Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
CALCULUS
Term 1 2019
A set is a collection of distinct objects. The objects in a set are called the elements
or members of the set.
N = {0, 1, 2, 3, 4, . . .}.
−3 −2 −1 0 1 2 3
Notation. If x is a member of a set A, then we write x ∈ A. If x is not a member
of A then we write x ∈
/ A.
Example.
22 22 22 22
∈
/ N, ∈
/ Z, ∈ Q, ∈R
7 7 7 7
4
Notation for intervals. Suppose that a and b are real numbers and that a < b.
Then
• (a, b) = {x ∈ R : a < x < b} (open)
a b
• [a, b] = {x ∈ R : a ≤ x ≤ b} (closed)
a b
• [a, b) = {x ∈ R : a ≤ x < b} (???)
a b
• (a, b] = {x ∈ R : a < x ≤ b} (???)
a b
An interval [a, b] that includes its endpoints a and b is called a closed interval, while
an interval (a, b) that excludes its endpoints is called an open interval. The intervals
[a, b) and (a, b] are neither open nor closed.
5
Rays of the real line using the symbol ∞.
• [a, ∞) = {x ∈ R : a ≤ x}
|
a b
• (−∞, b) = {x ∈ R : x < b}
|
a b
• (−∞, ∞) = R
Definition. We say that a set A is a subset of a set B if every element of A is an
element of B. If A is a subset of B then we also say that B contains the set A.
Examples.
• N is a subset of Z, and Z is a subset of Q, and Q is a subset of R.
• {0, 2, 3} is a subset of {0, 1, 2, 3, 5}.
• (−1, 2] is not a subset of [0, ∞).
• Any set is a subset of itself.
• (1, 3) is a subset of [1, 3).
6
1.2 Solving inequalities
Two types of inequalities deserve special attention: quadratic inequalities and rational
inequalities.
7
Examples. (a) Solve the quadratic inequality
x2 + 4x ≥ 21.
8
1.3 Absolute values
The magnitude, or absolute value, of a real number x is denoted by |x| and defined
by (
x if x ≥ 0
|x| =
−x if x < 0.
9
The following facts are useful for solving inequalities.
• For every real number x,
√
|x| = x2 , |x|2 = x2.
• Geometrically, |x−a| represents the distance between x and a on the real number
line.
• For any positive real number a,
|x| < a ⇔ x2 < a2 ⇔ −a < x < a.
10
Examples. Solve the following inequalities.
(a)
|3x + 1| ≥ 4
(b)
|x + 5|
<1
|x − 11|
11
1.4 Functions
A function
f :A→B
is a rule which assigns to every element x belonging to a set A exactly one element
f (x) belonging to a set B, that is
x 7→ f (x).
Terminology.
• A is called the domain of the function f , that is
A = Dom(f ) = {all allowable inputs}.
• B is called the codomain of f , that is
B = Codom(f ) = {all allowable outputs}.
• The range of f is
Range(f ) = {f (x) : x ∈ A}
= {all outputs that actually occur}.
12
Example.
f : [0, ∞) → R
√
x 7→ 2 + x
• Functions which are defined by the same rule but have different domains are not
the same.
13
Example.
f : R → R, f (x) = x2
g : [0, ∞) → R, g(x) = x2
f (x) g(x)
x x
g is invertible, while f is not!
Natural domain. If, for whatever reason, the domain of a function is not defined
then we may choose the natural domain or maximal domain, that is the largest pos-
sible domain for which the rule makes sense (for real numbers).
14
Examples. Find the maximal domain for
(a) p
f (x) = x − 1/x.
(b)
1
f (x) = 2
.
x +x−2
Remark. We distinguish between the range and the codomain of a function since
we may not actually be able to determine the range of a function.
15
Combining functions. If f and g are two functions with the same domain, then
one can combine f and g to form new functions.
16
Another way of constructing new functions is given below.
Example. Suppose that the functions f and g are given by the rules
p
f (x) = ln x, g(x) = 4 − x2
What is the maximal domain of the composition f ◦ g?
17
1.5 Polynomials and rational functions
Rational functions. Suppose that p and q are polynomials. The function f defined
by the rule
p(x)
f (x) =
q(x)
with
Dom(f ) = {x ∈ R : q(x) 6= 0}
is called a rational function.
18
Previous example. f defined by
1
f (x) = 2 , Dom(f ) = R \ {−2, 1}
x +x−2
is rational.
19
1.6 The trigonometric functions
x 1
c
• sin2 x + cos2 x = 1.
21
Other trigonometric functions with suitable domains are defined by
sin x cos x
tan x = , cot x =
cos x sin x
1 1
sec x = , cosec x = .
cos x sin x
The six trigonometric functions are related by various identities and formulae (which
you are supposed to know):
• complementary identities
π
sin − x = cos x
2
π
cos − x = sin x
2
• Pythagorean identities
cos2 x + sin2 x = 1
1 + tan2 x = sec2 x
cot2 x + 1 = cosec2 x
22
• the sum and difference formulae
sin(x ± y) = sin x cos y ± cos x sin y
cos(x ± y) = cos x cos y ∓ sin x sin y
tan x ± tan y
tan(x ± y) =
1 ∓ tan x tan y
• double-angle formulae
sin(2x) = 2 sin x cos x
cos(2x) = cos2 x − sin2 x
2 tan x
tan(2x) = .
1 − tan2 x
23
1.7 The elementary functions
The elementary functions are all those functions that can be constructed by combining
(a finite number of) polynomials, exponentials, logarithms, roots and trigonometric
functions (including the inverse trigonometric functions) via function composition,
addition, subtraction, multiplication and division. Hence the following expressions
give rise to elementary functions:
f (x) = esin x + x2,
ln x − tan x
g(x) = √ ,
x
p3
h(x) = x4 − 2x2 + 5,
√
k(x) = |x| = x2.
It also follows that every rational function is an elementary function.
24
1.8 Implicitly defined functions
Many curves on the plane can be described as all those points (x, y) on the plane
that satisfy some equation involving x and y. For example, consider the equation
(x2 + y 2 − 1)3 − x2y 3 = 0. (♥)
The set of points (x, y) satisfying this equation are shown on the graph below.
y
Properties.
• There exist several y-values for some x-values. Hence, the curve cannot be the
graph of one function of x.
• The curve may be decomposed into two curves which may be regarded as the
graphs of two functions, f and g, say.
25
y y
b y = f (x) b y = g(x)
|
| | | |
−a a x −a a x
• We say that the functions f and g are implicitly defined by the relation (♥).
[In this case, can one determine formulae for these functions f and g?]
26
Chapter 2
Limits
27
f (x) g(x) h(x)
b
bc bc
b b
| | |
a x a x a x
28
‘Counterexample’. Consider the function
tan : A → R
with
A = Dom(f ) = {x ∈ R : x 6= π2 + nπ, n ∈ Z}.
tan x
| | | |
−π − π2 π π x
2
The function tan is continuous everywhere! The break in the graph is merely due to
the ‘missing’ points in the domain A.
29
2.1 Limits of functions at infinity
In this section, we examine some techniques for calculating limits of the form lim f (x)
1 x→∞
and postpone their precise definition. x2 ln x
1
x
Notation. x
• If f (x) gets ‘arbitrarily close’ to some real number L as x tends to infinity, then
we write
lim f (x) = L
x→∞
or
f (x) → L as x → ∞.
• If f (x) gets ‘arbitrarily large’ (that is, ‘approaches’ ∞) as x tends to ∞ then we
write
f (x) → ∞ as x → ∞.
• We never write lim f (x) = ∞ since ∞ is not a real number.
x→∞
30
2.1.1 Basic rules for limits
‘Elementary’ rules.
• If f is a constant function, that is f (x) = c for all x, then
lim f (x) = c.
x→∞
• If f (x) → ∞ as x → ∞ then
1
lim = 0.
x→∞ f (x)
31
Theorem. Suppose that
lim f (x) = a, lim g(x) = b
x→∞ x→∞
Proof. Later.
32
Example. Determine the limit of
2
1+
f (x) = 3x + 4
5 − 6e−x
as x → ∞.
33
2.1.2 The pinching theorem
Idea. Assume that two functions f and h have the same limit as x → ∞ and the
graph of a function g lies between the graphs of f and h (if x is large enough). Then,
g has the same limit as f and h.
y
y = h(x)
y = g(x)
L
b
|
y = f (x)
34
The pinching theorem. Suppose that f , g and h are three functions such that
f (x) ≤ g(x) ≤ h(x)
on an interval (b, ∞) for some b ∈ R and
lim f (x) = lim h(x) = L.
x→∞ x→∞
Then
lim g(x) = L.
x→∞
35
2.1.3 Limits of the form f (x)/g(x)
Suppose that we want to calculate a limit of the form
f (x)
lim ,
x→∞ g(x)
Problem. We cannot apply the preceding rules since f and g do not have limits.
Idea. Divide both f and g by the leading term, that is the fastest growing term
appearing in the denominator g (if it exists).
36
Examples. Find the following limits (if they exist).
(a)
5x3 + 6x2 − 4 sin x
lim
x→∞ cos 3x + 5x − 2x3
(b)
x2 + 3x
lim √
x→∞ 2x4 + 3 − 4x
37
p p
2.1.4 Limits of the form f (x) − g(x)
In this case, one needs to be very careful since a similar ‘leading term argument’ may
be erroneous!
