Contents

1 Limit and continuity 1

1.1 Limit Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Some special limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 The concept of continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.1 Examples and properties of continuous functions . . . . . . . . . . . . . . . . 10

2 Differentiation 15
2.1 Derivative of a function at a point . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Continuity and differentiability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3 Techniques of Differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3.1 Differentiation of sums, differences, products and quotients . . . . . . . . . . 20
2.3.2 The chain rule of differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4 Differentiation of inverse functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.5 A collection of differentiation formulas . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.6 Differential and linear approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.7 L’Hôpital’s rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.8 Higher Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.9 Logarithmic and Exponential Functions . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.10 Implicit Differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.11 The Mean Value Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.11.1 Relative Extremum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.11.2 Absolute Extremum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.11.3 Rolle’s theorem and the mean value theorem . . . . . . . . . . . . . . . . . . 34
2.12 Applications of differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.12.1 Increasing and decreasing Functions . . . . . . . . . . . . . . . . . . . . . . . 36
2.12.2 Concavity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

I
 2.12.3 Some applications of differentiation in economics . . . . . . . . . . . . . . . . 39
2.12.4 Rate of Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.12.5 Maximum / minimum problems . . . . . . . . . . . . . . . . . . . . . . . . . 43

3 Indefinite and definite integrals 45

3.1 Indefinite integrals as anti-derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2 Definite integrals as limits of sums . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3 Properties of Definite Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.4 Techniques for Integrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.4.1 Integration by Substitution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.4.2 Integration by Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.4.3 Reduction Formulae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.4 Method of Partial Fractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.4.5 Miscellaneous Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

4 Applications of Integration 62
4.1 Area between curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.2 Average value of a function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

A Summation Notation 64
A.1 Changing the Index of Summation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
A.2 Properties of Summation Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.3 Summation Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

II
Chapter 1

Limit and continuity

1.1 Limit Concept

We are going to study calculus! Since the idea of a limit is essential to the study of calculus, we
must develop certain understanding of that concept. We shall first give you a ”feeling” for limits
by some simple examples.

Example 1.1.1 Let us examine the function

f (x) = 3x − 2

when x is “near” 2, but never equal to 2. Some values of f (x) for x less than 2 and then greater
than 2 are given in the following table.

x<2 x>2
f (1.8) = 3.4 f (2.2) = 4.6
f (1.9) = 3.7 f (2.1) = 4.3
.
f (1.99) = 3.97 f (2.01) = 4.03
f (1.995) = 3.985 f (2.005) = 4.015
f (1.999) = 3.997 f (2.001) = 4.003

It is evident from the table that as x takes on values closer to 2, regardless of whether x gets close
to 2 from the left (x < 2) or from the right (x > 2), the corresponding values of f (x) become closer
to one number, 4. To express this, we say that the limit of f (x) as x approaches 2 is equal to 4.
Symbollically, we write
lim (3x − 2) = 4.
x→2

In fact, we can make the number f (x) as close to 4 as we wish by taking x sufficiently close, but
not equal, to 2. For instance, suppose we want f (x) to be within 0.0001 from 4, i.e.,

−0.0001 ≤ (3x − 2) − 4 ≤ 0.0001.

1
This means
0.0001 0.0001
−0.0001 ≤ 3x − 6 ≤ 0.0001 or − ≤x−2≤ .
3 3
0.0001
In other words, as long as x is within from 2 (but not equal to 2), f (x) will be within 0.0001
3
from 4.
You may think that you can find the limit of a function as x approaches some number a by
just evaluating the function when x is a. For the function f (x) = 3x − 2, this is true because
f (2) = 3 × 2 − 2 = 4, which is also the limit. However, this method of substitution does not always
work. For instance, consider the function
(
3x − 2, if x 6= 2
g(x) =
5, if x = 2.

y
6

x
-2 -1 1 2 3

-2

Notice that g(2) = 5, while lim g(x) = lim (3x − 2) = 4. Therefore,

x→2 x→2

lim g(x) 6= g(2).

x→2

The general notion of limit of a function at a point can be stated as follows: Suppose f (x) is any
given function and a is a given real number. We say that “the limit of f (x), as x approaches a, is
L,” written
lim f (x) = L,
x→a

means that f (x) will be as close to the number L as we please for all x sufficiently close to the
number a, but not equal to a.
Again, here we are not concerned with what happens to f (x) when x = a, but only what happens
to f (x) when x is close to a.
More precisely, we have the following definition:
We say that lim f (x) = L if given any positive number ε, there corresponds a number δ > 0 such
x→a
that whenever x lies inside the interval (a − δ, a + δ), but x 6= a, one has f (x) lying inside (L − ε,
L + ε).

2
 y y
y = f(x)
y = f(x)
L

x x
a a

We have to emphasize that a limit is independent of the way in which x → a. The limit must be
the same whether x approaches a from the left or from the right (for x < a or x > a, respectively).
We shall now state some properties of limits which may seem reasonable to you.

1. If f (x) = K is a constant function, then lim f (x) = K.

x→a

2. If f (x) = xn , where n is a positive integer, then

lim f (x) = lim xn = an .

x→a x→a

3. If f (x) and g(x) are functions such that lim f (x) = L and lim g(x) = M , then
x→a x→a

(a) lim (f (x) + g(x)) = L + M ,

x→a
(b) lim (f (x) − g(x)) = L − M ,
x→a
(c) lim f (x) g(x) = LM , and
x→a
f (x) L
(d) lim = , if M 6= 0. In particular, lim kf (x) = kL for any constant k.
x→a g(x) M x→a

Example 1.1.2

1. lim 100 = 100 and lim x3 = (−5)3 = −125.

x→2 x→−5

2. Evaluate lim (5x4 − 2x3 + 3x − 24).

x→3
Solution Using the above properties, one has

lim (5x4 − 2x3 + 3x − 24)

x→3
= lim 5x4 + lim (−2x3 ) + lim 3x + lim (−24)
x→3 x→3 x→3 x→3
= 5 lim x4 − 2 lim x3 + 3 lim x − lim 24
x→3 x→3 x→3 x→3
4 3
= 5 × 3 − 2 × 3 + 3 × 3 − 24
= 336.

3
 x2 + x − 1 x3 + x2 − 3x + 1
3. Find (a) lim and (b) lim .
x→−2 x3 + 12 x→1 x−1
Solution In (a), the numerator approaches to 1 and the denominator approaches to 4.
Therefore, (3)(a) implies
x2 + x − 1 1
lim 3
= .
x→−2 x + 12 4
On the other hand, in (b), both the numerator and the denominator approach to zero as x
approaches to 1. Thus (3)(a) above is not applicable in this case. However, for x 6= 1, one
may use long division to obtain

x3 + x2 − 3x + 1
= x2 + 2x − 1.
x−1

It thus follows that

x3 + x2 − 3x + 1
lim = lim (x2 + 2x − 1) = 2.
x→1 x−1 x→1

Suppose that as x approaches a through values greater than a, (or x approaches a from above), we
have f (x) approaches a limit L1 . Then L1 is said to be the right hand limit of f at a. We denote
this by the symbol
lim f (x) = L1 or lim f (x) = L1 .
x→a+ xցa

Similarly, the left hand limit of f (x) at a, denoted by lim f (x) or lim f (x), is defined as the limit
x→a− xրa
of f (x) as x approaches a through values less than a, (or x approaches a from below).

Theorem 1.1.3 lim f (x) exists if and only if lim f (x) and lim f (x) both exist and are equal.
x→a x→a+ x→a−

y
f
f(x)

f(x)

x a x x

lim f(x) = L lim f(x) = L

lim f(x) = L

Example 1.1.4

1. Consider the function defined by

(
4 − x2 if x ≤ 1
f (x) = .
1 + x2 if x > 1

4
 6 y

x
-2 -1 1 2 3

-2

Since lim f (x) = lim (1 + x2 ) = 2 and lim f (x) = lim (4 − x2 ) = 3, we conclude that
x→ 1+ x→ 1+ x→ 1− x→ 1−
lim f (x) does not exist.
x→ 1

1 1
2. Discuss the limits lim and lim 2 .
x→0 x x→0 x
1
Solution For f (x) = (x 6= 0), consider the following table:
x

x<0 x>0
f (−0.1) = −10 f (0.1) = 10
.
f (−0.01) = −100 f (0.01) = 100
f (−0.001) = −1000 f (0.001) = 1000

It can be easily concluded that as x approaches to zero from above, f (x) is getting arbitrarily
large. We describe this by saying that f (x) approaches to infinity as x → 0+. Symbolically,

1
lim = ∞.
x→0+ x

Similarly, as x approaches to zero from below, f (x) is getting negatively large. We say that
f (x) approaches to negative infinity as x → 0−. Symbolically,

1
lim = −∞.
x→0− x

1
As a whole, we may say that lim does not exist.
x
x→0
1
We ask the student to discuss the limit lim 2 as an exercise.
x→0 x

5x3 + 4x − 1
3. Find lim .
x→∞ 1 + 3x − 2x3
Solution While we have not discussed limit of a function at ∞, this particular problem may
1
be dealt with informally by setting x = . We know that as x → ∞, y must approach to 0+.
y
Therefore, we conclude that

5 4
+ −1
5x3 + 4x − 1 y 3 y 5 + 4y 2 − y 3 5
lim = lim = lim =− .
x→∞ 1 + 3x − 2x 3 y→0+ 3 2 3 2
y→0+ y + 3y − 2 2
1+ − 3
y y

5
A function f (x) defined on an interval I is said to be a bounded function if there is a constant
M > 0 such that
|f (x)| ≤ M

for every x in I. Note that M must be independent of x.

For example, sin x is a bounded function on the whole x−axis because |sin x| ≤ 1 for any x, while
x3 is unbounded on the x−axis. However, if the function x3 is regarded as a function on the interval
[0, 2], then it is a bounded function because x3 ≤ 8 for every x between 0 and 2.
The following result on the limit of the product of two functions is useful:

If f (x) is bounded and lim g(x) = 0, then lim f (x) g(x) = 0.

x→a x→a

1 1
Example 1.1.5 Consider lim x sin . Since the function sin is bounded, and lim x = 0, we
x→0 x x x→0
1
conclude from the above result that lim x sin = 0.
x→0 x

-2 -1 1 2

-1

-2

1.2 Some special limits

In this section, we shall present a number of special limits which will be needed later. The following
“Sandwich Principle” for functions is required.

Theorem 1.2.1 Suppose f (x) ≤ g(x) ≤ h(x) for x in some open interval containing a (but x 6= a).
Suppose also that
lim f (x) = lim h(x) = L.
x→a x→a

Then lim g(x) = L.

x→a

6
 Sandwich Principle
y

g
L
f

x
0 a

Example 1.2.2

sin θ
1. Show that lim = 1.
θ
θ→0
Solution Consider a circle of radius 1 and centre at the origin O of the xy−plane. Let A be
the intersection of the circle with the positive x−axis and P be a point on the circle in the first
quadrant with ∠AOP = θ. Let Q be the point on the positive x−axis such that P Q ⊥ OA.
Finally, let T be the point such that AT ⊥ the x−axis and O, P and T are colinear.

y T
1

O A(1, 0)
x
Q

As 0 < θ < π/2, it is clear that P Q = sin θ, the length of arc P A = θ and AT = tan θ. Since
the area of ∆OAP ≤ the area of sector OAP ≤ the area of ∆OAT , we have

sin θ
sin θ ≤ θ ≤ tan θ, or 1 ≥ ≥ cos θ.
θ

sin θ
Since lim cos θ = 1, the sandwich principle now implies that lim = 1.
θ→0+ θ→0+ θ
sin(−θ) − sin θ sin θ sin θ
As = = , we conclude that lim = 1.
−θ −θ θ θ→0− θ
The result now follows from Theorem (1.2.1).
√
2. Find lim n p, where p is a positive constant.
n→∞
√
Solution Let us assume that p > 1. Suppose that hn = n p − 1 > 0. Binomial theorem then
implies
p
p = (1 + hn )n > n hn , or 0 < hn < .
n
p √
Since lim = 0, we conclude that lim hn = 0, or equivalently, lim n p = 1.
n→∞ n n→∞ n→∞
The student may consider the case 0 < p < 1 as an exercise.

7
What we just saw is something that we haven’t seen before, i.e., the limit of “a sequence of numbers”.
Specifically, a set of numbers a1 , a2 , a3 , · · · arranged in a particular order is known as a sequence.
It may be denoted by the symbol {a1 , a2 , a3 , · · · }, or simply by {an }.
Simple examples of sequences are
   
1 1 1
1. 1, , , · · · , or simply by ;
2 3 n
(−1)n+1
   
1 1 1
2. 1, − , , − , · · · , or simply ;
2 3 4 n
   
1 2 3 4 n
3. , , , , · · · , or simply .
2 3 4 5 n+1

It is often important to know the behaviour of a sequence {an } for large n. For instance, it can be
1 1
observed that, as n gets larger, the number gets smaller. And for “very large” n, the number
n n
becomes “very close” to zero. Following our earlier experience in handling limit of a function, we
may describe this by saying
1
lim = 0.
n→∞ n
In general, we say that the limit of a sequence {an } is A, we mean that an can get as close to A as
we please, provided that n is sufficiently large . Symbolically, we write

lim an = A,
n→∞

We also use the notation an → A as n → ∞. A sequence which has a limit is called a convergent
sequence.
When dealing with limits of sequences, the following rules are useful.

