0% found this document useful (0 votes)

27 views111 pages

Unit 1 Sem V

The document discusses metric topology, focusing on the definition and examples of metric spaces, including various metrics such as Euclidean, sum, and sup metrics. It covers concepts like open sets, closed sets, and properties of metric spaces, along with specific examples and proofs. The content is structured into units that systematically explore the properties and applications of metric spaces.

Uploaded by

8698136860

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

27 views111 pages

Unit 1 Sem V

Uploaded by

8698136860

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 111

METRIC TOPOLOGY

Dr. Abhaya Chitre

February 27, 2020

Contents

1 Unit I: Metric Spaces 1

1.1 Definition, examples of metric spaces . . . . . . . . . . . . . . . . . 1
1.1.1 Examples of metric spaces . . . . . . . . . . . . . . . . . . . . 1
1.1.2 Some More Examples of metric spaces . . . . . . . . . . . . . 16
1.1.3 Metrics derived from some other metrics defined on the
same space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.1.4 Product of two metric spaces . . . . . . . . . . . . . . . . . . 34
1.1.5 Some Examples of functions that are not metrics . . . . . . 40
1.2 Distance Metric Induced by a Norm . . . . . . . . . . . . . . . . . . 42
1.2.1 Examples of Normed Linear Spaces and Distance Induced
by Them . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
1.3 Metric Subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
1.4 Open balls in a metric space . . . . . . . . . . . . . . . . . . . . . . . 53
1.4.1 Examples of open balls in various metric spaces. . . . . . . 53
1.5 Hausdorff Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
1.5.1 Open Balls in a Subspace . . . . . . . . . . . . . . . . . . . . . 66
1.6 Open Sets In A Metric Space . . . . . . . . . . . . . . . . . . . . . . 67
1.7 Properties of open sets . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
1.7.1 Examples of Open Sets and Non Open Sets in Various Met-
ric Spaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
1.8 Open sets in Subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
1.9 Structure of an open set in R . . . . . . . . . . . . . . . . . . . . . . . 80
1.9.1 Translation invariance of the metric induced by norm . . . 82
1.10 Interior of a set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
1.11 Equivalent metric spaces . . . . . . . . . . . . . . . . . . . . . . . . . . 85
1.12 Distance of a point from a set in a metric space . . . . . . . . . . . 92
1.13 Distance between two sets in a metric space . . . . . . . . . . . . . 94
1.14 Diameter of a set in a metric space . . . . . . . . . . . . . . . . . . . 97
1.15 Closed balls in a metric space . . . . . . . . . . . . . . . . . . . . . . . 98
1.16 Closed Sets in a metric space . . . . . . . . . . . . . . . . . . . . . . 98
1.16.1 Closed Sets in a sub space . . . . . . . . . . . . . . . . . . . . 101
1.16.2 Examples of closed sets in various metric spaces. . . . . . . . 103
1.17 Limit Point of a set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
1.18 Closure of a set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
ii CONTENTS

1.19 Dense Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

1.20 boundary of a set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Chapter 1

Unit I: Metric Spaces

Definition, examples of metric spaces R, R2 ,Euclidean space Rn with its Eu-

clidean, sup and sum metric, C (complex numbers), the spaces l 1 and l 2 of
sequences and the space C[a, b], of real valued continuous functions on [a, b].
Discrete metric space.
Distance metric induced by the norm, translation invariance of the metric in-
duced by the norm. Metric subspaces, Product of two metric spaces. Open balls
and open set in a metric space, examples of open sets in various metric spaces.
Hausdorff property. Interior of a set. Properties of open sets. Structure of an
open set in IR. Equivalent metrics.
Distance of a point from a set, between sets ,diameter of a set in a metric space
and bounded sets. Closed ball in a metric space, Closed sets- definition, exam-
ples. Limit point of a set, isolated point, a closed set contains all its limit points,
Closure of a set and boundary of a set.

1.1 Definition, examples of met-

ric spaces
Definition 1.1.1 Let X be a non empty set. A function d : X × X −→ R is said to be a
metric on X if it satisfies the following conditions:
(i) d(x, y) ≥ 0 for all x, y ∈ X and d(x, y) = 0 if and only if x = y.
(ii) d(x, y) = d(y, x) for all x, y ∈ X.
(iii) d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ X.
The ordered pair (X, d) is called a metric space

1.1.1 Examples of metric spaces

Examples 1.1.2
2 CHAPTER 1. UNIT I: METRIC SPACES

(1) d : R × R −→ R defined as d(x, y) = |x − y|. Then d is a metric on R and d is called

as the usual metric on R.

(2) d1 : R2 × R2 −→ R defined as d1 (x, y) = |x1 − y1 | + |x2 − y2 | for all x = (x1 , x2 ), y =

(y1 , y2 ) ∈ R2 . Then d1 is a metric on R2 and d1 is called as the sum metric on R2 .
p
(3) d2 : R2 × R2 −→ R defined as d2 (x, y) = (x1 − y1 )2 + (x2 − y2 )2 for all x =
(x1 , x2 ), y = (y1 , y2 ) ∈ R2 . Then d2 is a metric on R2 and d2 is called as the
Euclidean metric on R2 .

(4) d∞ : R2 × R2 −→ R defined as d∞ (x, y) = max{|x1 − y1 |, |x2 − y2 |} for all x =

(x1 , x2 ), y = (y1 , y2 ) ∈ R2 . Then d∞ is a metric on R2 and d∞ is called as the sup
metric on R2 .
n
X
n n
(5) d1 : R × R −→ R defined as d1 (x, y) = |xi − yi | for all x = (x1 , x2 , · · · , xn ), y =
i=1
(y1 , y2 , · · · , cn ) ∈ Rn . Then d1 is a metric on Rn and d1 is called as the sum metric
on Rn .
s n
X
(6) d2 : Rn ×Rn −→ R defined as d(x, y) = (xi − yi )2 for all x = (x1 , x2 , · · · , xn ), y =
i=1
(y1 , y2 , · · · , yn ) ∈ Rn . Then d2 is a metric on R2 and d2 is called as the Euclidean
metric on Rn .

(7) d∞ : Rn × Rn −→ R defined as d∞ (x, y) = max{|xi − yi | : 1 ≤ i ≤ n} for all

x = (x1 , x2 , · · · , xn ), y = (y1 , y2 , · · · , yn ) ∈ R2 . Then d∞ is a metric on R2 and d∞
is called as the sup metric on Rn .

(8) d : C × C −→ R, d(z, w) = |z − w| is a metric on C.

∞
X
(9) Let l 1 = {(an )∞
n=1 : |an | < ∞}. Define d1 : l 1 × l 1 −→ R, d1 ((an ), (bn )) =
n=1
∞
X
|an − bn |. Then d1 is a metric on l 1 .
n=1

∞
X
(10) Let l = 2
{(an )∞
: n=1 an2 < ∞}. Define d2 : l 2 × l 2 −→ R, d2 ((an ), (bn )) =
s∞ n=1
X
(an − bn )2 . Then d2 is a metric on l 2 .
n=1

(11) Let C [a, b] denote the vecotr space of all real valued continuous functions on [a, b].
If f, g ∈ C [a, b] define d∞ (f, g) = sup{|(f − g)(t)| : t ∈ [a, b]}. Then d∞ is a metric
on C [a, b].
3

(12) Discrete metric Let X be a nonempty ( set and k > 0.

k if x 6= y,
Define d : X × X −→ R, d(x, y) =
0 if x = y.
Then d is called a discrete metirc on X and (X, d) is called a discrete metric space.
(Usually we take k = 1)

We will prove the above statements.

(1) d : R × R −→ R defined as d(x, y) = |x − y|. Then d is a metric on R and d is called

as the usual metric on R.

To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .

Since |a| ≥ 0 for all a ∈ R, |x − y| ≥ 0. Hence d(x, y) ≥ 0 for all x, y ∈ R.

d(x, y) = 0 if and only if |x − y| = 0.

if and only if x − y = 0
if and only if x = y

Thus non-negativity is proved. (I)

To show that d(x, y) = d(y, x) for all x, y ∈ R.

d(x, y) = |x − y|.
= |y − x|
= d(y, x) ∀ x, y ∈ R

Thus symmetry is proved. (II)

To show that d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ R .
First we will prove that |x + y| ≤ |x| + |y| for all x, y ∈ R.
|x + y| = max{x + y, −(x + y)}
If |x + y| = x + y then since x ≤ |x| and y ≤ |y|, we have x + y ≤ |x| + |y| and
therefore |x + y| ≤ |x| + |y|.
If |x + y| = −(x + y) = −x − y = −x + (−y) then since −x ≤ |x| and −y ≤ |y|,
we have −x + (−y) ≤ |x| + |y| and therefore |x + y| ≤ |x| + |y|.
Thus in any case |x + y| ≤ |x| + |y|.

d(x, z) = |x − z|
= |x − y + y − z|
≤ |x − y| + |y − z|
= d(x, y) + d(y, z)
∴ d(x, z) ≤ d(x, y) + d(y, z) ∀ x, y, z ∈ R. (III)
4 CHAPTER 1. UNIT I: METRIC SPACES

Hence from (I), (II) and (III) d is a metric on R.

(2) d1 : R2 × R2 −→ R defined as d1 (x, y) = |x1 − y1 | + |x2 − y2 | for all x = (x1 , x2 ), y =

(y1 , y2 ) ∈ R2 . Then d1 is a metric on R2 and d1 is called as the sum metric on R2 .

To show that d1 (x, y) ≥ 0 and d1 (x, y) = 0 if and only if x = y .

Since |a| ≥ 0 for all a ∈ R, |x1 − y1 | ≥ 0 and |x2 − y2 | ≥ 0.
Therefore |x1 − y1 | + |x2 − y2 | ≥ 0. Hence d1 (x, y) ≥ 0 for all x, y ∈ R2 .

d1 (x, y) = 0 if and only if |x1 − y1 | + |x2 − y2 | = 0.

if and only if x1 − y1 = 0 and x2 − y2 = 0
if and only if x1 = y1 and x2 = y2
if and only if x=y

Thus non-negativity is proved. (I)

To show that d1 (x, y) = d1 (y, x) for all x, y ∈ R2 .

d1 (x, y) = |x1 − y1 | + |x2 − y2 |.

== |y1 − x1 | + |y2 − x2 |.
= d1 (y, x) ∀ x, y ∈ R2

Thus symmetry is proved. (II)

To show that d1 (x, y) ≤ d1 (x, z) + d1 (z, y) for all x, y, z ∈ R2 .

d1 (x, z) = |x1 − z1 | + |x2 − z2 |

= |x1 − y1 + y1 − z1 | + |x2 − y2 + y2 − z2 |
≤ |x1 − y1 | + |y1 − z1 | + |x2 − y2 | + |y2 − z2 |
= |x1 − y1 | + |x2 − y2 | + |y1 − z1 | + |y2 − z2 |
= d1 (x, y) + d1 (y, z)
∴ d1 (x, z) ≤ d1 (x, y) + d1 (y, z) ∀ x, y, z ∈ R2 . (III)

Hence from (I), (II) and (III) d1 is a metric on R2 .

p
(3) d2 : R2 × R2 −→ R defined as d2 (x, y) = (x1 − y1 )2 + (x2 − y2 )2 for all x =
(x1 , x2 ), y = (y1 , y2 ) ∈ R . Then d2 is a metric on R2 and d2 is called as the
2

Euclidean metric on R2 .

To show that d2 (x, y) ≥ 0 and d2 (x, y) = 0 if and only if x = y .

Since (x1 −py1 )2 ≥ 0 and (x2 − y2 )2 ≥ 0, (x1 − y1 )2 + (x2 − y2 )2 ≥ 0.
Therefore (x1 − y1 )2 + (x2 − y2 )2 ∈ R.p
√
As a ≥ 0 whenever it exists, we have (x1 − y1 )2 + (x2 − y2 )2 ≥ 0.
5

Hence d2 (x, y) ≥ 0 for all x, y ∈ R2 .

p
d2 (x, y) = 0 if and only if (x1 − y1 )2 + (x2 − y2 )2 = 0.
if and only if (x1 − y1 )2 + (x2 − y2 )2 = 0.
if and only if x1 − y1 = 0 and x2 − y2 = 0.
if and only if x1 = y1 and x2 = y2 .
if and only if x = y.

Thus non-negativity is proved. (I)

To show that d2 (x, y) = d2 (y, x) for all x, y ∈ R2 .

p
d2 (x, y) = (x1 − y1 )2 + (x2 − y2 )2 .
p
= (y1 − x1 )2 + (y2 − x2 )2 .
= d2 (y, x) ∀ x, y ∈ R2 .

Thus symmetry is proved. (II)

To show that d2 (x, y) ≤ d2 (x, z) + d2 (z, y) for all x, y, z ∈ R2 .

p
d2 (x, z) = (x1 − z1 )2 + (x2 − z2 )2
p
= (x1 − y1 + y1 − z1 )2 + (x2 − y2 + y2 − z2 )2
p √ √
By Minkowski’s Inequality : (a + b)2 + (c + d)2 ≤ a2 + c2 + b2 + d2
p p
≤ (x1 − y1 )2 + (x2 − y2 )2 (y1 − z1 )2 + (y2 − z2 )2
= d2 (x, y) + d2 (y, z)
∴ d2 (x, z) ≤ d2 (x, y) + d2 (y, z) ∀ x, y, z ∈ R2 . (III)

Hence from (I), (II) and (III) d2 is a metric on R2 .

(4) d∞ : R2 × R2 −→ R defined as d∞ (x, y) = max{|x1 − y1 |, |x2 − y2 |} for all x =

(x1 , x2 ), y = (y1 , y2 ) ∈ R2 . Then d∞ is a metric on R2 and d∞ is called as the sup
metric on R2 .

To show that d∞ (x, y) ≥ 0 and d∞ (x, y) = 0 if and only if x = y .

Since |x1 − y1 | ≥ 0 and |x2 − y2 | ≥ 0, max{|x1 − y1 |, |x2 − y2 |} ≥ 0.
Hence d∞ (x, y) ≥ 0 for all x, y ∈ R2 .
6 CHAPTER 1. UNIT I: METRIC SPACES

d∞ (x, y) = 0 if and only if max{|x1 − y1 |, |x2 − y2 |} = 0.

if and only if |x1 − y1 | = 0 and |x2 − y2 | = 0
if and only if x1 − y1 = 0 and x2 − y2 = 0
if and only if x1 = y1 and x2 = y2
if and only if x=y

Thus non-negativity is proved. (I)

To show that d∞ (x, y) = d∞ (y, x) for all x, y ∈ R2 .

d∞ (x, y) = max{|x1 − y1 |, |x2 − y2 |}.

= max{|y1 − x1 |, |y2 − x2 |}.
= d∞ (y, x) ∀ x, y ∈ R2

Thus symmetry is proved. (II)

To show that d∞ (x, y) ≤ d∞ (x, z) + d∞ (z, y) for all x, y, z ∈ R2 .

d∞ (x, z) = max{|x1 − z1 |, |x2 − z2 |}

d∞ (x, y) = max{|x1 − y1 |, |x2 − y2 |} =⇒ |x1 − y1 | ≤ d∞ (x, y)and |x2 − y2
d∞ (y, z) = max{|y1 − z1 |, |y2 − z2 |} =⇒ |y1 − z1 | ≤ d∞ (y, z)and |y2 − z2 |
|x1 − z1 | ≤ |x1 − y1 | + |y1 − z1 |
≤ d∞ (x, y) + d∞ (y, z)
|x2 − z2 | ≤ |x2 − y2 | + |y2 − z2 |
≤ d∞ (x, y) + d∞ (y, z)
max{|x1 − z1 |, |x2 − z2 |} ≤ d∞ (x, y) + d∞ (y, z)
∴ d∞ (x, z) ≤ d∞ (x, y) + d∞ (y, z) ∀ x, y, z ∈ R2 . (III)

Hence from (I), (II) and (III) d∞ is a metric on R2 .

n
X
n n
(5) d1 : R × R −→ R defined as d1 (x, y) = |xi − yi | for all x = (x1 , x2 , · · · , xn ), y =
i=1
(y1 , y2 , · · · , cn ) ∈ Rn . Then d1 is a metric on Rn and d1 is called as the sum metric
on Rn .

To show that d1 (x, y) ≥ 0 and d1 (x, y) = 0 if and only if x = y .

n
X
d1 (x, y) = 0 if and only if |xi − yi | = 0.
i=1
if and only if xi − yi = 0 for all i, 1 ≤ i ≤ n.
if and only if xi = yi for all i, 1 ≤ i ≤ n.
if and only if x = y

Thus non-negativity is proved. (I)

To show that d1 (x, y) = d1 (y, x) for all x, y ∈ Rn .
n
X
d1 (x, y) = |xi − yi |.
i=1
Xn
= |yi − xi |.
i=1
= d1 (y, x) ∀ x, y ∈ Rn .

Thus symmetry is proved. (II)

To show that d1 (x, y) ≤ d1 (x, z) + d1 (z, y) for all x, y, z ∈ Rn .

Hence from (I), (II) and (III) d1 is a metric on Rn .

s n
X
(6) d2 : Rn ×Rn −→ R defined as d(x, y) = (xi − yi )2 for all x = (x1 , x2 , · · · , xn ), y =
i=1
(y1 , y2 , · · · , yn ) ∈ Rn . Then d2 is a metric on R2 and d2 is called as the Euclidean
metric on Rn .

To show that d2 (x, y) ≥ 0 and d2 (x, y) = 0 if and only if x = y .

8 CHAPTER 1. UNIT I: METRIC SPACES

n
X
2
Since (xi − yi ) ≥ 0 for all i, 1 ≤ i ≤ n, (xi − yi )2 ≥ 0.
s n i=1
X
Therefore (xi − yi )2 ∈ R.
i=1 s n
√ X
As a ≥ 0 whenever it exists, we have (xi − yi )2 ≥ 0.
i=1
Hence d2 (x, y) ≥ 0 for all x, y ∈ R2 .

v
u n
uX
d2 (x, y) = 0 if and only if t (xi − yi )2 = 0.
i=1
n
X
if and only if (xi − yi )2 = 0.
i=1
if and only if xi − yi = 0 for all i, 1 ≤ i ≤ n.
if and only if xi = yi for all i, 1 ≤ i ≤ n.
if and only if x = y.

Thus non-negativity is proved. (I)

To show that d2 (x, y) = d2 (y, x) for all x, y ∈ Rn .

v
u n
uX
d2 (x, y) = t (xi − yi )2 .
i=1
v
u n
uX
= t (yi − xi )2 .
i=1

= d2 (y, x) ∀ x, y ∈ Rn .

Thus symmetry is proved. (II)

To show that d2 (x, y) ≤ d2 (x, z) + d2 (z, y) for all x, y, z ∈ R2 .

v
u n
uX
d2 (x, z) = t (xi − zi )2
i=1
v
u n
uX
= t (xi − yi + yi − zi )2
i=1
v v v
u n u n u n
u X uX uX 2
2 2
By Minkowski’s Inequality :t (ai + bi ) ≤ t ai + t bi
i=1 i=1 i=1

Here xi − yi = ai and yi − zi = bi
v v
u n u n
uX uX
d2 (x, z) ≤ t (xi − yi ) + t (yi − zi )2
2

i=1 i=1

= d2 (x, y) + d2 (y, z)
∴ d2 (x, z) ≤ d2 (x, y) + d2 (y, z) ∀ x, y, z ∈ Rn . (III)

Hence from (I), (II) and (III) d2 is a metric on Rn .

(7) d∞ : Rn × Rn −→ R defined as d∞ (x, y) = max{|xi − yi | : 1 ≤ i ≤ n} for all

x = (x1 , x2 , · · · , xn ), y = (y1 , y2 , · · · , yn ) ∈ R2 . Then d∞ is a metric on R2 and d∞
is called as the sup metric on Rn .