In fact, it is the ‘lower order’ terms that may determine the limit!
as x → ∞.
Exercise. Does p √
lim x − x − x4 + 1
4 3
x→∞
exist?
38
2.1.5 Indeterminate forms
∞
The following limits have the form ” ∞ ” but each displays a very different limiting
behaviour as x → ∞:
x2
• →∞
x
x
• 2→0
x
2x2
• 2 →2
x
Since we cannot determine in advance what kind of limiting behaviour something of
∞
the form ” ∞ ” has, we say that ” ∞
∞ ” is an indeterminate form.
1
0 1 x
−1
40
• d(x) < 0.2 whenever x > 5.
y
y = f (x)
0.2
0 x
−0.2 5
ǫ
0 x
−ǫ M
41
Definition. Let f be a function defined on some interval (b, ∞) and let L be a real
number. We say that
lim f (x) = L
x→∞
if
for every ǫ > 0, there exists a real number M such that
if x > M then |f (x) − L| < ǫ.
y y = f (x)
L+ǫ
L
L−ǫ |
M x
42
2.3 Proving that lim f (x) = L using the limit definition
x→∞
This inequality gives an upper bound for the distance between f (x) and L!
Accordingly,
15
|f (x) − L| < ǫ whenever < ǫ.
x
The latter condition is equivalent to
15
x>
ǫ
and hence if we set
15
M=
ǫ
then
|f (x) − L| < ǫ whenever x > M,
as required.
44
General strategy. Given ǫ, we need to find a number M such that
|f (x) − L| < ǫ whenever x > M.
The number M can be found by following the procedure below.
1. Find a good upper bound for |f (x) − L|.
2. Find a simple condition on x such that this upper bound is less than ǫ.
3. Use this condition to state an appropriate value for M (in terms of ǫ).
45
Remarks.
• In general, M depends on ǫ but it is not uniquely defined.
• The definition of the limit does not tell you what the limit is. The definition may
be used to prove theorems which allow you to justify methods of finding limits.
• Applying the definition to verify an educated guess for a limit is usuallly the last
resort. Make use of the theorems unless you are specifically asked to apply the
definition.
46
2.4 Proofs of basic limit results
and set
ǫ
ǫ1 = ǫ2 =
2
for any given ǫ > 0.
and
|g(x) − L2| < ǫ2 whenever x > M2.
47
Hence, by the triangle inequality,
[f (x) + g(x)] − (L1 + L2) ≤ |f (x) − L1| + |g(x) − L2|
< ǫ1 + ǫ2
=ǫ
so that
lim [f (x) + g(x)] = L1 + L2.
x→∞
|
3 bc
|
2 bc
|
1 b b
|
| | | |
2 4 6 8 x
With reference to this graph, we will discuss the behaviour of f (x) when x is near
the points 2, 4, 6 and 8.
49
Notation.
• Let f be a function defined on an interval (c, a). We say that L is the left-hand
limit of f (x) as x approaches a if f (x) gets ‘closer and closer’ to L when x gets
‘closer and closer’ to a from the left. We write
lim f (x) = L.
x→a−
• Let f be a function defined on an interval (a, b). We say that L is the right-hand
limit of f (x) as x approaches a if f (x) gets ‘closer and closer’ to L when x gets
‘closer and closer’ to a from the right. We write
lim f (x) = L.
x→a+
50
• Let f be a function with the property
lim f (x) = lim f (x) = L
x→a− x→a+
for some point a ∈ Dom(f ). Then, we say that L is the (two-sided) limit of f (x)
as x approaches a and we write
lim f (x) = L.
x→a
51
• For a = 2:
lim f (x) = 3, lim f (x) = 1
x→2− x→2+
The two-sided limit does not exist.
• For a = 4:
f (x) → ±∞ as x → 4∓
No limit exists.
• For a = 6:
lim f (x) = lim f (x) = 2, f (6) = 4
x→6− x→6+
The two-sided limit exists but does not coincide with the value of f at a = 6.
• For a = 8:
lim f (x) = lim f (x) = f (8) = 1
x→8− x→8+
The two-sided limit exists and coincides with the value of f at a = 8.
52
2.4.2 Limits and continuous functions
Definition. Let f be defined on some open interval containing the point a. We say
that f is continuous at a if
lim f (x) = f (a);
x→a
otherwise we say that f is discontinuous at a.
Note that x = 4 is not part of the domain of f and hence asking whether or not f
is continuous at x = 4 does not make any sense.
Remark. A rigorous definition of the preceding limits exists but, here, we will focus
on rules for calculating limits.
53
2.4.3 Rules for limits at a point
Polynomials, rational functions, the trigonometric functions, exponentials and loga-
rithms are continuous at every point in the respective domains.
f (x) L1
• lim = provided that L2 6= 0.
x→a g(x) L2
54
... and interact nicely with function composition.
55
The pinching theorem. Suppose that f , g and h are all defined on an open interval
I containing the point a. If
and
lim f (x) = lim h(x) = L
x→a x→a
then
lim g(x) = L.
x→a
y = h(x)
L
a y = g(x)
|
y = f (x)
56
Example. Is the function f defined by
sin x if x 6= 0
f (x) = x
1 if x = 0
continuous?
57
58
Chapter 3
Question. If two continuous functions are combined via function addition, sub-
traction, multiplication, division or composition, when is the resulting function also
continuous?
Theorem. Suppose that the functions f and g are continuous at a point a. Then
f + g, f − g, f g
are continuous at a. If g(a) 6= 0 then
f /g
is also continuous at a.
59
Proof. Suppose that f and g are continuous at a. Then,
lim f (x) = f (a), lim g(x) = g(a)
x→a x→a
The proofs that the functions f − g, f g and f /g are continuous at a are similar.
60
Example. Based on the fact that constant functions and the function f : R → R
defined by f (x) = x are continuous at every point, show that polynomials and ratio-
nal functions are continuous at every point of their respective domains.
Solution. Any polynomial can be obtained from f and constant functions via addi-
tion and multiplication, e.g.
x3 − 4x2 + 5 = [(x × x × x)] + [(−4) × x × x] + 5,
and hence is continuous everywhere.
61
If we can prove that trigonometric functions are continuous at every point of their
domain then functions defined by rules such as
1 + tan x
f (x) =
1 + x2 + x6
are likewise continuous on the appropriate domain.
Lemma. Show that the functions sin and cos are continuous everywhere.
62
Interlude. Why is the function f : R → R defined by
f (x) = x
continuous everywhere?
63
Even larger classes of continuous functions may be obtained in the following manner:
Proof.
lim (g ◦ f )(x) = lim (g(f (x)) (def. of g ◦ f )
x→a x→a
64
Example. Suppose that the function g : R → R is continuous at the point a and
let f : R → R be the function defined by
f (x) = |g(x)|.
Show that f is continuous at the point a.
65
3.2 Continuity on intervals
66
Definition. Suppose that f is a real-valued function defined on a closed interval
[a, b]. We say that
• f is continuous at the endpoint a if
lim f (x) = f (a),
x→a+
√
Further generalisation. The function f : [0, ∞) → R defined by f (x) = x is
continuous on [0, ∞).
67
Example. Consider the functions f , g and h, whose graphs are shown below.
bc b b
| | | | | |
a b x a b x a b x
68
3.3 The intermediate value theorem
f (x)
f (b) b
|
z
|
f (a) b
|
| | |
a c b x
69
Proof. Omitted but based on the the least upper bound property of the real numbers:
(It is this property that distinguishes the real numbers from the rational numbers!)
70
Remarks.
• The number c in [a, b] may not be unique.
f (x)
f (a) b
|
z
f (b) b
|
| |
a c1 c2 c3 b x
71
• Continuity of f is crucial.
f (x)
f (b) b
|
b
|
bc
f (a) b
|
| |
a b x
72
• The intermediate value theorem is a theorem about real numbers. Consider, for
instance, the function
f : [0, 2] → R, f (x) = x2 − 2
so that, in particular, f (0) = −2 and f (2) = 2.
The intermediate value theorem says that there exists a real number c ∈ [0, 2]
such that f (c) = 0. In fact, √
c = 2 6∈ Q
and, hence, an analogous theorem for rational numbers cannot exist.
73
Application. Show that there exists a solution c ∈ [1, 2] of the equation
√
c = c2 − 1
and approximate its value.
74
3.4 The maximum-minimum theorem
An absolute maximum point and an absolute minimum point are sometimes referred
to as a global maximum point and a global minimum point.
75
Example. Consider the functions g and h, whose graphs are illustrated below.
g(x) h(x)
4 4
|
3 3
|
2 2
|
|
b
1 |
1 b
|
| | | | | | | | | |
1 2 3 4 5 x 1 2 3 4 5 x
The absolute minimum and maximum points of g and h on [1, 5] are recorded in the
following table.
g h
Absolute minimum points none 3
Absolute minimum value n.a. 1
Absolute maximum points 2, 4 none
Absolute maximum value 4 n.a.
76
The above example shows that a function f : [a, b] → R need not have an absolute
maximum point (or an absolute minimum point) on a closed interval [a, b].
Proof. Omitted but, once again, the least upper bound property of the real numbers
is used. The proof relies on the crucial fact that f is continuous on the closed interval
[a, b]!