1. If {an } and {bn } are sequences such that lim an = A and lim bn = B, then
n→∞ n→∞

(a) lim (an ± bn ) = A ± B ;

n→∞
(b) lim (an · bn ) = A · B ; in particular, lim k an = k · A for any constant k;
n→∞ n→∞
an A 1 1
(c) lim = , provided that B 6= 0. In particular, lim = .
n→∞ bn B n→∞ bn B
2. Let {an }, {bn } and {cn } be sequences such that

an ≤ cn ≤ bn for n = 1, 2, 3, · · · .

If lim an = lim bn = L, then lim cn exists and is equal to L. This is known as the Sandwich
n→∞ n→∞ n→∞
Principle for sequences.

2n2 − 10n + 4
Example 1.2.3 Consider limn→∞ . One may divide both the numerator and the
3n2 − 5n + 6
denominator by n2 to obtain
2n2 − 10n + 4 2 − 10
n + n2
4
= .
3n2 − 5n + 6 3 − n5 + n62

8
We then use the above rule to conclude that

lim 2 − 10 + n42


2n2 − 10n + 4 n 2
lim 2
= n→∞ 5 6

= .
n→∞ 3n − 5n + 6 lim 3 − n + n2 3
n→∞

We would like to conclude this section by introducing a very important limit in mathematics,
namely
1
lim (1 + x) x .
x→0
1
The following table tabulates the values of f (x) = (1 + x) x for x close to 0:

x<0 x>0
f (−0.5) = 4 f (0.5) = 2.25
f (−0.1) = 2.8680 f (0.1) = 2.5937 .
f (−0.01) = 2.7320 f (0.01) = 2.7048
f (−0.001) = 2.7196 f (0.001) = 2.7169

1
It can be proved that the limit of f (x), as x approaches to 0, exists. Traditionally, lim (1 + x) x
x→0
is denoted by the symbol e. Its value is approximately equal to 2.71828 . . . . It is one of the most
important constants in mathematics.

Example 1.2.4 If a principal of P dollars is invested and interest is compounded k times a year
r
at an annual rate of r, then the rate per conversion period is equal to . In t years there are kt
k
periods. It follows that the compound amount S at the end of t years is given by
h r ikt
S =P 1+ .
k

If k → ∞, the number of conversion periods increases indefinitely and the length of each period
approaches 0. In this case we say that interest is compounded continuously, that is, at every instant
of time. The compound amount is equal to
rt
r i kr

h r ikt h
P × lim 1 + = P × lim 1 + .
k→∞ k k→∞ k

r
If we put x = , then
k
h r i kr 1
lim 1 + = lim (1 + x) x = e.
k→∞ k x→0

We therefore conclude that the compound amount after t years with principal P at an annual interest
rate r, compounded continuously, is equal to S = P ert . As a simple illustration, if $100, 000 is
deposited into a bank for 10 years with an annual interest rate of 5%, the compound amount is
equal to S = 100, 000 × e0.05×10 ≃ $164, 872.

9
1.3 The concept of continuity

A function f (x) defined on an interval is continuous at a point x0 in the interval if lim f (x) exists
x→x0
and f (x0 ) = lim f (x).
x→x0
A function defined on an interval is said to be continuous if it is continuous at every point in the
interval over which it is defined.
By definition, f (x) is continuous at x0 means that

lim f (x) = lim f (x) = f (x0 ).

x→x0 + x→x0 −

y
y = f(x)

x0
0 x

Example 1.3.1

1. Constant functions are continuous everywhere. In fact, if f (x) ≡ C for every x, where C is
a given constant, then

lim f (x) = lim C = C = f (x0 ) for every x0 .

x→x0 x→x0

Hence the constant function f (x) is continuous everywhere.

2. The function f (x) = x for all x is continuous everywhere because

lim f (x) = lim x = x0 = f (x0 ) for every x0 .

x→x0 x→x0

The following theorem says that continuity is preserved under the usual algebraic manipulations.

f
Theorem 1.3.2 If f (x) and g(x) are continuous at a point x0 , then f + g, f − g, f g and are
g
continuous at x0 . (In the case of quotient, we have to assume that g(x0 ) 6= 0.)

As a consequence of Theorem (1.3.2), we may also conclude that if f (x) and g(x) continuous
f
functions in an interval, then f + g, f − g, f g and are continuous in the same interval. In the
g
case of quotient, we require g(x) 6= 0.

1.3.1 Examples and properties of continuous functions

Example 1.3.3

10
1. Polynomials are continuous functions. Recall that a polynomial is a function of the form

P (x) = a0 + a1 x + . . . + an xn ,

where n is a nonnegative integer and a0 , a1 , · · · , an are real constants. Continuity of P (x)

then follows from Theorem (1.3.2) and the fact that constant functions and f (x) = x are
continuous on R. .

2. Rational functions are continuous on R except at those points where the denominator vanishes.
In fact, a rational function is a function of form

P (x)
f (x) = ,
Q(x)

where P (x) and Q(x) are polynomials. Since P (x) and Q(x) are continuous by previous
example, f (x) is therefore continuous on R except at those points where Q(x) = 0.

3. The trigonometrical functions sin x and cos x are continuous everywhere. In fact, given any
x0 , we have    
x + x0 x − x0
sin x − sin x0 = 2 cos sin .
2 2
This implies that
   
x + x0 x − x0 x − x0
|sin x − sin x0 | ≤ 2 cos sin ≤2 = |x − x0 | .
2 2 2

Here, we have made use of the inequality |sin θ| ≤ |θ| for any θ. Therefore, one concludes
from the definition of limit that lim sin x = sin x0 . Hence sin x is everywhere continuous.
x→x0
Similarly cos x is also continuous. As a consequence of Theorem (1.3.2), other trigono-
metric functions are continuous except at those points where the corresponding denominator
is zero. For example, the function

cos x
f (x) = cot x =
sin x

is continuous except at x = nπ, where n = 0, ±1, ±2, ±3, · · · .

4. Discuss the continuity of the function

x2 − x − 2
f (x) = , x 6= −1.
x+1

Solution The function is undefined at x = −1. Hence it is not continuous at x = −1.

However,

(x + 1)(x − 2)
lim f (x) = lim = lim (x − 2) = −3.
x→−1 x→−1 (x + 1) x→−1

11
 4
y

2 f

x
-4 -2 2 4
-2

-4

(
f (x), when x 6= −1
Therefore if a function g(x) is defined by g(x) = , then g is contin-
−3, when x = −1
uous everywhere.

4
y

2 g

x
-4 -2 2 4

-2

-4

This type of discontinuity is commonly known as a removable discontinuity.

Theorem 1.3.4 Let f (x) be a function continuous on a domain D and g(y) be a function contin-
uous on a domain E, where f (D) ⊂ E, then the composite function h(x) = g(f (x)) is continuous
on D. In other words, the composition of two continuous functions is also continuous.

g f
Continuous at c

f g
Continuous at c Continuous at f(c)

c f(c) g(f(c))

As a simple illustration of Theorem (1.3.4), the function h(x) = sin(x2 ) is continuous because it
is the composition of two continuous functions, f (x) = x2 and g(y) = sin y.

Theorem 1.3.5 Let f (x) be a strictly increasing and continuous function on an interval [a, b]. Let
f (a) = c and f (b) = d, and let g(y) be the inverse function of f (x). Then

1. g(y) is a strictly increasing function on [c, d];

2. g(y) is continuous on [c, d].

12
Note that corresponding results are valid for strictly decreasing functions.

Example 1.3.6 f (x) = sin x is strictly increasing on [−π/2, π/2], and f (−π/2) = −1, f (π/2) = 1.
Thus f −1 (x) = sin−1 x is continuous on [−1, 1] and is strictly increasing there.

y
1.5 f -1

1.0
f
0.5

x
-1.5 -1.0 -0.5 0.5 1.0 1.5

-0.5

-1.0

-1.5

Theorem 1.3.7 (Intermediate Value theorem) Let f (x) be a continuous on a closed interval
[a, b] such that f (a) and f (b) are of opposite signs. Then there exists a point c ∈ (a, b) such that
f (c) = 0.

This theorem is geometrically obvious. It simply says that if the two ends of a continuous curve lie
on opposite sides of the x − axis, then the curve must cross the x−axis at least once.
As a corollary of Theorem (1.3.7), if f is a continuous function defined on a closed and bounded
interval [a, b] such that f (a) 6= f (b), and if γ is a number between f (a) and f (b), then there exists
a point c ∈ (a, b) such that f (c) = γ.

f ( a) f(c2) f(c3)
f(c1)
C

f(b)

Example 1.3.8 A polynomial of degree 3 has at least one real root.

Proof Let P (x) = x3 + ax2 + bx + c, where a, b and c are real constants. For x 6= 0, one has
 
a b c
3
P (x) = x 1 + + 2 + 3 .
x x x

We can find numbers x1 > 0 and x2 < 0 such that

1 a b c 1 1 a b c 1
− < + 2 + 3 < and − < + 2+ 3 < .
2 x1 x1 x1 2 2 x2 x2 x2 2

13
 1
Since 2 < 1 + xa1 + xb2 + xc3 < 23 , it is clear that P (x1 ) and x31 have the same sign. Thus P (x1 ) > 0.
1 1
Similarly, P (x2 ) and x2 have the same sign and thus P (x2 ) < 0. The intermediate value theorem
now implies that there is a number c between x1 and x2 such that P (c) = 0.
Note that the result of this example may be extended to all odd degree polynomials.

14
Chapter 2

Differentiation

2.1 Derivative of a function at a point

The rate of change of a function is an important concept in mathematics. Let f (x) be a function
defined on an interval I. Suppose x0 is a point in the interval. For any non-zero number ∆x (the
x−increment), we consider the corresponding y−increment ∆y = f (x0 + ∆x) − f (x0 ) and the
difference quotient
∆y f (x0 + ∆x) − f (x0 )
= .
∆x ∆x

y
x
x

Clearly, ∆y is the change in the variable y as a result of the change of x from x = x0 to x = x0 +∆x.
∆y
Therefore, the difference quotient is a rate of change of the function f (x) at x0 .
∆x
f (x0 + ∆x) − f (x0 )
If lim exists (and is finite), then we say that the function f (x) is differen-
∆x→0 ∆x
tiable x0 . The limit is called the derivative of f (x) at x0 , and is denoted by the symbol f ′ (x0 ).

Remark 2.1.1 Since

f (x0 + ∆x) − f (x0 ) ∆y

f ′ (x0 ) = lim = lim ,
∆x→0 ∆x ∆x→0 ∆x

 
dy
it is also natural to use the symbol to denote the derivative f ′ (x0 ).
dx x =x0

The notion of derivative has the following geometrical interpretation. The function y = f (x) has
a graph on the xy−plane. Denote the point (x0 , f (x0 )) by P and (x0 + ∆x, f (x0 + ∆x)) by Q. It

15
is clear that the slope of the straight line connecting P and Q is given by the difference quotient
f (x0 + ∆x) − f (x0 ) ∆y
= .
∆x ∆x

y y = f(x)

( x0+ x , f( x0+ x) )

f( x0+ x) -f( x0)

P P(x(x0,0 , ff(x
( x00))
))
x
f( x0)

x0 x0+ x
0 x
As ∆x varies, Q moves along the curve, and the line P Q changes its direction. When ∆x → 0, P Q
approaches to the tangential direction of the graph at P , and its slope approaches to the derivative
f ′ (x0 ). Therefore, f ′ (x0 ), if exists, is simply the slope of the tangent line to the graph of the
function y = f (x) at the point P .
A physical interpretation of these notions may be given as follows. Suppose an object moves
along a straight line (the y−axis). Its distance from a certain fixed point on the line at time t
is given by y = f (t). Over the time interval [t0 , t0 + ∆t], the object covers a distance equal to
∆y = f (t0 + ∆t) − f (t0 ), and the difference quotient

∆y f (t0 + ∆t) − f (t0 )

=
∆t ∆t

is the average velocity of the object over this time interval. As a result, the derivative f ′ (t0 ) is
simply the instantaneous velocity of the object at the instant t0 .
dy
Given a function y = f (x), the derivative of f (x) is the function f ′ (x) (or ), which is defined on
dx
′
the same interval I over which f (x) is defined, provided that f (x) exists for every x in I. In other
words,
f (x + ∆x) − f (x)
f ′ (x) = lim .
∆x→0 ∆x
Thus f ′ (x) is also known as the derived function obtained from f (x).
To summarize, the computation of the derivative of a function f (x) at any point x0 can be system-
atically divided into three steps :

1. Compute ∆y = f (x0 + ∆x) − f (x0 );

∆y
2. Form the difference quotient ;
∆x
∆y
3. Take the limit of as ∆x tends to zero.
∆x
The first two steps are purely mechanical and may be carried out in a routine manner. It is the
last step that requires some technical skills or even certain degree of ingenuity.

16
Example 2.1.2

1. If y = f (x) = K (constant), then ∆y = f (x + ∆x) − f (x) = 0. Thus

dy ∆y 0
= lim = lim =0
dx ∆x→0 ∆x ∆x→0 ∆x

for every x.