To show that d∞ (x, y) ≥ 0 and d∞ (x, y) = 0 if and only if x = y .

Since |xi − yi | ≥ 0 for all i, 1 ≤ i ≤ n, max{|xi − yi | : 1 ≤ i ≤ n} ≥ 0.
Hence d∞ (x, y) ≥ 0 for all x, y ∈ Rn .

d∞ (x, y) = 0 if and only if max{|xi − yi | : 1 ≤ i ≤ n} = 0.

if and only if |xi − yi | = 0 for all i, 1 ≤ i ≤ n.
if and only if xi − yi = 0 for all i, 1 ≤ i ≤ n.
if and only if xi = yi for all i, 1 ≤ i ≤ n.
if and only if x=y

Thus non-negativity is proved. (I)

To show that d∞ (x, y) = d∞ (y, x) for all x, y ∈ Rn .

d∞ (x, y) = max{|xi − yi | : 1 ≤ i ≤ n}.

= max{|yi − xi | : 1 ≤ i ≤ n.}.
= d∞ (y, x) ∀ x, y ∈ Rn
10 CHAPTER 1. UNIT I: METRIC SPACES

Thus symmetry is proved. (II)

To show that d∞ (x, y) ≤ d∞ (x, z) + d∞ (z, y) for all x, y, z ∈ Rn .

d∞ (x, z) = max{|xi − zi | : 1 ≤ i ≤ n}
d∞ (x, y) = max{|xi − yi |, 1 ≤ i ≤ n} =⇒ |xi − yi | ≤ d∞ (x, y)for all i, 1 ≤
d∞ (y, z) = max{|yi − zi |, 1 ≤ i ≤ n} =⇒ |yi − zi | ≤ d∞ (y, z)for all i, 1 ≤
|xi − zi | ≤ |xi − yi | + |yi − zi | for all i, 1 ≤ i ≤ n.
≤ d∞ (x, y) + d∞ (y, z) for all i, 1 ≤ i ≤ n.
max{|xi − zi | : 1 ≤ i ≤ n} ≤ d∞ (x, y) + d∞ (y, z)
∴ d∞ (x, z) ≤ d∞ (x, y) + d∞ (y, z) ∀ x, y, z ∈ Rn . (III)

Hence from (I), (II) and (III) d∞ is a metric on R2 .

p
(8) d : C × C −→ R, d(z, w) = |z − w| is a metric on C where for z ∈ C, |z| = x2 + y 2
if z = x + iy.

To show that d(z, w) ≥ 0 and d(z, w) = 0 if and only if z = w .

Since |a| ≥ 0 for all a ∈ C, |z − w| ≥ 0. Hence d(z, w) ≥ 0 for all z, w ∈ C.

d(z, w) = 0 if and only if |z − w| = 0.

if and only if z − w = 0
if and only if z = w

Thus non-negativity is proved. (I)

To show that d(z, w) = d(w, z) for all z, w ∈ C.

d(z, w) = |z − w|.
= |w − z|
= d(w, z) ∀ z, w ∈ C

Thus symmetry is proved. (II)

To show that d(z, w) ≤ d(z, u) + d(u, w) for all z, u, w ∈ C .

First we will prove that |z + w| ≤ |z| + |w| for all z, w ∈ C.

|z + w|2 = (z + w)(z + w)
= (z + w)(z + w)
= zz + zw + wz + ww
= |z|2 + zw + zw + |w|2
= |z|2 + zw + zw + |w|2
= |z|2 + 2Re(zw) + |w|2
≤ |z|2 + 2|zw| + |w|2
= |z|2 + 2|z||w| + |w|2
= |z|2 + 2|z||w| + |w|2
= (|z| + |w|)2
Thus |z + w|2 ≤ (|z| + |w|)2
Hence |z + w| ≤ |z| + |w|

d(z, w) = |z − w|
= |z − u + u − w|
≤ |z − u| + |u − w|
= d(z, u) + d(u, w)
∴ d(z, w) ≤ d(z, u) + d(u, w) ∀ z, u, w ∈ C. (III)

Hence from (I), (II) and (III) d is a metric on R.

∞
X
(9) Let l 1
= {(an )∞
n=1 : |an | < ∞}. Define d1 : l 1 × l 1 −→ R, d1 ((an ), (bn )) =
n=1
∞
X
|an − bn |. Then d1 is a metric on l 1 .
n=1
First we will show that d1 is well defined.
|ai − bi | ≤ |ai | + |bi | for all i ∈ N.

We will show that d∞ is well defined .

X
d1 ((an ), (bn )) = |an − bn | for all (an ), (bn ) ∈ l 1 .
n=1
For each n ∈ N, |an − bn | ≤ |an | + |bn |.
∞
X ∞
X
1
Since (an ), (bn ) ∈ l , |an | and |bn | are convergent series.
n=1 n=1
∞
X
Hence the series |an | + |bn | is also convergent.
n=1
12 CHAPTER 1. UNIT I: METRIC SPACES

∞
X
d1 ((an ), (bn )) = 0 if and only if |an − bn | = 0.
n=1
if and only if |an − bn | = 0 ∀ n ∈ N.
if and only if an = bn ∀ n ∈ N.
if and only if (an ) = (bn )

Thus non-negativity is proved. (I)

To show that d1 ((an ), (bn )) = d1 ((bn ), (an )) for all (an ), (bn ) ∈ l 1 .
∞
X
d1 ((an ), (bn )) = |an − bn |.
n=1
X∞
= |bn − an |.
n=1
= d1 ((bn ), (an )) ∀ (an ), (bn ) ∈ l 1 .

Thus symmetry is proved. (II)

To show that d1 ((an ), (cn )) ≤ d1 ((an ), (bn )) + d1 ((bn ), (cn )) for all (an ), (bn ), (cn ) ∈ l 1 .
.

|an − cn | ≤ |an − bn | + |bn − cn | for all n ∈ N.

Hence d1 ((an ), (cn )) ≤ d1 ((an ), (bn ))+d1 ((bn ), (cn )) ∀ (an ), (bn ), (cn ) ∈ l 1 . (III )

From (I), (II) and (III) d1 is a metric on l 1 .

∞
X
(10) Let l 2
= {(an )∞
n=1 : an2 < ∞}. Define d2 : l 2 × l 2 −→ R, d2 ((an ), (bn )) =
n=1
13

∞
s
X
(an − bn )2 . Then d2 is a metric on l 2 .
n=1

We will show that s d2 is well defined .

X∞
d2 ((an ), (bn )) = (an − bn )2 for all (an ), (bn ) ∈ l 2 .
vn=1 v v
u n n
u n 2
uX uX uX
2 2
u
For each n ∈ N, t (ak − bk ) ≤ t ak + t bk . (∗)
k=1 k=1 k=1
∞
X ∞
X
2
Now, (an ), (bn ) ∈ l =⇒ an2 < ∞ and b2n < ∞.
n=1 n=1
∞
X ∞
X
That is a2n = α and b2n = β for some α, β ∈ R.
n=1 n=1
Xn Xn
Since a2n , b2n ≥ 0 for all n ∈ N, we have a2k ≤ α and b2k ≤ β and α, β ≥ 0.
v v k=1 k=1
n n
√ √ √ √
uX uX
a2k ≤ α and t b2k ≤ β. (note that α, β ≥ 0 =⇒ α, β ∈ R.)
u u
t
k=1 k=1
v
u n √ √
uX
Now, From (∗), for each n ∈ N, t (ak − bk )2 ≤ α + β
k=1
n √
X p 2
That is (ak − bk )2 ≤ α+ β
k=1
∞
X
Thus the sequence of partial sums of the series (an − bn )2 is bounded above.
n=1
!
n
X
2
Since (an −bn ) ≥ 0, the sequence (ak − bk )2 of partial sums of the series
k=1
∞
X
(an − bn )2 is monotonic increasing.
n=1 !
Xn ∞
X
2
Hence the sequence (ak − bk ) of partial sums of the series (an − bn )2
k=1 n=1
is monotonic increasing and bounded above and therefore it is convergent.
X∞
Thus (an − bn )2 < ∞.
n=1
s∞
X
Hence (an − bn )2 exists and is unique. Therefore d2 ((an ), (bn )) is well de-
n=1
fined.

To show that d2 ((an ), (bn )) ≥ 0 and d2 ((an ), (bn )) = 0 if and only if (an ) = (bn )
14 CHAPTER 1. UNIT I: METRIC SPACES

. s∞
√ X
Since a ≥ 0 whenever it exists, (an − bn )2 ≥ 0.
n=1
v
u∞
uX
d2 ((an ), (bn )) = 0 if and only if t (an − bn )2 = 0.
n=1

if and only if (an − bn )2 = 0 ∀ n ∈ N.

if and only if (an − bn ) = 0 ∀ n ∈ N.
if and only if an = bn ∀ n ∈ N.
if and only if (an ) = (bn )

Thus non-negativity is proved. (I)

To show that d2 ((an ), (bn )) = d2 ((bn ), (an )) for all (an ), (bn ) ∈ l 2 .
v
u∞
uX
d2 ((an ), (bn )) = t (an − bn )2 .
n=1
v
u∞
uX
= t (bn − an )2 .
n=1

= d2 ((bn ), (an )) ∀ (an ), (bn ) ∈ l 2 .

Thus symmetry is proved. (II)

To show that d2 ((an ), (cn )) ≤ d2 ((an ), (bn )) + d2 ((bn ), (cn )) for all (an ), (bn ), (cn ) ∈ l 2 .
.
v v
u n u n
u X uX
t (ak − ck )2 = t (ak − bk + bk − ck )2 for all n ∈ N.
k=1 k=1
v v
u n u n
uX uX
≤ t (ak − bk ) + t (ak − ck )2
2
for all n ∈ N.
k=1 k=1

(Minkowski’s Inequality)
v v v
u∞ u∞ u∞
uX uX uX
∴ t (an − cn ) ≤ t (an − bn ) + t (bn − cn )2
2 2

n=1 n=1 n=1

(since all three series are convergent)

Hence d2 ((an ), (cn )) ≤ d2 ((an ), (bn ))+d2 ((bn ), (cn )) ∀ (an ), (bn ), (cn ) ∈ l 2 . (III )

From (I), (II) and (III) d2 is a metric on l 2 .

(11) Let C [a, b] denote the vector space of all real valued continuous functions on [a, b].
If f, g ∈ C [a, b] define d∞ (f, g) = sup{|(f − g)(t)| : t ∈ [a, b]}. Then d∞ is a metric
on C [a, b].
First we show that d∞ is well defined.
Since f, g ∈ C[a, b], f − g ∈ C[a, b]. So f − g is continuous on [a, b],
Hence f − g is bounded and attains its bounds.
Thus sup{|(f − g)(t)| : t ∈ [a, b]} exists. Also sup S is unique whenever it exists for
all S ⊂ R.
Therefore d∞ d(f, g) is well defined.

To show that d∞ (f, g) ≥ 0 and d∞ (f, g) = 0 if and only if f = g .

d∞ (f, g) = 0 if and only if sup{|(f − g)(t)| : t ∈ [a, b]} = 0.

Since 0 ≤ |f (t) − g(t)| ≤ sup{|(f − g)(t)| : t ∈ [a, (b]} = 0 ∀ t ∈ [a, b], we have,
d∞ (f, g) = 0 if and only if |f (t) − g(t)| = 0 ∀ t ∈ [a, b].
if and only if f (t) − g(t) = 0 ∀ t ∈ [a, b].
if and only if f (t) = g(t) ∀ t ∈ [a, b].
if and only if f = g

Thus non-negativity is proved. (I)

To show that d∞ (f, g) = d∞ (g, f ) for all f, g ∈ C [a, b].
Let f, g ∈ C [a, b].

d∞ (f, g) = sup{|(f − g)(t)| : t ∈ [a, b]}.

= sup{|f (t) − g(t)| : t ∈ [a, b]}
= sup{|g(t) − f (t)| : t ∈ [a, b]}.
= sup{|(g − f )(t)| : t ∈ [a, b]}.
= d∞ (g, f )

Thus symmetry is proved. (II)

To show that d∞ (f, h) ≤ d∞ (f, g) + d∞ (g, h) for all f, g, h ∈ C [a, b].

|f (t) − (t)| ≤ |f (t) − g(t)| + |g(t) − h(t)| ∀ t ∈ [a, b]

≤ {sup{|(f − g)(t)| : t ∈ [a, b]} + {sup{|(g − h)(t)| : t ∈ [a
= d∞ (f, g) + d∞ (g, h) ∀ t ∈ [a, b].
∴ sup{{|(f − h)(t)| : t ∈ [a, b]}d∞ (f, g) + d∞ (g, h).
Hence d∞ (f, h) ≤ d∞ (f, g) + d∞ (g, h).

Hence d∞ (f, h) ≤ d∞ (f, g) + d∞ (g, h) ∀ f, g, h ∈ C [a, b]. (III )

16 CHAPTER 1. UNIT I: METRIC SPACES

From (I), (II) and (III) d2 is a metric on C [a, b].

(12) Discrete metric Let X be a nonempty ( set and k > 0.

k if x 6= y,
Define d : X × X −→ R, d(x, y) =
0 if x = y.
Then d is called a discrete metric on X and (X, d) is called a discrete metric space.
(Usually we take k = 1) (
1 if x 6= y,
d : X × X −→ R, d(x, y) = .
0 if x = y.
We will show that d is a metric on X.

To show that d∞ (x, y) ≥ 0 and d∞ (x, y) = 0 if and only if x = y for all x, y ∈

X.
If x = y then d(x, y) = 0 and if x 6= y then d(x, y) = 1 > 0.
Thus d(x, y) ≥ 0 for all x, y ∈ X.
If x = y then d(x, y) = 0
If x 6= y then d(x, y) = 1 =⇒ d(x, y) 6= 0.
Thus d(x, y) = 0 if and only if x = y.
Thus non-negativity is proved. (I)
To show that d(x, y) = d(y, x) for all x, y ∈ X .
Case I: If x = y then d(x, y) = 0.
Also x = y =⇒ y = x and hence d(y, x) = 0.
Thus d(x, y) = d(y, x)
Case II: If x 6= y then d(x, y) = 1.
Also x 6= y =⇒ y 6= x and hence d(y, x) = 1.
Thus d(x, y) = d(y, x)
So, in any case d(x, y) = d(y, x) for all x, y ∈ X. Symmetry is proved. (II)
To show that d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ X .
Case 1: xneqy or y 6= z. This implies d(x, y) = 1 or d(z, y) = 1.
So, d(x, y) + d(y, z) ≥ 1. (as d(x, y) ≥ 0 and d(z, y) ≥ 0)
Now, d(x, z) = 1 or d(x, z) = 0. This means d(x, z) ≤ 1.
Hence d(x, z) ≤ d(x, y) + d(y, z).
Case 2: x = y and y = z. This implies d(x, y) = 0 or d(z, y) = 0.
Also, x = y, y = z =⇒ x = z. So, d(x, z) = 0.
Thus d(x, z) = d(x, y) + d(y, z).
Hence in any case d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ X. (III)

Hence from (I), (II) and (III) d is a metric on X.

1.1.2 Some More Examples of metric

spaces
17

(
0, if x = y
(1) Let d : R × R −→ R be defined as d(x, y) =
|x| + |y| 6 y, x, y ∈ R.
if x =
Then d is a metric on R.

(2) Let d : N × N −→ R be defined as

d(m, n) = 0 if m = n
1
= 1+ if m 6= n
m+n
Then d is a metric on N.
1 1
(3) Let d : N × N −→ R be defined as d(m, n) = − . Then d is a metric on N.
m n

1 1
(4) Let d : R \ {0} × R \ {0} −→ R be defined as d(x, y) = − . Then d is a metric
x y
on R \ {0}.

(5) Let X be the set of real sequences and d : X × X −→ R be defined by

d(x, y) = 0 if x = y
1
= if x 6= y
min {i : xi 6= yi }

where x = (xn ), y = (yn ) ∈ X is a metric on X.

(6) Let X be the set of real sequences. If d : X × X −→ R is defined by

∞
X |xi − yi |
d(x, y) =
i=1
2i (1 + |xi − yi |)

where x = (xn ), y = (yn ) ∈ X then d is a metric on X.

We will prove above statements.

(
0, if x = y
(1) Let d : R × R −→ R be defined as d(x, y) =
|x| + |y| 6 y, x, y ∈ R.
if x =
Then d is a metric on R.

To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .

Case 1: x = y .Then by definition of d, d(x, y) = 0.
Case 2: x 6= y . Then x and y both cannot be zero. Hence |x| + |y| > 0, that
is d(x, y) > 0.
18 CHAPTER 1. UNIT I: METRIC SPACES

Thus d(x, y) ≥ 0 for all x, y ∈ X.

Also, x = y =⇒ d(x, y) = 0 and x 6= y =⇒ d(x, y) > 0, that is d(x, y) 6= 0.
Hence d(x, y) = 0 if and only if x = y. Thus non-negativity is proved. (I)

To show that d(x, y) = d(y, x) for all x, y ∈ X. .

Case 1: x = y

x = y =⇒ y = x
Also x = y =⇒ d(x, y) = 0
And y = x =⇒ d(y, x) = 0
Hence x = y =⇒ d(x, y) = d(y, x)

Case 2: x 6= y .

x 6= y =⇒ y 6= x
Also x 6= y =⇒ d(x, y) = |x| + |y|
And y 6= x =⇒ d(y, x) = |y| + |x| = |x| + |y|
Hence x 6= y =⇒ d(x, y) = d(y, x)

Hence d(x, y) = d(y, x) for all x, y ∈ X.

Thus symmetry is proved. (II)

To show that d(x, y) ≤ d(x, z) + d(z, y) for all x, y, z ∈ R .

Case 1: x = y So, d(x, y) = 0. (by (I))

Since d(x, z) ≥ 0 and d(z, y) ≥ 0,
we have d(x, z) + d(z, y) ≥ 0.
Put 0 = d(x, y).
So d(x, z) + d(z, y) ≥ d(x, y), that is d(x, y) ≤ d(x, z) + d(z, y).
Case 2: x 6= y This implies d(x, y) = |x| + |y|.
6 y implies the case x = z and y = z does not arise.
Also x =
Case 2a: x = z, y 6= z. This implies d(x, z) = 0 and d(y, z) = |y| + |z|.
As x = z, |x| = |z|. So, d(y, z) = |y| + |z| =⇒ d(y, z) = |y| + |x| =
d(x, y).

d(x, y) = d(y, z)
= 0 + d(y, z)
= d(x, z) + d(y, z)

Case 2b: x 6= z, y = z. This implies d(x, z) = |x|+|z| and d(y, z) = 0. As

y = z, |y| = |z|. So, d(x, z) = |x| + |z| =⇒ d(x, z) = |x| + |y| = d(x, y).

d(x, y) = d(x, z)
= d(x, z) + 0
= d(x, z) + d(y, z)

Case 2c: x 6= z, y 6= z. This implies d(x, z) = |x| + |z| and d(y, z) =

|y| + |z|.

d(x, y) = |x| + |y|

≤ |x| + |y| + 2|z|
= |x| + |z| + |y| + |z|
= d(x, z) + d(y, z)
∴ d(x, y) ≤ d(x, z) + d(y, z)

Hence d(x, y) ≤ d(x, z) + d(z, y) ∀ x, y, z ∈ R.

Thus triangle inequality is proved. (III)

Hence from (I), (II) and (III) d is a metric on R.