77
Example. Suppose that f : R → R is continuous and
lim f (x) = lim f (x) = 0
x→−∞ x→∞
78
Definition. A function f is said to be bounded on an interval I if there exists some
positive number M such that
|f (x)| ≤ M for all x ∈ I.
In other words, f is bounded if the y-values of its graph lie between −M and M for
some positive number M .
79
Remark. A function may be bounded without having a maximum or minimum value.
80
Chapter 4
Differentiable functions
9
Typical question. A snowboarder acceler-
ates from rest in such a way that her displace-
ment s (in metres) from her starting position
after t seconds is given by displacement (m)
4 b
|
s = t2 ∀t ∈ [0, 3].
What is the snowboarder’s ‘instantaneous’
speed when t = 1? 1 b
|
| |
0 1 2 3
time (s)
81
‘Intuitive solution.’ Calculate the average velocity v of the snowboarder over any
time interval according to
∆s
v= ,
∆t
where ∆s denotes the change in displacement corresponding to a change ∆t in time ...
Thus, calculate the gradient of the secant intersecting the graph at (1, 1) and some
other point P ...
s = t2
secant
2.25 P b
|
s
∆s (1.5)2 −(1)2
grad. of secant = ∆t = 1.5−1 = 2.5
1 b
|
| |
1 1.5
t
83
Intersection points ∆t Gradient of secant
t = 1, t = 1.5 0.5 2.5
t = 1, t = 1.4 0.4 2.4
t = 1, t = 1.3 0.3 2.3
t = 1, t = 1.2 0.2 2.2
t = 1, t = 1.1 0.1 2.1
84
s = t2
secants
tangent
2.25 b
|
1.96 b
s
1.69 b
|
1.44 b
|
1.21 b
|
1 b
|
| | | | | |
Example. Find the gradient of the tangent to the graph of the function f : R → R
defined by f (x) = x3.
y
y = f (x)
secant
(x + h)3 − x3 (x + h, (x + h)3) b tangent
(grad. at x) = lim
h→0 h
3x2h + 3xh2 + h3
= lim
h→0 h
x3 (x, x3) b
h→0
2
= 3x .
| |
x x+h
86
Definition. Suppose that f is defined on some open interval containing the point x.
If
f (x + h) − f (x)
lim
h→0 h
exists then it is called the derivative of f at x and we say that f is differentiable at x.
87
Example. Show that
d
cos x = − sin x.
dx
Exercise. Write
cos2 h − 1 sin2 h
cos h − 1 = =−
cos h + 1 cos h + 1
and conclude that
cos h − 1
lim = 0.
h→0 h
88
Example. For which n ∈ N is the function f defined by
xn sin 1
if x 6= 0
f (x) = x
0 if x = 0
continuous or differentiable at x = 0?
90
Theorem. Suppose that g is differentiable at the point x and f is differentiable at
the point g(x). Then,
(f ◦ g)′(x) = f ′(g(x))g ′(x) (chain rule).
91
4.3 Proofs of results in Section 4.2
Proof of product rule. Suppose that f and g are differentiable at the point x.
The difference quotient of f g at x gives
(f g)(x + h) − (f g)(x)
h
f (x + h)g(x + h) − f (x)g(x)
=
h
f (x + h) − f (x) g(x + h) − g(x)
= g(x + h) + f (x) .
h h
We will show later that if a function is differentiable at a point then it must be
continuous at that point! Hence,
g(x + h) → g(x) as h → 0.
Accordingly,
(f g)(x + h) − (f g)(x)
→ g(x)f ′(x) + f (x)g ′(x)
h
as h → 0.
92
Proofs of the other differentiation rules are found in most undergraduate calculus
textbooks.
93
4.4 Implicit differentiation
Idea. On use of the chain rule, determine the derivative of a function which is
implicitly defined.
Example. Determine the tangent line to the curve defined by
x4 − x2y 2 + y 4 = 13
at the point (2, 1).
Write
x4 − x2y(x)2 + y(x)4 = 13
and differentiate both sides:
4x3 − 2xy(x)2 − 2x2y(x)y ′(x) + 4y(x)3y ′(x) = 0.
Evaluate at (x, y) = (2, 1) and solve for y ′(2):
32 − 4 − 8y ′(2) + 4y ′(2) = 0 ⇒ y ′(2) = 7.
Equation of tangent line at (2, 1):
y − 1 = 7(x − 2).
94
Theorem. Suppose that q is a rational number. Then
d q
x = qxq−1.
dx
Proof. Since q is a rational number, there exist integers m and n such that
m
q= .
n
Write
y = xq = xm/n
and take the nth power of both sides, leading to
y n = xm .
96
4.5 Differentiation, continuity and split functions
= f ′(a) × 0 = 0
since f is differentiable at a. Accordingly,
lim f (a + h) = f (a).
h→0
97
Previous example. At x = 0, the function f defined by
xn sin 1
if x =
6 0
f (x) = x
0 if x = 0
is continuous for n ≥ 1 and differentiable for n ≥ 2.
Extra care is required if one deals with ‘split functions’. Whether or not the function is
differentiable at the ‘split point’ can be determined by calculating left- and right-hand
limits of the difference quotient. ...
98
Example. Suppose that f : (0, ∞) → R is defined by
( √
4 x if 0 < x ≤ 1
f (x) =
bx2 + c if x > 1,
where b and c are real numbers. Find all possible values of b and c such that f is (i)
continuous at x = 1 and (ii) differentiable at x = 1.
... However, in many cases, this may be avoided by using the following theorem:
99
Theorem. Suppose that a is a fixed real number and that a function f is the rule
(
p(x) if x ≥ a
f (x) =
q(x) if x < a,
where p and q are defined on some open interval containing a. If f is continuous at
a and p′(a) = q ′(a) then f is differentiable at x = a.
100
4.6 Derivatives and function approximation
101
Example. Suppose that the horizontal distance L covered by a tennis ball is given
by
v2
L = sin 2ϕ,
g
where v and ϕ are the initial speed and angle respectively and g is the constant
vertical acceleration due to gravity. How does L change if either v or ϕ are changed
by a ‘small’ amount?
102
4.7 Derivatives and rates of change
Example. A spherical balloon is being inflated and its radius is increasing at a con-
stant rate of 6 mm/sec. At what rate is its volume increasing when the radius of the
balloon is 20 mm?
Let V (t) be the volume of the balloon and r(t) be its radius at time t. Alternatively,
let Ṽ (r) be the volume of the balloon as a function of its radius r given by
4
Ṽ = πr3
3
so that
dṼ
= 4πr2.
dr
103
Then, the chain rule implies that
dV dṼ dr 2 dr
= = 4πr .
dt dr dt dt
Hence, the volume is increasing at a rate of 9600π mm3/sec when the radius is 20 mm.
104
The above example illustrates an approach to solving such problems.
1. Define variables for the quantities involved.
2. Write down what is known in terms of these variables and their derivatives.
3. Write down what you need to find in terms of these variables and their derivatives.
4. Write down anything else you know that relates the variables (for example, a
volume or area formula).
5. Use the chain rule (or implicit differentiation) to find the relevant derivative.
105
4.8 Local maximum, local minimum and stationary points
| |
c
|
x
c−h c+h
Definition. Let f be defined on some interval I.
• We say that a point c in I is a local minimum point if there exists an h > 0 such
that
f (c) ≤ f (x) for all x ∈ (c − h, c + h) ∩ I.
• We say that a point d in I is a local maximum point if there exists an h > 0 such
that
f (x) ≤ f (d) for all x ∈ (d − h, d + h) ∩ I.
106
Theorem. Suppose that f is defined on (a, b) and has a local maximum or minimum
point at c for some c in (a, b). If f is differentiable at c then f ′(c) = 0.
Example. Find all the stationary, maximum and minimum points of the function
f : [−3, 4] → R defined by
f (x) = 4x5 − 5x4 − 40x3 − 2.
f (x)
200
|
| |
−2 3 x
−200
|
107
Differentiation yields
f ′(x) = 20x4 − 20x3 − 120x2
= 20x2(x2 − x − 6)
= 20x2(x + 2)(x − 3).
Result:
• x = −3: local minimum point
• x = −2: stationary point: local maximum point
• x = 0: stationary point: point of inflection
• x = 3: stationary point: global minimum point
• x = 4: global maximum point
The proof of the above theorem may be found in the Calculus Notes.
108
Chapter 5
f (x)
C
f (b) B
|
f (a) A
|
| | |
a c b x
The mean value theorem. Suppose that a function f is continuous on [a, b] and
differentiable on (a, b). Then, there exists at least one real number c in (a, b) such
that
f (b) − f (a)
= f ′(c).
b−a
110
Remark. In the above theorem, it is required that f is continuous on the closed
interval but differentiable only on the open interval!
Example. Find counterexamples which demonstrate that the continuity and differ-
entiability requirements must be met.
Application. A car enters a tunnel at a speed of 30km/h and after 1 minute leaves
the tunnel at a speed of 40 km/h. The length of the tunnel is 1km. Did the driver
break the speed limit of 50km/h?
111
Example. Apply the mean value theorem to the function f defined by f (x) = x5
and a = −1, b = 4. Find the value(s) of c which satisfy the conclusion of the theorem.
