . . .
π −π

Example 5.7 Expand f (x) = x2 , 0 < x < 2π in a Fourier series if (a) the period is 2π
Solution
(a) The graph of f (x) with period 2π is shown below

Fig. 5.4:

Period = 2L = 2π. Choosing c = 0, we have

1 c+2L
Z Z 2π
an = f (x) cos(nπx/L)dx = 1π x2 cos nxdx
L c 0
      2π
1 2 sin nx − cos nx − sin nx 4
= x − 2x 2
+2 3
= 2, n 6= 0
π n n n 0 n
1
R 2π 8π 2
If n = 0, a0 = π 0
x2 dx = 3

1 c+2L
Z Z 2π
bn = f (x) sin(nπx/L)dx = 1π x2 sin nxdx
L c 0
 
1 2 cos nx  
sin nx
  cos nx 2π −4π
= x − − (2x) − 2 + (2) =
π n n n3 0 n

Then f (x) = 4π3 + ∞

2 P 4 4π


n=1 n 2 cos nx − n
sin nx .
This is valid for 0 < x < 2π. At x = 0 and x = 2π the series converges to 2π 2 .

Example 5.8 Find the Fourier series of f (x) = x − x2 , −π < x < π

Solution

49
Let f (x) = x − x2 for − π ≤ x ≤ π. Here L = π. Compute

1 π
Z
2
a0 = (x − x2 )dx = − π 2 ,
π −π 3
Z π
1
an = (x − x2 ) cos nxdx
π −π
4 sin nπ − 4nπ cos nπ − 2n2 π 2 sin nπ
=
n3 π
4 4
= − 2 cos nπ = − 2 (−1)n
n n
n+1
4(−1)
=
n2
and
1 π
Z
bn = (x − x2 ) sin nxdx
π −π
=2 sin nπ − 2nπ cos nx
2 2
= − cos nx = − (−1)n
n n
2(−1)n+1
= .
n
We have used the facts that sin nx = 0 and cos nx = (−1)n if n is an integer. The Fourier
series of f (x) = x − x2 on [−π, π] is
∞ 
1 2 X 4(−1)n+1 2(−1)n+1

− π + 2
cos nx + sin nx .
3 n=1
n n

5.8 Products of odd and even functions

Theorem 5.1 If f (x) is defined over the interval −π < x < π and f (x) is even, then
the Fourier series for f (x) contains cosine terms only. Included in this is a0 which may be
regarded as an cos nx with n = 0.
R0 Rπ
Proof: Since f (x) is even, −π f (x)dx = 0 f (x)dx
Rπ Rπ Rπ
(a)a0 = π1 −π f (x)dx = π2 0 f (x) ∴ a0 = π2 0 f (x)dx.
Rπ
(b) an = π1 −π f (x) cos nxdx.
But f (x)R cos nx is the product Rof two even functions and therefore itself even.
1 π 2 π
an = π −π f (x) cos nxdx = π 0 f (x) cos nxdx.
Because, since f (x) sin nxdx is the product of an even function and an odd function, it is
itself odd. R
π
∴ bn = π1 −π f (x) sin nxdx = 0 ∴ bn = 0
Therefore, there are no sine terms in the Fourier series for f (x).

Theorem 5.2 If f (x) is an odd function defined over the interval −π < x < π, then the
Fourier series for f (x) contains sine terms only.
R0 Rπ
Proof: Since f (x) is an odd function, −π f (x)dx = − 0 f (x)dx.
Rπ
(a) a0 = π1 −π f (x)dx. But f (x) is odd ∴ a0 = 0
Rπ
(b) an = π1 −π f (x) cos nxdx = 0 because f (x) cos nx is odd function (ie product of odd
and even)

50
 Fig. 5.5:

Example 5.9 Determine the Fourier series for the function shown in figure 5.12nc
Solution: This is neither odd nor even. Therefore we must find a0 , an and bn .
∞
1 X
f (x) = a0 + (an cos nx + bn sin nx)
2 n=1

(a)
1 2π
Z Z π Z 2π 
1 2
a0 = f (x)dx = xdx + 2dx
π 0 π 0 π π
1  2 π 1
= x /π 0 + [2x]2π
π = (π + 4π − 2π) = 3 ∴ a0 − 3
π π
(b)
1 2π
Z Z π Z 2π 
1
an = f (x) cos nxdx = (2x/π) cos nxdx + 2 cos nxdx
π 0 π 0 π
  π Z π Z 2π 
2 1 x sin nx 1
= − sin nxdx + cos nxdx
π π n 0 nπ 0 π
(  2π )
2 1 1 h cos nx iπ sin nx
an = (0 − 0) − − +
π π nπ n 0 n π
 
2 1
= − 2 (−(−1)n + 1) + (0 − 0)
π πn
2
= − 2 2 (1 − (−1)n )
π n
and so a0 = 0 (n even) and an = − π24n2 (n odd)

1 2π
Z Z π Z 2π 
1
bn = f (x) sin nxdx = (2x/π) sin nxdx + 2 sin nxdx
π π 0 π
  π Z π Z 2x 
2 1 −x cos nx 1
= + cos nxdx + sin nxdx
π π n 0 πn 0 π
(  π  2π )
2 1 1 sin nx − cos nx
= (−π cos nπ) + +
π πn πn n 0 n π
 
2 1 1
= − cos nπ + (0 − 0) − (cos 2πn − cos nπ)
π n n
 
2 1 2
= − cos 2nπ = − cos 2nπ
π n πn
2
But cos 2nπ = 1 . ∴ bn = − πn
 
3 4 1 1
f (x) = − 2 cos x + cos 3x + cos 5x + . . .
2 π 9 25
 
2 1 1 1
− sin x + sin 2x + sin 3x + sin 4x . . .
π 2 3 4

51
Theorem 5.3 Fourier’s theorem
This theorem states that a periodic function that satisfies certain conditions can be expressed
as the sum of a number of sine functions of different amplitudes, phases and periods. That
is, if f (t) is a periodic function with period T then

f (t) =A0 + A1 sin(ωt + φ1 ) + A2 sin(2ωt + φ2 ) + . . .

(5.15)
+ An sin(nωt + φn ) + . . .

Where As and φs are constants and ω = 2π/T is the frequency of f (t). The term A1 sin(ωt+
φ1 ) is called the first harmonic or fundamental mode, and it has the same frequency ω as
the parent function f (t). The term An sin(nωt + φn ) is called the n-th harmonic, and it
has frequency nω, which is n times that of the fundamental. An denotes the amplitude of
the n-th harmonic and φn is its phase angle, measuring the lag or lead of the n-th harmonic
with reference to a pure sine wave of the same frequency.
Since
An sin(nωt + φn ) ≡ (An cos φn ) sin nωt + (An sin φn ) cos nωt
≡ bn sin nωt + an cos nωt

where
bn = An cos φn , an = An sin φn (5.16)
the expression 5.33 can be written as
∞ ∞
1 X X
f (t) = a0 + an cos nωt + bn sin nωt (5.17)
2 n=1 n=1

where a0 = 2A0 (we shall see later that taking the first term as 12 a0 rather than a0 is a
convenience that enables us to make a0 fit a general result). The expression (5.35) is called
the Fourier series expansion of the function f (t), and as and bs are called respectively
as the in-phase and phase quadrature components of the n-th harmonic, this terminology
arising from the use of the phasor notation einωt = cos nωt + i sin nωt. Clearly, 5.33 is the
alternative representation of the Fourier series with the amplitude and phase of the n-th
harmonic being determined from (5.34) as
p
An = (a2n + b2n ), φn = tan−1 (an /bn )

with care being taken over choice of quadrant.

The Fourier coefficients are given by

2 d+T
Z
an = f (t) cos nωtdt (n = 0, 1, 2, . . . ) (5.18)
T d
Z d+T
2
bn = f (t) sin nωtdt (n = 0, 1, 2, . . . ) (5.19)
T d

Example 5.10 A periodic function f (t) of period 4 (that is , f (t + 4) = f (t)) is defined in

the range −2 < t < 2 by 
0 (−2 < t < 0)
f (t) =
1 (0 < t < 2)
Sketch a graph of f (t) for −6 ≤ t ≤ 6 and obtain a Fourier series expansion for the function.
Solution we have

52
 Fig. 5.6:

Z 2 Z 0 Z 2 
1 1
a0 = f (t)dt = 0dt + 1dt = 1
2 −2 2 −2 0

1 2
Z
1
an = f (t) cos nπtdt (n = 1, 2, 3, . . . )
2 −2 2
Z 0 Z 2 
1 1
= 0dt + cos nπtdt = 0
2 −2 0 2

and
1 2
Z
1
bn = f (t) sin nπtdt (n = 1, 2, 3, . . . )
2 −2 2
Z 0 Z 2 
1 1 1 1
= 0dt + sin nπtdt = (1 − cos nπ) = [1 − (−1)n ]
2 2 nπ nπ
 −2 0
0 (even n)
=
2/nπ (odd n)

Thus, from (5.35), the Fourier series expression of f (t) is

1 2 1 1 3 1 5
f (t) = + (sin πt + sin πt + sin πt + . . . )
2 π 2 3 2 5 2
∞
1 2 X 1 1
= + sin (2n − 1)πt
2 π n=1 2n − 1 2

5.9 Half Range Fourier Sine and Cosine Series

A half range Fourier sine or cosine series is a series in which only sine terms or only cosine
terms are present, respectively. When a half range series corresponding to a given function
is desired, the function is generally defined in the interval (0, L) [which is half of the interval
(−L, L), thus accounting for the name half range] and then the function is specified as odd
or even, so that it is clearly defined in the other half of the interval, namely, (−L, 0). In
such case, we have
( RL
an = 0, bn = L2 0 f (x) sin nπx L
dx for half range sine series
2 L nπx
R
bn = 0, an = L 0 f (x) cos L dx for half range cosine series

Example 5.11 Expand f (x) = x, 0 < x < 2, in a half range (a) sine series, (b) cosine
series
Solution (a) Extend the definition of the given function to that of the odd function of
period 4 shown in Fig. 5.14. This is sometimes called the odd extension of f (x). Then
2L = 4, L = 2

53
 Fig. 5.7:

Thus, an = 0 and
2 L 2 2
Z Z
nπx nπx
bn = f (x) sin dx = (x) sin dx
L 0 L 2 0 2
     2
−2 nπx −4 nπx −4
= (x) cos − (1) sin = cos nπ
nπ 2 n2 π 2 2 0 nπ
Then
∞
X −4 nπx
f (x) = cos nπ sin
nπ
n=1
2
 
4 πx 1 2πx 1 3πx
= sin − sin + sin − ...
π 2 2 2 3 2
(b) Extend the function of f (x) to that of the even function of period 4 shown in figure
5.15. This is the even extension of f (x). Then 2L = 4, L = 2.

