Lectures 1141
Lectures 1141
Lectures 1141
Chapter 1
Sets, inequalities and functions
1.1
Sets of numbers
A set is a collection of distinct objects. The objects in a set are called the elements
or members of the set.
The set N of natural numbers is given by
N = {0, 1, 2, 3, 4, . . .}.
The set Z of integers is given by
Z = {. . . , 3, 2, 1, 0, 1, 2, 3, . . .}.
3
p
: p, q Z, q 6= 0 .
q
The set R of real numbers may be represented as the collection of points lying on
the number line.
|
22
/ N,
7
22
/ Z,
7
22
Q,
7
{x R : x
/ Q}
22
R
7
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
[a, b] = {x R : a x b}
(closed)
a
[a, b) = {x R : a x < b}
(???)
a
(a, b] = {x R : a < x b}
(???)
a
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
(, b) = {x R : x < 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
Solving inequalities
Two types of inequalities deserve special attention: quadratic inequalities and rational
inequalities.
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.
Properties. Suppose that x and y are real numbers. Then
| x| = |x|,
|xy| = |x||y|,
x |x|
=
provided that y 6= 0,
y
|y|
|x + y| |x| + |y|,
|x y| |x| |y|.
|x|2 = x2.
x2 ,
Geometrically, |xa| 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.
x2 > a2
10
x < a or x > a.
(b)
|x + 5|
<1
|x 11|
11
1.4
Functions
A function
f :AB
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.
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
Remarks.
Dom(f ) = [0, ),
Codom(f ) = R,
Range(f ) = [2, )
f (x) =
f (x) =
may be abbreviated as
x [0, ).
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)
14
(b)
f (x) =
x 1/x.
1
.
2
x +x2
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.
Exercise. What is the range of the function f defined by
f (x) = x2 + ex ?
f : R [0, ),
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.
Definition. Suppose that f : A B and g : A B are real-valued functions.
Then, the functions f + g, f g, f.g and f /g are defined by the rules
(f + g)(x) = f (x) + g(x)
x A
(f g)(x) = f (x) g(x)
x A
(f.g)(x) = f (x)g(x)
x A
f (x)
(f /g)(x) =
g(x)
where A = A \ {x A : g(x) = 0}.
x A,
16
and
g:AB
are functions such that Range(g) is a subset of Dom(f ). Then the composition
is defined by the rule
f g :AD
(f g)(x) = f (g(x))
x A.
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
where n N is the degree and the coefficients a0, a1, . . . , an are real numbers with
the leading coefficient an 6= 0.
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
Dom(f ) = R \ {2, 1}
19
1.6
sin : R R
cos : R R
and
sin x = s
and
cos x = c,
where s and c are defined in a geometric manner:
s
cos(x + 2) = cos x.
21
are defined by
cos x
sin x
1
.
sin x
The six trigonometric functions are related by various identities and formulae (which
you are supposed to know):
complementary identities
x = cos x
sin
2
cos
x = sin x
2
Pythagorean identities
cos2 x + sin2 x = 1
1 + tan2 x = sec2 x
cot2 x + 1 = cosec2 x
22
tan x tan 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 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
,
x
p
3
h(x) = x4 2x2 + 5,
24
1.8
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
x
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 = f (x)
b
|
y = g(x)
b
|
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
Question. How would you define continuity?
Intuitive (incorrect) answer. The function is continuous if its graph can be
drawn without lifting the pencil off the page.
Rigorous (correct) answer. Via limits.
27
f (x)
g(x)
h(x)
bc
bc
28
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
In this section, we examine some techniques for calculating limits of the form lim f (x)
x
1
and postpone their precise definition.
x2
ln x
1
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
Elementary rules.
If f is a constant function, that is f (x) = c for all x, then
lim f (x) = c.
If f (x) as x then
1
= 0.
x f (x)
lim
31
lim g(x) = b
lim [f (x)g(x)] = ab
x
f (x) a
=
x g(x)
b
lim
provided that b 6= 0.
Proof. Later.
32
as x .
2
1+
3x + 4
f (x) =
5 6ex
33
2.1.2
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.
Then
lim g(x) = L.
g(x) = ex 2sin 3x
as x .
35
2.1.3
36
(b)
x2 + 3x
lim
x
2x4 + 3 4x
37
2.1.4
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!
Example. Determine the limit of
p
2
f (x) = 2x + 3x 2x2 x
as x .
Exercise. Does
4
3
lim x x x4 + 1
exist?
38
2.1.5
Indeterminate forms
x
x
20
x
2x2
2 2
x
Since we cannot determine in advance what kind of limiting behaviour something of
0
and appear in various applications.
39
2.2
y = f (x)
1
0 1
1
x
40
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 > 0, there exists a real number M such that
if x > M then |f (x) L| < .
y y = f (x)
L+
L
L
42
2.3
5x
= 5.
x x + 3
lim
for x > 3
for x > 0.
In summary,
15
|f (x) L| < .
x
This inequality gives an upper bound for the distance between f (x) and L!
Accordingly,
15
< .
|f (x) L| < whenever
x
The latter condition is equivalent to
15
x>
then
|f (x) L| < whenever x > M,
as required.
44
2. Find a simple condition on x such that this upper bound is less than .
3. Use this condition to state an appropriate value for M (in terms of ).
Example. Show that
x2 cos x
lim
= 1.
x x2 + 1
45
Remarks.
In general, M depends on 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
lim g(x) = L2
and set
1 = 2 =
so that
2.4.1
bc
bc
|
|
f (x)
4
3
2
1
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.
xa
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.
xa+
50
xa
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.
xa
51
For a = 2:
lim f (x) = 3,
x2
lim f (x) = 1
x2+
f (x) as x 4
No limit exists.