112
5.2 Proof of the mean value theorem
If f (a) = f (b) then the mean value theorem reduces to Rolle’s theorem.
g(x)
C
A B
a c b x
113
Proof of Rolle’s theorem.
Moreover,
g(c) ≥ g(d) > g(a) = g(b) = 0
so that c must lie in (a, b).
114
Case 3: Suppose that
g(x) ≤ 0
for all x in [a, b] and that g is not constant on [a, b].
In order to prove the mean value theorem, we merely ‘deform’ the graph of f in such
a way that Rolle’s theorem applies:
115
Proof of the mean value theorem. Suppose that f is continuous on [a, b] and
differentiable on (a, b).
By virtue of Rolle’s theorem, there exists a c in (a, b) such that g ′(c) = 0, that is,
such that
f (b) − f (a)
f ′(c) − = 0.
b−a
Hence,
′ f (b) − f (a)
f (c) =
b−a
as required.
116
5.3 Proving inequalities using the mean value theorem
117
Example. It is known that any polynomial ‘grows faster than’ the logarithm. For,
instance, show that
ln x < x − 1 for all x > 1.
Suppose that x > 1 and consider the closed interval [1, x]. We define a function
f : [1, x] → R
by
f (t) = ln t.
Now, f is continuous on [1, x] and differentiable on (1, x) so that the mean value
theorem implies that
f (x) − f (1)
= f ′(c)
x−1
for some c in (1, x). Accordingly,
ln x 1
=
x−1 c
for some c between 1 and x.
118
An upper bound of f ′(c) on (1, x) is given by
1
f ′(c) = <1
c
since c > 1. We therefore conclude that
ln x
<1
x−1
or, equivalently,
ln x < x − 1.
119
5.4 Error bounds
120
Accordingly,
π π
π
0 < cos 1 − cos =
3 − 1 sin c ≤ − 1 < 0.05
3 3
since π < 3.15 so that
0.5 < cos 1 < 0.55.
121
5.5 The sign of a derivative
decreasing increasing
x
122
Theorem. Suppose that f is continuous on [a, b] and differentiable on (a, b).
• If f ′(x) > 0 for all x in (a, b) then f is increasing on [a, b].
• If f ′(x) < 0 for all x in (a, b) then f is decreasing on [a, b].
• If f ′(x) = 0 for all x in (a, b) then f is constant on [a, b].
Proof. Suppose that f ′(x) > 0 for all x in (a, b) and choose two points x1 and x2
in [a, b] such that x1 < x2.
Example. Find and classify all stationary points of the function f : R → R whose
derivative is given by
f ′(x) = (x − 4)(x − 1)(x + 5)2.
124
Result:
• x = 4: Local minimum point
• x = 1: local maximum point
• x = −5: horizontal point of inflexion
125
5.6 The second derivative and applications
Another (potential) method for classifying the stationary points of a function f in-
volves the second derivative of f , which is denoted by
′′ d2 y
f or 2
, or y ′′
dx
if we set y = f (x).
126
The proof of this theorem uses the following result.
Exercise. Find and classify the stationary points of the function f : R → R given
by
f (x) = x3 − 6x2 + 9x − 5.
127
Remark. If f ′(c) = f ′′(c) = 0, no conclusion may be drawn!
x4 −x4 x3
x x x
128
5.7 Critical points, maxima and minima
Definition. Suppose that f is defined on [a, b]. We say that a point c in [a, b] is a
critical point for f on [a, b] if c satisfies one of the following properties:
• c is an endpoint a or b of the interval [a, b]
• f is not differentiable at c
• f is differentiable at c and f ′(c) = 0.
f (x)
b
b
4 b
b
| |
3 4 5 x
2 129
Theorem. Suppose that f is continuous on [a, b]. Then, f has a global maximum
and global minimum on [a, b]. Moreover, the global maximum point and the global
minimum point are both critical points for f on [a, b].
130
5.8 Counting zeros
Differentiation yields
f ′(x) = (x − 2)(4x2 + 5x + 4)
so that f ′(2) = 0 and
f ′(x) < 0 on (−∞, 2) and f ′(x) > 0 on (2, ∞).
Now, since f (2) = −25, we can roughly sketch the graph (exercise!) and conclude
the following:
• f is decreasing on (−∞, 2) and therefore cannot have more than one zero on this
interval.
But f (−1) = 2 and f (0) = −5 and, hence, the intermediate value theorem
implies that f has at least one zero on (−1, 0).
131
• f is inreasing on (2, ∞) and therefore cannot have more than one zero on this
interval.
But f (3) = −2 and f (4) = 107 and, hence, the intermediate value theorem
implies that f has at least one zero on (3, 4).
Result: f has two real zeros, one in the interval (−1, 0) and one in the interval (3, 4).
132
5.9 Antiderivatives
133
Example. Let
2
fα(x) = xαex , α ≥ 0.
Show that if Fα−2 is an antiderivative of fα−2 for some α ≥ 2 then Fα defined by
xα−1 x2 α − 1
Fα(x) = e − Fα−2(x)
2 2
is an antiderivative of fα.
134
Theorem. Suppose that f is a continuous function on an open interval I and that
F and G are two antiderivatives of f on I. Then, there exists a real constant C such
that
G(x) = F (x) + C
for all x in I.
Function Antiderivative
1
xr , where r is rational and r 6= −1 r+1 x
r+1
+C
sin x − cos x + C
cos x sin x + C
1 ax
eax ae +C
f ′(x)
ln |f (x)| + C
f (x)
136
5.10 L’Hôpital’s rule
Theorem (l’Hôpital’s rule). Suppose that f and g are both differentiable functions
in a neighbourhood of some a ∈ R and that either one of the two following conditions
hold:
• f (x) → 0 and g(x) → 0 as x → a
• f (x) → ∞ and g(x) → ∞ as x → a.
If
f ′(x)
lim
x→a g ′ (x)
exists then
f (x) f ′(x)
lim = lim ′ .
x→a g(x) x→a g (x)
137
Remark. The theorem also holds for
• limits as x → ∞ or x → −∞
• one-sided limits (as x → a+ or x → a−).
138
Remark. l’Hôpital’s rule may be applied iteratively.
Example.
sin x − x cos x x sin x
lim = lim = ... see above.
x→0 x − sin x x→0 1 − cos x
Remark. It is important that the limit exists after a finite number of applications of
l’Hôpital’s rule!
139
140
Chapter 6
Inverse functions
Problem. Given a function
f : A → B,
if we set
y = f (x),
under what circumstances is it possible to express x as a function of y, that is, to
find a function
g:C→A
such that
x = g(y) ?
Remark. The above process of expressing the ‘input’ of a function in terms of its
‘output’ is commonly referred to as inverting a function.
141
6.1 Some preliminary examples
(b) p
f : [−1, 1] → R, y = f (x) = 1 − x2
142
Question. How do we make sure that a function is invertible and what is the domain
of any inverse function?
Remark. It is evident that the domain and range of a function play a crucial role in
answering this question! For example, formal inversion of c = c(t) given by
c = cme−κ(t−td)
leads to
1 cm
t = td + ln .
κ c
143
Standard example. Consider the rule
y = x2 .
Whether any function defined by this rule is invertible depends on the domain:
• f1 : [0, ∞) → R, y = f1(x) = x2
If we take into account that Range(f1) = [0, ∞) then the inverse function is given
by
√
g1 : [0, ∞) → [0, ∞), x = g1(y) = y.
• f2 : (−∞, 0] → R, y = f2(x) = x2
If we take into account that, again, Range(f2) = [0, ∞) then the inverse function
is given by
√
g2 : [0, ∞) → (−∞, 0], x = g2(y) = − y.
144
• f3 : R → R, y = f3(x) = x2
The latter is not invertible since for any ‘output’ y 6= 0 there exist two ‘inputs’
√ √
x = y and x = − y.
output output
146
Example. For which α ∈ R is the function
f : R → R, f (x) = x3 + αx + 1
one-to-one?
147
The horizontal line test. Suppose that f is a real-valued function defined on some
subset of R. Then, f is one-to-one if and only if every horizontal line in the Cartesian
plane intersects the graph of f at most once.
Examples.
x x x
• f1 is not one-to-one.
• f2 is one-to-one.
• f3 is one-to-one (even though it is neither strictly increasing nor strictly decreas-
ing).
148
Although not every one-to-one function is strictly increasing (or strictly decreasing),
it is true that every strictly increasing function is one-to-one.
Since
f ′(x) = 3x2 + α > 0
for α > 0, f is strictly increasing and is therefore one-to-one for α > 0.
149
Example. Is the function f : (−1, 1) → R defined by
π
2
f (x) = 3 + x + 2 tan x
2
one-to-one?
Remark. Not every function whose derivative is only positive (or only negative) on
its domain is one-to-one. For example,
d
tan x = sec2 x ≥ 1
dx
but tan is not one-to-one on its maximal domain!
150
6.3 Inverse functions
It is then left as an exercise to show that g has the properties listed above.
151
The theorem allows us to define the term inverse function.
152
Remark. Since f −1 is a function just like any other function, we regard it as a
function
x 7→ f −1(x)
so that we can graph f −1 in the usual manner.
Set
1
y = 4 − x3
3
so that p
x = 3 12 − 3y.