Fig. 5.8:

Thus, bn = 0 and
2 L 2 2
Z Z
nπx nπx
an = f (x) cos dx = (x) cos dx
L 0 L 2 0 2
     2
2 nπx −4 nπx
= (x) sin − (1) cos
nπ 2 n2 π 2 2 0
4
= 2 2 (cos nπ − 1) If n 6= 0
nπ
R2
If n = 0, a0 = 0
xdx = 2.
Then
∞
X 4 nπx
f (x) =1 + (cos nπ − 1) cos
n=1
n2 π 2 2
 
8 πx 1 3πx 1 5πx
=1 − 2 cos + 2 cos + 2 cos + ...
π 2 3 2 5 2

54
5.10 Complex Fourier Series
The Euler identities eix = cos x + i sin x and e−ix = cos x − i sin x allow us to write
eix + e−ix eix − e−ix
cos x = and sin x =
2 2i

Fig. 5.9: The function f (x) defined for 0 ≤ x < 2π

When these results are used in the real variable Fourier series representation of f (x) over
the interval −L ≤ x ≤ L, it becomes
∞ 
+ e−inπx/L − e−inπx/L
 inπx/L   inπx/L 
X e e
f (x) = a0 + an + bn ,
n=1
2 2i

and after grouping terms we have

∞   ∞  
X an − ibn X an + ibn
f (x) = a0 + einπx/L
+ e−inπx/L (5.20)
n=1
2 n=1
2

If we now define
an − ibn an + ibn
c0 = a0 , cn = , and c−n = for n = 1, 2 . . . , (5.21)
2 2
the Fourier series presentation in (5.38) becomes
k
X
f (x) = lim cn einπx/L for − L ≤ x ≤ L. (5.22)
k→∞
n=−k

This is the complex or exponential form of the Fourier series representation of f (x).
If real functions f (x) are considered, the Fourier coefficients an and bn are real, and (5.39)
then shows that cn and c−n are complex conjugates, because c−n = c¯n . To proceed further
we now make use of the fact that the functions exp(imπx/L) and exp(−inπx/L) are
orthogonal over the interval −L ≤ x ≤ L, because integration shows that
Z L 
imπx/L inπx/L 0, for m 6= −n
e e dx =
−L 2π for m = −n for m, n positive integers
Multiplication of (5.40) by exp(−imπx/L), following by integration over −L ≤ x ≤ L and
use of the above orthogonality condition gives
Z L
1
cn = f (x)e−inπx/L dx, for n = 0, ±1, ±2, . . . (5.23)
2L −L

Collecting these results we arrive at the following definition.

55
The complex form of a Fourier series
Let the real function f (x) be defined on the interval −L ≤ x ≤ L. Then the complex
Fourier series representation of f (x) is
k
X
f (x) = lim cn einπx/L for − L ≤ x ≤ L,
k→∞
n=−k

where Z L
1
cn = f (x)e−inπx/L dx for n = 0, ±1, ±2, . . .
2L −L

Example 5.12 Find the complex Fourier series representation of


 0, −π < x < −π/2
f (x) = 1, −π/2 < x < π/2
0, π/2 < x < π


Solution As the function of f (x) is defined on the interval −π < x < π, we have L = π,
so the coefficients cn are given by
Z π Z π/2
1 1 1
c0 = f (x)dx = 1dx =
2π −π 2π −π/2 2
and

π π/2
einπ/2 − e−inπ/2
Z Z  
1 −inx 1 −inx 1
cn = f (x)e dx = e dx = , for n = ±1, ±2, . . .
2π −π 2π −π/2 nπ 2i

The coefficients cn reduce to the real values

1 nπ
cn = sin for n = ±1, ±2, . . .
nπ 2
so cn = c−n because cn is an even function of n. Consideration of the function sin(nπ/2)
for integer values of n shows that

(−1)n−1
c2n−1 = and c2n = 0 for n = 1, 2, . . .
π(2n − 1)
Thus, the complex Fourier series representation of f (x) is
k
1 X
f (x) = + lim cn (einx + e−inx ).
2 k→∞ n=−k

with einx + e−inx = 2 cos nx, the complex Fourier series

∞
1 2X cos(2n − 1)x
f (x) = + (−1)n+1
2 π n=1 (2n − 1)

Example 5.13 Find the complex Fourier series representation of


0, 0 < x < 1
f (x) =
1, 1 < x < 4

56
Solution The function f (x) is defined on the interval 0 ≤ x ≤ 2L, with 2L = 4, so L = 2.
Thus, the complex Fourier coefficients cn are given by

1 4 1 4 −inπ/2
Z Z
−inπ/2
cn = f (x)e dx = e dx, for n = 0, ±1, ±2, . . .
4 0 4 1
Setting n = 0 gives
3
c0 = ,
4
whereas
i
cn = [1 − einπ/2 ], for n = ±1, ±2, . . .
2πn
So the complex Fourier series representation of f (x) is
k
X
f (x) = c0 + lim cn einπx/2 ,
k→∞
n=−k

with c0 and cn defined as shown.

57
5.11 The Fourier Series of a Function
Let f (x) be defined on [−L, L]. We want to choose numbers a0 , a2 , . . . and b1 , b2 , . . . such
that ∞
1 X
f (x) = a0 + [ak cos(kπx/L) + bk sin(kπx/L)] (5.24)
2 k=1

This decomposition of the function into a sum of terms , each representing the influence of
a different fundamental frequency on the behavior of the function.
To determine a0 , integrate equation (5.24) term by term to get

Z L Z L
1
f (x)dx = a0 dx
−L 2 −L
∞ 
X Z L Z L 
+ ak cos(kπx/L)dx + bk sin(kπx/L)dx
k=1 −L −L

1
= a0 (2L) = πa0
2
because all of the integrals in the summation are zero. Then

1 L
Z
a0 = f (x)dx (5.25)
L −L

To solve for the other coefficients in the proposed equation (5.24), we will use the following
three facts, which follow by routine integrations. Let m and n be integers. Then
Z L
cos(nπx/L) sin(mπx/L)dx = 0. (5.26)
−L

Furthermore, if n 6= m then
Z L Z L
cos(nπx/L) cos(mπx/L)dx = sin(nπx/L) sin(mπx/L)dx = 0. (5.27)
−L −L

And, if n 6= 0, then
Z L Z L
2
cos (nπx/L)dx = sin2 (nπx/L)dx = L. (5.28)
−L −L

Now let n be any positive integer. To solve for an , multiply equation (5.24) by cos(nπx/L)
and integrate the resulting equation to get

Z L Z L
1
f (x) cos(nπx/L)dx = a0 cos(nπx/L)dx
−L 2 −L
X∞  Z L Z L 
+ ak cos(kπx/L) cos(nπx/L)dx + bk sin(kπx/L) cos(nπx/L)dx
k=1 −L −L

Because equations (5.26) and (5.27), all of the terms on the right are zero except the
coefficient of an , which occurs in the summation when k = n. The last equation reduces to
Z L Z L
f (x) cos(nπx/L)dx = an cos2 (nπx/L)dx = an L
L −L

58
by equation (5.28). Therefore
Z L
1
an = f (x) cos(nπx/L)dx. (5.29)
L −L

This expression contains a0 if we let n = 0.

Similarly, if we multiply equation (5.24) by sin(nπx/L) and integrate, we obtain

1 L
Z
bn = f (x) sin(nπx/L)dx. (5.30)
L −L
The numbers Z L
1
an = f (x) cos(nπx/L)dx for n = 0, 1, 2, . . . (5.31)
L −L
Z L
1
bn = f (x) sin(nπx/L)dx for n = 0, 1, 2, . . . (5.32)
L −L

are called the Fourier coefficients of f on [-L,L]. When these numbers are used, the series
in (5.24) is called the Fourier series of f on [-L,L].

Example 5.14 (a) Find the Fourier coefficients corresponding to the function

0 −5 < x < 0
f (x) =
3 0 < x < 5 Period = 10

(b) Write the corresponding Fourier series

Solution
The graph of f (x) is

Fig. 5.10:

(a) Period = 2L = 10 and L = 5. Choose the interval c to c + 2L as −5 to 5, so that

c = −5. Then

1 c+2L 1 5
Z Z
an = f (x) cos(nπx/L)dx = f (x) cos(nπx/5)dx
L c 5 −5
Z 0 Z 5
3 5
 Z
1
= (0) cos(nπx/5)dx + (3) cos(nπx/5)dx = cos(nπx/5)
5 −5 0 5 0
 5
3 5
= sin(nπx/5) = 0 if n 6= 0
5 nπ 0

If n = 0 Z 5 Z 5
3 3
a0 = cos(0πx/5)dx = dx = 3.
5 0 5 0

59
 1 c+2L 1 5
Z Z
bn = f (x) sin(nπx/L)dx = f (x) sin(nπx/5)dx
L c 5 −5
Z 0 Z 5
3 5
 Z
1
= (0) sin(nπx/5)dx + (3) sin(nπx/5)dx = sin(nπx/5)
5 −5 0 5 0
 5
3 5 3(1 − cos nπ)
= − cos(nπx/5) =
5 nπ 0 nπ

(b) The corresponding Fourier series is

∞ ∞
a0 X 3 X 3(1 − cos nπ)
+ (an cos(nπx/L) + bn sin(nπ/L)) = + sin(nπx/5)
2 n=1
2 n=1 nπ

3 6 1 1
= + sin(xπ/5) + sin(3xπ/5) + sin(5πx/5) + . . .
2 π 3 5

5.12 Dirichlet Conditions

Suppose that

1. f (x) is defined except possibly at a finite number of points in (−L, L)

2. f (x) is periodic outside (−L, L) with period 2L

3. f (x) and f 0 (x) are piecewise continuous in −L, L.