2. Let y = f (x) = xn , where n is a positive integer. For any fixed x, Binomial Theorem gives
     
n n−1
n n n n−2 2 n
(x + ∆x) = x + x (∆x) + x (∆x) + . . . + (∆x)n .
1 2 n

Hence

∆y (x + ∆x)n − xn
=
∆x
∆x
n n−1 n
(∆x)2 + . . . + nn (∆x)n

n−2

1 x (∆x) +
x 2
=
    ∆x  
n n−1 n n−2 n
= x + x (∆x) + . . . + (∆x)n−1 .
1 2 n

Therefore,
dy ∆y
= lim = nxn−1 .
dx ∆x→0 ∆x
√
3. Let y = f (x) = x. Then for any given x > 0,
√ √ √ √  √ √ 
∆y x + ∆x − x x + ∆x − x x + ∆x + x
= = × √ √
∆x ∆x ∆x x + ∆x + x
(x + ∆x) − x
= √ √
∆x ( x + ∆x + x)
1
= √ √ .
x + ∆x + x

Hence
dy ∆y 1 1 1
= lim = lim √ √ =√ √ = √ .
dx ∆x→0 ∆x ∆x→0 x + ∆x + x x+ x 2 x
√
4. The function f (x) = 3
x is not differentiable at x = 0.
Proof In fact, for any ∆x 6= 0,
√
3
f (0 + ∆x) − f (0) ∆x − 0 1
= = 2 .
∆x ∆x (∆x) 3

Therefore,
f (0 + ∆x) − f (0) 1
lim = lim 2 .
∆x→0 ∆x ∆x→0 (∆x) 3

1
As lim 2 does not exist (= ∞), f ′ (0) does not exist in this case.
∆x→0 (∆x) 3

17
 y
1.5
1.25
1
0.75
0.5
0.25

-5 -2.5 0 2.5 5
x

d(sin x) d(cos x)
5. = cos x, = − sin x.
dx dx
Proof Using the identity

1 1
sin A − sin B = 2 cos (A + B) sin (A − B),
2 2

we have, for any fixed x and any ∆x 6= 0,

∆x ∆x ∆x
sin(x + ∆x) − sin x 2 cos(x + ) · sin sin
= 2 2 = 2 · cos(x + ∆x )
∆x ∆x ∆x 2
2

∆x
sin ∆x
Since lim 2 = 1 and lim cos(x + ) = cos x, we have
∆x→0 ∆x ∆x→0 2
2

d(sin x) sin(x + ∆x) − sin x

= lim = cos x.
dx ∆x→0 ∆x

d
Similarly, we have (cos x) = − sin x.
dx

Remark 2.1.3 If ∆x is restricted in the definition of the derivative so ∆x tends to zero through
only positive or negative values, we obtain one-sided derivatives of the function. In other words,
we define the right-side derivative of f (x) at the point x0 by

f (x0 + ∆x) − f (x0 )

f+′ (0) = lim ,
∆x→0+ ∆x

and the left-side derivative of f at the point x0 by

f (x0 + ∆x) − f (x0 )

f−′ (0) = lim .
∆x→0− ∆x

It is easy to see that f ′ (x0 ) exists if and only if f+′ (0) and f−′ (0) both exist and are equal, in which
case f ′ (x0 ) = f+′ (x0 ) = f−′ (x0 ).

18
 y = f(x)

P (x0, f(x0))
f( x0 - x )-f( x0) f( x0+ x ) -f( x0)

x0 - x x0 + x
x0
x

Example 2.1.4

1. If f (x) = |x|, then

f (0 + ∆x) − f (0) |∆x| ∆x

f+′ (0) = lim = lim = lim =1
∆x→0+ ∆x ∆x→0+ ∆x ∆x→0+ ∆x

and
f (0 + ∆x) − f (0) |∆x| −∆x
f−′ (0) = lim = lim = lim = −1.
∆x→0− ∆x ∆x→0− ∆x ∆x→0− ∆x

Since f+′ (0) 6= f−′ (0), f is not differentiable at x = 0.

This phenomenon is reflected by the graph of y = f (x) = |x| which has a “sharp turn” at
x = 0.

y
y = -x y =x

0
x

(
4x, if x ≤ 0;
2. Let f (x) = If f (x) is differentiable at x = 0, find the values of a
ax + b, if x > 0.
and b.
Solution Observe that

f (0 + ∆x) − f (0) a ∆x + b − 0 b
f+′ (0) = lim = lim = a + lim .
∆x→0+ ∆x ∆x→0 + ∆x ∆x→0 + ∆x

If f+′ (0) exists, we must have b = 0. Furthermore,

f (0 + ∆x) − f (0) 4∆x − 0

f−′ (0) = lim = lim = 4.
∆x→0− ∆x ∆x→0 − ∆x

The existence of f ′ (0) implies f+′ (0) = f−′ (0). We thus conclude that a = 4.

19
2.2 Continuity and differentiability

It is intuitively clear that a curve cannot have a tangent line at a point if the curve has a break at
that point. This suggests that differentiability implies continuity. The following theorem confirms
this observation.

Theorem 2.2.1 If f (x) is differentiable at x0 , then f (x) is continuous at x0 .

Proof For any nonzero number ∆x, one has

 
f (x0 + ∆x) − f (x0 )
f (x0 + ∆x) − f (x0 ) = ∆x.
∆x

Hence
 
f (x0 + ∆x) − f (x0 )
lim [f (x0 + ∆x) − f (x0 )] = lim lim ∆x = f ′ (x0 ) × 0 = 0.
∆x→0 ∆x→0 ∆x ∆x→0

Therefore, f (x) is continuous at x0 .

Note that the converse of the theorem is false. For instance, f (x) = |x| is continuous at x = 0, but
is not differentiable there.

2.3 Techniques of Differentiation

2.3.1 Differentiation of sums, differences, products and quotients

In the previous section, we obtained derivatives of certain functions such as y = xn , y = cos x and
√
y = x. The rules for differentiating sums, differences, products and quotients of functions are
summarized in the following theorems.

Theorem 2.3.1 Let f and g be functions differentiable at a point x. Then the same is true for
f
the functions f + g, f − g, f g and (with g(x) 6= 0). Specifically, we have :
g
1. the sum rule:
(f + g)′ (x) = f ′ (x) + g ′ (x);

2. the difference rule:

(f − g)′ (x) = f ′ (x) − g ′ (x);

3. the product rule:

(f g)′ (x) = f (x) g ′ (x) + f ′ (x) g(x),

in particular, (cf )′ = c f ′ for any constant c; and

4. the quotient rule:  ′

f g(x) f ′ (x) − f (x) g ′ (x)
(x) = .
g g 2 (x)

20
Proof We shall prove the product rule here and leave the proof of the remaining formulas to the
students. For any fixed x and ∆x 6= 0, we have

(f g)(x + ∆x) − (f g)(x)

∆x
f (x + ∆x)g(x + ∆x) − f (x)g(x)
=
∆x
[f (x + ∆x)g(x + ∆x) − f (x)g(x + ∆x)] + [f (x)g(x + ∆x) − f (x)g(x)]
=
  ∆x  
f (x + ∆x) − f (x) g(x + ∆x) − g(x)
= g(x + ∆x) + f (x)
∆x ∆x

When ∆x → 0, one has

f (x + ∆x) − f (x) g(x + ∆x) − g(x)

→ f ′ (x), → g ′ (x) and g(x + ∆x) → g(x).
∆x ∆x

This implies the existence of (f g)′ (x) and the product rule.

Example 2.3.2

1. If f (x) = x2 sin x, then the product rule gives

f ′ (x) = (x2 )′ sin x + x2 (sin x)′ = 2x sin x + x2 cos x.

ex
2. If f (x) = , then the quotient rule implies
x

x · (ex )′ − ex · (x)′ xex − ex

f ′ (x) = = .
x2 x2

3. Quotient rule gives

d1 d cos x
d d 1 cos x · −1·
sec x = = dx dx = sin x = sec x tan x.
dx dx cos x (cos x)2 cos2 x

2.3.2 The chain rule of differentiation

The chain rule is required when we differentiate the composition of two or more functions.

Theorem 2.3.3 (Chain rule) Let g and h be functions such that h is differentiable at x0 and g
is differentiable at y0 = h(x0 ). If f = g ◦h, i.e., f (x) = g(h(x)), then the function f is differentiable
at x0 , and
f ′ (x0 ) = g ′ (y0 ) × h′ (x0 ).

Remark 2.3.4 If we write

y = h(x), z = g(y) and z = g(h(x)),

21
then the chain rule simply asserts that
     
dz dz dy
= × .
dx at x=x0 dy at y=y0 dx at x=x0

Example 2.3.5

1. If f (x) = sin(x3 − 3x + 10), then f (x) = g(h(x)), where g(y) = sin y and h(x) = x3 − 3x + 10.
The chain rule then implies that

f ′ (x) = g ′ (y) × h′ (x) = cos y × 3x2 − 3 = cos(x3 − 3x + 10) × 3x2 − 3 .

2. Suppose that f (x) = g(sin2 x)+g(cos2 x), where g is a differentiable function of an independent
variable y. Determine f ′ (x) in terms of g ′ (y).
Solution Applying the Chain rule, we obtain

d  d 
f ′ (x) = g(sin2 x) + g(cos2 x)
 
dx dx
d d(cos2 x)
= g ′ (sin2 x) · (sin2 x) + g ′ (cos2 x) · .
dx dx

d d d d
Since (sin2 x) = 2 sin x (sin x) = 2 sin x cos x and (cos2 x) = 2 cos x (cos x) =
dx dx dx dx
−2 cos x sin x, we conclude that

f ′ (x) = 2 sin x cos x (g ′ (sin2 x) − g ′ (cos2 x) .

 

2.4 Differentiation of inverse functions

The relation between the derivative of a function and that of its inverse is contained in the following
theorem.

Theorem 2.4.1 Suppose f is a function differentiable on (a, b), and that g is the inverse of f . If
the derivative f ′ (x0 ) is nonzero at some point x0 in (a, b), then the derivative g ′ (y0 ) exists, where
y0 = f (x0 ), and
1
g ′ (y0 ) =
f ′ (x0 )

Proof Let x = g(y) be the inverse of y = f (x). It follows that f (g(y)) = y for all y in the
domain of g. By the chain rule, we have

f ′ (g(y)) g ′ (y) = 1

for all y in the domain of g. In particular, we have f ′ (g(y0 )) g ′ (y0 ) = 1. Since

f ′ (g(y0 )) = f ′ (x0 ) 6= 0,

22
 1
we conclude that g ′ (y0 ) = , as asserted.
f ′ (x0 )
Notes:

1. The above proof is incomplete because it assumes the differentiability of g at y0 . For a

complete proof, please refer to [3].

2. Theorem (2.4.1) may be stated as

 
dx 1
=   ,
dy y=y0 dy
dx x=x0

which appears as a trivial algebraic identity.

Example 2.4.2 If y = f (x) = sin x, −π/2 < x < π/2, then x = g(y) = sin−1 y. Using Theorem
(2.4.1) and the fact that cos x 6= 0 whenever −π/2 < x < π/2, one has

1 1 1 1
g ′ (y) = = = =p ,
f ′ (x) cos x cos(g(y)) 1 − y2

for every y in (−1.1). Changing the dummy variable y to x, we obtain the formula

d 1
sin−1 x = √


,
dx 1 − x2

for | x | < 1. Similarly, we have

d 1
cos−1 x = − √
dx 1 − x2
for | x | < 1.

2.5 A collection of differentiation formulas

dc d tan x
1. = 0 for any constant c; 7. = sec2 x;
dx dx
dxn d sec x
2. = nxn−1 for any integer n; more gen- 8. = sec x tan x;
dx dx
dxα
erally, = αxα−1 for any real number d csc x
dx 9. = − csc x cot x;
α, provided that x > 0; dx
d cot x
dex 10. = − csc2 x;
3. = ex ; dx
dx
d sin−1 x 1
d ln x 1 11. =√ ;
4. = ; dx 1 − x2
dx x
d sin x d cos−1 x 1
5. = cos x; 12. = −√ ;
dx dx 1 − x2
d cos x d tan−1 x 1
6. = − sin x; 13. = 2 .
dx dx x +1

23
2.6 Differential and linear approximation
dy
The symbol is admittedly a bit confusing, because it is not a quotient with numerator dy and
dx
∆y
denominator dx. Instead, it is defined as the limiting value of the quotient , as ∆x → 0, i.e.,
∆x
 
dy ∆y
= lim .
dx at x=x0 ∆x→0 ∆x

It is however useful to assign meanings to quantities dx and dy separately as follows.

Consider the function defined by the graph y = f (x) as shown in following figure.

Tangent line
y at (x0, f(x0))
y = f(x)
∆y = Actual change
f(x0+dx)
f(x0)+dy

f(x0)
dy = Differential change

dx

x
x0 x0+dx

Let x0 be a fixed point and suppose x0 changes by an amount dx = ∆x, i.e., from x = x0 to a
new value, x = x0 + dx. The corresponding change in y would then be from f (x0 ) will change to
f (x0 + dx). The difference between these y values is defined by

∆y = f (x0 + dx) − f (x0 )

Now consider the tangent line to the curve y = f (x) at the point (x0 , f (x0 )). The equation of the
tangent line is at (x0 , f (x0 )) is given by

y = f (x0 ) + f ′ (x0 ) (x − x0 ) .