153
Hence, (interchanging x and y),
−1 −1
√
3
f : R → R, f (x) = 12 − 3x.
y y=x
y = f −1(x)
y = f (x)
154
6.4 The inverse function theorem
156
Proof.
• Since f ′(x) 6= 0 on I, f is one-to-one (mean value theorem!).
• g is differentiable ... too hard!
• Differentiation of
f (g(y)) = y
with respect to y yields
f ′(g(y)) × g ′(y) = 1.
Remark. Once again, we usually write the derivative of the inverse function g as
1
g ′(x) = ′
f (g(x))
for x ∈ Range(f ).
157
Example. Consider the function
f : (0, ∞) → (0, ∞), f (x) = x3
and its inverse √
3
g : (0, ∞) → (0, ∞), g(x) = x.
Then,
′ 1 1 1
g (x) = ′ = =
f (g(x)) 3[g(x)]2 3x2/3
as expected.
158
6.5 Applications to the trigonometric functions
b
x x
(− π2 , −1) b
(−1, − π2 )
159
On (−1, 1), according to the inverse function theorem, the derivative of sin−1 is given
by
d −1 1
(sin x) = −1 .
dx cos(sin x)
Since cos is positive on (− π2 , π2 ), we conclude that
d −1 1 1
(sin x) = p =√ .
dx 2 −1
1 − sin (sin x) 1−x 2
d
Note. (sin−1 x) → ∞ as x → ±1.
dx
160
Table of inverse trigonometric functions.
161
(−1, π) b
cos−1 x tan−1 x
π
b
(0, π2 ) 2 b
(1, 0) x x
b
− π2
162
Example. Determine
(a)
−1 3
cos 2 sin .
5
(b)
5π
sin−1 sin .
6
163
164
Chapter 7
Curve sketching
y y
x
x
165
In this chapter, we study curves in a plane which are given in terms of
• a parameter such as
p
x(t) = sin t cos t ln |t|, y(t) = |t| cos t
t ∈ [−1, 1], t 6= 0.
• polar coordinates such as
x = r cos θ, y = r sin θ, where r = cos 4θ.
166
7.1 Curves defined by a Cartesian equation
In this section, we survey techniques for sketching curves which are described by a
Cartesian equation of the form y = f (x).
167
• Find x- and y-axis intercepts.
• Identify vertical asymptotes.
• Examine the behaviour of f (x) as x → ±∞ and identify existing (oblique)
asymptotes.
• If necessary, identify stationary points and other features using calculus.
168
Symmetries. Consider the function f : R → R defined by
| sin x|
f (x) = .
2 + cos(2x)
• f is even since f (−x) = f (x) for all x ∈ R.
• f is of period π since f (x + π) = f (x) for all x ∈ R.
169
Oblique asymptotes. The graph of a function f may be asymptotic to a line
y = ax + b as x → ∞ or x → −∞.
y y = ax + b
y = f (x)
x
Definition. Suppose that a 6= 0 and b are real numbers. We say that a straight line
given by the equation
y = ax + b
is an oblique asymptote for a function f if
lim f (x) − (ax + b) = 0
x→∞
or
lim f (x) − (ax + b) = 0.
x→−∞
170
Remark. If f is a rational function with
p(x)
f (x) = , deg(p) = deg(q) + 1
q(x)
then the oblique asymptotes of f may be determined by polynomial division.
If x > 0 then
(x − 2)x + x
f (x) =
x−2
(x − 2) + 2
=x+
x−2
2
=x+1+ .
x−2
171
Thus, the line y = x + 1 is an oblique asymptote.
If x < 0 then
−(x − 2)x + x
f (x) =
x−2
(x − 2) + 2
= −x +
x−2
2
= −x + 1 + .
x−2
Thus, the line y = −x + 1 is an oblique asymptote.
172
Example. Sketch the function f defined by
x2 − 3
f (x) =
2x − 4
on the maximal domain.
173
7.2 Parametrically defined curves
where t is the parameter and A is a given domain arise in various contexts such as
• the planar motion of a point particle (or body) under the influence of forces.
Newton’s 2nd law: m ẍ(t), ÿ(t) = F x(t), y(t)
It is evident that x(t) and y(t) assume any value in the ‘intervals’
x(t) ∈ [−1, ∞), y(t) ∈ (−∞, ∞)
respectively.
Intercepts:
175
Vector derivatives:
176
Here,
dy y ′(t) 3
γ ′(t) = (2t, 3t2), = = t
dx x′(t) 2
so that
γ ′(t) points տ for t<0
and
γ ′(t) points ր for t > 0.
In fact, there exists a cusp at γ(0) = (−1, −1) since the normalised tangent vector
has the property
γ ′(t) (2t, 3t2)
lim ′ = lim √ = (±1, 0).
t→0 |γ (t)|
± t→0 ± 2
4t + 9t 4
Conclusion: The curve does not have a ‘proper’ tangent vector at the point (−1, −1)!
177
Limiting behaviour:
As t → ±∞,
γ ′(t) (2t, 3t2)
lim ′
= lim √ = (0, 1).
t→±∞ |γ (t)| t→±∞ 2
4t + 9t 4
178
Justification of slope formula: By definition,
dy y(t + h) − y(t)
= lim
dx h→0 x(t + h) − x(t)
.
y(t + h) − y(t) x(t + h) − x(t)
= lim
h→0 h h
y(t + h) − y(t) . x(t + h) − x(t)
= lim lim
h→0 h h→0 h
y ′(t)
= ′
x (t)
provided that x(t) and y(t) are differentiable and x′(t) 6= 0.
179
7.2.1 Parametrisation of conic sections
The ellipse
x2 y 2
2
+ 2 =1
a b
with semi-axes a and b admits the parametrisation
x(t) = a cos t, y(t) = b sin t, 0 ≤ t < 2π.
y(t) π
π
t= 2 π
<t<π b
0<t<
2 b 2
t=π t=0
a x(t)
b b
−a
π<t< 3π −b b
3π
< t < 2π
2 3π 2
t= 2
y(t) = at2
y(t) = a sin t
x2 y 2
Ellipse + =1 x(t) = a cos t
a2 b2
y(t) = b sin t
x2 y 2
Hyperbola − =1 x(t) = a sec t
a2 b2
y(t) = b tan t
Note that there exist other useful parametrisations for each of these curves.
181
7.2.2 The cycloid and curve of fastest descent
Question. Find the shape of a curve that a particle should follow if it is to ‘slide’
without friction in the minimum time from a higher point A to a lower point B (not
directly beneath it) under the influence of gravity.
A b
b
B
182
Answer. The curve of fastest descent from A to B is the unique arc of an (inverted)
cycloid whose tangent at A is vertical.
y
P
P
P
P x
The cycloid.
Exercise. A circle of radius r rolls along the x-axis, starting from the origin as shown
above. Show that the locus (x(t), y(t)) of the point P on the edge of the circle which
satisfies (x(0), y(0)) = (0, 0) is given by
x(t) = r(t − sin t)
y(t) = r(1 − cos t),
where t ≥ 0.
Remark. If the particle moves on an inclined plane (‘surfer’) then the trajectory is
still given by an arc of a cycloid!
183
7.3 Curves defined by polar coordinates
Many problems in mathematics are easier to solve if one chooses a suitable coordinate
system. Here, we focus on polar coordinates.
x b
P b
P
y r
θ
O O
184
The pair (x, y) is called Cartesian coordinates of P .
Polar coordinates (r, θ) and Cartesian coordinates (x, y) of a point P are related by
x = r cos θ, y = r sin θ
and p y
r= x2 + y 2, tan θ =
x
provided that x 6= 0.
185
Example. Find all polar coordinates of the point P with Cartesian coordinates
√
P = (x, y) = (−3, 3).
186
7.3.2 Basic sketches of polar curves
Many curves can be described by equations of the form
r = f (θ) or θ = g(r)
so that we obtain the parametrically defined curves
γ(θ) = (r cos θ, r sin θ) = f (θ) cos θ, f (θ) sin θ
or
γ(r) = (r cos θ, r sin θ) = r cos g(r), r sin g(r) ,
where θ or r plays the role of the parameter.
Previous examples.
π
√
(a) θ = 3 corresponds to y = 3 x for x ≥ 0.
(b) r = 2 corresponds to x2 + y 2 = 4.
187
Remark. In order to sketch a curve represented by an equation in polar form, it may
be helpful to begin with an r vs θ sketch.
r
e
3 b
|
1
2 + √2 b b
|
d f
r = 2 − cos θ
2 b b
|
c g
2 − √12 b b
|
b h
1 b
a i
b
| | | | | | | |
π π 3π π 5π 3π 7π 2π θ
4 2 4 4 2 4
188
... and then mark the corresponding points on the (x, y)-plane:
y θ= π
2
3π π
θ= 4 θ= 4
b c′ r = 2
b
′
d
b′ b
θ=π b
r=3 a′ r = 1b
θ=0
e ′ i′ x
b
h′
b
f′ g′ r = 2
b
5π 7π
θ= 4 θ= 4
3π
θ= 2
189
Now:
• As θ increases from 0 to π, r increases from 1 to 3. Hence, the distance r from
the origin to points on the curve increases from 1 to 3.
• As θ increases from π to 2π, r decreases from 3 to 1. Hence, the distance r from
the origin to points on the curve decreases from 3 to 1.