Then the series (5.24) with Fourier coefficients converges to

(a) f (x) if x is a point of continuity

f (x+0)+f (x−0)
(b) 2
if x is a point of discontinuity

Here f (x + 0) and f (x − 0) are the right and left- hand limits of f (x) at x and represent
lim→+0 f (x + ) and lim→0+ f (x − ), respectively.

0 −5 < x < 0
Example 5.15 How should f (x) be defined at x = −5, x = 0, and
3 0<x<5
x = 5 in order that the Fourier series will converge to f (x) for −5 ≤ x ≤ 5?

Solution
Since f (x) satisfies the Dirichlet conditions, we can say that the series converges to f (x) at
all points of continuity and to f (x+0)+f
2
(x−0)
at points of discontinuity. At x = −5, 0, and 5,
which are points of discontinuity, the series converge to (3 + 0)/2 = 3/2 as seen from the
graph. If we redefine f (x) as follows,


 3/2 x = −5
 0 −5 < x < 0


f (x) = 3/2 x=0 Period = 10
3 0<x<5




3/2 x=5


Then the series will converge to f (x) for −5 ≤ x ≤ 5.

60
5.13 Fourier series for periodic functions of period
2π
Given that certain conditions are satisfied then it is possible to write a periodic function of
period 2π as a series expansion of the orthogonal periodic functions just discussed. That is,
if f (x) is defined on the interval −π ≤ x ≤ π where f (x + 2πn) = f (x) then
∞
a0 X
f (x) = + (an cos nx + bn sin nx)
2 n=1

The coefficients of an
Z π
f (x) cos mxdx
−π
∞
!
Z π
a0 X
= + (an cos nx + bn sin nx) cos mxdx
−π 2 n=1
∞
a0 π
Z X Z π
= cos mxdx + an cos nx cos mxdx
2 −π n=1 −π

X∞ Z π
+ bn sin nx cos mxdx
n=1 −π
∞∞
a0 X
X
= ×0+ an πδnm + bn × 0
2 n=1 n=1
=am π
and so
1 π
Z
am = f (x) cos mxdx
π −π
1 π
Z
an = f (x) cos nxdx, n = 0, 1, 2, . . .
π −π
The coefficient of bn
Z π
f (x) sin mxdx
−π
∞
!
Z π
a0 X
= (an cos nx + bn sin nx) sin mxdx
−π 2 n=1
∞
a0 π
Z X Z π
= sin mxdx + an cos nx sin mxdx
2 −π n=1 −π

X∞ Z π
+ bn sin nx sin mxdx
n=1 −π
∞ ∞
a0 XX
= ×0+ an × 0 + bn πδnm
2 n=1 n=1
=bm π
1 π
Z
∴ bn = f (x) sin nxdx, n = 0, 1, 2, . . .
π −π

61
 Fig. 5.11:

Example 5.16 Expand f (x) = x2 , 0 < x < 2π in a Fourier series if (a) the period is 2π
Solution
(a) The graph of f (x) with period 2π is shown below
Period = 2L = 2π. Choosing c = 0, we have
1 c+2L
Z Z 2π
an = f (x) cos(nπx/L)dx = 1π x2 cos nxdx
L c 0
      2π
1 2 sin nx − cos nx − sin nx 4
= x − 2x 2
+2 3
= 2 , n 6= 0
π n n n 0 n
R 2π 2
If n = 0, a0 = π1 0 x2 dx = 8π3

1 c+2L
Z Z 2π
bn = f (x) sin(nπx/L)dx = 1π x2 sin nxdx
L c 0

1 2  cos nx 

sin nx
  cos nx 2π −4π
= x − − (2x) − 2 + (2) =
π n n n3 0 n

Then f (x) = 4π3 + ∞

2 P 4 4π


n=1 n 2 cos nx − n
sin nx .
This is valid for 0 < x < 2π. At x = 0 and x = 2π the series converges to 2π 2 .

Example 5.17 Find the Fourier series of f (x) = x − x2 , −π < x < π

Solution
Let f (x) = x − x2 for − π ≤ x ≤ π. Here L = π. Compute
1 π
Z
2
a0 = (x − x2 )dx = − π 2 ,
π −π 3
Z π
1
an = (x − x2 ) cos nxdx
π −π
4 sin nπ − 4nπ cos nπ − 2n2 π 2 sin nπ
=
n3 π
4 4
= − 2 cos nπ = − 2 (−1)n
n n
4(−1)n+1
=
n2
and
1 π
Z
bn = (x − x2 ) sin nxdx
π −π
=2 sin nπ − 2nπ cos nx
2 2
= − cos nx = − (−1)n
n n
2(−1)n+1
= .
n
62
We have used the facts that sin nx = 0 and cos nx = (−1)n if n is an integer. The Fourier
series of f (x) = x − x2 on [−π, π] is
∞ 
1 2 X 4(−1)n+1 2(−1)n+1

− π + 2
cos nx + sin nx .
3 n=1
n n

5.14 Products of odd and even functions

Theorem 5.4 If f (x) is defined over the interval −π < x < π and f (x) is even, then
the Fourier series for f (x) contains cosine terms only. Included in this is a0 which may be
regarded as an cos nx with n = 0.
R0 Rπ
Proof: Since f (x) is even, −π f (x)dx = 0 f (x)dx
Rπ Rπ Rπ
(a)a0 = π1 −π f (x)dx = π2 0 f (x) ∴ a0 = π2 0 f (x)dx.
Rπ
(b) an = π1 −π f (x) cos nxdx.
But f (x)R cos nx is the product Rof two even functions and therefore itself even.
1 π 2 π
an = π −π f (x) cos nxdx = π 0 f (x) cos nxdx.
Because, since f (x) sin nxdx is the product of an even function and an odd function, it is
itself odd. R
π
∴ bn = π1 −π f (x) sin nxdx = 0 ∴ bn = 0
Therefore, there are no sine terms in the Fourier series for f (x).

Theorem 5.5 If f (x) is an odd function defined over the interval −π < x < π, then the
Fourier series for f (x) contains sine terms only.
R0 Rπ
Proof: Since f (x) is an odd function, −π f (x)dx = − 0 f (x)dx.
Rπ
(a) a0 = π1 −π f (x)dx. But f (x) is odd ∴ a0 = 0
Rπ
(b) an = π1 −π f (x) cos nxdx = 0 because f (x) cos nx is odd function (ie product of odd
and even)

Example 5.18 Determine the Fourier series for the function shown in figure 5.12nc

Fig. 5.12:

Solution: This is neither odd nor even. Therefore we must find a0 , an and bn .
∞
1 X
f (x) = a0 + (an cos nx + bn sin nx)
2 n=1

(a)

1 2π
Z Z π Z 2π 
1 2
a0 = f (x)dx = xdx + 2dx
π 0 π 0 π π
1  2 π 1
= x /π 0 + [2x]2π
π = (π + 4π − 2π) = 3 ∴ a0 − 3
π π

63
(b)
1 2π
Z Z π Z 2π 
1
an = f (x) cos nxdx = (2x/π) cos nxdx + 2 cos nxdx
π 0 π 0 π
  π Z π Z 2π 
2 1 x sin nx 1
= − sin nxdx + cos nxdx
π π n 0 nπ 0 π
(  2π )
2 1 1 h cos nx iπ sin nx
an = (0 − 0) − − +
π π nπ n 0 n π
 
2 1
= − 2 (−(−1)n + 1) + (0 − 0)
π πn
2
= − 2 2 (1 − (−1)n )
π n
and so a0 = 0 (n even) and an = − π24n2 (n odd)

1 2π
Z Z π Z 2π 
1
bn = f (x) sin nxdx = (2x/π) sin nxdx + 2 sin nxdx
π π 0 π
  π Z π Z 2x 
2 1 −x cos nx 1
= + cos nxdx + sin nxdx
π π n 0 πn 0 π
(  π  2π )
2 1 1 sin nx − cos nx
= (−π cos nπ) + +
π πn πn n 0 n π
 
2 1 1
= − cos nπ + (0 − 0) − (cos 2πn − cos nπ)
π n n
 
2 1 2
= − cos 2nπ = − cos 2nπ
π n πn
2
But cos 2nπ = 1 . ∴ bn = − πn

 
3 4 1 1
f (x) = − 2 cos x + cos 3x + cos 5x + . . .
2 π 9 25
 
2 1 1 1
− sin x + sin 2x + sin 3x + sin 4x . . .
π 2 3 4
Theorem 5.6 Fourier’s theorem
This theorem states that a periodic function that satisfies certain conditions can be expressed
as the sum of a number of sine functions of different amplitudes, phases and periods. That
is, if f (t) is a periodic function with period T then
f (t) =A0 + A1 sin(ωt + φ1 ) + A2 sin(2ωt + φ2 ) + . . .
(5.33)
+ An sin(nωt + φn ) + . . .