Since the tangent line is a good approximation to the graph y = f (x) for values of x close to x0 ,
the change in y along the tangent line will be approximately the same as the change in y along the
graph y = f (x). We now define the quantity dy by

dy = f ′ (x0 ) dx.

It then follows that

∆y ∼
= dy = f ′ (x0 ) dx,

provided that dx is small. This is a linear approximation formula. It should be noted that dy
depends on x0 and dx, and is in fact a linear function of dx.
Reason for an Interest in Differentials

24
Differentials approximate the exact change in the dependent variable y as a result of a change in the
independent variable x. There are a number of reasons, why we are satisfied with this approximate
change
dy = f ′ (x0 )dx

in contrast to the real change

∆y = f (x0 + dx) − f (x0 )

The major reason is that it is often impossible to compute the real change, owing to the fact that
we cannot evaluate the function at the point x0 + dx. The differential approximation dy uses
information only at the point x0 . Also, the differential change dy is generally easier to compute
than the real change y, and it is often a good enough approximation for the purpose at hand.
Differentials Used in Error Analysis
Differentials are often used in studying the propagation of errors. For example, suppose that x is
a variable that we measure and that y is a variable that is computed by means of a formula

y = f (x)

If there is an error in the measurement of x, this error will give rise to an error in y. It is important
to know how errors in the measured value of x will give rise to errors in y.
To determine these relationships, we introduce some of the language of error analysis. If the true
value of some quantity is x, but the value determined by some experiment is x + dx, then the error
in the measurement is dx, and the relative error in the measurement is

dx
Relative error =
x

If we multiply the relative error times 100, we get the percentage error:

dx
Percentage error = 100 %
x

Example 2.6.1 Suppose you approximate the length of the line using a ruler, and suppose your
estimate is 4.06 inches. Suppose that the real length of the line is exactly 4 inches. What are the
absolute error, relative error, and percentage error in your measurement?
Solution The error (or absolute error) is simply the difference between the measurement and the
real value, although you can never really determine this value in practice (since you never know the
exact real value). In this case it is

Absolute error = 0.06 inch

The absolute error is always taken to be a positive number. The relative error is the absolute error

25
divided by the real value. In this case it is

0.06
Relative error = = 0.015
4

If we multiply this quantity times 100, we get the percentage error:

Percentage error = 100(0.015)% = 1.5%

Remark 2.6.2 (Comment on Relative Error) The relative error is a much more useful mea-
sure of the accuracy of a measurement than the absolute error. Someone may tell you that the
absolute error in a measurement is 5 pounds but unless you know the size of the measurement, this
value is almost meaningless. After all, an error of 5 pounds could be the error in weighing a child
or an elephant. If an elephant were being weighed, an error of 5 pounds would not be as startling
as an error of 5 pounds in weighing a child.

Example 2.6.3

1. Let y = f (x) = x2 . Use the above linear approximation formula to find the change in y as x
changes from 1.0 to 1.1.
Solution Here we have x0 = 1 and dx = 0.1. Since f ′ (x) = 2x for every x, we conclude
that at x0 ,
dy = f ′ (x0 ) dx = 2 × 0.1 = 0.2.

Therefore, ∆y = 1.12 − 1.02 ∼

= dy = 0.2. See the following figure.

y ∆y = 0.21 (1.1, 1.21)

dy = 0.20 (1.1, 1.2)
y = x2 dy
(1,1)

dx
x
1 1.1

2. The area of a circle is given by the formula A = πr2 , where r is the radius of the circle.
Suppose you know that your measurement of the radius has a relative error not exceeding
0.02. What is the relative error in your estimate of the area?
Solution Since A = πr2 , one has dA = 2πr dr. Therefore,

∆A ∼ dA 2πr dr dr
= = 2
=2 .
A A πr r

dr
It is now given that ≤ 0.02. We therefore conclude that
r

∆A ∼ dA dr
= =2 ≤ 2 × 0.02 = 0.04.
A A r

26
 Therefore, if the radius of the circle can be measured with a relative error not exceeding 0.02,
then the area of the circle will have a relative error not exceeding 0.04. It is easy to see that
in general, the relative error in estimating the area of the circle is roughly twice the relative
error in the measurement of the radius.

2.7 L’Hôpital’s rules

f (x) A
Students may recall that if lim f (x) = A and lim g(x) = B, then lim = , provided that
x→a x→a x→a g(x) B
B=
6 0. Now the question is: what can we do when A = B = 0 ? It is customary to call such a
limit an “indeterminate forms”.

Example 2.7.1

sin x sin2 x sin x

1. Observe that lim = 1 and lim = 0, while lim 2 does not exist.
x→0 x x→0 x x→0 x

0
2. shows that the “indeterminate form” may take any value. The L’Hôpital’s rule as described
0
below is a useful technique to handle indeterminate form.

Theorem 2.7.2 Suppose f and g are differentiable on (a, a + δ) for some δ > 0. Assume further
f ′ (x) f (x)
that lim f (x) = lim g(x) = 0. If lim ′ exists, then lim also exists, and
x→a+ x→a+ x→a+ g (x) x→a+ g(x)

f (x) f ′ (x)
lim = lim ′ .
x→a+ g(x) x→a+ g (x)

f ′ (x)
Note: Observe that the existence of lim may be confirmed by further differentiations.
x→a+ g ′ (x)
Moreover, it is important to check the hypothesis of Theorem (2.7.2) before applying the rule.
x
For instance, if you apply L’Hôpital’s rule to evaluate lim x , you would obtain
x→0 e

x 1
lim = lim = 1,
x→0 ex x→0 ex

which is obviously wrong. This is because the denominator does not tend to zero as x → 0.

Example 2.7.3
√
3
1 + 2x + 1
1. Evaluate lim √ .
x→−1 2+x+x
Solution By the L’Hôpital’s rule (check the hypothesis), we obtain
√
3
1 + 2x + 1 2/3 4
lim √ = = .
x→−1 2+x+x (1/2) + 1 9

27
 2. Successive applications of Theorem (2.7.2) yields

x − sin x 1 − cos x sin x cos x 1

lim 3
= lim 2
= lim = lim = .
x→0 x x→0 3x x→0 6x x→0 6 6

2.8 Higher Derivatives

dy
If y = f (x) is a differentiable function of x, we use the notation = f ′ (x) to denote its derivative.
dx
f ′ (x) is called the first order derivative of f (x). If f ′ (x) is again a differentiable function of x, we
may compute its derivative to obtain the second order derivative of f (x), which is denoted by the
symbol
d2 y
f ′′ (x) or f (2) (x) or .
dx2
In general, for any positive integer n, the n th order derivative of f (x) may be defined. We use
the notation f (n) (x) to denote the n th order derivative of f (x). For convenience, we also define
dn y
f (0) (x) = f (x). The symbol is another popular notation for the n th order derivative of
dxn
y = f (x).
A function f (x) is said to be a C n −function on an open interval I if f has continuous derivatives
up to order n in I.
Two basic rules for n-th derivatives are:

1. (f + g)(n) = f (n) + g (n) ; 2. (cf )(n) = cf (n) , where c is a constant.

Example 2.8.1

1. The following formulas hold for any n ≥ 1:

nπ   (n)
(a) (sin x)(n)= sin x + ; 1 (−1)n × n!
2 (c) = ;
x xn+1
 nπ 
(b) (cos x)(n) = cos x + ; (d) (ax )(n) = (ln a)n ax , a > 0.
2
Pn
2. Let P (x) = a0 + a1 x + a2 x2 + · · · + an xn = k=0 ak x
k be a polynomial of degree n, where
a0 , a1 , · · · , an are real coefficients with an 6= 0. It is not difficult to see that these coefficients
may be expressed in terms of the higher derivatives of P (x) at x = 0. In fact, the student can
easily convince himself that

1 1
a0 = P (0) (0), a1 = P (1) (0), a2 = P (2) (0), a2 = P (3) (0),
2 6

and more generally, that

1 (k)
ak = P (0) for any 0 ≤ k ≤ n.
k!

28
 To summarize, we conclude that
n
X P (k) (0)
P (x) = xk .
k!
k=0

Remark 2.8.2 It is interesting to note that the binomial theorem is a simple consequence of this
formula. In fact, if we let P (x) = (1 + x)n , then P (x) is a polynomial of degree n. Furthermore, it
is evident that  
n
P (k) (0) = k! × for any 0 ≤ k ≤ n.
k
Therefore, we may conclude that
n n n

n  
n
X P (k) (0) k
X k! × k k
X n k
(1 + x) = x = x = x ,
k! k! k
k=0 k=0 k=0

from which the binomial theorem follows immediately.

We now prove a product formula for higher derivatives.

Theorem 2.8.3 If f (x) and g(x) are functions with derivatives up to order n, then

n  
(n)
X n
(f g) = f (n−k) g (k) .
k
k=0

Proof This is known as the Leibniz’s formula and may be proved by mathematical induction on
n. When n = 1, this is just the usual product rule. Assuming that the formula is true for n = r,
i.e.,
r  
(r)
X r
(f g) = f (r−k) g (k) ,
k
k=0

we then have

(f g)(r+1)
" r   #
d h (r)
i d X r (r−k) (k)
= (f g) = f g
dx dx k
k=0
r  
X r d h (r−k) (k) i
= f g
k dx
k=0
r  
X r 
= f (r−k+1) g (k) + f (r−k) g (k+1)
k
k=0
r   r  
X r (r−k+1) (k) X r (r−k) (k+1)
= f g + f g
k k
k=0 k=0
r   r+1  
X r (r+1−k) (k) X r
= f g + f (r+1−k) g (k)
k k−1
k=0 k=1
  r      
r (r+1) (0) X r r (r+1−k) (k) r (0) (r+1)
= f g + + f g + f g .
0 k k−1 r
k=1

29
Using the identities
       
r+1 r r+1 r
= = 1, = = 1 and
0 0 r+1 r
     
r+1 r r
= + ,
k k k−1

we conclude that
r+1  
(r+1)
X r + 1 (r+1−k) (k)
(f g) = f g .
k
k=0

This completes the proof of Leibniz’s formula.

Example 2.8.4

1. Suppose y = x2 sin x. Find y (25) (x).

Solution Taking f (x) = sin x and g(x) = x2 , we have g (k) (x) = 0 when k > 2. Leibniz’s
formula therefore implies that

y (25) = (f g)(25)
25 · 24 (23) ′′
= f (25) g + 25 f (24) g ′ + f g
2
= cos x · x2 + 50 sin x · x − 600 cos x
= x2 − 600 cos x + 50x sin x.

1
2. If f (x) = , compute f (n) (0) for any positive integer n.
1 + x2
Solution Since 1 = f (x) (1 + x2 ), Leibniz’s rule gives

n(n − 1)
0 = (1 + x2 ) f (n) (x) + n(1 + x2 )′ f (n−1) (x) + (1 + x2 )(2) f (n−2) (x)
2
= (1 + x2 ) f (n) (x) + 2nx f (n−1) (x) + n(n − 1) f (n−2) (x),

whenever n ≥ 2. Putting x = 0,we obtain

0 = f (n) (0) + n(n − 1)f (n−2) (0), for n ≥ 2.

As f (0) (0) = f (0) = 1 and f ′ (0) = 0, we conclude that

f (2n−1) (0) = 0 and f (2n) (0) = (−1)n (2n)! .

2.9 Logarithmic and Exponential Functions

1
In earlier section, we discuss lim (1 + x) x . The limit is called e, which has a value of approximately
x→0
2.71828. The function y = loge x (logarithm with base e) is known as the logarithmic function,

30
and is denoted by the symbol y = ln x. By definition, y = ln x if and only if ey = x. Thus x must
be positive. Students are requested to prove the formula

d 1
ln x = for every x > 0.
dx x

Since the exponential function y = ex is the inverse of the logarithmic function, we conclude from
Theorem (2.4.1) that
d x
e = ex for every x.
dx
In other words, the exponential function is unchanged under differentiation.
Properties of the logarithmic function give rise to a special technique known as ”logarthmic deriva-
tive”.

Example 2.9.1 Evaluate the derivative of y = xsin x .

Solution Taking ”log”, one has ln y = ln xsin x = sin x ln x. Differentiating the identity and using
the chain rule of differentiation, one obtains

1 dy d ln x d sin x sin x
= sin x + ln x = + ln x cos x.
y dx dx dx x
   
dy sin x sin x sin x
Therefore, =y + ln x cos x = x + ln x cos x .
dx x x

2.10 Implicit Differentiation

dy
If y is given as an explicit expression of x, we may then compute its derivative . However, y may
dx
sometime be expressed as a function of x by an equation. In this case, the technique of computing
dy
is known as ”implicit differentiation”. We shall only consider one illustrative example here.
dx
More can be found in the Exercise.

dy
Example 2.10.1 Evaluate where y and x are related by the equation x3 + y 3 − 2xy − 100 = 0.
dx
Solution Although we do not intend to solve y in terms of x from the euqation, we realize that
y is defined as a function of by virtue of the equation. As such, we may differentiate the equation
with respect to x, using the chain rule of differentiation to obtain

dy dy
3x2 + 3y 2 − 2x − 2y = 0,
dx dx

dy 2y − 3x2
from which it follows that = 2 .
dx 3y − 2x

2.11 The Mean Value Theorem

The mean value theorem is the most important theorem in differential calculus. We shall discuss
the theorem in this section together with some of its applications.