For a = 6:
x6
x6+
f (6) = 4
The two-sided limit exists but does not coincide with the value of f at a = 6.
For a = 8:
x8
x8+
52
2.4.2
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);
xa
53
2.4.3
Polynomials, rational functions, the trigonometric functions, exponentials and logarithms are continuous at every point in the respective domains.
Limits at a point also have nice arithmetic properties ...
Theorem. Suppose that
lim f (x) = L1,
lim g(x) = L2
xa
xa
f (x) L1
=
lim
xa g(x)
L2
provided that L2 6= 0.
54
xa
x5
2x + cos(x2) .
55
The pinching theorem. Suppose that f , g and h are all defined on an open interval
I containing the point a. If
f (x) g(x) h(x)
x I
and
lim f (x) = lim h(x) = L
xa
xa
then
lim g(x) = L.
xa
y
y = h(x)
L
a
|
y = g(x)
x
y = f (x)
56
sin x
x
f (x) =
1
continuous?
if x 6= 0
if x = 0
|x 9|
if x 6= 3
x3
f (x) =
6
if x = 3.
Discuss the limiting behaviour of f (x) as x approaches 3.
57
58
Chapter 3
Properties of continuous functions
3.1
Question. If two continuous functions are combined via function addition, subtraction, 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
xa
xa
xa
xa
(def. of f + g)
(limit rule)
= f (a) + g(a)
(f, g cont.)
xa
xa
= (f + g)(a)
(def. of f + g).
Hence f + g is continuous at 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 rational functions are continuous at every point of their respective domains.
Solution. Any polynomial can be obtained from f and constant functions via addition and multiplication, e.g.
x3 4x2 + 5 = [(x x x)] + [(4) x x] + 5,
and hence is continuous everywhere.
Any rational function is of the form
p(x)
,
q(x)
where p and q are two (continuous) polynomials, and is therefore continuous at every
point a for which q(a) 6= 0.
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.
Thus, it suffices to verify the following Lemma:
Lemma. Show that the functions sin and cos are continuous everywhere.
62
63
Even larger classes of continuous functions may be obtained in the following manner:
Theorem. Suppose that f is continuous at a and that g is continuous at f (a).
Then g f is continuous at a.
Proof.
lim (g f )(x) = lim (g(f (x))
(def. of g f )
= g(lim f (x))
(Th. of Ch. 2)
xa
xa
xa
= g(f (a))
(f cont.)
= (g f )(a).
(def. of g f )
Hence g f is continuous at a.
64
65
3.2
Continuity on intervals
66
xa+
xb
67
x is
Example. Consider the functions f , g and h, whose graphs are shown below.
f (x)
g(x)
h(x)
bc
bc
b x
b x
b x
g is continuous at the endpoints a and b but not continuous on the open interval
(a, b).
3.3
z
f (a)
f (b)
f (x)
b
69
Proof. Omitted but based on the the least upper bound property of the real numbers:
If a1 a2 a3 M is a non-decreasing bounded sequence of real numbers
then ak converges to some real number L M as k .
(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.
|
z
f (b)
f (a)
f (x)
c1
c2
c3
71
Continuity of f is crucial.
f (x)
f (a)
bc
f (b)
72
The intermediate value theorem is a theorem about real numbers. Consider, for
instance, the function
f : [0, 2] R,
f (x) = x2 2
c = 2 6 Q
73
c = c2 1
and approximate its value.
lim f (x) = 1.
74
3.4
for all
x [a, b].
The corresponding value f (c) is called the absolute minimum value of f on [a, b].
If f has an absolute minimum point on [a, b] then we say that f attains a minimum
on [a, b].
We say that a point d in [a, b] is an absolute maximum point for f on [a, b] if
f (x) f (d)
for all
x [a, b].
1 2 3 4 5 x
|
|
h(x)
4
3
2
1
g(x)
4
3
2
1
Example. Consider the functions g and h, whose graphs are illustrated below.
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].
The maximum-minimum theorem. If f is continuous on a closed interval [a, b]
then f attains a minimum and maximum on [a, b]. That is, there exist points c and
d in [a, b] such that
f (c) f (x) f (d)
for all x in [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
78
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 .
Theorem. If f : [a, b] R is continuous on [a, b] then f is bounded on [a, b].
Proof. Suppose that f is continuous on [a, b]. By the maximum-minimum theorem,
f attains a maximum and minimum on [a, b], that is
f (c) f (x) f (d)
79
1 sin 1
x
f (x) = x2 + 1
if x 6= 0
if x = 0
80
Chapter 4
Differentiable functions
9
s = t2
t [0, 3].
displacement (m)
1
0
81
Typical question. A snowboarder accelerates from rest in such a way that her displacement s (in metres) from her starting position
after t seconds is given by
1 2 3
time (s)
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 ...
Time interval t Average velocity
[1, 1.5]
0.5
2.5
[1, 1.4]
0.4
2.4
[1, 1.3]
0.3
2.3
[1, 1.2]
0.2
2.2
[1, 1.1]
0.1
2.1
[1, 1.01]
0.01
2.01
[1, 1.001] 0.001
2.001
... and let t approach 0. It appears that
v 2
as
t 0
Geometric interpretation. Find the slope (gradient) of the (tangent line to the)
graph at t = 1.