Where As and φs are constants and ω = 2π/T is the frequency of f (t). The term A1 sin(ωt+
φ1 ) is called the first harmonic or fundamental mode, and it has the same frequency ω as
the parent function f (t). The term An sin(nωt + φn ) is called the n-th harmonic, and it
has frequency nω, which is n times that of the fundamental. An denotes the amplitude of
the n-th harmonic and φn is its phase angle, measuring the lag or lead of the n-th harmonic
with reference to a pure sine wave of the same frequency.
Since
An sin(nωt + φn ) ≡ (An cos φn ) sin nωt + (An sin φn ) cos nωt
≡ bn sin nωt + an cos nωt

64
where
bn = An cos φn , an = An sin φn (5.34)
the expression 5.33 can be written as
∞ ∞
1 X X
f (t) = a0 + an cos nωt + bn sin nωt (5.35)
2 n=1 n=1

where a0 = 2A0 (we shall see later that taking the first term as 12 a0 rather than a0 is a
convenience that enables us to make a0 fit a general result). The expression (5.35) is called
the Fourier series expansion of the function f (t), and as and bs are called respectively
as the in-phase and phase quadrature components of the n-th harmonic, this terminology
arising from the use of the phasor notation einωt = cos nωt + i sin nωt. Clearly, 5.33 is the
alternative representation of the Fourier series with the amplitude and phase of the n-th
harmonic being determined from (5.34) as
p
An = (a2n + b2n ), φn = tan−1 (an /bn )

with care being taken over choice of quadrant.

The Fourier coefficients are given by

2 d+T
Z
an = f (t) cos nωtdt (n = 0, 1, 2, . . . ) (5.36)
T d
Z d+T
2
bn = f (t) sin nωtdt (n = 0, 1, 2, . . . ) (5.37)
T d

Example 5.19 A periodic function f (t) of period 4 (that is , f (t + 4) = f (t)) is defined in

the range −2 < t < 2 by 
0 (−2 < t < 0)
f (t) =
1 (0 < t < 2)
Sketch a graph of f (t) for −6 ≤ t ≤ 6 and obtain a Fourier series expansion for the function.
Solution we have

Fig. 5.13:

Z 2 Z 0 Z 2 
1 1
a0 = f (t)dt = 0dt + 1dt = 1
2 −2 2 −2 0

1 2
Z
1
an = f (t) cos nπtdt (n = 1, 2, 3, . . . )
2 −2 2
Z 0 Z 2 
1 1
= 0dt + cos nπtdt = 0
2 −2 0 2

65
and
1 2
Z
1
bn = f (t) sin nπtdt (n = 1, 2, 3, . . . )
2 −2 2
Z 0 Z 2 
1 1 1 1
= 0dt + sin nπtdt = (1 − cos nπ) = [1 − (−1)n ]
2 2 nπ nπ
 −2 0
0 (even n)
=
2/nπ (odd n)

Thus, from (5.35), the Fourier series expression of f (t) is

1 2 1 1 3 1 5
f (t) = + (sin πt + sin πt + sin πt + . . . )
2 π 2 3 2 5 2
∞
1 2 X 1 1
= + sin (2n − 1)πt
2 π n=1 2n − 1 2

5.15 Half Range Fourier Sine and Cosine Series

A half range Fourier sine or cosine series is a series in which only sine terms or only cosine
terms are present, respectively. When a half range series corresponding to a given function
is desired, the function is generally defined in the interval (0, L) [which is half of the interval
(−L, L), thus accounting for the name half range] and then the function is specified as odd
or even, so that it is clearly defined in the other half of the interval, namely, (−L, 0). In
such case, we have
( RL
an = 0, bn = L2 0 f (x) sin nπx L
dx for half range sine series
2 L nπx
R
bn = 0, an = L 0 f (x) cos L dx for half range cosine series

Example 5.20 Expand f (x) = x, 0 < x < 2, in a half range (a) sine series, (b) cosine
series
Solution (a) Extend the definition of the given function to that of the odd function of
period 4 shown in Fig. 5.14. This is sometimes called the odd extension of f (x). Then
2L = 4, L = 2

Fig. 5.14:

Thus, an = 0 and

2 L 2 2
Z Z
nπx nπx
bn = f (x) sin dx = (x) sin dx
L 0 L 2 0 2
     2
−2 nπx −4 nπx −4
= (x) cos − (1) sin = cos nπ
nπ 2 n2 π 2 2 0 nπ

66
Then
∞
X −4 nπx
f (x) = cos nπ sin
nπ
n=1
2
 
4 πx 1 2πx 1 3πx
= sin − sin + sin − ...
π 2 2 2 3 2

(b) Extend the function of f (x) to that of the even function of period 4 shown in figure
5.15. This is the even extension of f (x). Then 2L = 4, L = 2.

Fig. 5.15:

Thus, bn = 0 and

2 L 2 2
Z Z
nπx nπx
an = f (x) cos dx = (x) cos dx
L 0 L 2 0 2
     2
2 nπx −4 nπx
= (x) sin − (1) cos
nπ 2 n2 π 2 2 0
4
= 2 2 (cos nπ − 1) If n 6= 0
nπ
R2
If n = 0, a0 = 0
xdx = 2.
Then
∞
X 4 nπx
f (x) =1 + (cos nπ − 1) cos
n=1
n2 π 2 2
 
8 πx 1 3πx 1 5πx
=1 − 2 cos + 2 cos + 2 cos + ...
π 2 3 2 5 2

5.16 Complex Fourier Series

The Euler identities eix = cos x + i sin x and e−ix = cos x − i sin x allow us to write
eix + e−ix eix − e−ix
cos x = and sin x =
2 2i
When these results are used in the real variable Fourier series representation of f (x) over
the interval −L ≤ x ≤ L, it becomes
∞ 
+ e−inπx/L − e−inπx/L
 inπx/L   inπx/L 
X e e
f (x) = a0 + an + bn ,
n=1
2 2i

and after grouping terms we have

∞   ∞  
X an − ibn inπx/L X an + ibn −inπx/L
f (x) = a0 + e + e (5.38)
n=1
2 n=1
2

67
 Fig. 5.16: The function f (x) defined for 0 ≤ x < 2π

If we now define
an − ibn an + ibn
c0 = a0 , cn = , and c−n = for n = 1, 2 . . . , (5.39)
2 2
the Fourier series presentation in (5.38) becomes
k
X
f (x) = lim cn einπx/L for − L ≤ x ≤ L. (5.40)
k→∞
n=−k

This is the complex or exponential form of the Fourier series representation of f (x).
If real functions f (x) are considered, the Fourier coefficients an and bn are real, and (5.39)
then shows that cn and c−n are complex conjugates, because c−n = c¯n . To proceed further
we now make use of the fact that the functions exp(imπx/L) and exp(−inπx/L) are
orthogonal over the interval −L ≤ x ≤ L, because integration shows that
Z L 
imπx/L inπx/L 0, for m 6= −n
e e dx =
−L 2π for m = −n for m, n positive integers
Multiplication of (5.40) by exp(−imπx/L), following by integration over −L ≤ x ≤ L and
use of the above orthogonality condition gives
Z L
1
cn = f (x)e−inπx/L dx, for n = 0, ±1, ±2, . . . (5.41)
2L −L
Collecting these results we arrive at the following definition.

The complex form of a Fourier series

Let the real function f (x) be defined on the interval −L ≤ x ≤ L. Then the complex
Fourier series representation of f (x) is
k
X
f (x) = lim cn einπx/L for − L ≤ x ≤ L,
k→∞
n=−k

where Z L
1
cn = f (x)e−inπx/L dx for n = 0, ±1, ±2, . . .
2L −L

Example 5.21 Find the complex Fourier series representation of


 0, −π < x < −π/2
f (x) = 1, −π/2 < x < π/2
0, π/2 < x < π


Solution As the function of f (x) is defined on the interval −π < x < π, we have L = π,
so the coefficients cn are given by

68
 Z π Z π/2
1 1 1
c0 = f (x)dx = 1dx =
2π −π 2π −π/2 2
and

π π/2
einπ/2 − e−inπ/2
Z Z  
1 −inx 1 −inx 1
cn = f (x)e dx = e dx = , for n = ±1, ±2, . . .
2π −π 2π −π/2 nπ 2i

The coefficients cn reduce to the real values

1 nπ
cn = sin for n = ±1, ±2, . . .
nπ 2
so cn = c−n because cn is an even function of n. Consideration of the function sin(nπ/2)
for integer values of n shows that

(−1)n−1
c2n−1 = and c2n = 0 for n = 1, 2, . . .
π(2n − 1)
Thus, the complex Fourier series representation of f (x) is
k
1 X
f (x) = + lim cn (einx + e−inx ).
2 k→∞
n=−k

with einx + e−inx = 2 cos nx, the complex Fourier series

∞
1 2X cos(2n − 1)x
f (x) = + (−1)n+1
2 π n=1 (2n − 1)

Example 5.22 Find the complex Fourier series representation of


0, 0 < x < 1
f (x) =
1, 1 < x < 4

Solution The function f (x) is defined on the interval 0 ≤ x ≤ 2L, with 2L = 4, so L = 2.

Thus, the complex Fourier coefficients cn are given by

1 4 1 4 −inπ/2
Z Z
−inπ/2
cn = f (x)e dx = e dx, for n = 0, ±1, ±2, . . .
4 0 4 1
Setting n = 0 gives
3
c0 = ,
4
whereas
i
[1 − einπ/2 ], for n = ±1, ±2, . . .
cn =
2πn
So the complex Fourier series representation of f (x) is
k
X
f (x) = c0 + lim cn einπx/2 ,
k→∞
n=−k

with c0 and cn defined as shown.

69
Chapter 6

Functions of Several Variables

6.1 Functions of Two or More Variables

Definition 6.1 Suppose D is a set of n-tuples of real numbers (x1 , x2 , . . . , xn ). A real-
valued function f on D is a rule that assigns a unique (single) real number

w = f (x1 , x2 , . . . , xn )

each element in D. The set D is the function domain. The set of w-values taken on by f
is the function’s range. The symbol w is the dependent variable of f , and f is said to be a
function of the n independent variables x1 to xn . We also call the xj ’s the function’s input
variables and call w the function’s output variables.