(2) Let d : N × N −→ R be defined as

d(m, n) = 0 if m = n
1
= 1+ if m 6= n
m+n
Then d is a metric on N.
For m, n, p ∈ N

(i)
(
0if m = n
d(m, n) = 1
1+ 6 n
if m =
m+n
∴ d(m, n) ≥ 0 for m, n ∈ N
and d(m, n) = 0 if and only if m = n
(since if m 6= n then d(m, n) ≥ 1) [1 mark]
(ii) d(m, n) = 0 = d(n, m) if m = n
1 1
d(m, n) = 1 + =1+ = d(n, m) if m 6= n [2 marks]
m+n n+m
(iii) If m = p
d(m, p) = 0 ≤ d(m, n) + d(n, p)
If m 6= p then either n 6= m and/or n 6= p.
1
and d(m, p) = 1 +
m+p
20 CHAPTER 1. UNIT I: METRIC SPACES

case 1 n 6= m and n 6= p
1 1
d(m, n) = 1 + and d(n, p) = 1 +
m+n n+p
1 1
∴ d(m, n) + d(n, p) = 1 + +1+ ≥2
m+n n+p
1
∴ d(m, p) = 1 + ≤ 1 + 1 = 2 ≤ d(m, n) + d(n, p)
m+p
case 2 m = n and n 6= p

1
∴ d(m, p) = 1 +
m+p
1
=1+ +0
n+p
= d(n, p) + d(m, n)
= d(m, n) + d(n, p)

case 3 m 6= n and n = p

1
∴ d(m, p) = 1 +
m+p
1
=1+ +0
m+n
= d(m, n) + d(n, p)
= d(m, n) + d(n, p)

∴ in any case, d(m, p) ≤ d(m, n) + d(n, p) [3 marks]

∴ from (i), (ii) and (iii) d is a metric on N
1 1
(3) Let d : N × N −→ R be defined as d(m, n) = − for all m, n ∈ N. Then d is a
m n
metric on N.
To show that d(m, n) ≥ 0 and d(m, n) = 0 if and only if m = n .
1 1
Since |a| ≥ 0 for all a ∈ R, − ≥ 0. Hence d(m, n) ≥ 0 for all m, n ∈ N.
m n

1 1
d(m, n) = 0 if and only if − = 0.
m n
1 1
if and only if − =0
m n
1 1
if and only if =
m n
if and only if m = n
21

Thus non-negativity is proved. (I)

To show that d(m, n) = d(n, m) for all m, n ∈ N.

1 1
d(m, n) = − .
m n
1 1
= −
n m
= d(n, m) ∀ m, n ∈ N.

Thus symmetry is proved. (II)

To show that d(m, p) ≤ d(m, n) + d(n, p) for all m, n, p ∈ N .

1 1
d(m, p) = −
m p
1 1 1 1
= − + −
m n n p
1 1 1 1
≤ − + − = d(m, n) + d(n, p)
m n n p
∴ d(m, p) ≤ d(m, n) + d(n, p) ∀ m, n, p ∈ N. (III)

Hence from (I), (II) and (III) d is a metric on N.

1 1
(4) Let d : R \ {0} × R \ {0} −→ R be defined as d(x, y) = − . Then d is a metric
x y
on R \ {0}.

To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .

1 1
Since |a| ≥ 0 for all a ∈ R, − ≥ 0. Hence d(x, y) ≥ 0 for all x, y ∈ R\{0}.
x y

1 1
d(x, y) = 0 if and only if − = 0.
x y
1 1
if and only if − = 0
x y
1 1
if and only if =
x y
if and only if x = y

Thus non-negativity is proved. (I)

22 CHAPTER 1. UNIT I: METRIC SPACES

To show that d(x, y) = d(y, x) for all x, y ∈ R \ {0}.

1 1
d(x, y) = − .
x y
1 1
= −
y x
= d(y, x) ∀ x, y ∈ R \ {0}.

Thus symmetry is proved. (II)

To show that d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ R \ {0} .

1 1
d(x, z) = −
x z
1 1 1 1
= − + −
x y y z
1 1 1 1
≤ − + − = d(x, y) + d(y, z)
x y y z
∴ d(x, z) ≤ d(x, y) + d(y, z) ∀ x, y, z ∈ R \ {0}. (III)

Hence from (I), (II) and (III) d is a metric on R \ {0}.

(5) Let X be the set of real sequences and d : X × X −→ R be defined by

d(x, y) = 0 if x = y
1
= if x 6= y
min {i : xi 6= yi }

where x = (xn ), y = (yn ) ∈ X is a metric on X.

To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .

Case 1: x = y .Then by definition of d, d(x, y) = 0.
Case 2: x 6= y . Hence min{i : xi 6= yi } > 0, that is d(x, y) > 0.

Thus d(x, y) ≥ 0 for all x, y ∈ X.

Also, x = y =⇒ d(x, y) = 0 and x 6= y =⇒ d(x, y) > 0, that is d(x, y) 6= 0.
Hence d(x, y) = 0 if and only if x = y. Thus non-negativity is proved. (I)
To show that d(x, y) = d(y, x) for all x, y ∈ X. Case 1: x = y

x = y =⇒ y = x
Also x = y =⇒ d(x, y) = 0
And y = x =⇒ d(y, x) = 0
Hence x = y =⇒ d(x, y) = d(y, x)
23

Case 2: x 6= y .

x 6= y =⇒ y 6= x
1
Also x 6= y =⇒ d(x, y) =
min {i : xi 6= yi }
1
And y 6= x =⇒ d(y, x) = = d(x, y)
min {i : yi 6= xi }
Hence x 6= y =⇒ d(x, y) = d(y, x)

Hence d(x, y) = d(y, x) for all x, y ∈ X.

Thus symmetry is proved. (II)

To show that d(x, y) ≤ d(x, z) + d(z, y) for all x, y, z ∈ X .

Case 1: x = y So, d(x, y) = 0. (by (I))

Since d(x, z) ≥ 0 and d(z, y) ≥ 0,
we have d(x, z) + d(z, y) ≥ 0.
Put 0 = d(x, y).
So d(x, z) + d(z, y) ≥ d(x, y), that is d(x, y) ≤ d(x, z) + d(z, y).

1
Case 2: x 6= y This implies d(x, y) = .
min {i : xi 6= yi }
Also x 6= y implies the case x = z and y = z does not arise.

Case 2a: x = z, y 6= z.
1
This implies d(x, z) = 0 and d(y, z) = .
min {i : yi 6= zi }
As x = z, xi = zi ∀ i ∈ N.
1
d(y, z) =
min {i : yi 6= zi }
1
=
min {i : yi 6= xi }
= d(x, y).
Now, d(x, y) = d(y, z)
= 0 + d(y, z)
= d(x, z) + d(y, z)

Case 2b: x 6= z, y = z.
1
This implies d(x, z) = and d(y, z) = 0.
min {i : xi =
6 zi }
As y = z, yi = zi, ∀ i ∈ N.
24 CHAPTER 1. UNIT I: METRIC SPACES

1
d(x, z) =
min {i : xi =
6 zi }
1
=
min {i : xi =6 yi }
= d(x, y)

Now, d(x, y) = d(x, z)

= d(x, z) + 0
= d(x, z) + d(y, z)

Case 2c: x 6= z, y 6= z.
1 1
=⇒ d(x, z) = and d(y, z) = .
min {i : xi 6= zi } min {i : yi 6= zi }
1
Let min {i : xi 6= yi } = l. So l ≥ 1 and d(x, y) =
l
1
min {i : xi 6= zi } = m. So m ≥ 1, and d(x, z) =
m
1
min {i : zi 6= yi } = k. So k ≥ 1 and d(z, y) =
k

1 1
If l ≥ m then ≤
l m
1 1 1
=⇒ ≤ +
l m k
=⇒ d(x, y) ≤ d(x, z) + d(z, y)

If l < m then x1 = y1 , x2 = y2 , · · · xl−1 = yl−1 , xl 6= yl and

x1 = z1 , x2 = z2 , · · · xl−1 = zl−1 , xl = zl , · · · , xm−1 = zm−1 , xm = zm .
Therefore x1 = y1 and x1 = z1 =⇒ y1 = z1 ,
x2 = y2 and x2 = z2 =⇒ y2 = z2 ,
..
.
xl−1 = yl−1 and xl−1 = zl−1 =⇒ yl−1 = zl−1 ,
Also, xl 6= yl and xl = zl =⇒ yl 6= zl .
Hence min{i : yi 6= zi } = l.
That is k = l.
1 1 1
Hence d(x, y) = , d(x, z) = , d(y, z) = .
l m l
25

∴ d(x, y) = d(y, z)
< d(x, z) + d(y, z) (as d(x, z) > 0)
= d(x, z) + d(y, z)

Thus in any case d(x, y) ≤ d(x, z)+d(z, y) for all x, y, z ∈ X. Thus triangle
inequality is proved. (III)
Hence from (I), (II) and (III) d is a metric on X.
(6) Let X be the set of real sequences. If d : X × X −→ R is defined by
∞
X |xi − yi |
d(x, y) =
i=1
2i (1+ |xi − yi |)

where x = (xn ), y = (yn ) ∈ X then d is a metric on X.

We will show that ∞d is well defined .

Thus non-negativity is proved. (I)

Thus symmetry is proved. (II)

0 < |xn − yn | ≤ |xn − zn | + |zn − yn |.

1 1
0<1+ ≤1+ .
|xn − zn | + |zn − yn | |xn − yn |
1 + |xn − zn | + |zn − yn | 1 + |xn − yn |
0< ≤ .
|xn − zn | + |zn − yn | |xn − yn |
|xn − yn | |xn − zn | + |zn − yn |
≤ .
1 + |xn − yn | 1 + |xn − zn | + |zn − yn |
|xn − zn | |zn − yn |
= + .
1 + |xn − zn | + |zn − yn | 1 + |xn − zn | + |zn − yn |
|xn − zn | |zn − yn |
≤ n + n
2 (1 + |xn − zn |) 2 (1 + |zn − yn |)
27

|xn − yn | |xn − zn | |zn − yn |

From (I), (II) and (III) d is a metric on X.

28 CHAPTER 1. UNIT I: METRIC SPACES

1.1.3 Metrics derived from some other metrics defined

on the same space

(1) Let d1 and d2 be metrics on a non-empty set X. Each of the following functions is
also a metric on X.

(i) d, where d(x, y) = max {d1 (x, y), d2 (x, y)} for x, y ∈ X.
(ii) d, where d(x, y) = r1 d1 (x, y) + r2 d2 (x, y) for x, y ∈ X and r1 , r2 > 0.
p
(iii) d, where d(x, y) = (d1 (x, y))2 + (d2 (x, y))2 for x, y ∈ X

(2) Let (X, d) be a metric space. The following are metrics on X.

p
(i) d1 where d1 (x, y) = d(x, y)
d(x, y)
(ii) d1 , where d1 (x, y) =
1 + d(x, y)
(iii) d1 , where d1 (x, y) = min{1, d(x, y)}

We prove the above results.

(1) Let d1 and d2 be metrics on a non-empty set X. Each of the following functions is
also a metric on X.

(i) d, where d(x, y) = max {d1 (x, y), d2 (x, y)} for x, y ∈ X.
To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .
Since d1 , d2 are metrics on X, d1 (x, y) ≥ 0, d2 (x, y) ≥ 0 for all x, y ∈ X.
Hence max{d1 (x, y), d2 (x, y) ≥ 0}. That is d(x, y) ≥ 0 for all x, y ∈ X.

d(x, y) = 0 if and only if max {d1 (x, y), d2 (x, y)} = 0.

if and only if d1 (x, y) = 0 &d2 (x, y) = 0 (as d1 (x, y) ≥ 0 & d2 (x, y) ≥ 0)
if and only if x = y (as d1 , d2 are metrics on X)

Thus non-negativity is proved. (I)

To show that d(x, y) = d(y, x) for all x, y ∈ X.

d(x, y) = max{d1 (x, y), d2 (x, y)}.

= max{d1 (y, x), d2 (y, x)} (as d1 , d2 are metrics on X, di (x, y) = di (y, x))∀
= d(y, x) ∀ x, y ∈ X

Thus symmetry is proved. (II)

To show that d(x, y) ≤ d(x, z) + d(z, y) for all x, y, z ∈ X.

d1 (x, y) ≤ d1 (x, z) + d1 (z, y) (as d1 is a metric on X)

≤ max{d1 (x, z), d2 (x, z)} + max{d1 (z, y), d2 (z, y)}
= d(x, z) + d(z, y)
∴ d1 (x, y) ≤ d(x, z) + d(z, y) ∀ x, y, z ∈ X. · · · (∗)

Similarly,

d2 (x, y) ≤ d2 (x, z) + d2 (z, y) (as d2 is a metric on X)

≤ max{d1 (x, z), d2 (x, z)} + max{d1 (z, y), d2 (z, y)}
= d(x, z) + d(z, y)
∴ d2 (x, y) ≤ d(x, z) + d(z, y) ∀ x, y, z ∈ X. · · · (∗∗)

From (∗) and (∗∗), max{d1 (x, y), d2 (x, y)} ≤ d(x, z)+d(z, y) ∀ x, y, z ∈ X.
Hence d(x, y) ≤ d(x, z) + d(z, y) ∀ x, y, z ∈ X.
Thus triangle inequality is proved. (III)

Hence from (I), (II) and (III) d is a metric on X.

(ii) d, where d(x, y) = r1 d1 (x, y) + r2 d2 (x, y) for x, y ∈ X and r1 , r2 > 0.

To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .

Since d1 , d2 are metrics on X, d1 (x, y) ≥ 0, d2 (x, y) ≥ 0 for all x, y ∈ X.
Hence r1 d1 (x, y) + r2 d2 (x, y) ≥ 0} for all x, y ∈ X as r1 , r2 > 0. Hence
d(x, y) ≥ 0 for all x, y ∈ X.

d(x, y) = 0 if and only if r1 d1 (x, y) + r2 d2 (x, y) = 0.

if and only if r1 d1 (x, y) = 0 &r2 d2 (x, y) = 0 (as r1 d1 (x, y) ≥ 0 & r2 d2 (x,
if and only if d1 (x, y) = 0 &d2 (x, y) = 0 (as r1 , r2 6= 0)
if and only if x=y (as d1 , d2 are metrics on X)

Thus non-negativity is proved. (I)

To show that d(x, y) = d(y, x) for all x, y ∈ X.

d(x, y) = r1 d1 (x, y) + r2 d2 (x, y)}.

= r1 d1 (y, x) + r2 d2 (y, x)} (as d1 , d2 are metrics on X, di (x, y) = di (y, x))∀
= d(y, x) ∀ x, y ∈ X

Thus symmetry is proved. (II)

30 CHAPTER 1. UNIT I: METRIC SPACES

To show that d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ X.

d1 (x, z) ≤ d1 (x, y) + d1 (y, z) (as d1 is a metric on X)

r1 d1 (x, z) ≤ r1 d1 (x, y) + r1 d1 (y, z) ∀ x, y, z ∈ X (as r1 > 0)

d2 (x, z) ≤ d2 (x, y) + d2 (y, z) (as d2 is a metric on X)

r2 d2 (x, z) ≤ r2 d2 (x, y) + r2 d2 (y, z) ∀ x, y, z ∈ X (as r2 > 0)

Hence r1 d1 (x, z)+r2 d2 (x, z) ≤ r1 d1 (x, y)+r2 d2 (x, y)+r1 d1 (y, z)+r2 d2 (y, z).
Hence d(x, z) ≤ d(x, y) + d(y, z) ∀ x, y, z ∈ X.
Thus triangle inequality is proved. (III)
Hence from (I), (II) and (III) d is a metric on X.
p
(iii) d, where d(x, y) = (d1 (x, y))2 + (d2 (x, y))2 for x, y ∈ X.
To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .
As d21 (x, 2 2 2
py) ≥ 0, d2 (x, y) ≥ 0, d1 (x, y) + d2 (x, y) ≥ 0} for all x, y ∈ X.
Hence d21 (x, y) + d22 (x, y) ∈ R.
√ √ p
Also since a ≥ 0 whenever l a exists, d21 (x, y) + d22 (x, y) ≥ 0.
Hence d(x, y) ≥ 0, for all x, y ∈ X.

d(x, y) = 0 if and only if d21 (x, y) + d22 (x, y) = 0.

if and only if d21 (x, y) = 0 &d22 (x, y) = 0 (as d21 (x, y) ≥ 0 & d22 (x, y) ≥ 0)
if and only if d1 (x, y) = 0 &d2 (x, y) = 0
if and only if x=y (as d1 , d2 are metrics on X)

Thus non-negativity is proved. (I)

To show that d(x, y) = d(y, x) for all x, y ∈ X.

d(x, y) = d21 (x, y) + d22 (x, y).

= d21 (y, x) + d22 (y, x) (as d1 , d2 are metrics on X, di (x, y) = di (y, x))∀ x, y ∈
= d(y, x) ∀ x, y ∈ X

Thus symmetry is proved. (II)

To show that d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ X.

d1 (x, z) ≤ d1 (x, y) + d1 (y, z) (as d1 is a metric on X)

h i2
d21 (x, z) ≤ d1 (x, z) + d1 (y, z) ∀ x, y, z ∈ X
31

d2 (x, z) ≤ d2 (x, y) + d2 (y, z) (as d2 is a metric on X)

h i2
d22 (x, y) ≤ d2 (x, y) + d2 (y, z) ∀ x, y, z ∈ X

h i2 h i2
Hence d21 (x, z)
+ d22 (x, z)
≤ d1 (x, y) + d1 (y, z) + d2 (x, y) + d2 (y, z) .
rh i2 h i2
p
2 2
This means d1 (x, z) + d2 (x, z) ≤ d1 (x, y) + d1 (y, z) + d2 (x, y) + d2 (y, z) .
p p p
Therefore d21 (x, z) + d22 (x, z)p≤ d21 (x, y) + d22 (x, y)+ 2 2
√ d1 (y, z)√+ d2 (y, z).
(By Minkowski’s Inequality: (a + b)2 + (c + d)2 ≤ a2 + c2 + b2 + d2 )
Hence d(x, z) ≤ d(x, y) + d(y, z) ∀ x, y, z ∈ X.
Thus triangle inequality is proved. (III)

Hence from (I), (II) and (III) d is a metric on X.

(2) Let (X, d) be a metric space. The following are metrics on X.

p
(i) d1 where d(x, y) = d(x, y).