Thus, calculate the gradient of the secant intersecting the graph at (1, 1) and some
other point P ...
s = t2
secant
P
2.25
s
1
grad. of secant =
b
1 1.5
t
83
s
t
(1.5)2 (1)2
1.51
= 2.5
Intersection points
t = 1, t = 1.5
t = 1, t = 1.4
t = 1, t = 1.3
t = 1, t = 1.2
t = 1, t = 1.1
t Gradient of secant
0.5
2.5
0.4
2.4
0.3
2.3
0.2
2.2
0.1
2.1
84
s = t2
secants
tangent
1.21
1
|
|
1.44
1.96
s
1.69
2.25
4.1
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, (x + h)3)
(x + h)3 x3
(grad. at x) = lim
h0
h
3x2h + 3xh2 + h3
= lim
h0
h
x3
= lim 3x2 + 3xh
(x, x3)
h0
2
= 3x .
|
x
86
x+h
tangent
Definition. Suppose that f is defined on some open interval containing the point x.
If
f (x + h) f (x)
lim
h0
h
exists then it is called the derivative of f at x and we say that f is differentiable at x.
The derivative of f at x is denoted by
f (x)
or
df
(x)
dx
or
d
f (x).
dx
f (x + h) f (x)
h
is called the difference quotient for f at the point x.
87
Exercise. Write
and conclude that
d
cos x = sin x.
dx
cos2 h 1
sin2 h
cos h 1 =
=
cos h + 1
cos h + 1
cos h 1
lim
= 0.
h0
h
88
xn sin 1
if x 6= 0
x
f (x) =
0
if x = 0
continuous or differentiable at x = 0?
sin x = cos x.
dx
89
4.2
(product rule)
(quotient rule)
90
f (x) = sin
91
x
x2 + 1
2
4.3
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
g(x + h) g(x)
f (x + h) f (x)
+ f (x)
.
= g(x + h)
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
at the point (2, 1).
x4 x2y 2 + y 4 = 13
Write
and differentiate both sides:
x4 x2y(x)2 + y(x)4 = 13
32 4 8y (2) + 4y (2) = 0
y 1 = 7(x 2).
94
y (2) = 7.
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 .
Differentiation of both sides with respect to x yields
ny n1
dy
= mxm1.
dx
95
Hence,
dy mxm1
=
dx
ny n1
m xm1
=
n xq(n1)
= qx(m1)qn+q
= qxq1
as required.
96
4.5
f (a + h) f (a)
h
h0
h
f (a + h) f (a)
= lim
lim h
h0
h0
h
h0
= f (a) 0 = 0
since f is differentiable at a. Accordingly,
lim f (a + h) = f (a).
h0
xn sin 1
if x =
6 0
x
f (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
... 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 defined
by 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.
Remark. Note that the requirement of f being continuous at a is equivalent to
demanding that p(a) = q(a) since p and q are continuous at a.
Previous example. Consider the function f : (0, ) R defined by
(
if 0 < x 1
4 x
f (x) =
bx2 + c
if x > 1,
where b and c are real numbers.
100
4.6
f (a) = lim
so that
f (x) f (a)
f (a)
xa
if x is sufficiently close to a. Thus,
101
Example. Suppose that the horizontal distance L covered by a tennis ball is given
by
v2
L = sin 2,
g
where v and 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 are changed
by a small amount?
102
4.7
Many physical processes involve quantities (such as temperature, volume, concentration, velocity) that change with time but may not be independent of each other.
Their rates of change may then be obtained by careful application of the chain rule.
Example. A spherical balloon is being inflated and its radius is increasing at a constant 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 V (r) be the volume of the balloon as a function of its radius r given by
4
V = r3
3
so that
dV
= 4r2.
dr
103
dr
=6
dt
dV
= 4(20)2 6 = 9600
dt
Hence, the volume is increasing at a rate of 9600 mm3/sec when the radius is 20 mm.
104
105
4.8
ch
c+h
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.
Definition. If a function f is differentiable at a point c and f (c) = 0 then c is
called a stationary point of f .
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
3
200
107
Differentiation yields
f (x) = 20x4 20x3 120x2
= 20x2(x2 x 6)
= 20x2(x + 2)(x 3).
Result:
x = 3: local minimum point
The proof of the above theorem may be found in the Calculus Notes.
108
Chapter 5
The mean value theorem and its applications
The mean value theorem is one of the most important results for establishing the
theoretical structure of calculus.
Applications of the mean value theorem include
identifying where a function is increasing or decreasing,
identifying different types of stationary points,
and 00 ,
5.1
f (a)
|
|
f (b)
A
|
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).
ba
110
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
If f (a) = f (b) then the mean value theorem reduces to Rolles theorem.
g(x)
C
A
a
B
b x
113
115
Proof of the mean value theorem. Suppose that f is continuous on [a, b] and
differentiable on (a, b).
We consider the function g defined by
f (b) f (a)
(x a) + f (a) .
g(x) = f (x)
ba
It is evident that g is continuous on [a, b], differentiable on (a, b) and that
g(a) = g(b) = 0.
By virtue of Rolles theorem, there exists a c in (a, b) such that g (c) = 0, that is,
such that
f (b) f (a)
f (c)
= 0.
ba
Hence,
f (b) f (a)
f (c) =
ba
as required.
116
5.3
for all
x > 0.
Hint: Make the identification [a, b] = [0, x] for any fixed x > 0.
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)
x1
for some c in (1, x). Accordingly,
ln x
1
=
x1 c
for some c between 1 and x.
118
1
<1
c
119
5.4
Error bounds
f (x) = cos x
3 1
120
Accordingly,
121
5.5
implies that
implies that
increasing
decreasing
x
122
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.