If f is a function of two independent variables, we usually call the independent variables x

and y and the dependent variable z, and we picture the domain of f as a region in the xy-
plane (Figure 6.1). If f is a function of three independent variables, we call the independent
variables x, y and z and the dependent variable w, and we picture the domain as a region in
space.
In applications, we tend to use letters that remind us of what the variables stand for. To
say that the volume of a right circular cylinder is a function of its radius and height, we
might write V = f (r, h). To be more specific, we might replace the notation f (r, h) by the
formula that calculates the value of V from the values of r and h, and write V = πr2 h. In
either case, r and h would be the independent variables and V the dependent variables of
the function.

Fig. 6.1: An arrow diagram for the function z = f (x, y)

As usual, we evaluate functions defined by formulas by substituting the values of the inde-
pendent variables in the formula and calculating the corresponding value of the dependent

70
 p
variables. For example, the value of f (x, y, z) = x2 + y 2 + z 2 at the point (3, 0, 4) is
p √
f (3, 0, 4) = (3)2 + (0)2 + (4)2 = 25 = 5.

Domains and Ranges

In defining a function of more than one variable, we follow the usualppractice of excluding
inputs that lead to complex numbers or division by zero. If f (x, y) = y − x2 , y cannot be
1
less than x2 . If f (x, y) = xy , xy cannot be zero. The domain of a function is assumed to the
largest set for which the defining rule generates real numbers, unless the domain is otherwise
specified explicitly. The range consists of the set of output values for the dependent variable.

Example 6.1 (a) These are functions of two variables(Table 6.1). Note the restrictions that
may apply to their domains in order to obtain a real value for the dependent variable z
(b) These are functions of three variables with restrictions on some of the their domains

Table 6.1:
Function
p Domain Range
2
z = y − x2 y≥x [0, ∞)
1
z = xy xy 6= 0 (−∞, 0)U (0, ∞)
z = sin xy Entire plane [-1, 1]

(Table 6.2)

Table 6.2:
Function
p Domain Range
w = x2 + y 2 + z 2 Entire space [0, ∞)
w = x2 +y12 +z2 (x, y, z) 6= (0, 0, 0) (0, ∞)
w = xy ln z Half-space z > 0 (−∞, ∞)

6.2 Partial Derivatives

Functions of Two Variables
Let f be a function of x and y; take for example

f (x, y) = 3x2 y − 5x cos πy.

The partial derivatives of f with respect to x is the function fx obtained by differentiating

f with respect to x, keeping y fixed. In this case

fx (x, y) = 6xy − 5 cos πy.

The partial derivative of f with respect to y is the function fy obtained by differentiating f

with respect to y, keeping x fixed. In this case

fy (x, y) = 3x2 + 5πx sin πy.

These partial derivatives are limits:

71
Definition 6.2 Partial Derivatives (two variables)
Let f be a function of two variables x, y. The partial derivatives of f with respect to x and
with with respect to y are the functions fx and fy defined by setting

f (x + h, y) − f (x, y)
fx (x, y) = lim
h→0 h
f (x, y + h) − f (x, y)
fy (x, y) = lim
h→0 h
provided these limits exist.

Example 6.2 For the function f (x, y) = x arctan xy

x xy
fx (x, y) = x + arctan xy = + arctan xy
1 + (xy)2 1 + x2 y 2

x x2
fy (x, y) = x =
1 + (xy)2 1 + x2 y 2
In the one variable case, f 0 (x0 ) gives the rate of change of f (x) with respect to x at x = x0 .
In the two-variable case, fx (x0 , y0 ) gives the rate of change of f (x, y0 ) with respect to x at
x = x0 , and fy (x0 , y0 ) gives the rate of change of f (x0 , y) with respect to y at y = y0 .

Example 6.3 For the function f (x, y) = exy + ln(x2 + y),

2x 1
fx (x, y) = yexy + and fy (x, y) = xexy +
x2 +y x2 +y
The number
4 4
fx (2, 1) = e2 + = e2 +
4+1 5
gives the rate of change with respect to x of the function

f (x, 1) = ex + ln(x2 + 1) at x = 2;

the number
1 1
fy (2, 1) = 2e2 + = 2e2 +
4+1 5
gives the rate of change with respect to y of the function

f (2, y) = e2y + ln(4 + y) at y = 1

Example 6.4 Let f (x, y) = 4 − 2x2 − y 2 . Find the slope of the tangent line at the point
(1, 1) on the curve formed by the intersection of the surface z = f (x, y) and (a) the plane
y = 1 (b) the plane x = 1
Solution
(a) The slope of the tangent line at any point on the curve formed by the intersection of the
plane y = 1 and the surface z = 4 − 2x2 − y 2 is given by
∂f ∂
= (4 − 2x2 − y 2 ) = −4x
∂x ∂x
In particular, the slope of the required tangent line is

∂f
= −4(1) = −4
∂x (1,1)

72
(b) The slope of the tangent line at any point on the curve formed by the intersection of
the plane x = 1 and the surface z = 4 − 2x2 − y 2 is given by
∂f ∂
= (4 − 2x2 − y 2 ) = −2y
∂y ∂y
In particular, the slope of the required tangent line is

∂f
= −2(1) = −2
∂y (1,1)

(See Figure 6.2)

Fig. 6.2:

Example 6.5 Electrostatic Potential Figure 6.3 shows a crescent-shaped region R that
lies inside the disk D1 = {(x, y)|(x − y)2 + y 2 ≤ 4} and outside the disk D1 = {(x, y)|(x −
y)2 + y 2 ≤ 1}. Suppose that the electrostatic potential along the inner circle is kept at 50
volts and the electrostatic potential along the outer circle is kept at 100 volts. Then the
electrostatic potential at any point (x, y) in R is given by
200x
U (x, y) = 150 −
x2 + y 2
volts

Fig. 6.3: The electrostatic potential inside the crescent-shaped region is U (x, y)

(a) Compute Ux (x, y) and Uy (x, y)

(b) Compute Ux (3, 1) and Uy (3, 1) and interpret your results.

73
Solution
   
∂ 200x ∂ 200x
(a) Ux (x, y) = 100 − 2 =−
∂x x + y2 ∂x x2 + y 2
∂ ∂
(x2 + y 2 ) ∂x (200x) − 200x ∂x (x2 + y2)
=−
(x2 + y 2 )2
200(x2 + y 2 ) − 200x(2x) 200(x2 − y 2 )
=− =
(x2 + y 2 )2 (x2 + y 2 )2
   
∂ 200x ∂ 200x
Uy (x, y) = 150 − 2 =−
∂y x + y2 ∂y x2 + y 2
∂
= − 200x (x2 + y 2 )−1
∂y
∂
= − 200x(−1)(x2 + y 2 )−2 (x2 + y 2 )
∂y
400xy
=200x(x2 + y 2 )−2 (2y) = 2
(x + y 2 )2

200(9 − 1) 400(3)(1)
(b) Ux (3, 1) = 2
= 16 and Uy (3, 1) = = 12
(9 + 1) (9 + 1)2
This tell us that the rate of change of the electrostatic potential at the point (3,1) in the
x-direction is 16 volts per unit change in x with y held fixed at 1, and the rate of change of
the electrostatic potential at the point (3,1) in the y-direction is 12 volts per unit change in
y with x held fixed at 3.

6.3 Higher-Oder Derivatives

Consider the function z = f (x, y) of two variables. Each of the partial derivatives ∂f /∂x
and ∂f /∂y are functions of x and y. Therefore, we may take the partial derivatives of the
these functions to obtain the four second-order partial derivatives

∂ 2f ∂ 2f ∂ 2f ∂ 2f
       
∂ ∂f ∂ ∂f ∂ ∂f ∂ ∂f
= , = , = , and =
∂x2 ∂x ∂x ∂y∂x ∂y ∂x ∂x∂y ∂x ∂y ∂y 2 ∂y ∂y

(See Figure 6.4)

Fig. 6.4:

Notation
Partial derivatives of second and higher orders are calculated by taking partial derivatives of
already calculated partial derivatives. The order in which the differentiations are performed

74
is indicated in the notations used. If z = f (x, y), we can calculate four partial derivatives
of second order, namely, two pure second partial derivatives with respect to x or y,
∂ 2z ∂ ∂z
2
= = f11 (x, y) = fxx (x, y),
∂x ∂x ∂x
∂ 2z ∂ ∂z
2
= = f22 (x, y) = fyy (x, y),
∂y ∂y ∂y
and two mixed second partial derivatives with respect to x and y,
∂ 2z ∂ ∂z
= = f21 (x, y) = fyx (x, y),
∂x∂y ∂x ∂y
∂ 2z ∂ ∂z
= = f12 (x, y) = fxy (x, y).
∂y∂x ∂y ∂x
Again, we remark that the notations f11 , f12 , f21 and f22 are usually preferable to fxx , fxy , fyx ,
and fyy , although the latter are often used in partial differential equations. Note that f12
indicates differentiation of f first with respect to its first variable and then with respect to
its second variable; f21 indicates the opposite order of differentiation. The subscript closest
to f indicates which differentiation occurs first.
Similarly, if w = f (x, y, z), then
∂ 5w ∂ ∂ ∂ ∂ ∂w
2
= = f32212 (x, y, z) = fzyyxy (x, y, z).
∂y∂x∂y ∂z ∂y ∂x ∂y ∂y ∂z
Example 6.6 Find the four second partial derivatives of f (x, y) = x3 y 4
Solution
f1 (x, y) = 3x3 y 4 , f2 (x, y) = 4x3 y 3
∂ 2 4 4 ∂
f11 (x, y) = ∂x (3x y ) = 6xy , f21 (x, y) = ∂x (4x3 y 3 ) = 12x2 y 3 ,
∂ ∂
f12 (x, y) = ∂y (3x2 y 4 ) = 12x2 y 3 f22 = ∂y (4x3 y 3 ) = 12x3 y 2 .