31
2.11.1 Relative Extremum

Let f (x) be a given real valued function defined on an open interval I. A point x0 in I is said to
be a relative maximum of f (x) if there is a number δ > 0 such that

1. the open interval (x0 − δ, x0 + δ) is contained in I;

2. f (x) ≤ f (x0 ) for every x in (x0 − δ, x0 + δ).

Similarly, a point x1 in I is said to be a relative minimum of f (x) if there is a number ρ > 0 such
that

1. the open interval (x1 − ρ, x1 + ρ) is contained in I;

2. f (x) ≥ f (x1 ) for every x in (x1 − ρ, x1 + ρ).

A relative maximum or a relative minimum is called a relative extremum.

Example 2.11.1 The graph of f (x) = x3 − 3x + 1 is sketched below.

y
2.5

1.25
x
-2.5 -1.25 0 1.25 2.5
-1.25

-2.5

It may be observed that x = 1 is a relative minimum, while x = −1 is a relative maximum.

On the other hand, the function g(x) = x2 defined on the open interval (−1, 1) has x = 0 as its
relative minimum, but has no relative maximum.

We now prove an important theorem concerning relative extrema.

Theorem 2.11.2 (Fermat Lemma) Let x0 be a relative extremum of f (x) on an open interval
I. If f ′ (x0 ) exists, then f ′ (x0 ) = 0.

Proof Without loss of generality, suppose x0 is a relative maximum of f (x). Then there is a δ > 0
such that f (x) ≤ f (x0 ) for x in (x0 − δ, x0 + δ).
Since f (x) ≤ f (x0 ) for x ∈ (x0 − δ, x0 + δ), we conclude that

f (x0 + h) − f (x0 ) f (x0 + h) − f (x0 )

≤ 0 for 0 < h < δ and ≥ 0 for − δ < h < 0,
h h

32
 f (x0 + h) − f (x0 ) f (x0 + h) − f (x0 )
which imply lim ≤ 0 and lim ≥ 0. The existence of f ′ (x0 )
h→0+ h h→0− h
now implies that

f (x0 + h) − f (x0 ) f (x0 + h) − f (x0 )

f ′ (x0 ) = lim = lim .
h→0+ h h→0− h

We thus conclude that f ′ (x0 ) ≤ 0 and f ′ (x0 ) ≥ 0. Therefore, f ′ (x0 ) = 0.

Definition 2.11.3 Let f (x) be a differentiable function defined on an open interval I. A point x0
is said to be a critical point of f (x) if f ′ (x0 ) = 0. Theorem (2.11.2 ) asserts that every relative
extremum of a differentiable function is a critical point of the function. However, as we can see
from the following example, the converse of this statement is not true.

Example 2.11.4 If f (x) = x3 for any real number x, then f ′ (x) = 3x2 and therefore x = 0 is a
critical point of f (x). However, x = 0 is not a relative extremum. See the following figure.

7.5
y
5

2.5
x
-2 -1 0 1 2
-2.5

-5

-7.5

2.11.2 Absolute Extremum

Let f (x) be a real valued function defined on an interval I (closed or open). A point x0 in I is
called an absolute maximum of f (x) if f (x0 ) ≥ f (x) for every x in I. The concept of absolute
minimum can be defined in the same manner.

( x2,M)

y=f(x)
a b x
m

( x1,m)

Example 2.11.5 For the function f (x) = x2 , where −1 ≤ x ≤ 2, it is clear that x = 2 is an

absolute maximum, while x = 0 is an absolute minimum. From this example, it is evident that
absolute extremum may occur at critical points (such as x = 0 in this example), or else the end
points of the interval (such as x = 2 in this example).

33
It is evident that if x0 is an absolute extremum of f (x), but is not an end point of I, then x0 is a
relative extremum of f (x).

The existence of absolute extremum for a continuous function on a closed and bounded interval is
guaranteed by the following theorem.

Theorem 2.11.6 (Weierstrass) If f (x) is a continuous function on the interval [a, b], where
a < b, then f (x) has an absolute maximum and an absolute minimum in the interval.

Example 2.11.7 Let f (x) be defined by


1


 2 if x = 0;
f (x) = x if 0 < x < 1;

1
if x = 1.


2

It is clear that f has no absolute extremum in [0, 1]. This is because the function is discontinuous
at x = 0 and x = 1.

2.11.3 Rolle’s theorem and the mean value theorem

Theorem 2.11.8 (Rolle’s Theorem) Let f (x) be a function which is continuous on [a, b], dif-
ferentiable in (a, b) and satisfies f (a) = f (b) = 0. Then there is a point p in (a, b) such that
f ′ (p) = 0.

(b, f(b))
y=f(x)
f (p)=0
a
p p
x
b
f(b)-f(a)
(a, f(a)) f ( p) = b-a =0

Proof By Theorem (2.11.6), f (x) has an absolute maximum x0 and an absolute minimum x1
in the interval [a, b]. We now define

M = f (x0 ) and m = f (x1 ).

It follows that m ≤ f (x) ≤ M for every x in [a, b]. There are two cases:

Case (i) M = m: This implies that f (x) is identically equal to a constant on [a, b]. Therefore,
f ′ (x) = 0 for every x in [a, b].

Case (ii) m < M : Since f (a) = f (b) = 0, either M > 0 or m < 0. If M > 0, then x0 is inside the
open interval (a, b) because f (a) = f (b) = 0, and x0 must therefore be a relative maximum

34
 of f (x). By Theorem (2.11.2), we conclude that f ′ (x0 ) = 0. On the other hand, if m < 0,
then is inside the open interval (a, b), and x1 must be a relative minimum of f (x). Theorem
(2.11.2) then implies f ′ (x1 ) = 0.

Remark 2.11.9

1. If f (x) is not everywhere differentiable in (a, b), the conclusion of Rolle’s Theorem may be
false. For instance, the function f (x) = 1 − |x| is continuous on [−1, 1], differentiable in
(−1, 1) except at x = 0, and f (−1) = f (1) = 0. However, f ′ (x) 6= 0 in (−1, 1).

2. If a function f (x) satisfies the hypothesis of Rolle’s Theorem, then f ′ (x) may vanish at more
than one point between a and b. For instance, f (x) = sin x satisfies f (0) = 0 and f (2π) = 0,
π 3π
and f ′ (x) = cos x = 0 at x = and x = .
2 2

Example 2.11.10 If a, b and c are real constants such that

1 a b
+ + + c = 0,
4 3 2

prove that the equation x3 + ax2 + bx + c = 0 has a root between 0 and 1.

Solution Consider the function

x4 ax3 bx2
f (x) = + + + cx.
4 3 2

1 a b
It is clear that f (0) = 0 and f (1) = + + + c = 0. Rolle’s Theorem therefore implies the
4 3 2
existence of p in the interval (0, 1) such that

0 = f ′ (p) = p3 + ap2 + bp + c = 0.

We are now in a position to state and prove the main result in this section.

Theorem 2.11.11 (The Mean Value Theorem) Suppose f (x) is a real value function which
is continuous on [a, b] and differentiable in (a, b). Then there exists a point p in (a, b) such that

f (b) − f (a)
f ′ (p) = .
b−a

y=f(x) (b, f(b))

a
p p
b
(a, f(a)) f(b)-f(a)
b- a

35
Proof Define a function g(x) as follows:
 
f (b) − f (a)
g(x) = f (x) − f (a) − (x − a) for every x in [a, b].
b−a

It is clear that g(x) is continuous on [a, b], differentiable in (a, b), g(a) = 0 and g(b) = 0. By Rolle’s
Theorem, there is a point p in (a, b) such that g ′ (p) = 0. Since

f (b) − f (a)
g ′ (x) = f ′ (x) − ,
b−a

one concludes that

f (b) − f (a)
0 = g ′ (p) = f ′ (p) − .
b−a
This proves the theorem.

Remark 2.11.12 If we denote the points (a, f (a)) and (b, f (b)) by A and B respectively, then the
f (b) − f (a)
slope of the straight line segment joining A and B is given by . Therefore, the mean
b−a
value theorem simply says that there is a point P on the segment of the graph of between A and B
such that the tangent line to the graph at P is parallel to the line segment AB.

2.12 Applications of differentiation

2.12.1 Increasing and decreasing Functions

Definition 2.12.1 A function f (x) defined on an interval I is said to be increasing if f (x1 ) ≤

f (x2 ) for x1 < x2 in I. (Notation: f (x) ր on I.) If f (x1 ) < f (x2 ) for x1 < x2 in I, the function
is said to be strictly increasing in I. The concept of decreasing and strictly decreasing functions
may be defined in an entirely similar manner.

f(x)

f(x1)

f(x2)

x1 x2

Let f (x) be a differentiable function defined on an interval I. The following results are consequences
of the Mean Value Theorem.

1. If f ′ (x) ≥ 0 for every x in I, then f (x) is increasing on I. On the other hand, if f ′ (x) ≤ 0 for
every x in I, then f (x) is decreasing on I.
Furthermore, if “f ′ (x) ≥ 0” or “f ′ (x) ≤ 0”are replaced by “f ′ (x) > 0” or “f ′ (x) < 0”, then
f (x) is strictly increasing or strictly decreasing respectively.

36
 2. If f ′ (x) = 0 for every x in I, then there is a constant C such that f (x) = C for each x in I.

3. If g(x) and h(x) be differentiable functions on I such that

g ′ (x) = h′ (x) for every x in I,

then there is a constant C such that g(x) = h(x) + C for every x.

Example 2.12.2

4
1. Discuss the behaviour of the function f (x) = x + .
x+1
Solution The function is defined for all x 6= −1. A simple calculation shows that

4 x2 + 2x − 3 (x − 1)(x + 3)
f ′ (x) = 1 − 2
= 2
= .
(x + 1) (x + 1) (x + 1)2

The behaviour of f (x) and f ′ (s) may be summarized in the following table:

−∞ < x < −3 −3 < x < −1 x = −1 −1 < x < 1 x>1

f ′ (x) +ve −ve undefined −ve +ve
f (x) ր ց undefined ց ր

Note that f ′ (−3) = f ′ (1) = 0. We thus conclude that f (x) has a relative maximum at x = −3,
a relative minimum at x = 1.

20
-3<x<-1 -1<x<1

10

x = -1 f(x)

-4 (-3,-1) -2 (1,3)

-10

-20

-30

2. If g ′ (x) = g(x) for every real number x, then g(x) = C ex for some constant C.
Proof Define f (x) = g(x) e−x . By the product rule and the given condition, we have

f ′ (x) = g ′ (x) e−x − g(x) e−x = [g ′ (x) − g(x)] e−x = 0.

The result now follows from (2).

37
 ln x
3. Define f (x) = for x > 0. Then the quotient formula gives
x

1 − ln x
f ′ (x) = .
x2

Since ln x is a strictly increasing function of x for x > 0, we conclude that f ′ (x) < 0 if
0 < x < e, f ′ (x) = 0 if x = e and f ′ (x) > 0 if x > e. Therefore, f (x) is ր on (0, e), f (x)
is ց on (e, ∞) and x = e is the absolute minimum of the function. In particular, we have
ln π ln e
> , which implies π e > eπ .
π e

2.12.2 Concavity

We have just seen how the first derivative is used to determine when a function is increasing or
decreasing and to locate relative maxima and minima. However, if we wish to know the actual
shape of a curve, we must have additional information. That is why the second order derivatives
come in.

Definition 2.12.3 A function f (x) is said to be concave up (concave down) on an interval I if

f ′ (x) is ր on I (f ′ (x) is ց on I). By the results of the previous subsection, it is clear that if
f ′′ (x) ≥ 0 (f ′′ (x) ≤ 0) on an interval, then f (x) is concave up (concave down) on that interval.

Concave down Concave up

A point x0 is said to be a point of inflexion (or inflection) of f (x) if f (x) is concave up on one side
of x0 and concave down on another side of x0 .

Remark 2.12.4 Concavity has the following geometric meaning. If y = f (x) is concave up on an
interval I, then for any points P , Q on the graph of the function, the line segment P Q always lies
above the portion of the graph between P and Q. Similarly, if the function is concave down, then
the line segment P Q always lies below the portion of the graph between P and Q.

38
 y y

Concave Up Concave Down

f(x)

P Q
P Q

f(x)

0 x 0 x

The following theorem gives a sufficient condition for a critical point to be a relative extremum,
and is known as the second order derivative test.

Theorem 2.12.5 Letf (x) be a function with continuous second order derivative on an open inter-
val I. Suppose that x0 is a critical point of f (x). If f ′′ (x0 ) > 0, then x0 is a relative minimum. If
f ′′ (x0 ) < 0, then x0 is a relative maximum.

Example 2.12.6 Discuss the behaviour of the function y = 6x4 − 8x3 + 1.

dy d2 y
Solution Since = 24x2 (x − 1) and 2
= 72x2 − 48x = 24x(3x − 2), we conclude that the
dx dx
function is concave up on (−∞, 0) and on ( 23 , ∞), concave down on (0, 32 ) and has points of inflexion
d2 y
at x = 0 and x = 23 . The function has two critical points, x = 0 and x = 1. At x = 0, = 0.
dx2
2
d y
Thus 0 is a point of inflexion. At x = 1, > 0. By Theorem (2.12.5), x = 1 is a relative
dx2
minimum for the function.