Since f is differentiable on I, it is continuous on [x1, x2] and differentiable on (x1, x2).
Hence, by the mean value theorem,
f (x2) f (x1)
= f (c)
x2 x1
5 5+
+ +
+ +
0
+
0
+
124
+
+
0
+
0
1+
+
+
+
+
4
0
+
+
0
4+
+
+
+
+
Result:
x = 4: Local minimum point
x = 1: local maximum point
125
5.6
Another (potential) method for classifying the stationary points of a function f involves the second derivative of f , which is denoted by
f
or
d2 y
,
2
dx
or
if we set y = f (x).
Theorem (The second derivative test). Suppose that a function f is twice
differentiable on (a, b) and that c (a, b).
If f (c) = 0 and f (c) > 0 then c is a local minimum point of f .
126
If g (c) < 0 then g(c + h) < g(c) < g(c h) for all positive h sufficiently small.
Exercise. Find and classify the stationary points of the function f : R R given
by
f (x) = x3 6x2 + 9x 5.
127
x4
x3
5.7
4
b
3
2
4
129
5 x
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].
Example. Suppose that the function f : R R is given by the rule
f (x) = |x2 3x 4|.
Find the absolute maximum and absolute minimum values of f on the interval [0, 5]
(cf. previous graph).
130
5.8
Counting zeros
f (x) = x4 x3 3x2 8x 5
f (x) = (x 2)(4x2 + 5x + 4)
f (x) < 0 on (, 2)
and
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
for all
x I.
133
Example. Let
f(x) = xex ,
0.
134
Function
Antiderivative
1
r+1
r+1 x
+C
sin x
cos x + C
cos x
sin x + C
1 ax
ae
eax
f (x)
f (x)
+C
ln |f (x)| + C
136
5.10
LH
opitals rule
If
exists then
f (x)
lim
xa g (x)
f (x)
f (x)
= lim .
lim
xa g (x)
xa g(x)
137
138
sin x x cos x
x sin x
= lim
= ... see above.
x0
x0
x sin x
1 cos x
lim
Remark. It is important that the limit exists after a finite number of applications of
lHopitals rule!
Exercise. What is the limit
2x sin x
lim
x 3x + sin x
139
140
Chapter 6
Inverse functions
Problem. Given a function
if we set
f : A B,
y = f (x),
under what circumstances is it possible to express x as a function of y, that is, to
find a function
g:CA
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
(b)
f : [1, 1] R,
y = f (x) =
142
1 x2
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(ttd)
leads to
1 cm
t = td + ln .
143
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.
f1(x)
f2(x)
f3(x)
output
output
output
input x
input
input
input x
6.2
One-to-one functions
f (x1) = f (x2)
implies that
x1 = x2
for all
in the domain of f .
146
x1 6= x2
f (x) = x3 + x + 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.
f1(x)
f2(x)
f3(x)
f1 is not one-to-one.
f2 is one-to-one.
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.
Theorem. If a function f is either strictly increasing or strictly decreasing then f is
one-to-one.
Previous example. Show that the function
f : R R,
f (x) = x3 + x + 1
149
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
x Dom(f )
for all
and
f (g(y)) = y
for all
Moreover,
Dom(g) = Range(f ),
and g is one-to-one.
y Range(f ).
Range(g) = Dom(f )
for all
x Dom(f )
and
f (f 1(y)) = y
for all
y Range(f )
152
1 3
f (x) = 4 x .
3
1
y = 4 x3
3
p
x = 3 12 3y.
153
: R R,
(x) =
12 3x.
y=x
y = f 1(x)
x
y = f (x)
154
6.4
f (x) = x3.
f : R R,
g : R R,
but g is not differentiable at x = 0!
g(x) =
y
g(x)
f (x)
x
1
g (y) =
f (g(y))
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
157
f (x) = x3
g : (0, ) (0, ),
g(x) =
x.
1
1
1
g (x) =
=
=
f (g(x)) 3[g(x)]2 3x2/3
as expected.
Previous example. Determine the derivative of the inverse of the function f :
(1, 1) R defined by
f (x) = 3 + x2 + 2 tan x
2
at the point f (x) = 3.
158
6.5
sin1 x
( 2 , 1)
(1, 2 )
( 2 , 1)
x
b
x
(1, 2 )
159
On (1, 1), according to the inverse function theorem, the derivative of sin1 is given
by
1
d
1
(sin x) =
.
1
dx
cos(sin x)
Since cos is positive on ( 2 , 2 ), we conclude that
d
1
1
1
(sin x) = p
.
=
2
1
2
dx
1x
1 sin (sin x)
Note.
d
(sin1 x) as x 1.
dx
160
Domain
Range
Derivative
d
(sin x) = cos x
dx
d
1
1
[1, 1]
[ 2 , 2 ]
(sin x) =
dx
1 x2
d
[0, ]
[1, 1]
(cos x) = sin x
dx
d
1
1
(cos x) =
[1, 1]
[0, ]
dx
1 x2
d
( 2 , 2 ) (, )
(tan x) = sec2 x
dx
1
d
1
(, ) ( 2 , 2 )
(tan x) =
dx
1 + x2
[ 2 , 2 ]
[1, 1]
161
cos1 x
(1, )
b
tan1 x
(0, 2 )
b
(1, 0) x
x
b
sin1(sin x) 6= x
162
Example. Determine
(a)
cos 2 sin
(b)
sin1 sin
163
3
.
5
5
.
6
164
Chapter 7
Curve sketching
y
x
x
165
or
a parameter such as
y=
1
.
2
1x
y(t) =
t [1, 1],
t 6= 0.
y = r sin ,
where
|t| cos t
166
r = cos 4.