To show that d1 (x, y) ≥ 0 and d1 (x, y) = 0 if and only if x = y .

is a metric on X, d(x, y) ≥ 0 for all x, y ∈ X.
Since dp
Hence
√ d(x, y) ∈ R.
As a ≥ 0 for all a ≥ 0, we have, That is d1 (x, y) ≥ 0 for all x, y ∈ X.
p
d(x, y) = 0 if and only if d(x, y) = 0.
if and only if d(x, y) = 0
if and only if x = y (as d is a metric on X)

Thus non-negativity is proved. (I)

To show that d1 (x, y) = d1 (y, x) for all x, y ∈ X.

p
d1 (x, y) = d(x, y).
p
= d(y, x) (as d is a metric on X, d(x, y) = d(y, x))∀ x, y ∈ X)
= d1 (y, x) ∀ x, y ∈ X

Thus symmetry is proved. (II)

32 CHAPTER 1. UNIT I: METRIC SPACES

To show that d1 (x, z) ≤ d1 (x, y) + d1 (y, z) for all x, y, z ∈ X.

d(x, z) ≤ d(x, y) + d(y, z) (as d is a metric on X)

p p
≤ d(x, y) + d(y, z) + 2 d(x, y) d(y, z)
p p 2
= d(x, y) + d(y, z)
p p p
∴ d(x, z) ≤ d(x, y) + d(y, z) ∀ x, y, z ∈ X.
∴ d1 (x, z) ≤ d1 (x, y) + d1 (y, z) ∀ x, y, z ∈ X. (III)

Hence from (I), (II) and (III) d1 is a metric on X.

d(x, y)
(ii) d1 (x, y) = .
1 + d(x, y)
To show that d1 (x, y) ≥ 0 and d1 (x, y) = 0 if and only if x = y .
Since d is a metric on X, d(x, y) ≥ 0 for all x, y ∈ X.
This implies 1 + d(x, y) ≥ 1 > 0 for all x, y ∈ X.
d(x, y)
Hence ≥ 0 for all x, y ∈ X.
1 + d(x, y)
Thus d1 (x, y) ≥ 0 for all x, y ∈ X.

d(x, y)
d1 (x, y) = 0 if and only if =0
1 + d(x, y)
if and only if d(x, y) = 0
if and only if x = y (as d is a metric on X)

Thus non-negativity is proved. (I)

To show that d1 (x, y) = d1 (y, x) for all x, y ∈ X.

d(x, y)
d1 (x, y) = .
1 + d(x, y)
d(y, x)
= (as d is a metric on X, d(x, y) = d(y, x))∀ x, y ∈ X)
1 + d(y, x)
= d1 (y, x) ∀ x, y ∈ X

Thus symmetry is proved. (II)

To show that d1 (x, z) ≤ d1 (x, y) + d1 (y, z) for all x, y, z ∈ X. We have two
cases as follows:
Case 1 x = z. So d1 (x, z) = 0 (by (I))
Since d1 (x, y) ≥ 0 and d1 (y, z) ≥ 0,
we have d1 (x, y) + d1 (y, z) ≥ 0.
Put 0 = d1 (x, z).
So d1 (x, y) + d1 (y, z) ≥ d1 (x, z), that is d1 (x, z) ≤ d1 (x, y) + d1 (y, z).
33

Case 2 x 6= z. So d(x, z) > 0.

Since d is a metric on X, d(x, z) ≤ d(x, y) + d(y, z).
So, 0 < d(x, z) ≤ d(x, y) + d(y, z).
1 1
This implies 0 < ≤
d(x, y) + d(y, z) d(x, z)

1 1
=⇒ 0 < 1 + ≤1+ .
d(x, y) + d(y, z) d(x, z)
1 + d(x, y) + d(y, z) 1 + d(x, z)
=⇒ 0 < ≤ .
d(x, y) + d(y, z) d(x, z)
d(x, z) d(x, y) + d(y, z)
=⇒ 0 < ≤ .
1 + d(x, z) 1 + d(x, y) + d(y, z)
d(x, y) d(y, z)
= + ∀
1 + d(x, y) + d(y, z) 1 + d(x, y) + d(y, z)
d(x, y) d(y, z)
≤ + ∀ x, y ∈ X.
1 + d(x, y) 1 + d(y, z)
=⇒ d1 (x, z) ≤ d1 (x, y) + d1 (y, z)

Hence d1 (x, z) ≤ d1 (x, y) + d1 (y, z) ∀ x, y, z ∈ X. (III)

From (I), (II) and (III) d1 is a metric on X.

(iii) d1 , where d1 (x, y) = min{1, d(x, y)}.
To show that d1 (x, y) ≥ 0 and d1 (x, y) = 0 if and only if x = y .
d1 (x, y) = min{1, d(x, y)} =⇒ d1 (x, y) = 1 or d1 (x, y) = d(x, y).
If d1 (x, y) = 1 then d1 (x, y) > 0.
If d1 (x, y) = d(x, y) then since d(x, y) ≥ 0, d1 (x, y) ≥ 0.
Thus in any case, d1 (x, y) ≥ 0 for all x, y ∈ X.

d1 (x, y) = 0 if and only if min{1, d(x, y)} = 0

if and only if d(x, y) = 0
if and only if x = y (as d is a metric on X)

Thus non-negativity is proved. (I)

To show that d1 (x, y) = d1 (y, x) for all x, y ∈ X.

d1 (x, y) = min{1, d(x, y)}.

= min{1, d(y, x)} (as d is a metric on X, d(x, y) = d(y, x))∀ x, y ∈ X)
= d1 (y, x) ∀ x, y ∈ X

Thus symmetry is proved. (II)

34 CHAPTER 1. UNIT I: METRIC SPACES

To show that d1 (x, z) ≤ d1 (x, y) + d1 (y, z) for all x, y, z ∈ X .

Since d1 (x, z) = min{1, d(x, z)}, d1 (x, z) ≤ 1 & d1 (x, z) ≤ d(x, z). (∗)

Case 1 d1 (x, y) = 1.
From (∗), we have d1 (x, z) ≤ 1.
Put 1 = d1 (x, y)
So, d1 (x, z) ≤ d1 (x, y) (∗∗).
Now, 0 ≤ d1 (y, z), 0 + d1 (x, y) ≤ d1 (y, z) + d1 (x, y)
(by adding d1 (x, y) on both sides).
That is d1 (x, y) ≤ d1 (y, z) + d1 (x, y). Therefore from (∗∗), d1 (x, z) ≤
d1 (x, y) ≤ d1 (y, z) + d1 (x, y)..
That is d1 (x, z) ≤ d1 (x, y) + d1 (y, z).
Case 2 d1 (y, z) = 1.
From (∗), we have d1 (x, z) ≤ 1.
Put 1 = d1 (y, z)
So, d1 (x, z) ≤ d1 (y, z) (∗ ∗ ∗).
Now, 0 ≤ d1 (x, y) =⇒ 0 + d1 (y, z) ≤ d1 (x, y) + d1 (y, z)
(by adding d1 (y, z) on both sides).
That is d1 (y, z) ≤ d1 (x, y) + d1 (y, z). Therefore from (∗ ∗ ∗), d1 (x, z) ≤
d1 (y, z) ≤ d1 (x, y) + d1 (y, z)..
That is d1 (x, z) ≤ d1 (x, y) + d1 (y, z).
Case 3 d1 (x, y) 6= 1 and d1 (y, z) 6= 1.
Therefore, d1 (x, y) = d(x, y) and d1 (y, z) = d(y, z).
Since d is a metric on X, d(x, z) ≤ d(x, y) + d(y, z).
=⇒ d(x, z) ≤ d1 (x, y)+d1 (y, z) (∗∗∗∗) (as d(x, y) = d1 (x, y), d(y, z) =
d1 (y, z))
From (∗), d1 (x, z) ≤ d(x, z).
Therefore, from (∗ ∗ ∗∗), d1 (x, z) ≤ d1 (x, y) + d1 (y, z).
Hence d1 (x, z) ≤ d1 (x, y) + d1 (y, z) ∀ x, y, z ∈ X. (III)

From (I), (II) and (III) d1 is a metric on X.

1.1.4 Product of two metric

spaces

Example 1.1.3 Let (X, d1 ) and (Y, d2 ) be metric spaces.

(1) Define d : X × Y × X × Y −→ R as follows: d((x1 , y1 ), (x2 , y2 )) = d1 (x1 , x2 ) +

d2 (x2 , y2 ). Then d is a metric on X × Y.
p
(2) Define d : X×Y ×X×Y −→ R as follows: d((x1 , y1 ), (x2 , y2 )) = d21 (x1 , x2 ) + d22 (x2 , y2 ).
Then d is a metric on X × Y.

We will prove that the above product spaces are metric spaces.
35

(1) d : X × Y × X × Y −→ R as follows: d((x1 , y1 ), (x2 , y2 )) = d1 (x1 , x2 ) + d2 (x2 , y2 ).

Then d is a metric on X × Y.

To show that d((x1 , y1 ), (x2 , y2 )) ≥ 0 and d((x1 , y1 ), (x2 , y2 )) = 0 if and only if (x1 , y1 ) = (
.
Since d1 , d2 are metrics on X and Y respectively, d1 (x1 , x2 ) ≥ 0, d2 (y1 , y2 ) ≥ 0
for all x1 , x2 ∈ X and y1 , y2 ∈ Y .
So, d1 (x1 , x2 ) + d2 (y1 , y2 ) ≥ 0 for all x1 , x2 ∈ X and y1 , y2 ∈ Y .
Hence d((x1 , y1 ), (x2 , y2 )) ≥ 0 for all (x1 , y1 ), (x2 , y2 ) ∈ X × Y.

d((x1 , y1 ), (x2 , y2 )) = 0 if and only if d1 (x1 , x2 ) + d2 (y1 , y2 ) = 0.

if and only if d1 (x1 , x2 ) = 0 &d2 (y1 , y2 ) = 0 (as d1 (x1 , x2 ) ≥ 0 &
if and only if x1 = x2 and y1 = y2 (as d1 , d2 are metrics on X
if and only if (x1 , y1 ) = (x2 , y2 )

Thus non-negativity is proved. (I)

To show that d((x1 , y1 ), (x2 , y2 )) = d((x2 , y2 ), (x1 , y1 )) for all (x1 , y1 ), (x2 , y2 ) ∈ X × Y.

d((x1 , y1 ), (x2 , y2 )) = d1 (x1 , x2 ) + d2 (y1 , y2 ).

= d1 (x2 , x1 ) + d2 (y2 , y1 ) (as d1 , d2 are metrics on X and Y respectiv
= d((x2 , y2 ), (x1 , y1 )) ∀ (x1 , y1 ), (x2 , y2 ) ∈ X × Y

Thus symmetry is proved. (II)

To show that d((x1 , y1 ), (x3 , y3 ) ≤ d((x1 , y1 ), (x2 , y2 )) + d((x2 , y2 ), (x3 , y3 )) for all (x1 , y1 ), (x

d1 (x1 , x3 ) ≤ d1 (x1 , x2 ) + d1 (x2 , x3 ) (as d1 is a metric on X)

d2 (y1 , y3 ) ≤ d2 (y1 , y2 ) + d2 (y2 , y3 ) (as d2 is a metric on Y )

Hence d1 (x1 , x3 ) + d2 (y1 , y3 ) ≤ d1 (x1 , x2 ) + d1 (x2 , x3 ) + d2 (y1 , y2 ) + d2 (y2 , y3 ).

That is d1 (x1 , x3 ) + d2 (y1 , y3 ) ≤ d1 (x1 , x2 ) + d2 (y1 , y2 ) + d1 (x2 , x3 ) + d2 (y2 , y3 ).
=⇒ d((x1 , y1 ), (x3 , y3 )) ≤ d((x1 , y1 ), (x2 , y2 )) + d((x2 , y2 ), (x3 , y3 )) ∀ (x1 , y1 ), (x2 , y2

Thus triangle inequality is proved. (III)

Hence from (I), (II) and (III) d is a metric on X × Y .

p
(2) d : X × Y × X × Y −→ R as follows: d((x1 , y1 ), (x2 , y2 )) = d21 (x1 , x2 ) + d22 (x2 , y2 ).
Then d is a metric on X × Y.
36 CHAPTER 1. UNIT I: METRIC SPACES

To show that d((x1 , y1 ), (x2 , y2 )) ≥ 0 and d((x1 , y1 ), (x2 , y2 )) = 0 if and only if (x1 , y1 ) = (
.
Since d1 , d2 are metrics on X and Y respectively, d1 (x1 , x2 ) ∈ R, d2 (y1 , y2 ) ∈ R
for all x1 , x2 ∈ X and y1 , y2 ∈ Y .
So, d21 (x1 , x2 ) ≥ 0, d22 (y1 , y2 ) ≥ 0 for all x1 , x2 ∈ X and y1 , y2 ∈ Y .
Hence d21 (xp 2
1 , x2 ) + d2 (y1 , y2 ) ≥ 0} for all (x1 , y1 ) and (x2 , y2 ∈ X × Y.
√
2 2
Therefore
p d 1 (x 1 , x2 ) + d 2 (y 1 , y2 ) ∈ R and since a ≥ 0 whenever it exists,
d21 (x1 , x2 ) + d22 (y1 , y2 ) ≥ 0 for all (x1 , y1 ) and (x2 , y2 ∈ X × Y.
Hence d((x1 , y1 ), (x2 , y2 )) ≥ 0 for all (x1 , y1 ), (x2 , y2 ) ∈ X × Y.
q
d((x1 , y1 ), (x2 , y2 )) = 0 if and only if d21 (x1 , x2 ) + d22 (y1 , y2 ) = 0.
if and only if d21 (x1 , x2 ) + d22 (y1 , y2 ) = 0.
if and only if d21 (x1 , x2 ) = 0 &d22 (y1 , y2 ) = 0 (as d21 (x1 , x2 ) ≥ 0 &
if and only if d1 (x1 , x2 ) = 0 &d2 (y1 , y2 ) = 0
if and only if x1 = x2 and y1 = y2 (as d1 , d2 are metrics on X
if and only if (x1 , y1 ) = (x2 , y2 )

Thus non-negativity is proved. (I)

To show that d((x1 , y1 ), (x2 , y2 )) = d((x2 , y2 ), (x1 , y1 )) for all (x1 , y1 ), (x2 , y2 ) ∈ X × Y.

q
d((x1 , y1 ), (x2 , y2 )) = d21 (x1 , x2 ) + d22 (y1 , y2 )}.
q
= d21 (x2 , x1 ) + d22 (y2 , y1 )} (as d1 , d2 are metrics on X and Y respe
= d((x2 , y2 ), (x1 , y1 )) ∀ (x1 , y1 ), (x2 , y2 ) ∈ X × Y

Thus symmetry is proved. (II)

To show that d((x1 , y1 ), (x3 , y3 ) ≤ d((x1 , y1 ), (x2 , y2 )) + d((x2 , y2 ), (x3 , y3 )) for all (x1 , y1 ), (x

d1 (x1 , x3 ) ≤ d1 (x1 , x2 ) + d1 (x2 , x3 ) (as d1 is a metric on X)

d2 (y1 , y3 ) ≤ d2 (y1 , y2 ) + d2 (y2 , y3 ) (as d2 is a metric on Y )
37

h i2 h
Hence d21 (x1 , x3 )
+ d22 (y1 , y3 ) ≤ d1 (x1 , x2 ) + d1 (x2 , x3 ) + d2 (y1 , y2 )
rh
q 2 i h
This implies d21 (x1 , x3 ) + d22 (y1 , y3 ) ≤ d1 (x1 , x2 ) + d1 (x2 , x3 ) + d2 (y1 , y
p √ √
By Minkowski’s Inequality, (a + b)2 + (c + d)2 ≤ a2 + c2 + b2 + d2
q rh q
2 2
That is d1 (x1 , x3 ) + d2 (y1 , y3 ) ≤ d1 (x1 , x2 ) + d2 (y1 , y2 ) + d21 (x2 , x
2 2

=⇒ d((x1 , y1 ), (x3 , y3 )) ≤ d((x1 , y1 ), (x2 , y2 )) + d((x2 , y2 ), (x3 , y3

Thus triangle inequality is proved. (III)

Hence from (I), (II) and (III) d is a metric on X × Y .

Examples 1.1.4

(1) Let (X, d) be a metric space. Prove that |d(x, y) − d(x, z)| ≤ d(y, z) ∀x, y, z ∈ X.
Let x, y, z ∈ X.
d(x, y) ≤ d(x, z) + d(z, y) =⇒ d(x, y) − d(x, z) ≤ d(z, y) (∗)
d(x, z) ≤ d(x, y) + d(y, z) =⇒ d(x, z) − d(x, y) ≤ d(z, y) (∗∗)
By (∗) and (∗∗)
|d(x, y) − d(x, z)| ≤ d(y, z)

(2) Let d1 , d2 , d∞ : Rn × Rn −→ R be definedv as follows:

∞
X
u n
uX
d1 (x, y) = |xi − yi |, d2 (x, y) = t (xi − yi )2 , d∞ (x, y) = max{|xi − yi | : 1 ≤
k=1 k=1
i ≤ n} for all x, y ∈ Rn where x = (x1 ,√
x2 , · · · , xn ) and y = (y1 , y2 , · · · , yn ).
Then d∞ (x, y) ≤ d2 (x, y) ≤ d1 (x, y) ≤ nd2 (x, y) ≤ nd∞ (x, y) for all x, y ∈ Rn .

n
X
d1 (x, y) = |xi − yi |
i=1
n
hX i2
d21 (x, y) = |xi − yi |
i=1
n
X X
= |xi − yi |2 + 2 |xi − yi ||xj − yj |
i=1 1≤i<j≤n
X
= d22 (x, y) + 2 |xi − yi ||xj − yj |
1≤i<j≤n
X
≥ d22 (x, y) (as 2 |xi − yi ||xj − yj | ≥ 0)
1≤i<j≤n

d1 (x, y) ≥ d2 (x, y) (∗)

38 CHAPTER 1. UNIT I: METRIC SPACES

v v
u n u n
uX uX p
d2 (x, y) = t (xi − yi )2 t (xi − yi )2 = (x1 − y1 )2 + (x2 − y2 )2 · · · + (xn − yn
i=1 i=1
p
≥ (xi − yi )2 for all 1 ≤ i ≤ n
= |xi − yi | for all 1 ≤ i ≤ n
d2 (x, y) ≥ max {|xi − yi | : 1 ≤ i ≤ n}
d2 (x, y) ≥ d∞ (x, y) (∗∗)
n
√ d1 (x, y) ≥ d2 (x, y) ≥ d∞ (x, y) forn all x, y ∈ R .
From (∗) and (∗∗), we have
To prove that d1 (x, y) ≤ nd2 (x, y) ≤ nd∞ (x, y) for x, y ∈ R .

d∞ (x, y) = max{|xi − yi | : 1 ≤ i ≤ n}
d2∞ (x, y) ≥ |xi − yi |2 1 ≤ i ≤ n
= (xi − yi )2 1 ≤ i ≤ n
Xn
nd2∞ (x, y) ≥ (xi − yi )2
i=1
∴ nd2∞ (x, y)
≥ d22 (x, y)
√
∴ nd∞ (x, y) ≥ d2 (x, y)
1
∴ d∞ (x, y) ≥ √ d2 (x, y) (1)
n

n
X
d22 (x, y) = (xi − yi )2
i=1
n
X
= |xi − yi |2
i=1
n
!2
X X
d22 (x, y) = |xi − yi | −2 |xi − yi ||xj − yj | (∗)
i=1 1≤i<j≤n

X
2 |ai ||aj | =
1≤i<j≤n
2|a1 ||a2 | + 2|a1 ||a3 | + · · · + 2|a1 ||an | + 2|a2 ||a3 | + · · · 2|a2 ||an | + · · · + 2|an−1 ||an |.
X
2 |ai ||aj | ≤ (|a1 |2 + |a2 |2 ) + · · · + (|a1 |2 + |an |2 ) + (|a2 |2 + |a3 |2 ) + · · · + (|an−1 |2 + |an |2 )
1≤i<j≤n

= (n − 1)a21 + (n − 1)a22 + · · · + (n − 1)a2n

Xn
= (n − 1) a2i
i=1
39

X
Put ai = xi − yi , =⇒ 2 |xi − yi ||xj − yj | ≤ (n − 1)d22 (x, y)
1≤i<j≤n
X
∴ −2 |xi − yi ||xj − y| ≥ −(n − 1)d22 (x, y).
1≤i<j≤n
Substituting this in (∗), we get,
n
!2
X
d22 (x, y) ≥ |xi − yi | − (n − 1)d22 (x, y)
i=1
n
!2
X
nd22 (x, y) ≥ |xi − yi |
i=1
nd22 (x, y) ≥ 2
d1 (x, y)
√
nd2 (x, y) ≥ d1 (x, y)

1 1 1
√ d2 (x, y) ≥ d1 (x, y) (2) dividing both sides by
n n n
1 1
From (1) and (2), we get, d∞ (x, y) ≥ √ d2 (x, y) ≥ d1 (x, y) for all x, y ∈ Rn .
n n

(3) Let (X, d) be a metric space. Consider the metric d1 : X × X −→ R, defined as

d(x, y)
d1 (x, y) = . Show that d1 (x, y) ≤ d(x, y) for all x, y ∈ X.
1 + d(x, y)

0 ≤ d(x, y) < 1 + d(x, y) ∀ x, y ∈ X.

d(x, y)
0≤ < 1 ∀ x, y ∈ X
1 + d(x, y)
∴ 0 ≤ d1 (x, y) < 1 ∀ x, y ∈ X.
=⇒ −1 < −d1 (x, y) ≤ 0
=⇒ 1 − 1 < 1 − d1 (x, y) ≤ 1 (adding 1 on both sides)
∴ 0 < 1 − d1 (x, y) ≤ 1 ∀ x, y ∈ X. (∗)
d(x, y)
Now d1 (x, y) =
1 + d(x, y)
=⇒ d1 (x, y)(1 + d(x, y)) = d(x, y)
=⇒ d1 (x, y) + d1 (x, y)d(x, y) = d(x, y)
=⇒ d1 (x, y) = d(x, y)(1 − d1 (x, y))
From (∗), d1 (x, y) ≤ d(x, y) (as 0 < 1 − d1 (x, y) ≤ 1)

Hence d1 (x, y) ≤ d(x, y) for all x, y ∈ X.