2.12.3 Some applications of differentiation in economics

Differential calculus is useful in determining the value of a variable that will maximize or minimize a
function. Typical problems in economics is to maximize profit and minimize cost. We shall discuss
a number of examples in this subsection to illustrate the idea.

Example 2.12.7

1. In a certain factory, the total cost function for producing q units of goods is given by

1
c = q 2 + 3q + 4000.
4

At what level of output will average cost per unit be a minimum? What is this minimum?
Solution The average cost per unit is equal to

c 1 4000
c̄ = = q+3+ .
q 4 q

dc̄ 1 4000 dc̄

To minimize this function, we consider its derivative
= − 2 . It follows that <0
dq 4 q dq
√ dc̄ √
when 0 < x < 16000 (≃ 126.49), > 0 when x > 16000. Consequently, the average cost
dq

39
 √ √
function is ց on (0, 16000) and ր on ( 16000, ∞). Thus the minimum average cost is
√
attained at q = 16000. Bearing in mind that q must be an integer, we compute c̄ at q = 126
and q = 127 to obtain the values 66.246031 and 66.246062. Thus the minimum average cost
per unit is attained at q = 126.
Suppose p = f (q) is the demand function for a firm’s product, where p is the selling price
per unit and q is the number of units produced and sold. Then r = q × p = q × f (q) is the
total revenue. If the cost of producing these product is given by c = h(q), then the total profit
P is equal to
P = r − c = q f (q) − h(q).
dP d2 P
We thus conclude that the profit is maximized if = 0 and < 0. In other words, we
dq dq 2
dr dc d2 r d2 c
must have = (marginal revenue = marginal cost) and < (the slope of the
dq dq dq 2 dq 2
marginal revenue curve must be less than the slope of the marginal cost curve).

2. It is given that the demand function for a product is equal to p = 400 − 2q and the average
400
cost function is c̄ = 0.2q + 4 + , where q is the number of units produced and both p and
q
q are expressed in dollars.

(a) Find the level of output at which profit is maximized.

(b) Find the price at which maximum profit occurs, and what is the maximum profit?

Solution Revenue is given by r = q×p = 400q−2q 2 and total cost c = q×c̄ = 0.2q 2 +4q+400.
Thus profit P = r − c = −2.2q 2 + 396q − 400.
d2 P
 
dP
(a) We set = −4.4q + 396 = 0 to obtain q = 90. Since = −4.4 < 0. One
dq dq 2 q=90
concludes that q = 90 gives a maximum profit.
(b) At q = 90, the unit price of the product is equal to p = 400 − 2 × 90 = 220, and the
maximum profit is P = 17, 420.

2.12.4 Rate of Change

If Q is a quantity that varies with time t, then the derivative dQ/dt gives the rate of change of that
quantity per unit time.
5-step procedure for solving related rates problems

1. Draw a relevant diagram and indicate the quantities that vary.

2. Specify in mathematical form of the rate of change you are looking for, and record all given
information.

3. Find an equation involving the variable whose rate of change is to be found.

4. Differentiate with respect to time t the equation found in step 3.

40
 5. State the final answer in coherent form, specifying the units that you are using.

Example 2.12.8 Water runs into a conical tank at rate of 9ft3 / min . The tank stands point down
and has a height of 10 ft and a base radius of 5 ft. How fast is the water level rising when the water
is 6 ft deep?

Solution: The variables in the problem are

V = volume (ft3 ) of the water in the tank at time t (min)

x = radius (ft) of the surface of the water at time t
y = depth (ft) of water in tank at time t

dV
= 9 cm3/sec
dt
5 cm

dy x
=?
dt 10 cm
when y = 6 cm
y

At the time in question,

dV
y = 6 ft and = 9 ft3 /min
dt
The water forms a cone with volume
1
V = πx2 y
3
This equation involves x as well as V and y. Because no information is given about x and dx/dt at
time in question, we need to eliminate x. The similar triangles give us a way to express x in terms
of y :
x 5 y
= or x=
y 10 2
Therefore,
1  y 2 π
V = π y = y3
3 2 12
Differentiate with respect to t.

dV π dy 3 π dy 3 dy π dy
= · = · · = y2
dt 12 dt 12 dy dt 4 dt

Use y = 6 and dV /dt = 9 to solve for dy/dt.

π dy
9= (6)2
4 dt
dy 1
= = 0.32 ft/min
dt π

At the moment in question, the water level is rising at about 0.32 ft/min.

41
Example 2.12.9 A police cruiser, approaching a right-angled intersection from the north, is chas-
ing a speeding car that has turned the corner and is now moving straight east. When the cruiser is
0.6 mi north of the intersection and the car is 0.8 mi to the east, the police determine with radar
that the distance between them and the car is increasing at 20 mph. If the cruiser is moving at 60
mph at the instant of measurement, what is the speed of the car?

Solution We picture the car and cruiser in the coordinate plane, using the positive x-axis as the
eastbound highway and the positive y-axis as the southbound highway. We let t represent time and
set

x = position of car at time t

y = position of cruiser at timet
s = distance between car and cruiser at timet.

N
dy
= - 60 mph
dt
Cruiser ds
= 20 mph
dt

y = 0.6 mi s

car
Intersaction
x = 0.8 mi

At the instant in question,

dy ds
x = 0.8 mi, y = 0.6 mi, = −60 mph, = 20 mph
dt dt

The variables are related as follow:

s2 = x2 + y 2

Differentiate with respect to t.

ds2 dx2 dy 2
= +
dt dt dt
ds2 ds dx2 dx dy 2 dy
= +
ds dt dx dt dy dt
ds dx dy
2s = 2x + 2y
dt dt dt 
ds 1 dx dy
=p x +y
dt x2 + y 2 dt dt

42
Use x = 0.8, y = 0.6, dy/dt = −60, ds/dt = 20, and solve for dx/dt.
 
1 dx
20 = q 0.8 + (0.6) (−60)
dt
(0.8)2 + (0.6)2
 
1 dx
20 = 0.8 − 36
1 dt
dx
0.8 = 56
dt
dx
= 70
dt

At the moment in question, the car’s speed is 70 mph.

2.12.5 Maximum / minimum problems

Example 2.12.10 A lighthouse lies 3 miles offshore directly across from a point A of a straight
coastline. Five miles down the coast from A there is a general store. The lighthouse keeper can row
his boat at 4 miles per hour and he can walk at 6 miles per hour. To what point of the shore should
he row so as to reach the store as quickly as possible?

Row
Walk
Light house

Shortest Time
to the Store
3

Point A
x 5-x
Store

√
Solution If the lighthouse keeper lands x miles from A, he must row for x2 + 9 miles and walk
for 5 − x miles. The total time required on that route is
√
distance rowed distance walked x2 + 9 5 − x
+ = +
rowing rate walking rate 4 6

We want to find the value of x that minimizes the function

√
x2 + 9 5 − x
T (x) = + , 0≤x≤5
4 6

Differentiation gives
x 1
T ′ (x) = √ − , 0<x<5
4 x +9 6
2

43
Setting T ′ (x) = 0, we find that
p
6x = 4 x2 + 9
36x2 = 16 x2 + 9

20x2 = 144
6√
x=± 5
5
√ √
Since − 56 5 does not lie in the range 0 < x < 5, we reject it. Since T decreases for 0 < x < 65 5
√ √
and increases for 65 5 < x < 5, it follows that x = 56 5 gives the absolute minimum value for T .
The lighthouse keeper can reach the general store most quickly by rowing to the point that lies
6
√
5 5 miles down shore from A.

44
Chapter 3

Indefinite and definite integrals

3.1 Indefinite integrals as anti-derivatives

Anti-differentiation is to reverse the process of differentiation. Specifically, given a function f (x),

we wish to find another function F (x) such that

F ′ (x) = f (x).

Such a function F (x) is called an indefinite integral of f (x).

Z
Notation: We use the symbol f (x) dx to denote F (x). By definition,

d
Z
f (x) dx = f (x).
dx

d 3
Z
As a simple illustration, since x = 3x2 , we have 3x2 dx = x3 .
dx

Remark 3.1.1 It is clear that if F (x) is an indefinite integral of f (x), then F (x) + C, where C is
any constant, is also an indefinite integral of f (x). On the other hand, if f (x) has two indefinite
integrals F1 (x) and F2 (x), then the mean value theorem shows that there is a constant C such that

F1 (x) = F2 (x) + C.

Referring to above example, x3 + 100 is also an indefinite integral of 3x2 .

Differentiation
x3 + 2
x3 - 1
x3 + 2000 3x2
x3 + e Integration

45
 d
With every differentiation formula, there is an integration formula. For instance, since tan x =
Z dx
sec2 x, we have sec2 x dx = tan x.

We list a few useful integration formulas as follows. They form a so-called integration table. For
simplicity, the constant of integration is omitted.

1
Z Z
k
1. x dx = xk+1 (k 6= −1) 5. sin x dx = − cos x;
k+1
Z Z
2. x−1 dx = ln | x |; 6. sec2 x dx = tan x;

1
Z Z
3. eax
dx = eax (a 6= 0); 7. csc2 x dx = − cot x;
a
1 1 −1 x
Z Z  
4. cos x dx = sin x; 8. dx = tan (a > 0)
x2 + a2 a a

In general, if f (x) and g(x) are given functions and k is a constant, then
Z Z Z
1. [f (x) + g(x)] dx = f (x) dx + g(x) dx;
Z Z
2. k f (x) dx = k f (x) dx.

Example 3.1.2

1.
Z Z Z Z Z
(5 sin x + 3 cos x) dx = 5 sin x dx + 3 cos x dx = 5 sin x dx + 3 cos x dx

= −5 cos x + 3 sin x + C

2.
Z  
12
Z Z Z Z
2 2
2x − x + 1 − dx = 2 x dx − xdx + dx − 12 x−1 dx
x
2 1
= x3 − x2 + x − 12 ln | x | + C.
3 2

3.2 Definite integrals as limits of sums

Suppose f (x) is a continuous function on a closed interval [a, b]. We divide [a, b] into n subintervals
by inserting (n − 1) points x1 , x2 , . . ., xn−1 between a and b such that

a < x1 < x2 < · · · < xn−1 < b.

46
Let us denote a by x0 and b by xn . The set of points {x0 , x1 , x2 , . . . , xn−1 , xn } is then called a
partition of the interval [a, b]. It defines n subintervals of [a, b] as follows:

[x0 , x1 ] , [x1 , x2 ] , . . . , [xn−1 , xn ] .

The length of the kth subinterval [xk−1 , xk ] is given by ∆xk = xk − xk−1 .

x0 = a x1 x2 xk-1 xk xn-1 xn = b

For each 1 ≤ k ≤ n, we choose a point ck between xk−1 and xk . Then (ck , f (ck )) is a point on the
curve y = f (x). We now form a rectangle on each subinterval as shown in the following figure.
y
y = f(x)
(cn, f(cn))

(c2, f(c2))
(ck, f(ck))
(c1, f(c1))

f(ck)

x
x0 = a x1 x2 xk-1 xk xn-1 xn = b

Geometrically, the number f (ck ) ∆xk is equal to the area of the rectangle with base [xk−1 , xk ] and
height equal to f (ck ). We now add the products f (ck ) ∆xk to form the sum

n
X
S= f (ck ) ∆xk .
k=1

This sum, which depends on the partition and the choice of the numbers ck , is called a Riemann
sum for f (x) on the interval [a, b] .
If {x0 , x1 , x2 , . . . , xn−1 , xn } is a partition of [a, b], we define the norm of the partition by ∆ =
max ∆xk .
1≤k≤n

Definition 3.2.1 Let f (x) be a function defined on a closed interval [a, b]. We then consider the
limit of sums lim nk=1 f (ck ) ∆xk , where the limit is taken over all possible partitions with norm
P
∆→0
approaching to zero. This limit, if exists, is called the definite integral of f over [a, b]. We use the
Z b
symbol f (x) dx to denote the definite integral of f over [a, b]. In other words,
a

n
X Z b
lim f (ck ) ∆xk = f (x) dx.
∆→0 a
k=1

47
The numbers a and b are called the lower and the upper limit of the integral respectively.

Example 3.2.2
Z b
1. While the integral of f (x) from a and b is usually denoted by f (x) dx, it is important to
a
note that x is only a dummy variable, i.e., we may replace it by any letter we like, such as
Z b Z b
f (t) dt or f (u) du.
a a

Note: Definite Integral has the following geometric meaning: Let f (x) be continuous and
suppose f (x) ≥ 0 for a ≤ x ≤ b. The area of the region between the graph of f and the
Z b
interval [a, b] is then equal to f (x) dx.
a
Z 1
2. Using the definition of definite integral, evaluate x2 dx.
0
Solution
y y
y = x2 y = x2

x
0 1/n 2/n 3/n x 0 1/n 2/n 3/n 1
1

j
Introduce the partition 0, n1 , n2 , . . . , 1

of [0, 1]. Choosing cj = for every 1 ≤ j ≤ n, it
n
follows from the definition of definite integral that

1 n  2 n
k 1 1 X 2
Z X
2
x dx = lim × = lim 3 k .
0 n→∞ n n n→∞ n
k=1 k=1

n 1
n (n + 1) (2n + 1) 1
Z
k2 = x2 dx = .
P
As , we conclude that
k=1 6 0 3

3.3 Properties of Definite Integral

Definition 3.3.1 If b < a, one defines

Z b Z a
f (x) dx = − f (x) dx.
a b

The following properties of the definite integral are important:

48
 1. If f (x) and g(x) are functions on [a, b] and k is a constant , then
Z b Z b Z b
[f (x) + g(x)] dx = f (x) dx + g(x) dx, and
a a a
Z b Z b
kf (x) dx = k f (x) dx.
a a

2. If a < c < b, then

Z b Z c Z b
f (x) dx = f (x) dx + f (x) dx.
a a c

3. If f (x) ≤ g (x) for all x in the interval [a, b], then

Z b Z b
f (x) dx ≤ g (x) dx.
a a

In particular, one has

Z b Z b
f (x) dx ≤ |f (x)| dx.
a a

4. Let f (x) be a continuous function on [a, b] and m, M be numbers such that m ≤ f (x) ≤ M
for all x in the interval. Then
Z b
m (b − a) ≤ f (x) dx ≤ M (b − a)
a

We shall now discuss the most important theorem in calculus.