7.1
In this section, we survey techniques for sketching curves which are described by a
Cartesian equation of the form y = f (x).
Checklist. Consider the following items:
Identify the domain of f .
Identify any symmetries:
f is even if
f is odd if
f (x) = f (x)
f (x) = f (x)
for all
for all
f is periodic of period T if
f (x + T ) = f (x)
for all
167
x Dom(f )
x Dom(f )
x, x + T Dom(f )
168
| sin x|
.
2 + cos(2x)
sin x for x 6= 0
x
f (x) =
1
for x = 0.
169
y = f (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
deg(p) = deg(q) + 1
(x 2)|x| + x
x2
If x > 0 then
(x 2)x + x
x2
(x 2) + 2
=x+
x2
2
=x+1+
.
x2
f (x) =
171
172
x2 3
f (x) =
2x 4
173
7.2
x(t), y(t) ,
t A,
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.
Newtons 2nd law: m x(t), y(t) = F x(t), y(t)
a suitable parametrisation of a curve defined implicitly.
Ellipse: x(t), y(t) = a cos(t), b sin(t)
y(t) (, )
respectively.
Intercepts:
x(t) = 0 if and only if t = 1 so that we obtain the points
(1) = (0, 2),
In addition, y(t) = 0 if and only if t = 1.
175
Vector derivatives:
Consider the tangent vector
We will justify later that the slope of a parametrised curve (t) = x(t), y(t) is
given by
dy y (t)
=
dx x(t)
at all points with x(t) 6= 0.
176
Here,
dy y (t) 3
=
= t
dx x(t) 2
(t) points
for
t<0
(t) points
for
t > 0.
2
4
t0
t0 | (t)|
4t + 9t
Conclusion: The curve does not have a proper tangent vector at the point (1, 1)!
177
Limiting behaviour:
As t ,
(2t, 3t2)
(t)
= lim
lim
= (0, 1).
2
4
t
t | (t)|
4t + 9t
178
179
7.2.1
The ellipse
x2 y 2
+ 2 =1
2
a
b
with semi-axes a and b admits the parametrisation
x(t) = a cos t,
y(t) = b sin t,
y(t)
<t<
t=
<t<
t=
0<t<
t=0
a x(t)
b
a
3
2
0 t < 2.
b
b
t=
3
2
3
2
< t < 2
The table below lists some commonly used parametrisations of conic sections.
Conic section Cartesian equation Parametric equation
Parabola
4ay = x2
x(t) = 2at
y(t) = at2
Circle
x2 + y 2 = a2
x(t) = a cos t
y(t) = a sin t
Ellipse
x2 y 2
+ =1
a2 b2
x(t) = a cos t
y(t) = b sin t
Hyperbola
x2 y 2
=1
a2 b2
x(t) = a sec t
y(t) = b tan t
Note that there exist other useful parametrisations for each of these curves.
181
7.2.2
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
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
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
Many problems in mathematics are easier to solve if one chooses a suitable coordinate
system. Here, we focus on polar coordinates.
7.3.1
Polar coordinates
P
b
y
O
O
184
x2
y 2,
y = r sin
y
tan =
x
185
Example. Find all polar coordinates of the point P with Cartesian coordinates
3 x.
(b) circle
x2 + y 2 = 4.
186
7.3.2
or
= g(r)
corresponds to y =
3 x for x 0.
(b) r = 2 corresponds to x2 + y 2 = 4.
187
0 2.
2
2 12
1
d
b
2 + 2
r
3
1
b
b
r = 2 cos
h
b
3
4
5
4
3
2
7
4
188
... and then mark the corresponding points on the (x, y)-plane:
y =
3
4
c r = 2
b
a r = 1
i
r=3
=0
g r = 2
b
5
4
=
=
3
2
189
7
4
Now:
As increases from 0 to , r increases from 1 to 3. Hence, the distance r from
the origin to points on the curve increases from 1 to 3.
As increases from to 2, r decreases from 3 to 1. Hence, the distance r from
the origin to points on the curve decreases from 3 to 1.
These considerations lead to the final sketch.
y
=
3
4
2
b
3
b
1
b
r = 2 cos
5
4
190
7
4
Symmetries.
If f () = f () then the polar curve is symmetric about the x-axis.
191
7.3.3
() = x (), y ()
dr
dr
=
cos r sin , sin + r cos .
d
d
Thus, horizontal tangents are obtained by solving
y () = 0
but
x() 6= 0,
but
y () 6= 0.
192
193
(r, + )
and hence (r, ) is obtained from (r, ) by reflection in the origin. This is consistent
with the formulae
x = r cos = (r) cos( + )
y = r sin = (r) sin( + ).
194
Chapter 8
Integration
Problem. How does one find/measure/define areas (of regions with curved boundaries)?
The main ideas can be traced back to
Archimedes (the great Greek mathematician of antiquity!) and then
Isaac Barrow (Isaac Newtons mentor)
Isaac Newton (one of the greatest mathematicians and physicists of all time)
Gottfried Leibniz (Newtons contemporary)
One (but not the only) method for calculating the area of regions with curved boundaries is known as Riemann integration.
Note. A priori, calculating areas (integration) and antidifferentiation are two separate 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
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.
But what are the rules that we take for granted?
The area of a rectangle is the product of its length and height.
Areas of congruent regions are equal.
The area of a whole region is the sum of the areas of its parts.
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
b
b1
b2
197
Formally, we demand that any definition of an area satisfy the following axioms:
198
2
1
1 2
2
1
b
a
199
8.1.1
Question. How could one calculate the area of the region below?