−3 b b
1 x
b
b
b
θ= 5π −2 θ= 7π
r = 2 − cos θ 4 4
190
Symmetries.
• If f (−θ) = f (θ) then the polar curve is symmetric about the x-axis.
• If f (π − θ) = f (θ) then the polar curve is symmetric about the y-axis.
• If f is 2π–periodic then it suffices to consider θ in the range 0 ≤ θ < 2π.
191
7.3.3 Sketching polar curves using calculus
Suppose that a curve can be expressed in polar form as
r = f (θ).
Since the curve’s parametric form is given by
γ(θ) = x(θ), y(θ) = (r cos θ, r sin θ),
the tangent vector reads
′ ′ ′
γ (θ) = x (θ), y (θ)
dr dr
= cos θ − r sin θ, sin θ + r cos θ .
dθ dθ
193
Concluding remark. In connection with representing and sketching curves in terms
of polar coordinates, it is often convenient to allow negative values of r. Thus, if r
is negative, we make the identification
(r, θ) ↔ (−r, θ + π)
and hence (r, θ) is obtained from (−r, θ) by reflection in the origin. This is consistent
with the formulae
x = r cos θ = (−r) cos(θ + π)
y = r sin θ = (−r) sin(θ + π).
194
Chapter 8
Integration
Problem. How does one find/measure/define areas (of regions with curved bound-
aries)?
Note. A priori, calculating areas (integration) and antidifferentiation are two sepa-
rate problems but the remarkable fundamental theorem of calculus shows that these
are essentially the same!
This theorem was actually known to Barrow but its implications were developed by
Newton, Leibniz and their disciples.
Remark. The problem of calculating area is much the same as that of calculating
mass, volume, work and probability. The unifying feature is the so-called integral
calculus.
Here, we confine ourselves to the determination of ‘the area under the graph of a
function’ via the Riemann integral. This may then be generalised to the area of sets
the boundary of which cannot be represented in this simple manner.
196
8.1 Area and the Riemann integral
We all (think that we) know how to calculate the area of a rectangle, a triangle or
even general polygons by partitioning the polygon into triangles.
Thus, from the area of a rectangle (bh), we can ‘derive’ the formula for
• the area of a right-angled triangle ( 21 bh), then
• the area of an arbitrary triangle ( 21 (b1 + b2)h) and then
• the area of an arbitrary polygon.
h h
b b1 b2
197
Formally, we demand that any definition of an area satisfy the following axioms:
198
(A4) If Ω1 and Ω2 are congruent regions then
area(Ω1) = area(Ω2).
Ω2
Ω1
(A5) If Ω is a rectangle of length a and height b then
area(Ω) = ab. Ω
b
a
199
8.1.1 Area of regions with curved boundaries
Question. How could one calculate the area of the region below?
Ω1
Ω Ω3 Ω2
Then, rotate and translate each subregion, so that application of axiom (A4) implies
that each of area(Ω1), area(Ω2) and area(Ω3) is equal to the area under the graph of
a function.
y f1 y y
f2 f3
x x x
200
This procedure can be done for any region in the plane with a ‘reasonable’ boundary.
Hence, we are left with defining what is meant by ‘the area under the graph of a
function’.
201
8.1.2 Approximations of area using Riemann sums
Example. Consider the function f : R → R given by the rule
f (x) = x2.
Let Ω denote the region bounded by the graph of f , the x-axis and the lines x = 0
and x = 1.
Ω Ω1 Ω2
1x 1x 1x
Idea. Find lower and upper bounds for area(Ω) by choosing appropriate ‘approxima-
tions’ Ω1 and Ω2 of the region Ω in terms of n rectangles.
202
It is evident that
area(Ω1) ≤ area(Ω) ≤ area(Ω2).
Remark. One of Archimedes’ best ideas was to estimate areas in circles and parabo-
las by approximating them by ‘inner’ and ‘outer’ polygons!
203
y y = f (x)
Rk
··· ···
1 2 3 4 · · ·k−1 k · · ·n−1 n x
n n n n n n n n
which divides the interval [0, 1] into these subintervals is called a partition of [0, 1].
Let Rk denote the area of the kth rectangle. Then
1 k k2
Rk = width × height = × f ( n ) = 3 .
n n
204
If S Pn (f ) denotes the total area of the shaded region then
n n
X 1 X 2
S Pn (f ) = Rk = 3 k.
n
k=1 k=1
The quantity S Pn (f ) is called the upper Riemann sum of f with respect to the par-
tition Pn.
y y = f (x)
Rk
··· ···
1 2 3 4 · · ·k−1 k · · ·n−1 n x
n n n n n n n n
so that
1 1 1
S Pn (f ) = − + 2.
3 2n 6n
207
Hence, for every positive integer n, the inequality
1 1 1 1 1 1
− + ≤A≤ + +
3 2n 6n2 3 2n 6n2
gives an upper and a lower bound for A.
Remark. The process of calculating upper and lower Riemann sums and taking a
limit of the above type (provided it exists) is called integration.
208
8.1.3 The definition of area under the graph of a function and the Riemann
integral
Suppose that f is a bounded function on [a, b] and that f (x) ≥ 0 for all x in [a, b].
209
Suppose that P is a partition of [a, b]:
f (x)
··· ···
The upper Riemann sum S P (f ) for f with respect to the partition P is defined by
n
X
S P (f ) = (ak − ak−1)f k
k=1
Likewise, the lower Riemann sum S P (f ) for f with respect to the partition P is
defined by
n
X
S P (f ) = (ak − ak−1)f k , (8.1)
k=1
where
f k = the minimum value of f on the subinterval [ak , ak−1]
or the greatest lower bound (MATH1241).
211
Definition. Suppose that a function f is bounded on [a, b] and that f (x) ≥ 0 for
all x in [a, b]. If there exists a unique real number A such that
S P (f ) ≤ A ≤ S P (f )
for every partition P of [a, b] then we say that A is the area under the graph of f
from a to b.
In general, if we remove the condition f (x) ≥ 0 then the following definition is still
sensible:
212
Definition. Suppose that a function f is bounded on [a, b]. If there exists a unique
real number I such that
S P (f ) ≤ I ≤ S P (f )
for every partition P of [a, b] then we say that f is Riemann integrable on the interval
[a, b].
The unique real number I is called the definite integral of f from a to b and we write
Z b
I= f (x) dx.
a
Remark. The function f is called the integrand of the definite integral, while the
points a and b are called the limits of the definite integral.
Exercise. Ponder about why this definition of area is consistent with the axioms
(A1)-(A5).
213
Historical remark. The notation
Z b
f (x) dx
a
is due to Leibniz. It evolved from a slightly different way of writing down lower and
upper Riemann sums. For example, S P (f ) may be written as
n
X
S P (f ) = f (xk )∆xk ,
k=1
where ∆xk = ak −ak−1 and f attains its maximum value on [ak−1, ak ] at the point xk .
P
When taking a limit as before, ∆xk was replaced with dx and the symbol was
replaced with an elongated ‘S’ (‘S’ stands for ‘sum’).
214
8.2 Integration with Riemann sums
Even in the case f (x) = C, where C is a positive constant, the actual determination
of the Riemann integral is somewhat tedious.
Question. Can we find simple sufficient conditions which guarantee that a Riemann
integral exists?
216
Definition. A function f : [a, b] → R is said to be piecewise continuous if it is
continuous on [a, b] at all except perhaps a finite number of points.
Examples.
f (x) b
g(x)
b b
bc
| | | |
a b x a b x
Both functions f and g are piecewise continuous but f is bounded while g is not!
Proof. Difficult!
217
Remark. If we know for some reason that f is Riemann integrable on [a, b] (e.g.
because the above conditions are satisfied) then it is sufficient to show that
lim S Pn (f ) = lim S Pn (f )
n→∞ n→∞
for some sequence of partitions Pn of [a, b]. If I denotes this common limit, then
Z b
f (x) dx = I.
a
218
Exercise. Find a Riemann integrable function which has infinitely many discontinu-
ities.
219
Remark. There exist more sophisticated ways of ‘measuring’ areas, volumes etc.
such as Lebesque integration.
• If f is Riemann integrable then it is Lebesque integrable.
• If f is not Riemann integrable, it may still be Lebesque integrable. The Lebesque
integral of the above example is 0!
• However, there exist regions in R2 to which it is impossible to assign an area in
any meaningful way! (Cf. Banach-Tarski paradox in three dimensions).
220
8.3 The Riemann integral and signed area
221
Consider the following example:
f (x)
Ω1
c
a b x
Ω2
222
8.4 Basic properties of the Riemann integral
f (x) y g y f
f M
m
a c bx a bx a bx
For brevity, from now on, we refer to Riemann integrable functions as merely ‘inte-
grable’.
224
Sketch of proof.
(vi) From
−|f (x)| ≤ f (x) ≤ |f (x)| ∀x ∈ [a, b]
and (iv), we deduce that
Z b Z b Z b
− |f (x)| dx ≤ f (x) dx ≤ |f (x)| dx.
a a a
225
It is convenient to introduce the following definition ...
Definition. Suppose that b < a and that f is integrable on [b, a]. Then,
Z b Z a
f (x) dx = − f (x) dx
a b
and Z a
f (x) dx = 0.
a
... since, for instance, the following version of (ii) is still valid.