(4) Let (X, d) be a metric space. Consider the metric d1 : X × X −→ R, defined as

d1 (x, y) = min{1, d(x, y)}. Show that d1 (x, y) ≤ d(x, y) for all x, y ∈ X.
40 CHAPTER 1. UNIT I: METRIC SPACES

Since d1 (x, y) = min{1, d(x, y)}, d1 (x, y) ≤ d(x, y) for all x, y ∈ X.

(5) Prove or disprove: Let (X,p d) be a metric space. Consider the metric d1 : X ×X −→
R, defined as d1 (x, y) = d(x, y). Then d1 (x, y) ≤ d(x, y) for all x, y ∈ X or
d(x, y) ≤ d1 (x, y) for all x, y ∈ X. p
Let
p X = R and d : R × R −→ R, d(x, y) = |x − y| and d 1 (x, y) = d(x, y) =
|x − y|. √
Take x1 = 2, y1 = 6. So, d(x1 , y1 ) = 4 and d1 (x1 , y1 ) = 4 = 2. So, d1 (x1 , y1 ) <
d(x, y). √
Take x2 = 0.02, y2 = 0.06. So, d(x2 , y2 ) = 0.04 and d1 (x1 , y1 ) = 0.04 = 0.2. So,
d(x2 , y2 ) < d1 (x2 , y2 ).

1.1.5 Some Examples of functions that are

not metrics
Examples 1.1.5

(1) Consider the following maps d : R × R −→ R defined below. None of them is a

metric on R.

(i) d(x, y) = |x − 2y| (ii) d(x, y) = |x2 − y 2 | (iii) d(x, y) = |x − y|2

(2) Let d1 and d2 be metrics on a non-empty set X. Then none of the following functions
is a metric on X.
2 2
(i) d : X ×X −→ R defined as d(x, y) = (d21 +d22 )(x, y) = d1 (x, y) + d2 (x, y)
(ii) d : X × X −→ R defined as d(x, y) = (ad1 + bd2 )(x, y) = ad1 (x, y) + bd2 (x, y).
(iii) d : X × X −→ R defined as d(x, y) = min{d1 (x, y), d2 (x, y)}.

(3) Let (X, d1 ) and (Y, d2 ) be metric spaces. d : (X × Y ) × (X × Y ) −→ R defined as

d00 ((x1 , y1 ), (x2 , y2 )) = [(d1 (x1 , x2 ))2 + (d2 (y1 , y2 ))2 ] is not a metric on X × Y.

We will prove that above functions are not metrics on the given space.

(1) (i) d(x, y) = |x − 2y|

Consider x = 2, y = 1. So x 6= y. But d(2, 1) = |x − 2y| = |2 − 2(1)| = 0. Thus
here x 6= y but still d(x, y) = 0.
Hence d is not a metric on R.
(ii) d(x, y) = |x2 −y 2 |. Let x = 1, y = −1. So x 6= y. But d(x, y) = |12 −(−1)2 | = 0.
Thus here x 6= y but still d(x, y) = 0.
Hence d is not a metric on R.
41

(iii) d(x, y) = |x − y|2 . Take x = 1, y = 2, z = 3.

So, d(x, z) = |1 − 3|2 = 4, d(x, y) = |1 − 2|2 = 1, d(y, z) = |2 − 3|2 = 1..
Hence d(x, z) d(x, y) + d(y, z0. Thus d is not a metric on R.
(2) Let d1 and d2 be metrics on a non-empty set X. Then none of the following functions
is a metric on X.
2 2
(i) d : X ×X −→ R defined as d(x, y) = (d21 +d22 )(x, y) = d1 (x, y) + d2 (x, y) .
Let X = R and d1 , d2 : R × R −→ R, d1 (x, y) = d2 (x, y) = |x − y|.
2 2
2 2
So, d(x, y) = (d1 + d2 )(x, y) = d1 (x, y) + d2 (x, y) =⇒ d(x, y) = 2|x − y|2 .
Take x = 1, y = 2, z = 3.
So, d(x, z) = 2|1 − 3|2 = 8, d(x, y) = 2|1 − 2|2 = 2, d(y, z) = 2|2 − 3|2 = 2..
Hence d(x, z) d(x, y) + d(y, z). Thus d is not a metric on R.
(ii) d : X × X −→ R defined as d(x, y) = (ad1 + bd2 )(x, y) = ad1 (x, y) + bd2 (x, y).
Let X = R, d1 , d2 : R × R −→ R, d1 (x, y) = d2 (x, y) = |x − y|.
Take a = 1 and b = −1. So, for any x, y ∈ R, x 6= y, d(x, y) = d1 (x, y) −
d2 (x, y) = |x − y| − |x − y| = 0.
Thus x 6= y but d(x, y) = 0. Hence d is not a metric on X.
(iii) d : X × X −→ R defined as d(x, y) = min{d1 (x, y), d2 (x, y)}.
Let X = R \ {0}, d1 , d2 : R \ {0} × R \ {0} −→ R, d1 (x, y) = |x − y| and
1 1
d2 (x, y) = − .
x y
1
Take x = 2, y = 1, z = .
3
1
x = 2, y = 1, z =
3
5 2
d1 (x, z) = |x − z| = d1 (x, y) = |x − y| = 1 d(y, z) = |y − z| =
3 3
1 5 1 1
d2 (x, z) = −3 = d2 (x, y) = −1 = d2 (y, z) = |1 − 3| = 2
2 2 2 2
5 1 2
d(x, z) = d(x, y) = d(y, z) =
3 2 3
5 1 2 7
d(x, z) = = 1.66 d(x, y) + d(y, z) = + = = 1.166
3 2 3 6
d(x, z) d(x, y) + d(y, z)

(3) Let (X, d1 ) and (Y, d2 ) be metric spaces. d : (X × Y ) × (X × Y ) −→ R defined as

d((x1 , y1 ), (x2 , y2 )) = [(d1 (x1 , x2 ))2 + (d2 (y1 , y2 ))2 ] is not a metric on X × Y.
Let X = R, Y = R, d : R2 ×R2 −→ R, d1 (x1 , x2 ) = |x1 −x2 | and d2 (y1 , y2 ) = |y1 −y2 |.
d((x1 , y1 ), (x2 , y2 )) = d21 (x1 , x2 ) + d22 (y1 , y2 ) = |x1 − x2 |2 + |y1 − y2 |2 .
Take (x1 , y1 ) = (0, 1), (x2 , y2 ) = (0, 2) = (x3 , y3 ) = (0, 3).
d((x1 , y1 ), (x3 , y3 )) = 4.
d((x1 , y1 ), (x2 , y2 )) = 1.
d((x2 , y2 ), (x3 , y3 )) = 1.
d(x, z) d(x, y) + d(y, z).
42 CHAPTER 1. UNIT I: METRIC SPACES

1.2 Distance Metric Induced by a

Norm
Definition 1.2.1 Let V be a vector space over R or C. A norm on V is a function
k k : V −→ R satisfying the following conditions:

(i) kxk ≥ 0 for all x ∈ V and kxk = 0 if and only if x = 0.

(ii) kαxk = |α|kxk for any scalar α and x ∈ V.

(iii) kx + yk ≤ kxk + kyk for all x, y ∈ V. (This is known as the triangle inequlity of the
norm)

Theorem 1.2.2 Let (V.k k) be a normed linear space. Define d : V × V −→ R, d(x, y) =

kx − yk. Then d is a metric on V .

Proof:

To show that d(x, y) ≥ 0 and d(x, y) = 0 if and only if x = y .

Since kak ≥ 0 for all a ∈ V, kx − yk ≥ 0. Hence d(x, y) ≥ 0 for all x, y ∈ V.

d(x, y) = 0 if and only if kx − yk = 0.

if and only if x − y = 0
if and only if x = y

Thus non-negativity is proved. (I)

To show that d(x, y) = d(y, x) for all x, y ∈ V.

d(x, y) = kx − yk.
= ky − xk
= d(y, x) ∀ x, y ∈ V

Thus symmetry is proved. (II)

To show that d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ R .

d(x, z) = kx − zk
= kx − y + y − zk
≤ kx − yk + ky − zk
= d(x, y) + d(y, z)
∴ d(x, z) ≤ d(x, y) + d(y, z) ∀ x, y, z ∈ V. (III)
43

Hence from (I), (II) and (III) d is a metric on V.

(Note: the metric d defined by d(x, y) = kx − yk on a normed linear vector space is called
as the metric associated with the norm k k.)

1.2.1 Examples of Normed Linear Spaces and Distance In-

duced by Them

Examples 1.2.3

(1) Let Rn = {(x1 , x2 , . . . , xn ) : xi ∈ R for 1 ≤ i ≤ n} and x = (x1 , x2 , . . . , xn ),.

i=n
X
(i) Define k k1 : Rn −→ R, kxk1 = |xi |. Then k k1 is a norm on the vector
i=1
space Rn over R.
v
u i=n
uX 2
n
(ii) Define k k2 : R −→ R, kxk2 = kxk2 = t xi then k k2 is a norm on the
i=1
vector space Rn over R.
(iii) Define k k∞ : Rn −→ R, kxk∞ = max {|x1 | : 1 ≤ i ≤ n}. Then k k∞ is a norm
on the vector space Rn over R.
Z 1
(2) Consider the vector space C[0, 1]. Define k k1 : C[0, 1] −→ R as kf k1 = |f (t)| dt.
0
Then k k1 is a norm on the vector space C[0, 1] over R.

(3) Consider the vector space C[0, 1]. Define k k2 : C[0, 1] −→ R as kf k∞ = sup {|f (t)| :
t ∈ [0, 1]}. Then k ksup is a norm on the vector space C[0, 1] over R.
n ∞
X o
1
(4) Consider the vector space l = (xn ) : xn ∈ R, |xn | < ∞ over R. Define
n=1
infty
X
k k1 : l 1 −→ R, kx k1 = |xn |. Then k k1 is a norm on the vector space l 1 over R.
n=1

∞
X
(5) Let l 2 = {(xn ) : (xn ) is a sequence of real numbers such that x2n < ∞}. If
n=1
∞
X 21
kxk2 = x2n for x = (x1 , x2 , . . . , xn , . . .) ∈ l 2 , then k k2 is a norm on the
n=1
vector space l 2 over R.

(6) Consider the vector space M2 (R) over R. Define k k : M2 (R) −→ R given by
kAk = max 1≤i,j≤2 |aij | for A = (aij ) is a norm on M2 (R).

We will prove above statements.

44 CHAPTER 1. UNIT I: METRIC SPACES

(1) Let Rn = {(x1 , x2 , . . . , xn ) : xi ∈ R for 1 ≤ i ≤ n} and x = (x1 , x2 , . . . , xn ),.

i=n
X
n
(i) Define k k1 : R −→ R, kxk1 = |xi |. Then k k1 is a norm on the vector
i=1
space Rn over R.
Let Rn = {(x1 , x2 , . . . , xn ) : xi ∈ R for 1 ≤ i ≤ n}.
i=n
X
For x = (x1 , x2 , . . . , xn ), kxk1 = |xi |.
i=1
We prove that k k1 is a norm on Rn .
We have to prove that
(i) kxk1 ≥ 0 for all x ∈ Rn and kxk1 = 0 if and only if x = 0.
(ii) kαxk1 = |α|kxk for all α ∈ R, x ∈ Rn .
(iii) kx + yk1 ≤ kxk1 + kyk1 for all x, y ∈ Rn .
Let x = (x1 , x2 , · · · , xn ) ∈ Rn .
Xn
kxk1 = |xi |.
i=1
n
X
As |xi | ≥ 0 for all i, 1 ≤ i ≤ n, we have, |xi | ≥ 0.
i=1
Hence kxk1 ≥ 0 for all x ∈ Rn .
n
X
kxk1 = 0 if and only if |xi | = 0.
i=1
n
X
As |xi | ≥ 0 for all i, 1 ≤ i ≤ n, |xi | = 0 =⇒ |xi | = 0 for all i, 1 ≤ i ≤ n,
i=1
that is xi = 0 for all 1 ≤ i ≤ n.
Hence kxk1 = 0 if and only if x = 0.
Now we prove that kαxk1 = |α|kxk1 for all α ∈ R, x ∈ Rn .
αx = α ∗ (x1 , x2 , · · · , xn ) = (α ∗ x1 , α ∗ x2 , · · · , α ∗ xn ).
Xn X n X n
kαxk1 = |α ∗ xi | = |α| ∗ |xi | = |α| ∗ |xi | = |α| ∗ kxk1 for all x ∈ Rn .
k=1 k=1 k=1
We prove that kx + yk1 ≤ kxk1 + kyk1 for all x, y ∈ Rn .
x = (x1 , x2 , · · · , xn ) and y = (y1 , y2 , · · · , yn ).
This implies x + y = (x1 + y1 , x2 + y2 , · · · , xn + yn ).
X n
So, kx + yk1 = |xi + yi |.
i=1

|xi + yi | ≤ |xi | + |yi | ∀i, 1 ≤ i ≤ n

n
X Xn n
X
|xi + yi | ≤ |xi | + |yi |
i=1 i=1 i=1
kx + yk1 ≤ kxk1 + kyk1 ∀ x, y ∈ Rn
45

Thus k k1 is a norm on Rn .
v
u i=n
uX 2
n
(ii) Define k k2 : R −→ R, kxk2 = kxk2 = t xi then k k2 is a norm on the
i=1
vector space Rn over R.
Let x =s (x1 , x2 , · · · , xn ) ∈ Rn .
X n
kxk2 = x2i .
i=1
n
X
As x2i ≥ 0 for all i, 1 ≤ i ≤ n, we have, x2i ≥ 0.
s n i=1 s n
X √ X
So, x2i ∈ R and since a ≥ 0 for all a ≥ 0, we have x2i ≥ 0.
i=1 i=1
Hence kxk2 ≥ 0 for all xs∈ Rn .
Xn n
X
2
kxk2 = 0 if and only if xi = 0 if and only if x2i = 0
i=1 i=1
n
X
As x2i ≥ 0 for all i, 1 ≤ i ≤ n, x2i = 0 =⇒ x2i = 0 for all i, 1 ≤ i ≤ n, that is
i=1
xi = 0 for all 1 ≤ i ≤ n.
Hence kxk2 = 0 if and only if x = 0.
Now we prove that kαxk2 = |α|kxk2 for all α ∈ R, x ∈ Rn .
αx = α ∗ (x1 , x2 , · · · , xn ) = (α ∗ x1 , α ∗ x2 , · · · , α ∗ xn ).
We prove that kx + yk2 ≤ kxk2 + kyk2 for all x, y ∈ Rn .
x = (x1 , x2 , · · · , xn ) and y = (y1 , y2 , · · · , yn ).
This implies x +sy = (x1 + y1 , x2 + y2 , · · · , xn + yn ).
X n
So, kx + yk2 = (xi + yi )2 .
i=1
v v v
u n u n u n
uX u X uX 2
t (xi + yi )2 ≤ t 2
xi + t yi ∀i, 1 ≤ i ≤ n Minkowski’s Inequality
i=1 i=1 i=1

kx + yk2 ≤ kxk2 + kyk2 ∀ x, y ∈ Rn

Thus k k2 is a norm on Rn .
(iii) Define k k∞ : Rn −→ R, kxk∞ = max {|x1 | : 1 ≤ i ≤ n}. Then k k∞ is a norm
on the vector space Rn over R.
Let x = (x1 , x2 , · · · , xn ) ∈ Rn .
kxk∞ = max{|xi | : 1 ≤ i ≤ n}.
As |xi | ≥ 0 for all i, 1 ≤ i ≤ n, we have, max{xi : 1 ≤ i ≤ n} ≥ 0.
Hence kxk∞ ≥ 0 for all x ∈ Rn .
kxk∞ = 0 if and only if max{|xi | : 1 ≤ i ≤ n} = 0.
Now, 0 ≤ |xi | ≤ 0 for all i, 1 ≤ i ≤ n.
46 CHAPTER 1. UNIT I: METRIC SPACES

Therefore kxk∞ = 0 if and only if |xi | = 0 for all i, 1 ≤ i ≤ n.

This implies kxk∞ = 0 if and only if xi = 0 for all i, 1 ≤ i ≤ n.
Now we prove that kαxk∞ = |α|kxk∞ for all α ∈ R, x ∈ Rn .
αx = α ∗ (x1 , x2 , · · · , xn ) = (α ∗ x1 , α ∗ x2 , · · · , α ∗ xn ).

kαxk∞ = max{|α ∗ xi | : 1 ≤ i ≤ n}
= max{|α||xi | : 1 ≤ i ≤ n}
= |α| ∗ max{|xi | : 1 ≤ i ≤ n} (as |α| ≥ 0)
= |α| ∗ kxk∞

We prove that kx + yk∞ ≤ kxk∞ + kyk∞ for all x, y ∈ Rn .

x = (x1 , x2 , · · · , xn ) and y = (y1 , y2 , · · · , yn ).
This implies x + y = (x1 + y1 , x2 + y2 , · · · , xn + yn ).
So, kx + yk∞ = max{|xi + yi | : 1 ≤ i ≤ n}.

|xi + yi | ≤ |xi | + |yi | for all i, 1 ≤ i ≤ n

=⇒ |xi + yi | ≤ max{|xi | : 1 ≤ i ≤ n} + max{|yi | : 1 ≤ i ≤ n} for all i, 1
=⇒ |xi + yi | ≤ kxk∞ + kyk∞ for all i, 1 ≤ i ≤ n
∴ max{|xi + yi | : 1 ≤ i ≤ n} ≤ kxk∞ + kyk∞
∴ kx + yk∞ ≤ kxk∞ + kyk∞

Thus k k∞ is a norm on Rn .

Z 1
(2) Consider the vector space C[0, 1]. Define k k1 : C[0, 1] −→ R as kf k1 = |f (t)| dt.
0
Then k k1 is a norm on the vector space C[0, 1] over ZR.
1
Let X = C [0 , 1 ], k k1 : X −→ R defined by, kf k1 = |f (t)| dt
0
We prove that k k1 is a norm on C [0 , 1 ].
We have to prove that

(i) kf k1 ≥ 0 for all f ∈ C [0 , 1 ] and kf k1 = 0 if and only if f = 0.

(ii) kαf k1 = |α|kf k1 for all α ∈ R, f ∈ C [0 , 1 ].

(iii) kf + gk1 ≤ kf k1 + kgk1 for all f, g ∈ C [0 , 1 ].