1
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
Let denote the region bounded by the graph of f , the x-axis and the lines x = 0
and x = 1.
f (x)
f (x)
f (x)
1
1x
2
1x
1x
Idea. Find lower and upper bounds for area() by choosing appropriate approximations 1 and 2 of the region in terms of n rectangles.
202
It is evident that
area(1) area() area(2).
If limn area(1) = limn area(2) then
A = area() = lim area(1) = lim area(2).
n
Remark. One of Archimedes best ideas was to estimate areas in circles and parabolas by approximating them by inner and outer polygons!
203
y = f (x)
4 k1 k n1 n
1 2 3
n n n n
n n
n n
Pn = {0, n1 , n2 , n3 , . . . , n1
n , 1}
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
k2
k
Rk = width height = f ( n ) = 3 .
n
n
204
k=1
and hence
S Pn (f ) =
1
1
1
+
+ 2.
3 2n 6n
The quantity S Pn (f ) is called the upper Riemann sum of f with respect to the partition Pn.
Axiom (A2) now implies that
1
1
1
A S Pn (f ) = +
.
+
3 2n 6n2
205
y = f (x)
Rk
4 k1 k n1 n
1 2 3
n n n n
n n
n n
The sum of all the areas of the rectangles is called the lower Riemann sum for the
function f over the partition Pn and is denoted by S Pn (f ). We obtain
S Pn (f ) =
so that
n
X
k=1
1 X
(k 1)2
Rk = 3
n
S Pn (f ) =
k=1
1
1
1
+ 2.
3 2n 6n
207
+
A +
+
3 2n 6n2
3 2n 6n2
gives an upper and a lower bound for A.
Conclusion. In the limit n , we obtain
1
A= ,
3
regardless of the actual definition of A as long as it is compatible with the axioms
(A1)-(A5)!
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].
Definition. A finite set P of points in R is said to be a partition of [a, b] if
P = {a0, a1, a2, . . . , an}
and
a = a0 < a1 < a2 < . . . < an = b.
209
a0 a1 a2
n
X
k=1
(ak ak1)f k
where
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
f (x) dx.
I=
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
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
S P (f ) =
n
X
f (xk )xk ,
k=1
where xk = ak ak1 and f attains its maximum value on [ak1, ak ] at the point xk .
When taking a limit as before, xk was replaced with dx and the symbol
replaced with an elongated S (S stands for sum).
214
was
8.2
Even in the case f (x) = C, where C is a positive constant, the actual determination
of the Riemann integral is somewhat tedious.
Indeed, let P = {a0 = a, a1, . . . , an = b} be a partition of [a, b]. Since f is constant,
we have
fk = fk = C
for every k between 1 and n and hence
S P (f ) =
n
X
k=1
=C
f k (ak ak1)
n
X
k=1
(ak ak1)
= C(b a).
A similar calculation leads to
S P (f ) = C(b a).
215
Accordingly,
S P (f ) = C(b a) = S P (f )
for every partition P of [a, b] so that we conclude that f is Riemann integrable and
Z b
f (x) dx = C(b a).
a
Question. Can we find simple sufficient conditions which guarantee that a Riemann
integral exists?
216
g(x)
b
bc
b
b x
b x
Both functions f and g are piecewise continuous but f is bounded while g is not!
Theorem. If f is bounded and piecewise continuous on [a, b] then f is Riemann
integrable on [a, b].
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 )
for some sequence of partitions Pn of [a, b]. If I denotes this common limit, then
Z b
f (x) dx = I.
a
x4 dx,
218
Exercise. Find a Riemann integrable function which has infinitely many discontinuities.
Example. The function f : [0, 1] R defined by
(
1 if x Q
f (x) =
0 if x
/ Q.
fk = 0
S Pn (f ) = 0.
S P (f ) I S P (f )
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.
220
8.3
as the signed area under the graph of the function f from a to c and to
Z c
area() =
|f (x)| dx
a
as the unsigned area under the graph of the function f from a to c provided that the
latter integral exists.
221
c
b
b
a
f (x) dx
222
f (x) dx.
b
8.4
f (x)
f M
m
a
bx
bx
bx
For brevity, from now on, we refer to Riemann integrable functions as merely integrable.
Theorem. Suppose that f, g : [a, b] R are integrable. Then,
(i) (Linearity) f + g is integrable for any , R with
Z
(f + g)(x) dx =
a
223
f (x) dx +
a
g(x) dx.
a
f (x) dx =
a
f (x) dx +
a
f (x) dx.
c
f (x) dx 0.
Sketch of proof.
(iii) If f (x) 0 then
0 S P (f )
f (x) dx.
a
(vi) From
|f (x)| f (x) |f (x)|
x [a, b]
225
and
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
226
8.5
Idea (Leibniz, Newton). How does the integral (area) change as a boundary
changes?
y
y = f (t)
y = f (t)
F (x)
|
b t
x
b t
x+h
Suppose that a function f is continuous and therefore integrable on an interval [a, b].
We define the area function F by
F (x) =
for all x [a, b].
f (t) dt
a
227
Then,
F (x + h) F (x)
f (x),
h
where h is a small positive number.
Thus, we are tempted to believe that F is differentiable and
F (x + h) F (x)
= f (x).
h0
h
F (x) = lim
Implications.
Every continuous function f : [a, b] R has an antiderivative F on (a, b) given
by
Z x
F (x) =
f (t) dt.
a
Since two antiderivatives of f differ by a constant, integration and antidifferentiation are essentially the same procedures!