Remark. Suppose that a, b and c are real numbers and that f is integrable over
some interval containing a, b and c. Then
Z b Z c Z b
f (x) dx = f (x) dx + f (x) dx.
a a c
226
8.5 The first fundamental theorem of calculus
Idea (Leibniz, Newton). How does the integral (area) change as a boundary
changes?
y y = f (t) y y = f (t)
F (x)
| | | | | | |
a x b t a x b t
x+h
Suppose that a function f is continuous and therefore integrable on an interval [a, b].
227
Then,
F (x + h) − F (x)
≈ f (x),
h
where h is a ‘small’ positive number.
228
Implications.
229
Proof of the 1. FTC. y
Mh y = f (x)
mh
xx+h x
Differentiability.
Suppose that x ∈ (a, b) and choose an h > 0 such that x + h ∈ (a, b). Then,
Z x+h
F (x + h) − F (x) = f (t) dt.
x
Since f is continuous on [a, b], it attains a minimum value mh and a maximum value
Mh on [x, x + h], that is,
mh ≤ f (t) ≤ Mh
for all t ∈ [x, x + h].
230
We therefore conclude that
Z x+h
mh h ≤ f (t) dt ≤ Mhh
x
and hence
mhh ≤ F (x + h) − F (x) ≤ Mhh
so that
F (x + h) − F (x)
mh ≤ ≤ Mh
h
since h > 0.
and hence
F (x + h) − F (x)
lim = f (x)
h→0 + h
by virtue of the pinching theorem.
231
In a similar manner, one can show that
F (x + h) − F (x)
lim = f (x).
h→0− h
Continuity.
Since f is continuous on [a, b], it is bounded on [a, b], that is, there exists an M > 0
such that
|f (x)| ≤ M
for all x in [a, b].
232
Hence, for x ∈ (a, b), we obtain:
Z x Z a
|F (x) − F (a)| = f (t) dt − f (t) dt
a a
Z x
= f (t) dt
a
Z x
≤ |f (t)| dt
Za x
≤ M dt
a
= M |x − a|.
Accordingly,
|F (x) − F (a)| ≤ M |x − a| → 0
as x → a+. Thus, F is continuous at a.
233
Remark. The above proof of continuity of F at the endpoints could be applied to
any point on [a, b]. Thus, for continuity of F on [a, b], it is actually only required
that f is integrable and bounded!
Note. Many important functions are actually defined as area functions such as
Z x
1
ln(x) = dt
1 t
Z x
2 2
erf(x) = √ e−t dt
π 0
Z x
sin t
Si(x) = dt.
0 t
The first fundamental theorem of calculus then implies that these functions are con-
tinuous and differentiable on the respective (closed or open) intervals.
234
8.6 The second fundamental theorem of calculus
235
Notation. We frequently use the notation
b h ib
F (x) or F (x)
a a
for the expression F (b) − F (a).
236
Example. Calculate the area under the curve
3
y = f (x) = 3x2ex
between x = 0 and x = 1.
237
8.7 Indefinite integrals
238
Examples. It is easy to verify that
Z
1 n+1
xn dx = x + C, n 6= −1
n+1
Z
sin x dx = − cos x + C
Z
ax 1 ax
e dx = e + C, a 6= 0
a
Z
2x + 1 2√
√ dx = 3x + 2 (1 + 6x) + C
3x + 2 27
Z ′
f (x)
dx = ln |f (x)| + C.
f (x)
239
8.8 Integration by substitution
since
d
F (g(x)) = F ′(g(x))g ′(x) = f (g(x))g ′(x).
dx
Previous example.
Z Z
2 x3 x3 d 3 x3
3x e dx = e x dx = e + C.
dx
240
Example. Find Z
1
dx.
x ln x
241
Thus, if we set
u = g(x)
then
Z
f (g(x))g ′(x) dx = F (g(x)) + C
= F (u) + C
Z
= f (u) du.
242
Example. Find Z
2x + 1
I= √ dx.
3x + 2
Set
√ 3 u2 − 2
u= 3x + 2 ⇒ du = √ dx, x= .
2 3x + 2 3
Hence,
Z 2
2(u − 2) 2
I= +1 du
3 3
Z
4 2 2
= u − du
9 9
4 3 2
= u − u+C
27 9
4 3/2 2√
= (3x + 2) − 3x + 2 + C.
27 9
Note. The last step always consists of rewriting everything in terms of the original
variable!
243
The precise statement in the case of definite integrals is the following:
holds.
Proof. Exercise.
244
f (x)
−a| |
a x
246
8.9 Integration by parts
247
Sometimes, it may be relatively easy to apply the above formulae...
It is evident that
Z
d
I= x sin x dx
dx
Z
d
= x sin x − x sin x dx
dx
= x sin x + cos x + C.
248
Example. Find the definite integral
Z e
I= ln x dx.
1
249
8.10 Improper integrals
f (x)
| |
a R x
250
Definition.
(a) Suppose that there exists a real number L such that
Z R
f (x) dx → L
a
as R → ∞. Then, f is said to be integrable over [a, ∞). We say that the
improper integral Z ∞
f (x) dx
a
is convergent and write Z ∞
f (x) dx = L.
a
(b) Suppose that
Z R
f (x) dx
a
does not have a limit as R → ∞. Then, we say that f is not integrable over
[a, ∞) and the improper integral
Z ∞
f (x) dx
a
is said to be divergent.
251
Rb
Remark. A similar definition for improper integrals of the form −∞ f (x) dx applies.
Example. The volume of a solid bounded by a curve y = f (x) rotated about the
x-axis and the planes x = a and x = b is given by
Z b
V =π [f (x)]2 dx.
a
Show that the volume of the body generated by the curve y = 1/x between 1 and
”∞” is Z ∞
1
V =π 2
dx = π
1 x
and, hence, finite.
252
A ”paradox”. The surface area of a solid of the above type (excluding the ‘vertical’
boundaries) is given by
Z b q
A = 2π f (x) 1 + [f ′(x)]2 dx.
a
Hence, for the above example, we obtain
Z b r
1 1
A = 2π 1 + 4 dx
1 x x
" √ #b
2
√ x4 + 1
= π ln x + x + 1 −4
x2
1
[Verify this!] →∞ as b → ∞.
If f is not integrable on either of the intervals (−∞, 0] or [0, ∞) then we say that
the improper integral Z ∞
f (x) dx
−∞
diverges.
254
Example. Does the improper integral
Z ∞
x dx
−∞
converge?
Since Z R
R
1 R2
x dx = x2 = →∞
0 2 0 2
is irrelevant
255
Theorem (Convergence and divergence of p-integrals). The improper integral
Z ∞
1
p
dx
1 x
Proof. If p 6= 1 then
Z R 1−p
R
−p x
x dx =
1 1−p 1
R1−p − 1
=
1−p
(
1
p−1 when 1 − p < 0
→
∞ when 1 − p > 0
as R → ∞.
256
If p = 1 then
Z R h iR
1
dx = ln x = ln R − ln 1 → ∞ as R → ∞.
1 x 1
257
8.11 Comparison tests for improper integrals
Theorem (Comparison test). Suppose that f and g are integrable functions and
that
0 ≤ f (x) ≤ g(x)
for x ≥ a.
Z ∞ Z ∞
(i) If g(x) dx converges then f (x) dx converges.
a a
Z ∞ Z ∞
(ii) If f (x) dx diverges then g(x) dx diverges.
a a
258
Sketch of proof. This follows from
Z R Z R
0≤ f (x) dx ≤ g(x) dx
a a
and the Least Upper Bound Axiom (MATH1241).
259
Example. Discuss the convergence of
Z ∞
−x2
I= e dx.
−∞
260
√
Note. One may show (2nd year) that I = π.
261
Instead of using inequalities to estimate integrands, one often uses a ‘dominant term
analysis’ such as √
sin x + x2
f (x) =
2x4 − 1
‘behaves like’
1
g(x) = 3
2x
for large x and hence one expects the convergence of the two associated improper
integrals to be the same.
262
Theorem (Limit form of the comparison test.) Suppose that f and g are
nonnegative and bounded on [a, ∞). If
f (x)
lim =L
x→∞ g(x)
263
Example. Discuss the convergence of
Z ∞√
sin x + x2
I= 4
dx.
2 2x − 1
Set √
sin x + x2 1
f (x) = , g(x) = 3 .
2x4 − 1 x
Since
f (x) 1
lim =
x→∞ g(x) 2
and the p-integral Z ∞
1
dx
2 x3
converges, we conclude that I converges.
264
Example. Discuss the convergence of
Z ∞
√ √
I= ( x − x − 3 ) dx.
3
265
266
Chapter 9
267
Question. Consider the functional equation
f (st) = f (s) + f (t), (9.1)
where s and t are independent variables. It is evident that f = ln constitutes one so-
lution of this functional equation. Are there other functions f obeying this functional
equation?
If we now demand that f ′(1) = 1 then f is uniquely determined via integration since
f (1) = 0.
268
Conclusion. A function f is uniquely defined by the functional equation
f (st) = f (s) + f (t),
subject to
f ′(1) = 1.
It is given by Z x
1
f (x) = dt.
1 t
269
9.1 Powers and logarithms
ln : (0, ∞) → R
is defined by the formula
Z x
1
ln x = dt.