Theorem 3.3.2 (Fundamental Theorem of Calculus) Let f (t) be a continuous function on

[a, b]. For any x in [a, b], we define
Z x
F (x) = f (t) dt.
a

Then F ′ (x) = f (x), i.e.,

x
d
Z
f (t) dt = f (x).
dx a

y
y = f(x)

m M

a x x

49
Proof For any number ∆x > 0, we have
Z x+∆x Z x Z x+∆x
F (x + ∆x) − F (x) = f (t) dt − f (t) dt = f (t) dt.
a a x

Suppose m and M are respectively the maximum and the minimum values of f (t) in the interval
[x, x + h]. It follows from Property (4) above that
Z x+∆x
m · ∆x ≤ f (t) dt ≤ M · ∆x.
x

We thus conclude that

F (x + ∆x) − F (x)
m≤ ≤ M.
∆x
When ∆x → 0+, it follows from the continuity of the function f that both m and M approach to
f (x), i.e.,
F (x + ∆x) − F (x)
f (x) = lim .
∆x→0+ ∆x
F (x + ∆x) − F (x)
Similarly, lim also exists and is equal to f (x). This proves the theorem.
∆x→0− ∆x

Theorem 3.3.3 (Second Fundamental Theorem of Calculus) Let f (x) be continuous on [a, b].
If G(x) is an indefinite integral of f (x), i.e., the function G(x) satisfies G′ (x) = f (x), then
Z b
f (x) dx = G(b) − G(a) ≡ G(x)|ba .
a

Proof By Theorem (3.3.2), we know that the function

Z x
F (x) = f (t)dt
a

is an indefinite integral of f (x) on [a, b] . If G is also an indefinite integral of f (x) on [a, b] , then
there exists a constant C such that
F (x) = G (x) + C

for all x in [a, b]. Since F (a) = 0, one concludes that G (a) + C = 0, and thus C = −G (a) . It
follows that F (x) = G (x) − G (a) for all x in [a, b]. In particular, we have
Z b
f (x) dx = F (b) = G (b) − G (a) .
a

It is clear that Theorem (3.3.3) is a very powerful tool to evaluate definite integrals. It converts a
problem of finding limits of sums into a problem of finding indefinite integrals.

Example 3.3.4
Z b
1. Evaluate xk dx, where k is any real number 6= −1.
a

50
 xk+1
Z
Solution Since xk dx = , Theorem (3.3.3) leads to
k+1

b b
xk+1 bk+1 − ak+1
Z
xk dx = = .
a k+1 a k+1

Z π
2
2. Compute [2 sin x − e−2x ] dx.
0
Solution
Z π Z π Z π
2 2 2
2 sin x − e−2x dx = 2 e−2x dx
 
sin x dx −
0 0 0
−2x x= π2
e e−π + 3
= −2 cos x + = .
2 x=0 2

3.4 Techniques for Integrations

3.4.1 Integration by Substitution

Evaluation of an indefinite integral may sometime be simplified by a suitable change of variable.

The process is known as integration by substitution. If the independent variable x in the indefinite
integral Z
J= f (x) dx

dJ
is changed to t by the substitution x = φ(t) so that by definition, = f (x), then the chain rule
dx
of differentiation yields
dJ dJ dx dx
= = f (x) = f (φ(t)) φ′ (t).
dt dx dt dt
As a result, we have Z Z
J= f (x)dx = f (φ(t)) φ′ (t) dt.

On the other hand, if the integral to be evaluated is of the form

du
Z
S= g(u) dx
dx

where u is some function of x, then Z

S= g(u) du

because the chain rule of differentiation gives

Z  Z 
d d du du
g(u) du = g(u) du × = g(u) .
dx du dx dx

These two formulas are commonly known as the change of variable formulas.
Skill in integration by the method of substitution is mainly derived from practice. The following

51
are typical examples.
Z
1. To evaluate integrals of the type f (ax + b) dx, we put u = ax + b to obtain a dx = du and

1
Z Z
f (ax + b) dx = f (u) du.
a

1
Z
For instance, if b is a constant, then dx = ln |x + b| + C.
x+b
Z
2. For integrals of the form g(f (x))f ′ (x) dx, where g is some function of one variable, we may
put u = f (x) to obtain du = f ′ (x) dx and thus conclude that
Z Z
′
g(f (x)) f (x) dx = g(u) du.

Example 3.4.1

1. By putting u = 1 + 3x, we have

1 sin u sin(1 + 3x)

Z Z
cos(1 + 3x) dx = cos u du = = + C.
3 3 3

dx
Z
2. Evaluate .
x2 + 4x + 13
Solution Completing square, we have x2 + 4x + 13 = (x + 2)2 + 9. Thus

dx dx du
Z Z Z
2
= 2 2
= (where u = x + 2)
x + 4x + 13 (x + 2) + 3 u + 32
2
 
1   1
−1 u −1 x + 2
= tan = tan + C.
3 3 3 3

dx
Z
3. Evaluate √ .
x+ x
Solution By putting x = t2 , one has

dx 2t dt 2 dt √
Z Z Z
√ = = = 2 ln |t + 1| = 2 ln( x + 1) + C.
x+ x t2 + t t+1

Z
2
4. Evaluate xe−x dx.

Solution We put u = x2 to obtain

1 1 1 1
Z Z Z
−x2 −x2 2 2
e−u du = − e−u = − e−x + C.


xe dx = e d x =
2 2 2 2

52
 Z
5. Evaluate sec x dx.

Solution since

tan x + sec x sec x tan x + sec2 x

Z Z Z
sec x dx = sec x × dx = dx.
tan x + sec x tan x + sec x

By putting u = tan x + sec x, one has du = (sec2 x + sec x tan x)dx. As such, we conclude that

du
Z Z
sec x dx = = ln | u | = ln | tan x + sec x | + C.
u

6.

4x + 7 2(2x + 4) dx
Z Z Z
√ dx = √ dx − √
2
x + 4x + 2 2
x + 4x + 2 2
x + 4x + 2
d(x2 + 4x + 2) dx
Z Z
= 2 √ − q √
2
x2 + 4x + 2 (x + 2)2 − 2
p  p 
= 4 x2 + 4x + 2 − ln x + 2 + x2 + 4x + 2 + C.

Remark 3.4.2 In the evaluation of a definite integral by substitution, the limits of the integral
have to be changed accordingly. For instance, consider the change of variable x = φ(t), where φ(t)
is either a strictly increasing function of t or a strictly decreasing function of t. Suppose a = φ(α)
and b = φ(β). Then
Z b Z β Z φ−1 (b)
′
f (x)dx = f (φ(t)) φ (t) dt = f (φ(t)) φ′ (t) dt.
a α φ−1 (a)

Z 1 p
Example 3.4.3 Evaluate x2 1 − x2 dx.
0
Solution Using the substitution x = sin t, we have
π Z π
1
1 2
Z Z
p
2 2 p
2
2
x 1 − x dx = 2
sin t 1 − sin t cos t dt = sin2 2t dt
0 0 4 0
π
1 π
Z
2
= (1 − cos 4t ) dt = .
8 0 16

3.4.2 Integration by Parts

The product rule of differentiation implies f g ′ = (f g)′ − f ′ g. Integration of this relation with
respect to x gives Z Z
′
f g dx = f g − f ′ g dx.

This is the formula for integration by parts.

53
Using the differential notation df = f ′ dx and dg = g ′ dx, the integration by parts formula can be
re-written as Z Z
f dg = f g − g df.

Example 3.4.4
Z
1. Evaluate xn ln x dx where n = 0, 1, 2, 3, · · · .

Solution

xn+1 xn+1 xn+1

Z Z   Z
n
x ln x dx = ln x d = ln x · − d ln x.
n+1 n+1 n+1

1
Since d ln x = dx, we have
x

xn+1 xn xn+1
Z Z
d ln x = dx = .
n+1 n+1 (n + 1)2

We therefore conclude that

xn+1 xn+1
Z
xn ln x dx = ln x · − + C.
n + 1 (n + 1)2

Z
2. Evaluate sec3 x dx.

Solution Since sec3 x = sec x sec2 x = sec x d tan x, we may integrate by parts to obtain

54
 Z Z
3
sec x dx = sec x tan x − tan x d sec x
Z
= sec x tan x − sec x tan2 x dx
Z
= sec x tan x − sec x (sec2 x − 1) dx
Z Z
3
= sec x tan x − sec x dx + sec x dx.

We therefore obtain that

sec x tan x 1 sec x tan x + ln |sec x + tan x|

Z Z
3
sec x dx = + sec x dx = + C.
2 2 2

3. Evaluate the integrals

Z Z
ax
I= e cos bx dx and J = eax sin bx dx,

where a, b are nonzero real constants.

Solution Integration by parts gives

eax eax cos bx


 Z ax
e
Z
I = cos bx d = − d cos bx
a a a
eax cos bx b eax cos bx b
Z
= + eax sin bx dx = + J.
a a a a

Another integration by parts leads to

eax eax sin bx


 Z ax
e
Z
J = sin bx d = − d sin bx
a a a
eax sin bx b eax sin bx b
Z
= − eax cos bx dx = − I.
a a a a

Combining these two results, one obtains a system of two linear equations in two unknowns:
(
aI − bJ = eax cos bx
,
bI + aJ = eax sin bx

from which it follows that

   
ax a cos bx + b sin bx ax a sin bx − b cos bx
I=e + C and J = e + C.
a2 + b2 a2 + b2

55
3.4.3 Reduction Formulae

Using integration by parts, it is often possible to reduce an integral Im (where m is a nonnegative

integer) to an expression which includes, among other things, an integral Ik where k is an integer
less than m. Such a relation is usually referred to as a reduction formula.

Example 3.4.5
Z
1. Let Im = cosm x dx, where m ≥ 2. Since d sin x = cos x dx, one has

Z Z
Im = cosm x dx = cosm−1 x d sin x
Z
= cosm−1 x sin x − sin x d cosm−1 x
Z
= cosm−1 x sin x + (m − 1) sin2 x cosm−2 x dx
Z Z
= cosm−1 x sin x + (m − 1) cosm−2 x dx − (m − 1) cosm x dx

= cosm−1 x sin x + (m − 1) Im−2 − (m − 1) Im .

It follows that
cosm−1 x sin x
 
m−1
Im = + Im−2 .
m m
This enables us to calculate Im for any positive integer m. For instance,

1 8
I9 = cos8 x sin x + I7
9 9
8 cos6 x sin x 6

1 8
= cos x sin x + + I5
9 9 7 7
16 cos4 x sin x 4
 
1 8 8 6
= cos x sin x + cos x sin x + + I3
9 63 21 5 5
1 8 16
= cos8 x sin x + cos6 x sin x + cos4 x sin x
9 63 105
64 cos2 x sin x 2
 
+ + I1 + C.
315 3 3
Z
Since I1 = cos x dx = sin x, we conclude that

1 8 16 64 128
I9 = cos8 x sin x + cos6 x sin x + cos4 x sin x + cos2 x sin x + sin x + C.
9 63 105 945 945
Z
2. Suppose I(m, n) = sinm x cosn x dx, where m, n are nonnegative integers. For m ≥ 2, show
that  
1 m−1
I(m, n) = − sinm−1 x cosn+1 x + I(m − 2, n).
m+n m+n

56
 Solution For any positive integers p and q,

d
(sinp x cosq x) = p sinp−1 x cosq+1 x − q sinp+1 x cosq−1 x
dx
= p sinp−1 x cosq−1 x(1 − sin2 x) − q sinp+1 x cosq−1 x
= p sinp−1 x cosq−1 x − (p + q) sinp+1 x cosq−1 x.

We may now integrate the above identity to obtain

Z Z
p+1 q−1 p q
(p + q) sin x cos x dx = − sin x cos x + p sinp−1 x cosq−1 x dx.