Differentiation undoes what integration does to f since
f (x) =
d
dx
Z
f (t) dt .
a
y
Mh
mh
y = f (x)
xx+h
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
x+h
x
f (t) dt Mhh
since h > 0.
Now, since f is continuous on [x, x + h], we know that
lim mh = f (x) = lim Mh
h0+
h0+
and hence
F (x + h) F (x)
= f (x)
+
h
h0
by virtue of the pinching theorem.
lim
231
232
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
= M |x a|.
Accordingly,
|F (x) F (a)| M |x a| 0
The first fundamental theorem of calculus then implies that these functions are continuous and differentiable on the respective (closed or open) intervals.
234
8.6
235
236
F (x)
ib
between x = 0 and x = 1.
It is easy to verify that F given by
F (x) = ex
237
8.7
Indefinite integrals
238
1 n+1
x
+ C,
n+1
n 6= 1
sin x dx = cos x + C
Z
1 ax
e dx = e + C,
a
ax
a 6= 0
2x + 1
2
3x + 2 (1 + 6x) + C
dx =
27
3x + 2
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
2 x3
3x e dx =
x3
240
d 3
x3
x dx = e + C.
dx
Example. Find
1
dx.
x ln x
241
Thus, if we set
u = g(x)
then
Z
du = g (x) dx
x = h(u),
dx = h(u) du.
or, if appropriate,
242
Example. Find
I=
Set
u=
3x + 2
2x + 1
dx.
3x + 2
3
du =
dx,
2 3x + 2
u2 2
.
x=
3
Hence,
I=
Z
2(u 2)
2
+1
du
3
3
4 2 2
u du
=
9
9
4 3 2
= u u+C
27
9
2
4
3/2
3x + 2 + C.
= (3x + 2)
27
9
Note. The last step always consists of rewriting everything in terms of the original
variable!
243
holds.
Proof. Exercise.
Example. Evaluate the definite integral
Z 1
1
I=
dx.
2
4x
0
244
f (x)
a
|
f (x) dx = 2
a
f (x) dx.
0
f (x) dx = 0.
a
245
f (x) dx.
0
Example. Find
Z
37
51
17
sin99 3x dx.
246
8.9
Integration by parts
247
d
I= x
sin x dx
dx
Z
d
= x sin x
x sin x dx
dx
= x sin x + cos x + C.
ln x dx.
1
249
8.10
Improper integrals
R
250
Definition.
(a) Suppose that there exists a real number L such that
Z R
f (x) dx L
a
f (x) dx = L.
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
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 =
[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 =
dx =
2
x
1
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 = 2
f (x) 1 + [f (x)]2 dx.
a
x4 + 1
2
4
= ln x + x + 1
x2
[Verify this!]
b .
as
1
x
1+
1
dx
4
x
diverges.
254
R
0
x dx
1
x dx = x2
2
diverges by definition.
R
0
R2
=
x dx
is irrelevant
x dx = lim 0 = 0
R
R
255
x
1
1p
x
dx =
1p
R
1
R1p 1
=
1p
(
1
p1 when 1 p < 0
when 1 p > 0
as R .
256
If p = 1 then
Z
R
1
h
iR
1
dx = ln x = ln R ln 1 as R .
1
x
257
8.11
(ii) If
f (x) dx converges.
g(x) dx diverges.
a
258
g(x) dx
a
259
x2
dx.
e >x
for x 6= 0 and that the p-integral
converges.
1
dx
x2
x2
1
< 2
x
ex dx.
1
dx.
ln x
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.
The precise formulation of this idea is as follows:
262
Theorem (Limit form of the comparison test.) Suppose that f and g are
nonnegative and bounded on [a, ). If
f (x)
=L
lim
x g(x)
or
both
f (x) dx and
263
g(x) dx converge
g(x) dx diverge.
sin x + x2
,
2x4 1
f (x) 1
=
lim
x g(x)
2
Z
1
dx
x3
264
1
g(x) = 3 .
x
I=
( x x 3 ) dx.
3
265
266
Chapter 9
The logarithmic and exponential functions
In the preceding, we have manipulated functions such as
ln x,
ex ,
267
(9.1)
where s and t are independent variables. It is evident that f = ln constitutes one solution of this functional equation. Are there other functions f obeying this functional
equation?
Answer. We first note that (9.1) evaluated at s = t = 1 yields
f (1) = 0.
Moreover, differentiation with respect to s leads to
tf (st) = f (s)
so that
1
f (t) = f (1)
t
at s = 1.
If we now demand that f (1) = 1 then f is uniquely determined via integration since
f (1) = 0.
268
x
1
269
1
dt.
t
9.1
1/t
ln : (0, ) R
is defined by the formula
ln x =
x
1
1
dt.
t
It is evident that ln x is simply the area of the shaded region shown below in the case
when x 1.
270
1
d
ln x = .
dx
x
ln(xy) = ln x + ln y.
x
ln
= ln(x) ln(y).
y
ln(xr ) = r ln x.
271
Proof.
(i) Apply the first fundamental theorem of calculus to the definition of ln.
(ii) This follows from the definition of ln and the fact that
1
t
2
1
1
dt
1 ,
t
2
y=
1
t
4
2
1
dt
2 ,
t
4
8
4
8
272
1
dt
4 .
t
8
In general,
Z
as n .
Hence the improper integral
2n
1
1
dt 1 1
+ + +
t
2}
|2 2 {z
n terms
n
=
2
1
dt
t
1
diverges to infinity and, therefore, ln x as x .
t
n
2
as n . Hence ln x as x 0+.
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
275
9.3
Question. It is recalled that we have already defined xr for r Q. Can this definition be extended to r 6 Q?