Example 6.7 Calculate f223 (x, y, z), f232 (x, y, z), and f322 (x, y, z) for the function f (x, y, z) =
ex−2y+3z
Solution

∂ ∂ ∂ x−2y+3z
f223 (x, y, z) = e
∂z ∂y ∂y
∂ ∂
= (−2ex−2y+3z )
∂z ∂y
∂
= (4ex−2y+3z ) = 12ex−2y+3z ,
∂z
∂ ∂ ∂ x−2y+3z
f232 (x, y, z) = e
∂y ∂z ∂y
∂ ∂
= (−2ex−2y+3z )
∂y ∂z
∂
= (−6ex−2y+3z ) = 12ex−2y+3z ,
∂y
∂ ∂ ∂ x−2y+3z
f322 (x, y, z) = e
∂y ∂y ∂z
∂ ∂
= (3ex−2y+3z )
∂y ∂y
∂
= (−6ex−2y+3z ) = 12ex−2y+3z .
∂y

75
6.4 The Total Differential
Definition 6.3 Differentials
Let z = f (x, y), and let 4x and 4y be increments of x and y, respectively. The differen-
tials dx and dy of the independent variables x and y are

dx = 4x and dy = 4y

The differential z, or total differential, of the dependent variable z is

∂f ∂f
dz = dx + dy = fx (x, y)dx + fy (x, y)dy
∂x ∂y
Example 6.8 Let z = f (x, y) = 2x2 − xy
(a) Find the differential dz
(b) Compute the value of dz if (x, y) changes from (1, 1) to (0.98, 1.03), and compare your
result with the value of 4z obtained in Example 6.7
Solution
(a) dz = ∂f
∂x
dx + ∂f
∂y
dy = (4x − y)dx − xdy
(b) Here x = 1, y = 1, dx = 4x = −0.02 , and dy = 4y = 0.03. Therefore, dz =
[4(1) − 1](−0.02) − 1(0.03) = −0.09
The value of 4z obtained in Example 6.7 was −0.0886, so dz is a good approximation of
4z in this case. Observe that it is easier to compute dz than to compute 4z.

Example 6.9 A storage tank has the shape of a right circular cylinder. Suppose that the
radius and height of the tank are measured at 1.5m and 5m, respectively, with a possible
error of 0.05m and 0.1m, respectively. Use differentials to estimate the maximum error in
calculating the capacity of the tank.
Solution The capacity (volume) of the tank is V = πr2 h. The error in calculating the
capacity of the tank is given by
∂V ∂V
4 ≈ dV = dr + dh = 2πrhdr + πr2 dh
∂r ∂h
Since the errors in the measurement of r and h are at most 0.05m and 0.1, respectively, we
have dr = 0.05 and dh = 0.1. Therefore, taking r = 1.5, h = 5, dr = 0.05, and dh = 0.1,
we obtain
dV =2πrhdr + πr2 dh
≈2π(1.5)(5)(0.05)2 (0.1) = 0.975π
Thus, the maximum error in calculating the volume of the storage tank is approximately
0.975π, or 3.1m3 .

Theorem 6.1 Theorems on Differentials

In the following we shall assume that all functions have continuous first partial derivatives
in s region <, i.e., the functions are continuously differentiable in <.
1. If z = f (x1 , x2 , . . . , xn ), then
∂f ∂f ∂f
df = dx1 + dx2 + · · · + dxn (6.1)
∂x1 ∂x2 ∂xn
regardless of whether the variables x1 , x2 , . . . , xn are independent or dependent on
other variables. In (6.1) we often use z in place of f .
2. If f (x1 , x2 , . . . , xn ) = c, a constant, then df = 0. Note that in this case x1 , x2 , . . . , xn
cannot all be independent variables.

76
 3. The expression P (x, y)dx+Q(x, y)dy or briefly P dx+Qdy is the differential of f (x, y)
if and only if ∂P
∂y
= ∂Q
∂x
In such case P dx + Qdy is called an exact differential.
∂P ∂Q ∂2f ∂2f
Note: Observe that ∂y
= ∂x
implies that ∂y∂x
= ∂x∂y
.

4. The expression P (x, y, z)dx + Q(x, y, z)dy + R(x, y, z)dz or briefly P dx + Qdy + Rdz
is the differential of f (x, y, z) if and only if ∂P
∂y
= ∂Q , ∂Q = ∂R
∂x ∂z
, ∂R = ∂P
∂y ∂x ∂z
. In such
case P dx + Qdy + Rdz is called an exact differential.
Example 6.10 (a) Let U = x2 ey/x . Find dU . (b) Show that (3x2 y − 2y 2 )dx + (x3 − 4xy +
6y 2 )dy can be written as an exact differential of a function φ(x, y) and find this function.
(a) Method 1:
∂U ∂U
= x2 ey/x (−y/x2 ) + 2xey/x , = x2 ey/x (1/x)
∂x ∂y
Then
∂U ∂U
dU = dx + dy = (2xey/x − yey/x )dx + xey/x dy
∂x ∂y
Method 2:
dU =x2 d(ey/x ) + ey/x d(x2 ) = x2 ey/x d(y/x) + 2xey/x dx
xdy − ydx
=x2 ey/x ( ) + 2xey/x dx = (2xey/x − yey/x )dx + xey/x dx
x2
(b) Method 1
Suppose that
(3x2 y − 2y 2 )dx + (x3 − 4xy + 6y 2 )dy = dφ = ∂φ ∂x
dx + ∂φ
∂y
dy.
∂φ ∂φ
Then (1) ∂x = 3x2 y − 2y 2 , (2) ∂y = x3 − 4xy + 6y 2
From (1), integrating with respect to x keeping y constant, we have
φ = x3 y = 2xy 2 + F (y)
where F (y) is the ”constant” of integration. Substituting this into (2) yields
x3 − 4xy + F 0 (y) = x3 − 4xy + 6y 2 from which F 0 (y) = 6y 2 , i.e., F (y) = 2y 3 + c
Hence, the required function is φ = x3 y − 2xy 2 + 2y 3 + c, where c is an arbitrary constant.
Note that the existence of such a function is guaranteed, since if P = 3x2 y − 2y 2 and
Q = x3 − 4xy + 6y 2 , then ∂P/∂y = 3x2 − 4y = ∂Q/∂x identically. If ∂P/∂y 6= ∂Q/∂x
this function would not exist and the given expression would not be an exact differential.
Method 2
(3x2 y − 2y 2 )dx + (x3 − 4xy + 6y 2 )dy =(3x2 ydx + x3 dy) − (2y 2 dx + 4xydy) + 6y 2 dy
=d(x3 y) − d(2xy 2 ) + d(2y 3 ) = d(x3 y − 2xy 2 + 2y 3 )
=d(x3 y − 2xy 2 + 2y 3 + c)
Then the required function is x3 y − 2xy 2 + 2y 3 + c.

6.5 Chain Rule

Case 1: z = f (x, y), x = g(t), y = h(t) and compute dz dt
.
In this case we are going to compute an ordinary derivative since z really would be a function
of t only if we were to substitute in for x and y.
The chain rule of this case is,
dz ∂f dx ∂f dy
= +
dt ∂x dt ∂y dt
So, basically what we are doing here is differentiating f with respect to each variable in it
and then multiplying each of these by the derivative of that variable with respect to t. The
final step is to then add all this up.

77
 dz
Example 6.11 Compute dt
for each of the following

(a) z = xexy , x = t2 , y = t−1

(b) z = x2 y 3 + y cos x, x = ln(t2 ), y = sin(4t)

Solution

(a) z = xexy , x = t2 , y = t−1

There really isn’t all that much to do here other than using the formula.
dz ∂f dx ∂f dy
= +
dt ∂x dt ∂y dt
=(e + yxexy )(2t) + x2 exy (−t−2 )
xy

=2t(exy + yxexy ) − t− 2x2 exy

So, technically we have computed the derivative. However, we should probability go

head and substitute in for x and y as well at this point since we have already got t’s in
the derivative. Doing this gives,
dz
= 2t(et + tet ) − t−2 t4 et = 2tet + t2 et
dt
Note that in this case it might actually have been easier to just substitute in for x and
y in the original function and just compute the derivative as we normally would. For
comparisons sake let’s do that.
dz
z = t2 et ⇒ = 2tet + t2 et
dt

(b) z = x2 y 3 + y cos x, x = ln(t3 ), y = sin(4t)

Okay, in this case it would almost definitely be more work to do the substitution first so
we’ll use the chain rule first and then substitute
dz
=(2xy 3 − y sin x)(2/t) + (3x2 y 2 + cos x)(4 cos(4t))
dt
4 sin3 (4t) ln t2 − 2 sin(4t) sin(ln t2 )
= + 4 cos(4t)(3 sin2 (4t)[ln t2 ]2 + cos(ln t2 ))
t
Note that sometimes, because of the significant mess of the final answer, we will only
simplify the first step a little and leave the answer in terms of x, y, and t. This is
dependent upon the situation, class and instructor however and for this class we will
pretty much always be substituting in for x and y.
Let’s suppose that we have the following situation,

z = f (x, y) y = g(x)
dz
In this case the chain rule for dx
becomes,

dz ∂f dx ∂f dy ∂f ∂f dy
= + = +
dx dx dx ∂y dx ∂x ∂y dx
dx d
In the first term we are using the fact that, dx
= dx
(x) = 1.