(i) To prove: kf k1 ≥ 0 for all f ∈ C [0 , 1 ] and kf k1 = 0 if and only if f = 0.

We prove that kf k1 ≥ 0 for all fZ∈ C[0, 1].
1
Since |f (t)| ≥ 0 for all t ∈ [0, 1], |f (t)|dt ≥ 0 for all f ∈ C[0, 1].
0
That is kf k1 ≥ 0 for all f ∈ C[0, 1].
47

Now we prove that kf k1 = 0 if and only if f = 0.

Z 1
kf k1 = 0if and only if |f (t)|dt = 0
0
if and only if |f (t)| = 0 for all t ∈ [0, 1]
if and only if f (t) = 0 for all t ∈ [0, 1]

if and only if f = 0

( Since if |f (t0 )| > 0 for some t0 ∈ [0, 1], then by continuity

Z 1 of |f |, ∃ > 0 such
that |f (t)| > 0 for all t ∈ (t0 − , t0 + ) and then |f (t)|dt > 0)
0

(ii) kαf k1 = |α|kf k1 forall α ∈ R, f ∈ C [0 , 1 ].

Z 1
kαf k1 = |αf (t)|dt
0
Z 1
kαf k1 = |α| ∗ |f (t)|dt
0
Z 1
kαf k1 = |α| |f (t)|dt
0
kαf k1 = |α|kf k1

(iii) We prove that kf + gk1 ≤ kf k1 + kgk1 for all f, g ∈ C [0 , 1 ].

Z 1
kf + gk1 = |(f + g)(t)|dt
0
Z 1
= |f (t) + g(t)|dt
0
Z 1
≤ |f (t)| + |g(t)|dt (since |f (t) + g(t)| ≤ |f (t)| + |g(t)| ∀ t ∈ [0, 1])
0
Z 1 Z 1
= |f (t)|dt + |g(t)|dt
0 0
∴ kf + gk1 ≤ kf k1 + kgk1

From (i), (ii) and (iii) k k1 is a norm on C[0, 1].

(3) Consider the vector space C[0, 1]. Define k k2 : C[0, 1] −→ R as kf k∞ = sup {|f (t)| :
t ∈ [0, 1]}. Then k ksup is a norm on the vector space C[0, 1] over R.
48 CHAPTER 1. UNIT I: METRIC SPACES

kαf k∞ = sup{|α ∗ f (t)| : t ∈ [0, 1]}

= sup{|α||f (t)| : t ∈ [0, 1]}
= |α| ∗ sup{|f (t)| : t ∈ [0, 1]} (as |α| ≥ 0)
= |α| ∗ kf k∞

To prove: kf + gk∞ ≤ kf k∞ + kgk∞ for all f, g ∈ C [0 , 1 ].

|(f + g)(t)| = |f (t) + g(t)| ∀ t ∈ [0, 1].

≤ |f (t)| + |g(t)| ∀ t ∈ [0, 1].
≤ sup{|f (t)| : t ∈ [0, 1]} + sup{|g(t)| : t ∈ [0, 1]} ∀ t ∈ [0, 1].
|(f + g)(t)| ≤ kf k∞ + kgk∞ ∀ t ∈ [0, 1]
sup{|(f + g)(t)| : t ∈ [0, 1]} ≤ kf k∞ + kgk∞
kf + gk∞ ≤ kf k∞ + kgk∞

From (i), (ii) and (iii) k k2 is a norm on C[0, 1].

n ∞
X o
1
(4) Consider the vector space l = (xn ) : xn ∈ R, |xn | < ∞ over R. Define
n=1
∞
X
k k1 : l 1 −→ R, kx k1 = |xn |. Then k k1 is a norm on the vector space l 1 over R.
n=1
∞
X
Let l 1 = {(xn ) : |xn | < ∞}.
i=1
∞
X
For x = (xn ), kxk1 = |xn |.
n=1
We prove that k k1 is a norm on l 1 .
We have to prove that

(i) kxk1 ≥ 0 for all x ∈ l 1 and kxk1 = 0 if and only if x = 0.

(ii) kαxk1 = |α|kxk1 forall α ∈ R, x ∈ l 1 .
(iii) kx + yk1 ≤ kxk1 + kyk1 for all x, y ∈ l 1 .

∞
X
kxk1 = |xn |.
n=1
∞
X
As |xn | ≥ 0 for all n ∈ N, we have, |xn | ≥ 0.
n=1
Hence kxk1 ≥ 0 for all x ∈ l 1 .
∞
X ∞
X
kxk1 = 0 if and only if |xn | = 0 if and only if |xn | = 0
n=1 n=1
49

∞
X
As |xn | ≥ 0 for all n, n ∈ N, |xn | = 0 =⇒ |xn | = 0 for all n, n ∈ N, that is xn = 0
n=1
for all n, n ∈ N.
Hence kxk1 = 0 if and only if x = 0.
Now we prove that kαxk1 = |α|kxk1 for all α ∈ R, x ∈ l 1 .
αx = α ∗ (xn , x2 , · · · , xn , · · · ) = (α ∗ x1 , α ∗ x2 , · · · , α ∗ xn , · · · ).
X∞ X∞ X∞
kαxk1 = |α ∗ xn | = |α| |xn | = |α| ∗ |xn | = |α| ∗ kxk1 for all x ∈ l 1 .
n=1 n=1 n=1
We prove that kx + yk1 ≤ kxk1 + kyk1 for all x, y ∈ l 1 .
x = (x1 , x2 , · · · , xn , · · · ) and y = (y1 , y2 , · · · , yn , · · · ).
This implies x + y = (x1 + y1 , x2 + y2 , · · · , xn + yn , · · · ).
n
X n
X n
X
|xi + yi | ≤ |xi | + |yi | ∀i, 1 ≤ i ≤ n
i=1 i=1 i=1
≤ kxk1 + kyk1 for all n ∈ N
n
X
|xi + yi | ≤ kxk1 + kyk1 for all n ∈ N
i=1
n
X
|xi + yi | ≤ kxk1 + kyk1 for all n ∈ N
i=1
∞
X
|xn + yn | ≤ kxk1 + kyk1
n=1
kx + yk1 ≤ kxk1 + kyk1

Thus k k1 is a norm on Rn .
∞
X
(5) Let l 2
= {(xn ) : (xn ) is a sequence of real numbers such that x2n < ∞}. If
n=1
∞
X 21
kxk2 = x2n for x = (x1 , x2 , . . . , xn , . . .) ∈ l 2 , then k k) is a norm on the
n=1
vector space l 2 over R.
X ∞
2
Let l = {(xn ) : xn2 < ∞}.
i=1 s
X∞
For x = (xn ), kxk2 = x2n .
n=1
We prove that k k2 is a norm on l 2 .
We have to prove that

(i) kxk2 ≥ 0 for all x ∈ l 2 and kxk2 = 0 if and only if x = 0.

(ii) kαxk2 = |α|kxk2 forall α ∈ R, x ∈ l 2 .
(iii) kx + yk2 ≤ kxk1 + kyk2 for all x, y ∈ l 2 .
50 CHAPTER 1. UNIT I: METRIC SPACES

s∞
X
kxk2 = x2n .
n=1
∞
X
As x2n ≥ 0 for all n ∈ N, we have, x2n ≥ 0.
s∞ n=1 s∞
X √ X
So, x2n ∈ R and since a ≥ 0 for all a ≥ 0, we have x2n ≥ 0.
n=1 n=1
Hence kxk2 ≥ 0 for all xs∈ l 2 .
X∞ ∞
X
2
kxk2 = 0 if and only if xn = 0 if and only if x2n = 0
n=1 n=1
X∞
As x2n ≥ 0 for all n, n ∈ N, x2n = 0 =⇒ x2n = 0 for all n, n ∈ N, that is xn = 0 for
n=1
all n, n ∈ N.
Hence kxk2 = 0 if and only if x = 0.
Now we prove that kαxk2 = |α|kxk2 for all α ∈ R, x ∈ l 2 .
αx = α ∗ s (xn , x2 , · · · , xn , · ·s
· ) = (α ∗ x1 , α ∗ x2 ,s
· · · , α ∗ xn , · · · ).
X∞ X∞ X∞
2 2
kαxk2 = (α ∗ xn ) = α 2 xn = |α| ∗ x2n = |α| ∗ kxk2 for all x ∈ l 2 .
n=1 n=1 n=1
We prove that kx + yk2 ≤ kxk2 + kyk2 for all x, y ∈ l 2 .
x = (x1 , x2 , · · · , xn , · · · ) and y = (y1 , y2 , · · · , yn , · · · ).
This implies x + y = (x1 + y1 , x2 + y2 , · · · , xn + yn , · · · ).

v v v
u n u n u n
uX u X uX 2
t (xi + yi )2 ≤ t 2
xi + t yi ∀i, 1 ≤ i ≤ n Minkowski’s Inequality
i=1 i=1 i=1

≤ kxk2 + kyk2 for all n ∈ N

v
u n
uX
t (xi + yi )2 ≤ kxk2 + kyk2 for all n ∈ N
i=1
n
X 2
(xi + yi )2 ≤ kxk2 + kyk2 for all n ∈ N
i=1
∞
X 2
2
(xn + yn ) ≤ kxk2 + kyk2
n=1
2
kx + yk22 ≤ kxk2 + kyk2
kx + yk22 ≤ kxk2 + kyk2

Thus k k2 is a norm on Rn .
51

(6) Consider the vector space M2 (R) over R. Define k k : M2 (R) −→ R given by
kAk = max 1≤i,j≤2 |aij | for A = (aij ) is a norm on M2 (R).
X = M2 (R) and kAk = max1≤i,j≤2 {|aij |} for A = (aij )
Clealy |aij | ≥ 0, ∀ i, j 1 ≤ i, j ≤ 2
=⇒ max {|aij | : 1 ≤ i, j ≤ 2} ≥ 0
=⇒ kAk ≥ 0, ∀ A ∈ M2 (R)
∴ k k is nonnegative on M2 (R).
Now if
0 0
A=
0 0
then kAk = max {|aij | : 1 ≤ i, j ≤ 2} = 0
Conversely, kAk = 0
∴ max {|aij | : 1 ≤ i, j ≤ 2} = 0
∴ 0 ≤ |aij | ≤ max1≤i,j≤2 {|aij |} = 0
∴ |aij | = 0 ∀i, j = 1, 2
∴ aij = 0 ∀i, j = 1, 2
∴ A=0
kαAk = max1≤i,j≤2 {|αaij |}
= max1≤i,j≤2 {|α| ∗ |aij |}
= |α| ∗ max1≤i,j≤2 {|aij |}
∴ kα ∗ Ak = |α|kAk, ∀ α ∈ R, ∀A ∈ M2 (R) If A = (aij )2×2 and B = (bij )
then |aij + bij | ≤ |aij | + |bij | ≤ kAk + kBk ∀ i, j = 1, 2
∴ max {|aij + bij | : 1 ≤ i, j ≤ 2} ≤ kAk + kBk
∴ kA + Bk ≤ kAk + kBk, ∀ A, B ∈ M2 (R)
Thus k k satisfies triangle inequality.

(7) C (set of complex numbers) is a normed linear space where norm is the absolute
value of a complex number.

√ 1.2.4 Further, show that nkxk∞ ≤ kxk2 ≤ kxk1 and

Example
kxk1 ≤ nkxk2 ≤ nkxk∞ for x ∈ R .

We show that kxk∞ ≤ kxk2 ≤ kxk1 for all x ∈ Rn .

n
X 2
kxk21 = |xi |
i=1
n
X X
= |xi |2 + 2 |xi ||xj |
i=1 1≤i<j≤n
X
= kxk22 + 2 |xi ||xj |
1≤i<j≤n
X
≥ kxk22 (as 2 |xi ||xj | ≥ 0)
1≤i<j≤n

kxk1 ≥ kxk2 (∗)

52 CHAPTER 1. UNIT I: METRIC SPACES

v v 
u n u n q
uX 2 uX
kxk2 = t xi t x2i = x21 + x22 · · · + x2n 
i=1 i=1
q
≥ x2i for all 1 ≤ i ≤ n
= |xi | for all 1 ≤ i ≤ n
kxk2 ≥ max{|xi | : 1 ≤ i ≤ n}
kxk2 ≥ kxk∞ (∗∗)

From (∗) and (∗∗), we have kxk1 ≥ kxk2 ≥ kxk∞ for all x ∈ Rn . To prove that kxk1 ≤
√
nkxk2 ≤ nkxk∞ for x ∈ Rn .

kxk∞ = max{|xi | : 1 ≤ i ≤ n}
kxk2∞ ≥ |xi |2 1 ≤ i ≤ n
= x2i 1 ≤ i ≤ n
Xn
2
nkxk∞ ≥ x2i
i=1
∴ nkxk2∞ ≥ kxk22
√
∴ nkxk∞ ≥ kxk2
1
∴ kxk∞ ≥ √ kxk2 (1)
n

n
X
kxk22 = x2i
i=1
Xn
= |xi |2
i=1
n
!2
X X
kxk22 = |xi | −2 |xi ||xj | (∗)
i=1 1≤i<j≤n
X
Now,2 |xi ||xj | = 2|x1 ||x2 | + 2|x1 ||x3 | + · · · + 2|x1 ||xn | + 2|x2 ||x3 | + · · · 2|x2 ||xn | + · · · + 2|xn−1
1≤i<j≤n

≤ (|x1 |2 + |x2 |2 ) + · · · + (|x1 |2 + |xn |2 ) + (|x2 |2 + |x3 |2 ) + · · · + (|xn−1 |2 + |xn |2 )

= (n − 1)x21 + (n − 1)x22 + · · · + (n − 1)x2n
X n
= (n − 1) x2i
i=1
X
∴2 |xi ||xj | ≤ (n − 1)kxk22
1≤i<j≤n
X
−2 |xi ||xj | ≥ −(n − 1)kxk22
1≤i<j≤n
53

Substituting this in (∗), we get,

n
!2
X
kxk22 ≥ |xi | − (n − 1)kxk22
i=1
n
!2
X
nkxk22 ≥ |xi |
i=1
nkxk22 ≥ kxk21
√
nkxk2 ≥ kxk1

1 1 1
√ kxk2 ≥ kxk1 (2) dividing both sides by
n n n
1 1
From (1) and (2), we get, kxk∞ ≥ √ kxk2 ≥ kxk1 for all x ∈ Rn .
n n

1.3 Metric Subspaces

Definition 1.3.1 Let (X, d) be a metric space and Y 6= ∅, Y ⊆ X. Then (Y, dY ) is said to
be a subspace of (X, d) if dY (x, y) = d(x, y) for all x, y ∈ Y.

For example: Consider (R, d) where d is the usual distance. Z ⊆ R. So (Z, dZ ) is a

subspace of (R, d) if dZ (m.n) = d(m, n) for all (m, n ∈ Z.

1.4 Open balls in a metric space

Definition 1.4.1 Let (X, d) be a metric space. Let p ∈ X and r > 0. The open ball with
center at p and radius r is denoted by B(p, r) and given by B(p, r) = {x ∈ X : d(x, p) < r.

1.4.1 Examples of open balls in various met-

ric spaces.

(1) Consider (R, d) where d is the usual distance. Then B(p, r) = (p − r, p + r)

That is B(p, r) in (R, usual) is an open interval with center at p and radius r.
p
(2) In (R2 , d2 ) where d2 is the distance induced by the norm kxk2 = x21 + x22 , the
Bd2 (p, r) = {(x1 , x2 ) ∈ R2 : (x1 − p1 )2 + (x2 − p2 )2 < r2 }. That is B(p, r) in
(R2 , Euclidean) is the interior part of the circle with center at p = (p1 , p2 ) and
radius r not including the border.
(3) Consider (R2 , d1 ) where d1 is the distance induced by the norm kxk1 = |x1 | + |x2 |.
Then Bd1 (p, r) = {(x1 , x2 ) ∈ R2 : |x1 − p1 | + |x2 − p2 | < r}. That is Bd1 (p, r) in
(R2 , d1 ) is the interior part of the square with sides along x1 + x2 = r + p1 + p2 , x1 −
x2 = r − p1 + p2 , −x1 + x2 = r + p1 − p2 , x1 + x2 = −(r + p1 + p2 ) and not including
the border.
54 CHAPTER 1. UNIT I: METRIC SPACES

(4) Consider (R2 , d∞ ) where d∞ is the distance induced by the norm kxk∞ = sup{|x1 |, |x2 |}.
Then Bd∞ (p, r) = {(x1 , x2 ) ∈ R2 : sup{|x1 − p1 |, |x2 − p2 |} < r}. That is Bd∞ (p, r)
in (R2 , d∞ ) is the interior part of the square with sides along x2 = r + p2 , −x2 =
r − p2 , x1 = r + p1 , −x1 = −r − p1 and not including the border.

(5) Let V = C [0 , 1 ] be the normed linear space of continuous real valued functions on
[0, 1] under the sup norm that is kf k∞ = sup{|f (t)| : t ∈ [0, 1]} ∀ f ∈ C [0 , 1 ]. Let
f ∈ C[0, 1] and r > 0. Then ψ ∈ B∞ (f, r) if and only if its graph lies in the region
bounded by the curves y = f ± r and the lines x = 0, x = 1.
Z 1
(6) Consider the normed linear space (C [0 , 1 ], k k1 ), where kf k1 = |f (t)|dt for all
0
f ∈ C [0 , 1 ].
Let f ∈ C [0 , 1 ] and > 0. Then ψ ∈ B1 (f, ) if and only if the area under the
graph of |ψ − f | is less than .

(7) Let (X, d) be a discrete metric space. Let p ∈ X and r > 0.

If r > 1 then B(p, r) = X.

If r ≤ 1 then B(p, r) = {p}.

(8) Consider (Z, d) where d is the induced usual metric on R. Let p ∈ Z, r > 0.

If r ≤ 1. Then BZ (p, r) = {p}.

If r > 1. Then BZ (p, r) = {[p − r] + 1, [p − r] + 2, · · · , p, p + 1, · · · [p + r]}.
For exampleB(2, 3.5) = {x ∈ Z : |x − 2| < 3.5} = {x ∈ Z : −3.5 + 2 < x <
2 + 3.5} = {x ∈ Z : −1.5 < x < 5.5} = {−1, 0, 1, 2, 3, 4, 5}.

(9) Let (X, d) be a finite metric space, that is X is finite. Let p ∈ X.

If r ≤ min{d(p, x) : x ∈ X, p 6= x} then B(p, r) = {p}.

(1) Consider (R, d) where d is the usual distance. Then B(p, r) = (p − r, p + r)

That is B(p, r) in (R, usual) is an open interval with center at p and radius r.

B(p, r) = {x ∈ R : d(x, p) < r}

= {x ∈ R : |x − p| < r}
= {x ∈ R : −r < x − p < r}
= {x ∈ R : p − r < x < p + r}
= (p − r, p + r)

Note that any open interval in R can be expressed as an open ball in (R, d) where
d is the usual metric.
Let (a, b) be any open interval in R.
a+b
Let p − r = a and p + r = b.This implies 2p = a + b. So p = .
2
b−a a+b b−a
Also 2r = b − a =⇒ r = . We can verfify that B , = (a, b).
2 2 2
55

p
(2) In (R2 , d2 ) where d2 is the distance induced by the norm kxk2 = x21 + x22 .
Then Bd2 (p, r) = {(x1 , x2 ) ∈ R2 : (x1 − p1 )2 + (x2 − p2 )2 < r2 }. That is B(p, r)
in (R2 , Euclidean) is the interior part of the circle with center at p = (p1 , p2 ) and
radius r excluding the border.
n o
Bd2 (p, r) = (x1 , x2 ) ∈ R2 : d2 (x1 , x2 ), (p1 , p2 ) < r
p
= {(x1 , x2 ) ∈ R2 : (x1 − p1 )2 + (x2 − p2 )2 < r}
= {(x1 , x2 ) ∈ R2 : (x1 − p1 )2 + (x2 − p2 )2 < r2 }

For example: Consider Bd2 (0, 0), 1 .