Definition. The function
exp : R (0, )
y = exp x
1
b
(e, 1)
1
y = ln x
276
and
er
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.
(v) For any x and y, we have
as required.
279
280
Note. The above definition merely means that ex is the unique real number R such
that
Z R
1
dt = x
t
1
for any x R. By construction, the function
f : R R,
f (x) = exp x = ex
281
9.4
fb(x) = bx = exp(x ln b)
283
y = 3x
y = 2x
y = 8x
y
y = log2 x
y = log3 x
y = log8 x
x
In particular, loge x = ln x.
Remark. The above definition may be stated as
y = bx
x = logb y.
The following theorem demonstrates that all logarithm functions are just scaled versions of the natural logarithm function ...
284
ln x
ln b
Remark. The above theorem implies that the class of logarithm functions represents
the general solution of the functional equation (9.1).
286
9.5
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.
Moreover, if a function f is positive and differentiable then
so that
f (x)
d
ln f (x) =
dx
f (x)
Z
f (x)
dx = ln f (x) + C.
f (x)
287
f (x)
dx = ln |f (x)| + C
f (x)
provided that f is differentiable and does not vanish on the interval of integration.
1
dx.
2 sec x + tan x
288
9.6
Logarithmic differentiation
Logarithms are powerful in that they transform powers into products, products into
sums and quotients into differences.
Example. Find the derivative of
2
3/5
(3x + 4)(x + 2)
.
y=
x3 + 5x
The idea is to take ln of both sides of the equation to obtain
2
3
(3x + 4)(x + 2)
ln y = ln
5
x3 + 5x
3
2
3
= ln(3x + 4) + ln(x + 2) ln(x + 5x) .
5
Differentiating both sides with respect to x is relatively easy and leads to
2
1 dy 3
1
3x + 5
6x
=
+
.
y dx 5 3x2 + 4 x + 2 x3 + 5x
289
Hence, we obtain
dy 3
=
dx 5
(3x + 4)(x + 2)
x3 + 5x
3/5
6x
1
3x + 5
+
.
2
3
3x + 4 x + 2 x + 5x
Remark. The above procedure is only valid for intervals on which y > 0.
290
f (x) = xsin x.
9.7
Example. Find
L = lim xsin x.
x0+
x0+
Now,
lim | ln x
sin x
x0+
| = lim | sin x ln x|
x0+
lim |x ln x|
x0+
=0
so that
ln L = 0
291
L = 1.
t
1+
x
292
x
= et ,
Chapter 10
The hyperbolic functions
10.1
cosh 0 = 1.
y = cosh x
y = 21 ex
x
295
y = 12 ex
y = sinh x
x
ex + ex
2
2
ex ex
2
2
1 2x
2x
2x
2x
= (e + 2 + e ) (e 2 + e )
4
= 1.
296
x > 0.
10.2
Other hyperbolic functions are defined in analogy with the trigonometric functions
according to
sinh x
tanh x =
,
cosh x
1
sech x =
,
cosh x
cosh x
coth x =
,
sinh x
1
cosech x =
.
sinh x
298
Recall that
sinh x ex ex
tanh x =
= x
cosh x e + ex
lim tanh x = 1.
x
299
tanh x
1
x
1
300
10.3
Hyperbolic identities
1 tanh2 x = sech2 x
coth2 x 1 = cosech2 x
tanh x tanh y
1 tanh x tanh y
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
d
cosh x = sinh x
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
Corresponding indefinite integrals are, for instance,
Z
Z
sinh x dx = cosh x + C,
sech2 x dx = tanh x + C.
303
I=
dx.
x
0
304
10.5
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
1
x
sinh1
cosh1
305
tanh1
sinh1 : R R
tanh1 : (1, 1) R
sinh : R R
306
Example. Evaluate
sinh
cosh1 34
Evaluation of
sinh2 t = cosh2 t 1
at t = cosh1 34 yields
sinh
cosh1 34
Accordingly,
sinh
cosh1 34
since t > 0.
307
4 2
3
1 = 97 .
=+
7
3
x
1
+
,
2 2x
308
x > 0.
The relation
ex ex
y = sinh x =
2
may be written as
so that
(ex)2 2yex 1 = 0
4y 2 + 4
e =
2
p
x = ln y + y 2 + 1 .
x
2y +
In a similar manner, the remaining identities of the following theorem may be proven.
309
1
sinh x = ln x + x2 + 1
p
cosh1 x = ln x + x2 1
1
1+x
tanh1 x = ln
.
2
1x
The derivatives of these functions are therefore given by
1
d
sinh1 x =
dx
x2 + 1
1
d
1
cosh x =
dx
x2 1
d
1
tanh1 x =
.
2
dx
1x
310
1 + sinh2 t,
x) =
1 + x2
dy
1
.
=
2
dx
x +1
311
10.6
dx
x
= sinh1 + C
x2 + a2
a
= ln x + x2 + a2 + C,
dx
1 x
+C
= cosh
2
2
a
x a
p
= ln x + x2 a2 + C,
312
a>0
xa>0
1 x
Z
+ C, |x| < a
tanh
dx
a
a
=
x
1
a2 x2
x2 6= a2.
[These formulae are included in the table of standard integrals which is issued at the
final examination.]
313
.
2
x 2x + 10
If we complete the square, that is,
x2 2x + 10 = (x 1)2 + 9 = u2 + 32,
where u = x 1, then
Z
du
u2 + 3 2
1 u
+C
= sinh
3
x1
= sinh1
+ C.
3
dx
=
2
x 2x + 10
314