78
 dz
Example 6.12 Compute dx for z = x ln(xy) + y 3 , y = cos(x2 + 1)
Solution
We’ll just plug into the formula.
   
dz y x 2
−2x sin x2 + 1



= ln(xy) + x + x + 3y
dx xy xy
 
2

2 x 2 2
= ln x cos(x + 1) + 1 − 2x sin(x + 1) + 3 cos (x + 1)
cos(x2 + 1)
= ln(x cos(x2 + 1)) + 1 − 2x2 tan(x2 + 1) − 6x sin(x2 + 1) cos2 (x2 + 1)

Case 2: z = f (x, y), x = g(s, t), y = h(s, t) and compute ∂z ∂s

and ∂z
∂t
In this case if we were to substitute in for x and y we would get that z is a function of s
and t and so it makes sense that we would be computing partial derivatives here and that
there would be two of them.
Here is the chain rule for both these cases.
∂z ∂f ∂x ∂f ∂y ∂z ∂f ∂x ∂f ∂y
= + = +
∂s ∂x ∂s ∂y ∂s ∂t ∂x ∂t ∂y ∂t
√
Example 6.13 Find ∂z ∂s
and ∂z∂t
for z = e2r sin(3θ), r = st − t2 , θ = s2 + t2
Solution
Here is the chain rule for ∂z
∂s
.

∂z s
=(2e2r sin(3θ))(t) + (3e2r cos(3θ)) √
∂s s + t2
2
 √  3se2(st−t2 ) cos(3√s2 + t2 )
2(st−t2 )
=t 2e sin(3 s2 + t2 ) + √
s2 + t 2
∂z
Now the chain rule for ∂t
.

∂z t
=(2e2r sin(3θ))(s − 2t) + (3e2r cos(3θ)) √
∂t s + t2
2
 √  3te2(st−t2 ) cos(3√s2 + t2 )
2(st−t2 )
=(s − 2t) 2e sin(3 s2 + t2 ) + √
s2 + t2
Case 3: Suppose that z is a function of n variables, x1 , x2 , . . . , xn , and that each of
these variables are in turn functions of m variables, t1 , t2 , . . . , tm . Then for any variable
ti , i = 1, 2, . . . , m we have the following,
∂z ∂z ∂x1 ∂z ∂x2 ∂z ∂xn
= + + ··· +
∂ti ∂x1 ∂ti ∂x2 ∂ti ∂xm ∂ti
Wow. That’s a lot to remember. There is actually an easier way to construct all the chain
rules that we have discussed in the section or will look at in later examples. We can build up
a tree diagram that will give us the chain rule for any situation. To see how these works
∂z
let’s go back and take a look at the chain rule for ∂x given that z = f (x, y), x = g(s, t), y =
h(s, t). We already know what this is, but it may help to illustrate the tree diagram if we
already know the answer. For reference here is the chain rule for this case,
∂z ∂f ∂x ∂f ∂y
= +
∂s ∂x ∂s ∂y ∂s
Here is the tree diagram for this case (see figure 6.5 (a) ).
Or
w = f (x, y, z), x = g1 (s, t, r), y = g2 (s, t, r), and z = g3 (s, t, r)

79
 (a) (b)

Fig. 6.5:

Here is the tree diagram for this situation (see figure 6.5 (b) ).
From this it looks like the derivative will be
∂w ∂f ∂x ∂f ∂y ∂f ∂z
= + +
∂r ∂x ∂r ∂y ∂r ∂z ∂r

Example 6.14 (Laplace’s equation in polar coordinates) If z = f (x, y) has continuous

partial derivatives of second order, and if x = r cos θ and y = r sin θ, show that
∂ 2 z 1 ∂z 1 ∂ 2z ∂ 2z ∂ 2z
+ + = +
∂r2 r ∂r r2 ∂θ2 ∂x2 ∂y 2
First note that
∂x ∂x ∂y ∂y
= cos θ, = −r sin θ, = sin θ, = r cos θ.
∂r ∂θ ∂r ∂θ
Thus,
∂z ∂z ∂x ∂z ∂y ∂z ∂z
= + = cos θ + sin θ .
∂r ∂x ∂r ∂y ∂r ∂x ∂y
Now differentiate with respect to r again. Remember that r and θ are independent variables,
∂z ∂z
so the factors cos θ and sin θ can be regarded as constants. However, ∂x and ∂y depend on
x and y and, therefore, on r and θ.
∂ 2z
   
∂ ∂z ∂ ∂z
= cos θ + sin θ
∂r2 ∂r ∂x ∂r ∂y
2
∂ 2z ∂ 2z ∂ 2z
   
∂ z
= cos cos θ 2 + sin θ + sin θ cos θ + sin θ 2 .
∂x ∂y∂x ∂x∂y ∂y
2 2 2
∂ z ∂ z ∂ z
= cos2 θ 2 + 2 cos θ sin θ + sin2 θ 2
∂x ∂x∂y ∂y
we have used the equality of mixed partials in the last line. Similarly,
∂z ∂z ∂z
= −r sin θ + r cos θ .
∂θ ∂x ∂y
When we differentiate a second time with respect to θ, we can regard r as constant, but
each term above is still a product of two functions that depend on θ. Thus,
∂ 2z
   
∂z ∂ ∂z ∂z ∂ ∂z
= − r cos θ + sin θ + r − sin θ + cos θ
∂θ2 ∂x ∂θ ∂x ∂y ∂θ ∂y
2 2
 
∂z ∂ z ∂ z
=−r − r sin θ −r sin θ 2 + r cos θ
∂r ∂x ∂y∂x
2 2
 
∂ ∂ z
+ r cos θ −r sin θ + r cos θ 2
∂x∂y ∂y
2
∂ 2z 2
 
∂z 2 2 ∂ z 2 ∂ z
=−r + r sin θ 2 − 2 sin θ cos θ + cos θ 2 .
∂r ∂x ∂x∂y ∂y

80
Combining these results, we obtain the desired formula:

∂z 1 ∂z 1 ∂ 2z ∂ 2z ∂2
+ + = + .
∂r2 r ∂r r2 ∂θ2 ∂x2 ∂y 2

6.6 Applications of Partial Derivatives

Local Extreme Values
Two Variables
We suppose for the moment that f = f (x, y) is defined on an open connected set and is
continuously differentiable there. The graph of f is surface

z = f (x, y).

Where f has a local maximum, the surface has a local high point. Where f has a local
minimum, the surface has a local low point. Where f has either a local maximum or a local
minimum, the gradient is 0 and therefore the tangent plane is horizontal. See Figure 6.6
(a).
A zero gradient signals the possibility of a local extreme value; it does not guarantee it. For
example, in the case of the saddle-shaped surface of Figure 6.6 (b), there is a horizontal
tangent plane at the origin and therefore the gradient is zero there, yet the origin gives
neither a local maximum nor a local minimum.
Critical points at which the gradient is zero are called stationary points. The stationary
points that do not give rise to local extreme values are called saddle points.

(a) (b)

Fig. 6.6:

Definition 6.4

1. A function f (x, y) has a relative minimum at the point (a, b) if f (x, y) ≥ f (a, b)
for all points (x, y) in some region around (a, b).

2. A function f (x, y) has a relative maximum at the point (a, b) if f (x, y) ≤ f (a, b)
for all points (x, y) in some region around (a, b)

Note that this definition does not say that a relative minimum is the smallest value that the
function will ever take. It only says that in some region around the point (a, b) the function
will always be larger than f (a, b). Outside of that region it is completely possible for the
function to be smaller. Likewise, a relative maximum only say that around (a, b) the function

81
will always be smaller than f (a, b). Again, outside of the region it is completely possible
that the function will be larger.
Fact
If the point (a, b) is a relative extrema of the function f (x, y) then (a, b) is also a critical
point of f (x, y) and in fact we’ll have ∇f (a, b) = ~0.
Fact
Suppose that (a, b) is a critical point of f (x, y) and that the second order partial derivatives
are continuous in some region that contains (a, b). Next define,

D = D(a, b) = fxx (a, b)fyy (a, b) − [fxy (a, b)]2

We then have the following classifications of the critical point.

1. If D > 0 and fxx (a, b) > 0 then (a, b) is a relative minimum.

2. If D > 0 and fxx (a, b) < 0 then (a, b) is relative maximum.

3. If D < 0 then (a, b) is a saddle point.

4. If D = 0 then (a, b) may be a relative minimum, relative maximum or a saddle point.

Other techniques would be need to be used to classify the critical point.

Example 6.15 Find and classify all the critical points of f (x, y) = 4 + x3 + y 3 − 3xy.
Solution
We first need all the first order (to find the critical points) and second order ( to classify the
critical points) partial derivatives so, let’s get those.

fx = 3x2 − 3y fy = 3y 2 − 3x
fxx = 6x fyy = 6y fxy = −3

Let’s first find the critical points. Critical points will be solutions to the system of equations

fx = 3x2 − 3y = 0
fy = 3y 2 − 3x = 0

This is a non-linear system of equations and these can, on occasion, be difficult to solve.
However, in this case it’s not too bad. We can solve the first equation for y as follows,

3x2 − 3y = 0 ⇒ y = x2

Plugging this into the second equation gives,

3(x2 )2 − 3x = 3x(x3 − 1) = 0

From this we can see that we must have x = 0 or x = 1. Now use the fact that y = x2 to
get the critical points
x = 0 : y = 02 = 0 ⇒ (0, 0)
x = 1 : y = 11 = 1 ⇒ (1, 1)
So, we get two critical points. All we need to do now is classify them. To do this we need
D. Here is the general formula for D.

D(x, y) =fxx (x, y)fyy (x, y) − [fxy (x, y)]2

=(6x)(6y) − (−3)2
=36xy − 9

82
To classify the critical points all that we need to do is plug in the critical points and use the
fact above to classify them.
(0, 0) : D = D(0, 0) = −9 < 0
So, for (0, 0) D is negative and so this must a saddle point.
(11) : D = D(1, 1) = 36 − 9 = 27 > 0 fxx (1, 1) = 6 > 0
For (1, 1) D is positive and fxx is positive and so we must have a relative minimum.