1 t
1 x t
It is evident that ln x is simply the area of the shaded region shown below in the case
when x ≥ 1.
270
Theorem. The function ln : (0, ∞) → R has the following properties:
(i) ln is differentiable on (0, ∞) and
d 1
ln x = .
dx x
(ii) ln x > 0 for x > 1,
ln 1 = 0,
ln x < 0 for 0 < x < 1.
(iii) ln x → −∞ as x → 0+,
ln x → ∞ as x → ∞.
271
Proof.
(i) Apply the first fundamental theorem of calculus to the definition of ln.
1
(ii) This follows from the definition of ln and the fact that t > 0 when t > 0.
1
y y=
t
1 2 4 8 t
272
In general,
Z 2n
dt 1 1 1
≥ + + ··· +
1 t |2 2 {z 2}
n terms
n
= →∞
2
as n → ∞.
273
(iv) Suppose that y is some fixed positive number and that x > 0. Then, the chain
rule implies that
d 1 d y 1 d
[ln(xy)] = (xy) = = = ln x.
dx xy dx xy x dx
Accordingly,
ln(xy) = ln(x) + C
for some constant C.
Evaluation at x = 1 leads to
ln(y) = C
and hence
ln(xy) = ln(x) + ln(y).
(v) Similar technique.
(vi) Similar technique.
274
Remark. The above properties imply that
• Range(ln) = R, and
• ln is increasing and hence invertible so that
• ln x = 1 has a unique solution.
275
9.3 The exponential function
Question. It is recalled that we have already defined xr for r ∈ Q. Can this defini-
tion be extended to r 6∈ Q?
b
(1, e)
1 b
y = exp x (e, 1)
1 x
y = ln x
276
Remark. For any rational number r, we can evaluate both
exp r and er
but are these two numbers the same?
(ii) exp(1) = e,
exp(0) = 1.
(iii) exp x → ∞ as x → ∞,
exp x → 0 as x → −∞.
277
(iv) exp is differentiable on R with
d
exp x = exp x.
dx
278
Proof.
(i) - (iii) Follows from the definition of exp.
(iv) The function exp is differentiable on R by virtue of the inverse function theorem
and differentiation of
ln(exp x) = x
produces
1 d
exp x = 1
exp x dx
as required.
279
(vi) Suppose that r is a rational number and x is a real number. Then,
exp(rx) = exp r ln exp x)
r
= exp ln (exp x)
= (exp x)r .
280
Definition. For any x 6∈ Q, we define the number ex to be
ex = exp x.
Note. The above definition ‘merely’ means that ex is the unique real number R such
that Z R
1
dt = x
1 t
for any x ∈ R. By construction, the function
f : R → R, f (x) = exp x = ex
is differentiable (and continuous) and is called the exponential function.
281
9.4 Exponentials and logarithms with other bases
282
Example. It is seen that
f3(2) = 32 = exp(2 ln 3) = exp(ln 3) exp(ln 3) = 3 × 3 = 9
as one would expect!
283
Definition. Suppose that b is a positive real y = 8x y = 3x
y y = 2x
number with b 6= 1. Then, the logarithm
function to the base b
y = log2 x
logb : (0, ∞) → R y = log3 x
1 y = log8 x
is defined to be the inverse of the function
1 x
x
fb : R → (0, ∞), fb(x) = b = exp(x ln b).
In particular, loge x = ln x.
The following theorem demonstrates that all logarithm functions are just scaled ver-
sions of the natural logarithm function ...
284
Theorem. Suppose that b is a positive real number with b 6= 1. Then
ln x
logb x =
ln b
for all x > 0.
Proof. Since
x = blogb x = exp(logb x ln b),
we conclude that
ln x = logb x ln b.
285
Remark. The above theorem implies that the class of logarithm functions represents
the general solution of the functional equation (9.1).
286
9.5 Integration and the ln function
Since
d 1
ln(−x) =
dx x
for x < 0, the function ln(−x) is also an antiderivative of 1/x. Thus,
Z
1
dx = ln |x| + C
x
provided that x is restricted to an interval which does not contain 0.
287
This may be generalised to
Z
f ′(x)
dx = ln |f (x)| + C
f (x)
provided that f is differentiable and does not vanish on the interval of integration.
288
9.6 Logarithmic differentiation
Logarithms are powerful in that they ‘transform’ powers into products, products into
sums and quotients into differences.
Remark. The above procedure is only valid for intervals on which y > 0.
290
9.7 Indeterminate forms with powers
Example. Find
L = lim xsin x.
x→0+
≤ lim |x ln x|
x→0+
=0
so that
ln L = 0 ⇒ L = 1.
291
The following example is of the indeterminate form ”1∞”.
292
Chapter 10
293
Note. cosh and sinh are differentiable with
d d
(sinh x) = cosh x, (cosh x) = sinh x
dx dx
so that cosh x and sinh x obey the differential equation
d2 y
2
= y.
dx
294
Properties of the sinh function. y
• sinh is an odd function.
• sinh 0 = 0. y = sinh x
y = 12 ex
• sinh is increasing on (−∞, ∞) and has a point
of inflexion at 0. x
295
Theorem. The hyperbolic functions are related by
cosh2 x − sinh2 x = 1.
Proof. By definition,
2 2
ex + e−x ex − e−x
cosh2 x − sinh2 x = −
2 2
1 2x −2x 2x −2x
= (e + 2 + e ) − (e − 2 + e )
4
= 1.
296
The term hyperbolic is motivated in the following manner:
297
10.2 Other hyperbolic functions
Other hyperbolic functions are defined in analogy with the trigonometric functions
according to
sinh x cosh x
tanh x = , coth x = ,
cosh x sinh x
1 1
sech x = , cosech x = .
cosh x sinh x
298
Recall that
sinh x ex − e−x
tanh x = = x
cosh x e + e−x
299
Proof of the derivative of tanh. By definition,
d d sinh x
(tanh x) =
dx dx cosh x
d d
cosh x sinh x − sinh x cosh x
= dx dx
cosh2 x
cosh2 x − sinh2 x
=
cosh2 x
1
=
cosh2 x
= sech2 x.
300
10.3 Hyperbolic identities
301
‘Double-angle’ formulae.
sinh(2x) = 2 sinh x cosh x
cosh(2x) = cosh2 x + sinh2 x
2 tanh x
tanh(2x) = 2 .
1 + tanh x
Exercise. Prove the first two ‘sum and difference’ formulae and, hence, derive the
third.
302
10.4 Hyperbolic derivatives and integrals
d d
sinh x = cosh x, cosh x = sinh x
dx dx
d d
tanh x = sech2 x, coth x = − cosech2 x
dx dx
d
sech x = − sech x tanh x
dx
d
cosech x = − cosech x coth x.
dx
303
Example. Determine the definite integral
Z (ln 2)2 2√
sech x
I= √ dx.
0 x
304
10.5 The inverse hyperbolic functions
The hyperbolic sine and tan functions are increasing and therefore invertible. In the
case of the hyperbolic cosine function, one usually restricts the domain to [0, ∞) to
guarantee invertibility.
y y y
1 −1
x 1 x 1 x
305
Definition. The functions
• cosh−1 : [1, ∞) → [0, ∞)
• sinh−1 : R → R
• tanh−1 : (−1, 1) → R
are defined to be the inverse functions of the (restricted) hyperbolic functions
• cosh : [0, ∞) → [1, ∞)
• sinh : R → R
• tanh : R → (−1, 1).
306
Example. Evaluate
sinh cosh−1 34 .
Evaluation of
sinh2 t = cosh2 t − 1
at t = cosh−1 34 yields
2
4 2
sinh cosh−1 34 = 3 − 1 = 97 .
Accordingly, √
sinh cosh−1 34 =+ 3
7
since t > 0.
307
Remark. Note that
cosh−1(cosh x) = x
for x ≥ 0 only!
Example. Simplify
−1 x 1
cosh + , x > 0.
2 2x
308
The relation
ex − e−x
y = sinh x =
2
may be written as
(ex)2 − 2yex − 1 = 0
so that p
x 2y + 4y 2 + 4
e =
2
since ex > 0. Hence, p
x = ln y + y 2 + 1 .
In a similar manner, the remaining identities of the following theorem may be proven.
309
Theorem. The following identities hold:
−1
√
sinh x = ln x + x2 + 1
p
cosh−1 x = ln x + x2 − 1
1 1+x
tanh−1 x = ln .
2 1−x
d 1
sinh−1 x = √
dx x2 + 1
d −1 1
cosh x = √
dx x2 − 1
d 1
tanh−1 x = 2
.
dx 1−x
310
Example. Use the inverse function theorem to confirm that
d −1 1
sinh x = √ .
dx 2
x +1
311
10.6 Integration leading to the inverse hyperbolic functions
Z
dx −1 x
√ = cosh +C
2
x −a2
a
p
= ln x + x2 − a2 + C̃, x≥a>0
312
1 −1 x
Z tanh
+ C, |x| < a
dx a a
=
a2 − x2 1 x
coth−1 + C, |x| > a > 0
a a
1 a + x
= ln + C, x2 6= a2.
2a a−x
[These formulae are included in the table of standard integrals which is issued at the
final examination.]
313
Example. Determine the indefinite integral
Z
dx
√ .
2
x − 2x + 10
314