1
x21 + x22 = 1

−1 0 1 2

−1

−2

(3) Consider (R2 , d1 ) where d1 is the distance induced by the norm kxk1 = |x1 | + |x2 |.
Then Bd1 (p, r) = {(x1 , x2 ) ∈ R2 : |x1 − p1 | + |x2 − p2 | < r}. That is Bd1 (p, r) in
(R2 , d1 ) is the interior part of the square with sides along x1 + x2 = r + p1 + p2 , x1 −
x2 = r − p1 + p2 , −x1 + x2 = r + p1 − p2 , x1 + x2 = −(r + p1 + p2 ) excluding the
border.

Bd1 (p, r) = {(x1 , x2 ) ∈ R2 : |x1 − p1 | + |x2 − p2 | < r}

If x1 − p1 ≥ 0 and x2 − p2 ≥ 0. Then |x1 − p1 | = x1 − p1 , |x2 − p2 | = x2 − p2 .

Hence |x1 − p1 | + |x2 − p2 | < r becomes x1 − p1 + x2 − p2 < r =⇒ x1 + x2 <
r + p1 + p 2 .
If x1 − p1 ≥ 0 and x2 − p2 < 0. Then |x1 − p1 | = x1 − p1 , |x2 − p2 | = −x2 + p2 .
Hence |x1 − p1 | + |x2 − p2 | < r becomes x1 − p1 − x2 + p2 < r =⇒ x1 − x2 <
r + p1 − p2 .
If x1 − p1 < 0 and x2 − p2 ≥ 0. Then |x1 − p1 | = −x1 + p1 , |x2 − p2 | = x2 − p2 .
Hence |x1 − p1 | + |x2 − p2 | < r becomes −x1 + p1 + x2 − p2 < r =⇒ −x1 + x2 <
56 CHAPTER 1. UNIT I: METRIC SPACES

r − p1 + p2 .
If x1 − p1 < 0 and x2 − p2 < 0. Then |x1 − p1 | = −x1 + p1 , |x2 − p2 | = −x2 + p2 .
Hence |x1 − p1 | + |x2 − p2 | < r becomes −x1 + p1 − x2 + p2 < r =⇒ −x1 − x2 <
r − p1 − p2 .
57

Bd1 (p1 , p2 ), r .

y = x + r − p1 + p2 y = −x + r + p1 + p2

(p1 , p2 )

r
y = −x − r + p1 + p2 y = x − r − p1 + p2

(0, 0)
58 CHAPTER 1. UNIT I: METRIC SPACES

B∞ ((p1 , p2 ), r) = {(x1 , x2 ) ∈ R2 : k(x1 , x2 ) − (p1 , p2 )k∞ < 1}

= {(x1 , x2 ) ∈ R2 : max{|x1 − p1 |, |x2 − p2 |} < 1}

If max{|x1 − p1 |, |x2 − p2 |} = |x1 − p1 | then max{|x1 − p1 |, |x2 − p2 |} < r =⇒ |x1 −

p1 | < r =⇒ p1 − r < x1 < p1 + r.
If max{|x1 − p1 |, |x2 − p2 |} = |x2 − p2 | then max{|x1 − p1 |, |x2 − p2 |} < r =⇒ |x2 −
p2 | < r =⇒ p2 − r < x2 < p2 + r.

B∞ B (p1 , p2 ), r .

y = p2 + r

x = p1 − r (p1 , p2 ) x = p1 + r
r
y = p2 − r

(0, 0)

Z 1
(6) Consider the normed linear space (C [0 , 1 ], k k1 ), where kf k1 = |f (t)|dt for all
0
f ∈ C [0 , 1 ].
Let f ∈ C [0 , 1 ] and > 0. Then ψ ∈ B1 (f, ) if and only if the area under the
graph of |ψ − f | is less than .

(7) Let (X, d) be a discrete metric space. Let p ∈ X and r > 0.

Case 1: 1 < r B(p, r) = {x ∈ X : d(x, p) < r}

We will show that B(p, r) = X.
Clearly B(p, r) ⊆ X.
We will prove that X ⊆ B(p, r).
Let x ∈ X. Then d(x, p) = 0 or d(x, p) = 1.
If d(x, p) = 0 then d(x, p) < 1 < r. This implies x ∈ B(p, r).
If d(x, p) = 1 then d(x, p) = 1 < r. This implies x ∈ B(p, r).
In any case x ∈ B(p, r). Hence X ⊆ B(p, r) and thus B(p, r) = X.
Case 2: r ≤ 1 B(p, r) = {x ∈ X : d(x, p) < r}
We will prove that B(p, r) = {p}.
Since d(p, p) = 0 < r, p ∈ B(p, r). Hence {p} ⊆ B(p, r).
Now, let x ∈ B(p, r). So d(x, p) < r. Since r < 1, d(x, p) < r < 1. This implies
d(x, p) 6= 1. Hence d(x, p) = 0 and thus x = p.
Hence x ∈ {p}. Thus B(p, r) ⊆ {p}.
Therefore B(p, r) = {p}.

(8) Consider (Z, d) where d is the induced usual metric on R. Let p ∈ Z, r > 0.
BZ (p, r) = {x ∈ Z : d(x, p) < r} = {x ∈ Z : |x − p| < r} = {x ∈ Z : p − r < x <
p + r} = Z ∩ (p − r, p + r). (∗)

Case 1: r ≤ 1.
We will prove that BZ (p, r) = {p}.
Since d(p, p) = 0 < r, p ∈ B(p, r). Hence {p} ⊆ BZ (p, r).
Let x ∈ BZ (p, r). So x ∈ (p − r, p + r) ∩ Z. (from (∗))
Since r < 1, (p − r, p + r) ⊂ (p − 1, p + 1).
Therefore (p − r, p + r) ∩ Z ⊆ (p − 1, p + 1) ∩ Z.
As (p − 1, p + 1) ∩ Z = {p}, (p − r, p + r) ∩ Z ⊆ {p} and hence BZ (p, r) ⊆ {p}..
Thus BZ (p, r) = {p}.
Case 2: r > 1.
Then BZ (p, r) = (p−r, p+r)∩Z = {[p−r]+1, [p−r]+2, · · · , p, p+1, · · · [p+r]}.

For example B(2, 3.5) = {x ∈ Z : |x − 2| < 3.5} = {x ∈ Z : −3.5 + 2 < x <

2 + 3.5} = {x ∈ Z : −1.5 < x < 5.5} = {−1, 0, 1, 2, 3, 4, 5}.

(9) Let (X, d) be a finite metric space. Therefore X is finite. Let X = {a1 , a2 , · · · , an }.
Let p ∈ X. Then p = ai for some i, 1 ≤ i ≤ n.
Let r ≤ min{d(p, aj ) : 1 ≤ j ≤ n, j 6= i}. So r > 0.
We will show that B(p, r) = {p}.
60 CHAPTER 1. UNIT I: METRIC SPACES

Since d(p, p) = 0 < r, p ∈ B(p, r). Hence {p} ⊆ BZ (p, r).

Let x ∈ B(p, r). Since x ∈ X, x = aj for some j, 1 ≤ j ≤ n.
So, aj ∈ B(p, r) and hence d(aj , p) < r.
But r ≤ min{d(p, aj ) : 1 ≤ j ≤ n, j 6= i}.
Therefore d(aj , p) < r ≤ d(aj , p) for all j 6= i.
So, j 6= i =⇒ d(aj , p) < d(aj , p). This is a contradiction.
Therefore j = i and hence aj = ai = p.
Thus x ∈ B(p, r) =⇒ x = p.
That is B(p, r) ⊆ {p}.
Hence B(p, r) = {p}.

Examples 1.4.2

(1) Consider the metric spaces (Z, d) and (Z, d1 ) where d is the distance induced usual
metric on R and d1 is the discrete metric. Show that for every r > 0, there exists
r0 > 0 such that Bd (x, r0 ) ⊆ Bd1 (x, r) and for every r > 0, there exists r00 > 0 such
that Bd1 (x, r00 ) ⊆ Bd (x, r) for all x ∈ X.
Also show that there exists r1 > 0 such that Bd1 (x, r1 ) 6= Bd (x, r) for any r > 0 for
all x ∈ X.

Part I Let r > 0 and consider Bd1 (x, r).

Choose r0 = 1.

Bd (x, 1) = {y ∈ Z : d(y, x) < 1}

= {y ∈ Z : |y − x| < 1}
= {y ∈ Z : −1 + x < y < x + 1}
= {y ∈ Z : y ∈ (x − 1, x + 1)
= Z ∩ (x − 1, x + 1)
= {x} (as x ∈ Z)

Now, consider Bd1 (x, r).

Since x ∈ Bd1 (x, r), {x} ⊆ Bd1 (x, r).
Put {x} = Bd (x, 1).
Therefore Bd (x, 1) ⊆ Bd1 (x, 1). (I)

Part II Let r > 0 and consider Bd (x, r).

Choose r00 = 1.

Bd1 (x, 1) = {y ∈ Z : d1 (y, x) < 1}

= {y ∈ Z : d1 (y, x) = 0} (as d1 is the discrete metric on Z)
= {y ∈ Z : y = x}
= {y ∈ Z : y ∈ (x − 1, x + 1)
= {x}
61

Now, consider Bd (x, r).

Since x ∈ Bd (x, r), {x} ⊆ Bd (x, r).
Put {x} = Bd1 (x, 1).
Therefore Bd1 (x, 1) ⊆ Bd (x, 1). (II)
Part III Choose r1 = 2.
Since r1 > 1, Bd1 (x, r1 ) = Z.
But Bd (x, r) = (x − r, x + r) for every r > 0.
There does not exist any r > 0 such that (x − r, x + r) = Z = Bd1 (x, 2).
Thus there does not exist any r > 0 such that Bd (x, r) = Bd1 (x, 2).

(2) Let (X, d) be a metric space. Consider the metric space d1 : X ×X −→ R, d1 (x, y) =
d(x, y)
. Show that for every r > 0, there exists r0 > 0 such that Bd (x, r0 ) ⊆
1 + d(x, y)
Bd1 (x, r) and for every r > 0, there exists r00 > 0 such that Bd1 (x, r00 ) ⊆ Bd (x, r) for
all x ∈ X.
r
Also show that for 0 < r < 1, Bd1 (x, r) = Bd (x, 1−r ).

Part I Let r > 0 and consider Bd1 (x, r).

Since d1 (x, y) ≤ d(x, y) for all x, y ∈ X, choose r0 = r. (from examples 1.1.4
no. 4)
Claim: Bd (x, r0 ) ⊆ Bd1 (x, r).
Let y ∈ Bd (x, r0 ). So, d(y, x) < r0 = r.
Then d1 (y, x) ≤ d(y, x) =⇒ d1 (y, x) < r and hence y ∈ Bd1 (x, r).
Thus Bd (x, r0 ) ⊆ Bd1 (x, r) where r0 = r. (I)
Part II Let r > 0 and consider Bd (x, r).
r
Choose r00 = .
1+r
00
Claim: Bd1 (x, r ) ⊆ Bd (x, r).
r d(y, x) r
Let y ∈ Bd1 (x, r0 ). So d1 (y, x) < r00 = . This implies < .
1+r 1 + d(y, x) 1+r
So, (1 + r)d(y, x) < (1 + d(y, x))r. Hence d(y, x) + rd(y, x) < r + rd(y, x). This
implies d(y, x) < r.
Hence y ∈ Bd (x, r) and Bd1 (x, r00 ) ⊆ Bd (x, r).
r
Part III show that for 0 < r < 1, Bd1 (x, r) = Bd (x, 1−r ).

Let y ∈ Bd1 (x, r). r > 0)

r
=⇒ d1 (y, x) < r. Hence y ∈ Bd (x, 1−r )
d(y, x) r
=⇒ Bd1 (x, r) ⊆ Bd (x, 1−r )
∴ < r.
1 + d(y, x)
∴ d(y, x) < r(1 + d(y, x))
∴ d(y, x) < r + rd(y, x)
∴ d(y, x) − rd(x, y) < r
∴ d(y, x)(1 − r) < r
r
∴ d(y, x) < (as 0 < r < 1, 1 −
1−r
62 CHAPTER 1. UNIT I: METRIC SPACES

r
Let y ∈ Bd (x, 1−r ). d(y, x)
r
∴ <r
=⇒ d(y, x) < 1−r . 1 + d(y, x)
∴ (1 − r)d(y, x) < r. (as1 − r > 0) ∴ d1 (y, x) < r
∴ d(y, x) − rd(y, x) < r Hence y ∈ Bd1 (x, r)
r
∴ d(y, x) < r + rd(y, x) =⇒ Bd (x, ) ⊆ Bd1 (x, r)
1−r
∴ d(y, x) < r(1 + d(x, y))

r
Hence Bd1 (x, r) = Bd (x, 1−r ) for all 0 < r < 1.

R2 induced by k k1 , k k2 and k k∞ respec-

(3) Let d1 , d2 , d∞ be three metrics defined on p
tively, where kxk1 = |x1 | + |x2 |, kxk2 = x21 + x22 , kxk∞ = max{|x1 |, |x2 |} ∀x =
(x1 , x2 ) ∈ R2 . Prove the following:

(i) For every r > 0, there exists r0 > 0 such that Bd1 (x, r0 ) ⊆ Bd2 (x, r)
(ii) For every r > 0, there exists r00 > 0 such that Bd2 (x, r00 ) ⊆ Bd1 (x, r) for all
x ∈ X.
(iii) For every r > 0, there exists r1 > 0 such that Bd1 (x, r1 ) ⊆ Bd∞ (x, r)
(iv) For every r > 0, there exists r2 > 0 such that Bd∞ (x, r2 ) ⊆ Bd1 (x, r) for all
x ∈ X.
(v) For every r > 0, there exists r3 > 0 such that Bd2 (x, r3 ) ⊆ Bd∞ (x, r)
(vi) For every r > 0, there exists r4 > 0 such that Bd∞ (x, r4 ) ⊆ Bd2 (x, r) for all
x ∈ X.

(i) For every r > 0, there exists r0 > 0 such that Bd1 (x, r0 ) ⊆ Bd2 (x, r)
(ii) For every r > 0, there exists r00 > 0 such that Bd2 (x, r00 ) ⊆ Bd1 (x, r) for all
x ∈ X.
√
From example 1.1.4 no.2, we have, d∞ (x, y) ≤ d2 (x, y) ≤ d1 (x, y) ≤ nd2 (x, y) ≤
nd∞ (x, y) for all x, y ∈ Rn . √
Here n = 2. Hence d∞ (x, y) ≤ d2 (x, y) ≤ d1 (x, y) ≤ 2d2 (x, y) ≤ 2d∞ (x, y)
for all x, y ∈ R2 .

(i) Let r > 0. We show that So Bd1 (x, r0 ) ⊆ Bd2 (x, r)

∃ r0 > 0 such that
Bd1 (x, r0 ) ⊆ Bd2 (x, r)
From (*) d2 (x, y) ≤ d1 (x, y) (I)
Choose r0 = r.
(We can choose any r0 ≤ r.)
Claim: Bd1 (x, r0 ) ⊆ Bd2 (x, r)
Let y ∈ Bd1 (x, r0 ).
=⇒ d1 (y, x) < r0 = r.
From (I), d2 (y, x) ≤ d1 (y, x) < r.
So, d2 (y, x) < r.
=⇒ y ∈ Bd2 (x, r).
63

(ii) Let r > 0. We show that Let y ∈ Bd2 (x, r00 ).

∃ r00 > 0 such that r
=⇒ d2 (y, x) < r00 = √ .
Bd2 (x, r00 ) ⊆ Bd1 (x, r)√ √ 2
From (*), d1 (x, y) ≤ 2d2 (x, y) (II) That is , 2 ∗ d2 (y, x)√< r.
d1 (x, y) From (II), d1 (y, x) ≤ 2d2 (y, x) <
That is, √ ≤ d2 (x, y) r.
2
r Therefore, d1 (y, x) < r
Choose r00 = √ (any r00 ≤ √r2 ).
2 =⇒ y ∈ Bd1 (x, r).
Claim: Bd2 (x, r00 ) ⊆ Bd1 (x, r) Hence Bd2 (x, r00 ) ⊆ Bd1 (x, r)

√r
2
r(p1 , p2 ) r(p1 , p2 )

(0, 0) (0, 0)

Figure 1.1: Bd1 (x, r0 ) ⊆ Bd2 (x, r) Figure 1.2: Bd2 (x, r00 ) ⊆ Bd1 (x, r)

(iii) For every r > 0, there exists r1 > 0 such that Bd1 (x, r1 ) ⊆ Bd∞ (x, r)
(iv) For every r > 0, there exists r2 > 0 such that Bd∞ (x, r2 ) ⊆ Bd1 (x, r) for all
x ∈ X.

(iii) Let r > 0. We show that (iv) Let r > 0. We show that
∃ r1 > 0 such that ∃ r2 > 0 such that
Bd1 (x, r1 ) ⊆ Bd∞ (x, r) Bd∞ (x, r2 ) ⊆ Bd1 (x, r)
From (*) d∞ (x, y) ≤ d1 (x, y) (I) From (*), d1 (x, y) ≤ 2d∞ (x, y) (II)
(We can choose any r1 ≤ r.) d1 (x, y)
That is, ≤ d∞ (x, y)
Choose r1 = r. 2
r
Claim: Bd1 (x, r1 ) ⊆ Bd∞ (x, r) Choose r2 = (any r2 ≤ 2r .).
2
Let y ∈ Bd1 (x, r1 ). Claim: Bd∞ (x, r2 ) ⊆ Bd1 (x, r)
=⇒ d1 (y, x) < r1 = r. Let y ∈ Bd∞ (x, r2 ).
From (I), d∞ (y, x) ≤ d1 (y, x) < r. r
=⇒ d∞ (y, x) < r2 = .
So, d∞ (y, x) < r. 2
=⇒ y ∈ Bd∞ (x, r). That is , 2 ∗ d∞ (y, x) < r.
So Bd1 (x, r1 ) ⊆ Bd∞ (x, r) From (II), d1 (y, x) ≤ 2d∞ (y, x) <
r.
Therefore, d1 (y, x) < r
64 CHAPTER 1. UNIT I: METRIC SPACES

=⇒ y ∈ Bd1 (x, r). Hence Bd∞ (x, r2 ) ⊆ Bd1 (x, r)

r
2
(p1 , p2 ) (p1 , p2 )
r

Figure 1.3: Bd1 (x, r1 ) ⊆ Figure 1.4: Bd∞ (x, r2 ) ⊆ Bd1 (x, r)
Bd∞ (x, r), r1 = r
65

(v) For every r > 0, there exists r3 > 0 such that Bd2 (x, r3 ) ⊆ Bd∞ (x, r)
(vi) For every r > 0, there exists r4 > 0 such that Bd∞ (x, r4 ) ⊆ Bd2 (x, r) for all
x ∈ R2 ..