Example 6.16 Find and classify all the critical points for f (x, y) = 3x2 y+y 3 −3x2 −3y 2 +2
Solution
As with the first example we will first need to get all the first and second order derivatives.
fx = 6xy − 6x fy = 3x2 + 3y 2 − 6y
fxx = 6y − 6 fyy = 6y − 6 fxy = 6x
we will first need the critical points. The equations that we will need to solve this time are,
6xy − 6x = 0
3x + 3y 2 − 6y = 0
2

These equation are a little trickier to solve than the first set, but once you see what to do
they really are’nt terribly bad.
First, let’s notice that we can factor out a 6x for the first equation to get
6x(y − 1) = 0
So, we can see that the first equation will be zero if x = 0 or y = 1. Be careful to not just
cancel the x from both sides. If we had done that we would have missed x = 0.
To find the critical points we can plug these (individually) into the second equation and solve
for the remaining variable.

x = 0 : 3y 2 − 6y = 3y(y − 2) = 0 ⇒ y = 0, y = 2
Y = 1 : 3x2 − 3 = 3(x2 − 1) = 0 ⇒ x = −1, x = 1
So, if x = 0 we have the following critical points,
(0, 0), (0, 2)
and if y = 1 the critical points are,
(1, 1) (−1, 1)
Now all we need to do is classify the critical points. To do this we’ll need the general formula
for D.
D(x, y) = (6y − 6)(6y − 6) − (6x)2 = (6y − 6)2 − 36x2
(0, 0) : D = D(0, 0) = 36 > 0 fxx (0, 0) = −6 < 0
(0, 2) : D = D(0, 2) = 36 > 0 fxx (0, 2) = 6 > 0
(1, 1) : D = D(1, 1) = −36 < 0
(−1, 1) : D = D(−1, 1) = −36 < 0
So, it looks like we have the following classifications of each of these critical points
(0, 0) : Relative Maximum
(0, 2) : Relative Minimum
(1, 1) : Saddle point
(−1, 1) : Saddle point

83
 Exercise One
1. Use  − δ definition of limits to prove that:

(i) limx→1 (3x + 1) = 4 (ii) limx→2 (5 − 2x) = 1 (iii) limx→0 x2 = 0

x−2 1
(iv) limx→2 1+x2 = 0 (v) limx→1 x+1 = 12 (vi) limx→2 x3 = 8
(vii) limx→−1 xx+1 1
2 −1 = − 2 (viii) limx→3 x−2
x+1
= 14 (ix) limx→2 7x + 4 = 18

2. State with reason(s) whether the function is continuous or discontinuous.

  1
x x<0 x2
x 6= 0
(i) f (x) = 2 (ii) f (x) =
 x2 x ≥ 0 0 x=0
x x≤1
(iii) f (x) =
0.987 x > 1


x2 x≤2
3. (i) Find k so that f (x) = 2 is a continuous function.
k − x x > 2
x−m x≤
(ii) Find m so that f (x) = is continuous for all x.
1 − mx x > 2
4. Evaluate the following limits if exist
2
(i) limx→4 x −7x+12
x−4  2
x + 5 if x 6= 3
(ii) limx→3 f (x) where f (x)
 27 if x=3
x + 9 if x<2
(iii) limx→2 g(x) where g(x)
x2 − 3 if x≥2
3x−15
(iv) limx→5 √x2 −10x+25

84
 Exercise Two
1. Find the Laplace transform of the following functions
(ii) f (x) = 5 sin 3x − 17e−2x
(i) f (x) = 3 + 2x2
(iv) f (x) = xe4x
(iii) f (x) = 2 sin x + 3 cos 2x
(vi) f (x) = sinh x cos x
(v) f (x) = e−2x sin 5x 7
(viii) f (x) = x 2
(vii) f (x) = e−x x cos 2x  x
e x≤2
(ix) f (x) = sinx3x (x) f (x) =
3 x>2
2. Find the Laplace transforms of the following functions
(i) f (t) = (t2 +t+1)u(t−1) (ii) f (t) = e3t u(t−2) (iii) (sin t+cos t)u(t−π/2)

3. Express f (x) in terms

 of2 Heavisides unit step function and find its Laplace transform
 t o<t<2  −t
e 0<t<3
(i) f (t) = 4t 2 < t < 4 (ii) f (t) =
0 t>3
8 t>4


4. Find (i) L{2δ(t − 1) + 3δ(t − 2) + 4δ(t − 3)} (ii) L{cosh 3tδ(t − 2)}
n o n o n o
5. Find (i) L−1 (s25s −1 √1 −1 −1
 s+1 s
+1)2
(ii)L (iii) L 2 (iv) L (s−2)2 +9
s
n s −9 o
(v) L−1 s2 −2s+9
1
(vi) L−1 s2 +4s+8
s+4
(vii) L−1 (s2 +1)+(s1 2 +4s+8)
 

6. Use Laplace transforms to solve

(i) y 0 − 5y = 0; y(0) = 2 (ii) y 0 + y = sin x; y(0) = 1
(iii) y 00 − 3y 0 + 4y = 0; y(0) = 1, y 0 (0) = 5 (iv) y 00 − y 0 − 2y = 4x2 ; y(0) =
0
1, y (0) = 4
(v) y 000 + y 0 = ex ; y(0) = y 0 (0) = y 00 (0) = 0

7. A resistance R in series with inductance L is connected e.m.f E(t). The current is

di
given by L dt + Ri = E(t). If the switch is connected at t = 0 and disconnected at
t = a, find the current i in terms of t.

8. Use Laplace transforms  to solve

v0 − u + v = 0
(i) 0 with u(0) = 1, v(0) = 2
u + u − v = 0
y0 + z = x
(ii) 0 with y(0) = 1, z(0) = −1
z + 4y = 0

w0 − y + 2z = 3e−x 
0 0
(iii) −2w + 2y + z = 0 with w(0) = 1, w0 (0) = 1, y(0) = 2, z(0) = 2,
2w0 − 2y + z 0 + 2z 00 = 0

z 0 (0) = −2

85
 Exercise Three
1. Find the Fourier series expansion of the functions on the specified intervals.

i. f (x) = ex , on [−1, 1]
4
ii. f (x) = x , on [−1, 1]. Sketch the periodic function.
iii. The rectified sine wave f (t) = | sin t| on 0 < t < π. Sketch the sine wave function.
1 1


2. i. Give reasons why the functions: (a) 3−t (b) sin t−2 do not satisfy Dirich-
let’s conditions in the interval 0 < t < 2π .
ii. How should f (x) be defined at discontinuities for convergence?


 5 sin x −2π < x < − π2
 4 x = − π2


f (x) = x2 − π2 < x < 2
8 cos x 2<x<π




4x π < x < 2π


3. i. Find the Fourier Series expansion of


−x −5 < x < 0
f (x) =
1 + x2 0 < x < 5

ii. A periodic function f (x) of period 2π is defined within the period 0 < t < 2π by
t

2− π
0<t<π
f (x) = t
π
π < t < 2π
Sketch a graph for −4π < t < 4π and obtain its Fourier series expansion.
iii. A periodic function f (x) of period 2π is defined within the period 0 < t < 2π by

 t 0 < t < π2
π π
f (x) = 2 2
<t<π
t
π− π < t < 2π

2
Sketch a graph for −2π < t < 3π. Find the Fourier series expansion of the
function.
iv. Determine the Fourier series representation of the periodic voltage e(t) shown in
figure 6.7:

Fig. 6.7:

4. i. Write the Fourier cosine and Fourier sine expansion of f (x) = e2x for x0 < x < 1.
ii. A tightly stretched flexible uniform string has its ends fixed at the points x = 0
and x = l. The midpoint of the string is displaced a distance a, shown in figure
6.8. If f (x) denotes the displaced profile of the string, express f (x) as a Fourier
series expansion consisting only of sine terms.

86
 Fig. 6.8:

5. Find the complex form of the Fourier series expansion of:

i. f (x) = x for −1 < x < 1, f (x + 2) = f (x).

ii. f (t) = cos(t/2) for −π < t < π, f (t + 2π) = f (t)

87
 Exercise Four
1. Specify the domain of the given functions.
√ xy
a. f (x, y) = xy b. f (x, y) =
x2 − y2
1
c. f (x, y) = p d. f (x, y) = sin−1 (x + y)
x2 − y 2
exyz
e. f (x, y, z) = √
xyz

2. Show that the given functions satisfy the corresponding partial differential equations.
x + y ∂z ∂z
a. z = ,x +y =0
x − y ∂x ∂y
∂w ∂w ∂w
b. w = x2 + yz, x +y +z = 2w
∂x partialy ∂z
c. z = f (x2 + y 2 ), where f is any differentiable function of one variable
∂z ∂z
y −x =0
∂x ∂y

3. Find and classify the critical points of the given functions.

a. f (x, y) = x4 + y 4 − 4xy
b. f (x, y) = cos x + cos y
1
4. Find the minimum value of f (x, y) = x + 8y + xy
in the first quadrant x > 0, y > 0.
How do you know that a minimum exists?

5. The material used to make the bottom of a rectangular box is twice as expensive per
unit area as the material used to make the top or side walls. Find the dimension of
the box of given volume V for which the cost of the materials is minimum.

6. Find the three positive numbers a, b and c, whose sum is 30 and for which the
expressionab2 c3 is maximum.

88
References

The list of references is not exhaustive. Any other book on the topics above may be suitable
as well.

1. Duffy, D. G. (1998): Advanced Engineering Mathematics, CRC Press.

2. Greenberg, M. D. (1998): Advanced Engineering Mathematics, Prentice Hall, New York,

2nd Edition.

3. Kaplan, W. (2003): Advanced Calculus, Addison-Wesley Higher Mathematics, 5th Edi-

tion.

4. Kreyszing, E. (2006): Advanced Engineering Mathematics, Wiley, 8th Edition.

5. Jeffery, A. (2001): Advanced Engineering Mathematics, Academic Press.

6. Ramana, B. V. (2007): Higher Engineering Mathematics, Tata McGraw-Hill.

7. Smith, R. T. and R. B. Minton (2002): Multivariable Calculus, McGraw-Hill.

8. Stroud, K. A. (2003): Advanced Engineering Mathematics, Palgrave Macmillian, 4th

Edition.

89