Putting p = m − 1 and q = n + 1, we obtain the required identity.

dx
Z
3. For any integer n ≥ 1, define In = 2
.
(x + 1)n
If n ≥ 2, integration by parts gives,
 
x 1
Z
In−1 = 2 − xd .
(x + 1)n−1 (x2 + 1)n−1

 
1 2(1 − n)x
Since d 2 n−1
= 2 dx, we have
(x + 1) (x + 1)n

x x2 dx
Z
In−1 = + 2(n − 1) .
(x2 + 1)n−1 (x2 + 1)n

Finally, we use the identity

x2 (x2 + 1) − 1 1 1
2 n
= 2 n
= 2 n−1
− 2
(x + 1) (x + 1) (x + 1) (x + 1)n

to conclude that

x dx dx
Z Z
In−1 = 2 + (2n − 2) 2
− (2n − 2) .
(x + 1)n−1 (x + 1)n−1 (x2 + 1)n

By a simple calculation, we may then conclude that

x (2n − 3)
In = 2 n−1
+ In−1
(2n − 2)(x + 1) (2n − 2)

for any integer n ≥ 2.

57
 dx
Z
Let us now use the formula in Example 3.4.5(3) to calculate the integral . By the
(x2 + 9)2
substitution x = 3t, we obtain

dx 1 dt
Z Z
2 =
(x2 + 9) 27 (t2
+ 1)2
t 1 dt
Z
= 2
+ 2
54(t + 1) 54 t +1
t 1
= + tan−1 t
54(t2 + 1) 54
x 1 −1 x
 
= + tan + C.
18x2 + 162 54 3

3.4.4 Method of Partial Fractions

We now discuss integrals of the form

P (x)
Z
J= dx,
Q(x)

where P (x) and Q(x) are polynomials in x. Before dealing with the general case, let us first consider
some special situations.

Case (1) When P (x) = 1 and Q(x) = (x + a)s , where s is an integer ≥ 1, then

dx
 Z

 = ln |x + a| + C, if s = 1;
 x+a
J=
dx −1
Z
= + C, if s ≥ 2.


s (s − 1)(x + a)s−1

(x + a)

Case (2) When P (x) = 1 and Q(x) = x2 + px + q, where p, q are constants and p2 < 4q, we may
complete square to obtain

p2
  
2 p 2
 p 2
x + px + q = x + + q− = x+ + k2 ,
2 4 2

p2 p2 4q − p2
where k 2 = q − . This is possible because q − = > 0. Using the change of
4 4 4
p
variable u = x + , we conclude that
2
 
dx du 1 −1 u
  1
−1 2x + p
Z Z
J= = = tan = tan + C.
x2 + px + q u2 + k 2 k k k 2k

Case (3) Suppose P (x) = x and Q(x) = x2 + px + q, where p, q are real numbers, with p2 < 4q.
 p 2 p2
We complete square to obtain x2 + px + q = x + + k 2 , where k 2 = q − . Using the
2 4

58
 p
substitution u = x + , we have
2
 p p p
x x+− u−
= 2 2 = 2 .
x2 + px + q p 2 u 2 + k2
x+ + k2
2

Therefore,

x dx
Z
J =
x2+ px + q
u du p du
Z Z
= 2 2
−
u +k 2 u + k2
2

1 d(u2 + k 2 ) p du
Z Z
= −
2 u2 + k 2 2 u2 + k 2
1 p u
= ln u2 + k 2 − tan−1
2 2k k 
1 2 p −1 2x + p
= ln x + px + q − tan + C.
2 2k 2k

r
Case (4) Let P (x) = 1 and Q(x) = x2 + px + q , where p and q are constants as given in Case
p
(3) and r is an integer ≥ 2. We complete square and put u = x + to obtain
2
1 1 1
=  r = .
(x2 + px + q)r p 2
 (u2 + k 2 )r
x+ + k2
2

Therefore,
dx du
Z Z
J= = ,
(x + px + q)r
2 [u2 + k 2 ]r
which can be evaluated by the method used in Example 3.4.5(3).

Case (5) Suppose P (x) = x and Q(x) = (x2 + px + q)r , where p, q are real numbers, with p2 < 4q
and r is an integer ≥ 2. We ask the students to work out this case as an exercise, following
ideas used in Case (3).
In general, given polynomials P (x) and Q(x) with degree of P (x) less than degree of Q(x), it
P (x)
can be shown that the rational function can be resolved into a sum of partial fractions
Q(x)

X A X Bx + C
s
+ 2
,
(x − a) (x + px + q)r

P (x)
Z
where s and r are positive integers. Thus dx may be evaluated using the previous
Q(x)
cases.
Finally, let us note that if the degree of P (x) is greater than or equal to the degree of Q(x),
P (x)
we may express as the sum of a polynomial and a rational function whose numerator
Q(x)
has degree less than the degree of its denominator.

59
Example 3.4.6
Z 3
x +1
1. Evaluate dx.
x2 + 4
x3 + 1 3
Solution Using long division, we have =x− 2 . We thus conclude that
x2 + 4 x +4

x3 + 1 x2 3
Z  
3 x
Z
dx = x− dx = − tan−1 + C.
x2 + 4 2
x +4 2 2 2

x2 + x + 1
Z
2. Evaluate J = dx.
(x2 − 1)(x2 − 4)
Solution Setting

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

where A, B, C, D are undetermined coefficients, we obtain

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

+C(x2 − 1)(x + 2) + D(x2 − 1)(x − 2).

A simple calculation gives

1 1 7 1
A=− ,B= ,C= and D=− .
2 6 12 4

Therefore,

1 dx 1 dx 7 dx 1 dx
Z Z Z Z
I = − + + −
2 x−1 6 x + 1 12 x−2 4 x+2
1 1 7 1
= − ln | x − 1| + ln | x + 1| + ln | x − 2| − ln | x + 2| + C.
2 6 12 4

3.4.5 Miscellaneous Methods

We shall describe here some methods for handling certain integrals of special types.

1. Integrals of the form Z

R(cos x, sin x) dx,

where R(u, v) is a rational function of two independent variables u and v, may be simplified
x
by the substitution t = tan . In fact, since
2

2 2t 1 − t2
dx = dt, sin x = and cos x = ,
1 + t2 1 + t2 1 + t2

we have
1 − t2
 
2t dt
Z Z
R(cos x, sin x) dx = 2 R , .
1 + t 1 + t2
2 1 + t2

60
 which may be evaluated using methods described in the previous subsection.

Example 3.4.7
x
(a) With the substitution t = tan , we have
2
π
1   Z 1
dx 2 dt 2 dt
Z Z
2
= 2 2
= 2+3
0 2 + cos x 0 1 − t 1 + t 0 t
2+
1 + t2
t t=1
 
2 −1 2 π π
= √ tan √ =√ × = √ .
3 3 t=0 3 6 3 3

In some cases, integrals involving sine and cosine functions may be handled by using
appropriate trigonometric identities, as illustrated in the following example.
(b) If m, n are nonnegative integers with m 6= n, then
i.

1
Z Z
cos mx cos nx dx = [cos(m + n)x + cos(m − n)x] dx
2
sin(m + n)x sin(m − n)x
= + + C;
2(m + n) 2(m − n)

ii.

1
Z Z
sin mx sin nx dx = [cos(m − n)x − cos(m + n)x] dx
2
sin(m − n)x sin(m + n)x
= − + C;
2(m − n) 2(m + n)

iii.

1
Z Z
sin mx cos nx dx =
[sin(m + n)x + sin(m − n)x] dx
2
cos(m + n)x cos(m − n)x
=− − + C.
2(m + n) 2(m − n)
√ √ √
2. Integrals involving x2 + a2 , x2 − a2 or a2 − x2
√ √ √
To evaluate integrals involving x2 + a2 , x2 − a2 or a2 − x2 , we may use the substitutions

x = a tan t, x = a sec t or x = a sin t

respectively. We illustrate these with the following example:

Example 3.4.8 With x = a tan t, so that dx = a sec2 t dt, one has

1
Z Z p 
√ dx = sec t dt = ln |sec t + tan t| = ln tan2 t + 1 + tan t + C.
x + a2
2

x 1
Z  p 
Since tan t = , one has √ dx = ln x + x2 + a2 + C.
a x2 + a2

61
Chapter 4

Applications of Integration

4.1 Area between curves

In the previous section, a definite integral is interpreted as area below a curve. In the following
examples, we shall try to find area of a region enclosed by several curves. Our procedure will be to
draw a sample element of area and use integration to add together the areas of all such elements.

f(x)

g(x)

a b

Example 4.1.1

√
1. Find the area of the region bounded by the curves y = x and y = x2 .
Solution A sketch of the region shows that area of the given region is given by

1 √ 2 3 1 3 x=1 1
Z  
x − x2 dx =


A= x2 − x = .
0 3 3 x=0 3

2. Find the area of the region bounded by the curve y 2 = x, the x−axis and the line y = 3.
(
y2 = x
Solution Solving the equations , we discover that the “vertices” of the region
y = 3
are (0, 0), (0, 3) and (9, 3). Area of the region is thus given by

9 √
2 3 x=9
Z  
A= 3 − x dx = 3x − x 2 = 9.
0 3 x=0

62
4.2 Average value of a function

Suppose we are given a function f (x) defined over an interval [a, b]. Let x1 , x2 , · · · , xn be points
in the interval. Then the average value of the n corresponding values f (x1 ), f (x2 ), · · · ,f (xn ) is
given by Pn
f (x1 ) + f (x2 ) + · · · + f (xn ) j=1 f (xj )
= .
n n
Assume that we divide [a, b] into n subintervals of equal length, and we choose x1 in the first
subinterval, x2 in the second, etc. Since the interval has length equal to b − a, each subinterval has
b−a
length equal to , which we shall called ∆x. Thus the above average is equal to
n
n n
1 X 1 X
ȳ = f (xj )∆x = f (xj )∆x.
n ∆x b−a
j=1 j=1

As n → ∞ (so ∆x → 0), we have the average value of the function f (x) over the interval [a, b]
given by
b
1
Z
ȳ = f (x) dx.
b−a a

Example 4.2.1 The profit P (in dollars) of a business is given by

P = 396q − 2.1q 2 − 400,

where q is the number of units of the product sold. Find the average profit on the interval from
q = 0 to q = 500.
Solution Average profit is equal to

100
1
Z
396q − 2.1q 2 − 400 dq = 12400.


100 − 0 0

63
Appendix A

Summation Notation

In this appendix, we introduce a notation that can be used to write lengthy sums in a compact
form. The notation uses the uppercase Greek letter Σ (sigma) and is called summation notation
or sigma notation.
Generally, if f (k) is a function of k, and m and n are integers such that m ≤ n, then

n
X
f (k)
k=m

denotes the sum of the terms that result when we substitute successive integers for k, starting with
k = m and ending with k = n.
The numbers m and n are called, respectively, the lower and upper limits of the summation; and
the letter k is called the index of summation. It is not essential to use k as the index of summation;
any letter will do. That is,
n
X n
X
f (i) = f (k).
i=m k=m

For example,
5
X 5
X 5
X
i2 , j2, k2
i=1 j=1 k=1

all denote the sum

12 + 22 + 32 + 42 + 52 .

A.1 Changing the Index of Summation

On occasion we shall want to change the sigma notation for a given sum to a sigma notation with
different limits of summation. The following example illustrates a method for doing this.
Example Given that
7
X
5k−2 = 5 + 52 + 53 + 54 + 55
k=3

64
If we define a new summation index j by mean of the formula j = k − 3. Then k = j + 3, j = 0
when k = 3, and j = 4 when k = 7,

7
X 4
X 4
X
k−2 (j+3)−2
5 = 5 = 5j+1
k=3 j=0 j=0

A.2 Properties of Summation Notation

The following properties will help to manipulate sums.

Theorem A.2.1 Given that ak , bk and c are real numbers.

n
X n
X n
X
1. (ak + bk ) = ak + bk
k=1 k=1 k=1
n
X n
X n
X
2. (ak − bk ) = ak − bk
k=1 k=1 k=1
n
X n
X
3. cak = c ak
k=1 k=1

A.3 Summation Formulas

The following are some commonly use formulas

n
X n(n + 1)
k = 1 + 2 + ··· + n =
2
k=1
n
X n(n + 1)(2n + 1)
k 2 = 12 + 22 + · · · + n2 =
6
k=1
n
n(n + 1) 2
X  
k3 3 3 3
= 1 + 2 + ··· + n =
2
k=1

65
Bibliography

[1] H. Anton and A. Herr. Calculus with Analytic Geometry. John Wiley & Sons, Inc., 5th edition,
1995.

[2] F. Ayres. Schaum’s Outline of Theory and Problems of Differential and Integral Calculus.
McGraw Hill, 3rd edition, 1990.

[3] Richard Courant and Fritz John. Introduction to Calculus and Analysis, volume I. Springer,
1st edition, 1998.

[4] C.H. Edwards and D.E. Penney. Calculus with Analytic Geometry. Prentice-Hall, 6th edition,
2000.

[5] E. Kreyszig. Advanced Engineering Mathematics. John Wiley & Sons, 8th edition, 1999.

[6] J. E. Marsden and A. J. Tromba. Basic Multivariate Calculus. W. H. Freeman & Co., 5th
edition, 2003.

[7] G. B. Thomas and R. L. Finney. Thomas′ Calculus. Addison Wesley, 10th edition, 2000.

66