(v) Let r > 0. We show that (vi) Let r > 0. We show that
∃ r3 > 0 such that ∃ r4 > 0 such that
Bd2 (x, r3 ) ⊆ Bd∞ (x, r) Bd∞ (x, r4 ) ⊆ Bd2√ (x, r)
From (*) d∞ (x, y) ≤ d2 (x, y) (I) From (*), 2d2 (x, y) ≤
Choose r3 = r. 2d∞ (x, y) (II)
(We can choose any r3 ≤ r.) d2 (x, y)
That is, √ ≤ d∞ (x, y)
Claim: Bd2 (x, r3 ) ⊆ Bd∞ (x, r) 2
Let y ∈ Bd2 (x, r3 ). r
Choose r4 = √ (any r4 ≤ √r2 ).
=⇒ d2 (y, x) < r3 = r. 2
From (I), d∞ (y, x) ≤ d2 (y, x) < r. Claim: B d∞ (x, r4 ) ⊆ Bd2 (x, r)

So, d∞ (y, x) < r. Let y ∈ Bd∞ (x, r4 ).

r
=⇒ y ∈ Bd∞ (x, r). =⇒ d∞ (y, x) < r4 = √ .
√ 2
So Bd2 (x, r3 ) ⊆ Bd∞ (x, r)
That is , 2 ∗ d∞ (y, x) √ < r.
From (II), d2 (y, x) ≤ 2d∞ (y, x) <
r.
Therefore, d2 (y, x) < r
=⇒ y ∈ Bd2 (x, r).
Hence Bd∞ (x, r4 ) ⊆ Bd2 (x, r)

r4 = √r
(p1 , p2 ) r 2

(p1 , p2 )
r3 = r r

(0, 0) (0, 0)

Figure 1.5: Bd2 (x, r0 ) ⊆ Bd∞ (x, r) Figure 1.6: Bd∞ (x, r00 ) ⊆ Bd2 (x, r)
66 CHAPTER 1. UNIT I: METRIC SPACES

(4) Let d1 and d2 be metrics on a non-empty set X such that there exist k1 , k2 > 0 such
that k1 d1 (x, y) ≤ d2 (x, y) ≤ k2 d1 (x, y), ∀x, y ∈ X. Show that for every r > 0 there
exists r0 > 0 such that Bd2 (x, r0 ) ⊆ Bd1 (x, r) and for every r > 0 there exists r00 > 0
such that Bd1 (x, r00 ) ⊆ Bd2 (x, r).

Let r > 0. We show that Let r > 0. We show that

∃ r0 > 0 such that ∃ r00 > 0 such that
Bd2 (x, r) ⊆ Bd1 (x, r) Bd1 (x, r00 ) ⊆ Bd2 (x, r)
Given: k1 d1 (x, y) ≤ d2 (x, y) (I) Given: d2 (x, y) ≤ k2 d1 (x, y) (II)
Choose r0 = k1 r. d2 (x, y)
That is, ≤ d1 (x, y)
(We can choose any r0 ≤ k1 r.) k2
r
Claim: Bd2 (x, r0 ) ⊆ Bd1 (x, r) Choose r00 = (any r00 ≤ kr2 ).
Let y ∈ Bd2 (x, r0 ). k2
Claim: Bd1 (x, r00 ) ⊆ Bd2 (x, r)
=⇒ d2 (y, x) < r0 = k1 r.
Let y ∈ Bd1 (x, r00 ).
From (I), k1 d1 (y, x) ≤ d2 (y, x) < k1 r. r
So, d1 (y, x) < r. (as k1 > 0) =⇒ d1 (y, x) < r00 = .
k2
=⇒ y ∈ Bd1 (x, r). That is , k2 ∗ d1 (y, x) < r.
So Bd2 (x, r0 ) ⊆ Bd1 (x, r) From (II), d2 (y, x) ≤ k2 d1 (y, x) < r.
Therefore, d2 (y, x) < r
=⇒ y ∈ Bd2 (x, r).
Hence Bd1 (x, r00 ) ⊆ Bd2 (x, r).

1.5 Hausdorff Property

Hausdorff property: Let (X, d) be a metric space and x, y ∈ X, x 6= y. Then x and y have
disjoint neighbourhoods. [2 marks]
Since x 6= y ∴ d(x, y) > 0 .
d(x, y)
Let r =
2
Let N1 = B(x, r) and N2 = B(y, r).
∴ N1 andTN2 are neighbourhoods of x and y respectively. [1 mark]
TST N1 N T2 = ∅
Let p ∈ N1 N2 .
∴ d(x, p) < r and d(y, p) < r.
By triangle inequality d(x, y) ≤ d(x, p) + d(p, y) < r + r
∴ d(x, y) < 2r.
But d(x, y) = 2r.
∴ 2r < 2r.
Since r > 0 ∴ 2 < 2 which is a contradiction.

1.5.1 Open Balls in a Subspace

Theorem 1.5.1 Let (Y, d) be a subspace of the metric space (X, d) and let p ∈ Y and
r > 0. Let BY (p, r) = {x ∈ Y : d(x, p) < r}. Then BY (p, r) = B(p, r) ∩ Y where
B(p, r) = {x ∈ X : d(x, p) < r}.

Proof:

BY (p, r) = {x ∈ Y : d(x, p) < r}

= Y ∩ {x ∈ X : d(x, p) < r} (as Y ⊆ X)
= Y ∩ B(x, r).

1.6 Open Sets In A Metric Space

Definition 1.6.1 Let (X, d) be a metric space. Let G 6= ∅, G ⊆ X.G is said to be open
if for each x ∈ G, there exists r > 0 such that B(x, r) ⊆ G. An empty set is defined to be
open.

Definition 1.6.2 Let (X, d) be a metric space and let et G ⊆ X.G is said to be not open
if there exists x ∈ G, such that, for every r > 0, B(x, r) 6⊆ G.

1.7 Properties of open sets

Theorem 1.7.1 (i) Arbitrary union of open sets is open.

(ii) Finite intersection of open sets is open.

(iii) Arbitrary intersection of open sets need not be open.

Proof:

(i) Arbitrary union of open sets is open.

Let {Gα }[
α∈I be an arbitrary collection of open subsets of X.
Let G = Gα .
α∈I
If G = ∅ thenG is open.
Suppose G 6= ∅.
Let x ∈ G.
∴ x ∈ Gα0 for some α0 ∈ I.
Gα0 is open,
∴ ∃ r > 0 such that B(x, r) ⊆ Gα) .
But Gα0 ⊆ G.
∴ B(x, r) ⊂ G and G is open.

(ii) Finite intersection of open sets is open.

Let {G1 , G2 , · · · , Gn }α∈I be a finite family of open subsets of X.
68 CHAPTER 1. UNIT I: METRIC SPACES

\
Let G = Gi .
1≤i≤n
If G = ∅ then G is open.
Suppose G 6= ∅.
Let x ∈ G.
∴ x ∈ Gi for all i, 1 ≤ i ≤ n.
Gi is open for all i, 1 ≤ i ≤ n.
=⇒ ∃ ri > 0 such that B(x, ri ) ⊆ Gi) for all i, 1 ≤ i ≤ n. · · · (I)
Choose r = min{ri : 1 ≤ i ≤ n}.
Therefore r ≤ ri ,for all i, 1 ≤ i ≤ n.
Hence, B(x, r) ⊆ B(x, ri ) for all i, 1 ≤ i ≤ n.
Hence, B(x, r) ⊆ B(x, ri ) ⊆ Gi for all i, 1 ≤ i ≤ n. · · · (from (I))
Therefore B(x, r) ⊆ Gi for all i, 1 ≤ i ≤ n.
Thus B(x, r) ⊆ ∩1≤i≤n Gi
That is, B(x, r) ⊆ G.
Thus, for every x ∈ G, ∃ r > 0 such that B(x, r) ⊆ G.
Hence G = ∩1≤i≤n Gi is an open set.

(iii) An arbitrary intersection

of open sets need not be open.
−1 1
Consider Gn = , in (R, d) where d is the usual distance.
n\ n
We will show that Gn = {0}.
n∈N
−1 1 \
Since < 0 < for all n ∈ N, 0 ∈ Gn .
n n n∈N
\
We will show that only 0 ∈ Gn .
n∈N
Consider x ∈ R such that x 6= 0. This means 0 < |x|.
1
So, by Archimedean Property, there exists n0 ∈ N such that < |x|.
n0

1 1 −1 1
If x > 0 then |x| = x. So < |x| =⇒ < x =⇒ x ∈
/ , .
n0 n0 n0 n0
1 1 −1
If x < 0 then |x| = −x. So < |x| =⇒ < −x =⇒ x < =⇒ x ∈
/
n0 n0 n0
−1 1
, .
n0 n0
\
In any case, x ∈
/ Gn .
n∈N \ \
Thus we have proved that 0 ∈ Gn and if x 6= 0 then x ∈
/ Gn .
\ n∈N n∈N
Hence Gn = {0}.
n∈N
Now, we will show that {0} is not open.
69

0 ∈ {0}. For any r > 0, B(0, r) = (−r, r).

Since (−r, r) contains infinitely many real numbers, (−r, r) 6⊆ {0}.
Thus there does not exist any r > 0 such that B(0, r) ⊆ {0}.
Hence {0} is not open.
\ 1 1
Therefore − , is not open.
n n

1.7.1 Examples of Open Sets and Non Open Sets in Var-

ious Metric Spaces.

(1) Open balls are open sets in every metric space. But an open set need not be an
open ball.

(2) Open intervals are open sets in (R, d) where d is the usual metric.

(3) Every subset of a discrete metric space is open.

(4) Every subset of a finite metric space is open.

(5) Every subset of (Z, d) were d is the induced metric from R with usual metric.

(6) Let p ∈ X and r > 0. Let A = {x ∈ X : d(x, p) > r} is an open set.

(7) U = {(x, y) ∈ R2 : xy 6= 0}. Then U is open in (R2 , d) where d is the Euclidean

distance.

(8) (a, b) × (c, d) is an open subset of (R2 , d) where d is the Euclidean distance is an
open set.

(9) U = {(x, y) ∈ R2 : 2x + 3y < 1} is an open subset of R2 with Euclidean metric.

(10) U = {(x , y) ∈ R2 : −1 < x + y < 1 } is an open subset of (R2 , d) where d is the

Euclidean metric.

(11) Let (X, d1 ) and (Y, d2 ) be metric spaces. Consider the metric d : (X × Y ) ×
(X × Y ) −→ R defined by d((x1 , y1 ), (x2 , y2 )) = max {d1 (x1 , x2 ), d2 (y1 , y2 )}. Let
p ∈ X, q ∈ Y and r, s > 0. B(p, r) × B(q, s) is an open set in (X × Y, d). item Let
(X, d1 ), (Y, d2 ) be metric spaces. Consider
the product space X ×Y with the product
metric d : (X × Y ) × (X × Y ) −→ R, d (x1 , y1 ), (x2 , y2 ) = max{d1 (x1 , y1 ), (x2 , y2 )}.
Let U, V be open subsets of X, Y respectively. Show that U × V is an open subset
of X × Y.

(12) Let (X, k k) be a normed linear space and A 6= ∅, A ⊆ X. Show that if U 6= ∅, U ⊆ X

is an open set then U + A is open.

(13) The open ball B1 (0, 1) = {f ∈ C [0 , 1 ] : kf k1 < 1 } in (C [0 , 1 ], k k1 ) is open in

(C [0 , 1 ], k k∞ ). (Hint kf k1 ≤ kf k∞ for any f ∈ C[0, 1].)
70 CHAPTER 1. UNIT I: METRIC SPACES

(14) The open ball B∞ (0, 1) = {f ∈ C [0 , 1 ] : kf k∞ < 1 } in ([0 , 1 ], k k∞ ), is not open

in C [0 , 1 ], k k1 ). (Hint: Construct a positive continuous function f on [0, 1] such
that the area under the grap is less than any positive δ but whose maximum could
be as large as we want. This implies that for no δ > 0, the open ball B1 (0, δ) can
be a subset of B∞ (0, 1). )

(1) Open balls are open sets in every metric space. But an open set need not be an
open ball.
B(x, r) = {y ∈ X : d(x, y) < r}. [2 marks]

Since d(x, x) = 0 < r, so x ∈ B(x, r) =⇒ B(x, r) 6= ∅. [1 mark]

Let y ∈ B(x, r).
∴ d(x, y) < r.
So r − d(x, y) > 0.
Let δ = r − d(x, y) ∴ δ > 0.
Claim B(y, δ) ⊆ B(x, r).
Let p ∈ B(y, δ) =⇒ d(p, y) < δ.
Consider d(p, x).
By triangle inequality, d(p, x) ≤ d(p, y) + d(y, x)
∴ d(p, x) < δ + d(y, x).
∴ d(p, x) < r − d(x, y) + d(y, x) (since δ = r − d(x, y))
∴ d(p, x) < r (since d(x, y) = d(y, x)) ∴ p ∈ B(x, r).
So, B(y, δ) ⊆ B(x, r).
Thus, ∀y ∈ B(x, r), ∃δ > 0 such that B(y, δ) ⊆ B(x, r).
Hence B(x, r) is an open set. [3 marks]

In (R, d) where d is the usual metric, R is an open set, but R 6= B(x, r) for any
x, r ∈ R and r > 0.
For if R = B(x, r) for some x, r ∈ R, r > 0,
then R = (x − r, x + r) (In (R, usual), B(x, r) = (x − r, x + r))
So, R is bounded.
and hence a contradiction.
So, R 6= B(x, r) for any x, r ∈ R and r > 0.
Thus R is an open set but it is not an open ball in (R, usual) [2 marks]

(2) Open intervals are open sets in (R, d) where d is the usual metric.

(3) Every subset of a discrete metric space

is open.
Let p ∈ X. B(p, 1) = {x ∈ X : d(x, p) < 1}.
= {x ∈ X : d(x, p) = 0}.
= {x ∈ X : x = p}.
= {p}.
71

Let G ⊆ X. (5) Every subset of (Z, d) were d is the in-

If G = ∅ then G is open. duced metric from R with usual met-
S G 6= ∅.
So, suppose ric.
∴ G = Sp∈G {p} Let p ∈ Z.
∴ G = p∈G B(p, 1)
Since open balls are open sets, BZ (p, 1) = {x ∈ Z : d(x, p) < 1}.
we have expressed G as a union of = {x ∈ Z : |x − p| < 1}.
open sets. = {x ∈ Z : x ∈ (p − 1, p + 1)}.
G is open in (X, d)
= Z ∩ (p − 1, p + 1).
(as arbitrary union of open sets is
open) = {p}. (as p ∈ Z.)

Let G ⊆ Z. If G = ∅ then G is open.

S G 6= ∅.
So, suppose
∴ G = Sp∈G {p}
∴ G = p∈G BZ (p, 1)
Since open balls are open sets,
we have expressed G as a union of
open sets.
G is open in (X, d)
(as arbitrary union of open sets is
open)

(4) Every subset of a finite metric space is open.

Let (X, d) be a finite metric space.
So, X = {a1 , a2 , · · · , an }.
We show that every singleton set is open.
Let p ∈ X.
This implies p = ai for some i, 1 ≤ i ≤ n.
Let r = min{d(p, aj ) : 1 ≤ j ≤ n, j 6= i}
Since j 6= i, aj 6= ai =⇒ aj 6= p.
=⇒ d(p, aj ) > 0 ∀ j, j 6= i, 1 ≤ j ≤ n.
Therefore r > 0.
Claim: B(p, r) = {p} Let x ∈ B(p, r). So d(x, p) < r.
But r ≤ d(p, aj ), for all 1 ≤ j ≤ n, j 6= i.
This implies d(x, p) < d(aj , p), for all 1 ≤ j ≤ n, j 6= i.
Hence x 6= aj for all j, 1 ≤ j ≤ n, j 6= i.
Thus we have, x ∈ {a1 , a2 , · · · , ai = p, · · · , an } and x 6= aj for all j, 1 ≤ j ≤ n, j 6= i.
Hence x = ai . That is x = p.
Hence B(p, r) ⊆ {p}.
As d(p, p) = 0 < r, p ∈ B(p, r). So {p} ⊆ B(p, r).
Hence B(p, r) = {p}.
Let G ⊆ X.
If G = ∅ then G is open.
So, suppose G 6= ∅.
72 CHAPTER 1. UNIT I: METRIC SPACES

S
∴ G = Sp∈G {p}
∴ G = p∈G Bd (p, 1)
Since open balls are open sets,
we have expressed G as a union of open sets.
G is open in (X, d)
(as arbitrary union of open sets is open)

(6) Let p ∈ X and r > 0. Let A = {x ∈ X : d(x, p) > r} is an open set.

If A = ∅ then A is open.
Let x ∈ A be any element.
Therefore d(x, p) > r. Hence d(x, p) − r > 0. Take δ = d(x, p) − r. So, δ > 0.
Claim: B(x, δ) ⊆ A.

Let y ∈ B(x, δ).

So, d(y, x) < δ = d(x, p) − r.
Thus, r < d(x, p) − d(y, x) (∗)
By Triangle Inequality,
d(x, p) ≤ d(x, y) + d(y, p)
Hence from (∗)r < d(x, y) + d(y, p) − d(y, x)
= d(x, y) + d(y, p) − d(x, y) (as d(y, x) = d(x, y)
Thus r < d(y, p)
Hence y ∈ A.
=⇒ B(x, δ) ⊆ A

Hence A is open.

(7) U = {(x, y) ∈ R2 : xy 6= 0}. Then U is open in (R2 , d) where d is the Euclidean

distance.

(8) (a, b) × (c, d) is an open subset of (R2 , d) where d is the Euclidean distance is an
open set.
Let (p1 , p2 ) ∈ (a, b) × (c, d). So, a < p1 < b and c < p2 < d.
Choose r = min{b − p1 , p1 − a, d − p2 , p2 − c}.
Therefore r > 0 and (p1 − r, p1 + r) ⊆ (a, b), (p2 − r, p2 + r) ⊆ (c, d).
Claim: Bd (p1 , p2 ), r ⊆ (a, b) × (c, d).

Let (x1 , y1 ) ∈ Bd (p1 , p2 ), r . This implies d (x1 , y1 ), (p1 , p2 ) < r.
p
(x1 − p1 )2 + (x2 − p2 )2 < r. Hence |x1 − p1 | < r and |x2 − p2 | < r.
x1 ∈ (p1 − r, p1 + r) ⊆ (a, b) and x2 ∈ (p2 − r, p2 + r) ⊆ (c, d).
Thus (x1 ,x2 ) ∈ (a, b)× (c, d).
Hnece Bd (p1 , p2 ), r ⊆ (a, b) × (c, d).

(9) U = {(x, y) ∈ R2 : 2x + 3y < 1} is an open subset of R2 with Euclidean metric.

Let p = (p1 , p2 ) ∈ U .
So, 2p1 + 3p2 < 1.
=⇒ 1 − 2p2 − 3p2 > 0
1 − 2p1 − 3p2
Choose r = , so r > 0.
5
Claim: B(p, r) ⊆ U .
Let (x1 , x2 ) ∈ B(p, r).
=⇒ d (x1 , x2 ), (p1 , p2 < R.
p
=⇒ (x1 − p1 )2 + (x2 − p2 )2 < r.
=⇒ |x1 − p1 | < r, |x2 − p2 | < r.
=⇒ −r < x1 − p1 < r, −r < x2 − p2 < r.
=⇒ p1 − r < x1 < r + p1 , p2 − r < x2 < r + p2 1.
=⇒ p1 − r < x1 < r + p1 , p2 − r < x2 < r + p2 1.
=⇒ 2p1 − 2r < 2x1 < 2r + 2p1 , 3p2 − 3r < 3x2 < 3r + 3p2 1.
=⇒ 2p1 − 2r + 3p2 − 3r < 2x1 + 3x2 < 2r + 2p1 + 3p2 + 3r.
=⇒ 2p1 + 3p2 − 5r < 2x1 + 3x2 < 2p1 + 3p2 + 5r.
=⇒ 2x1 + 3x2 < 2p1 + 3p2 + 1 − 2p1 − 3p2 .
=⇒ 2x1 + 3x2 < 1.
Hence (x1 , x2 ) ∈ U .
Thus B(p, r) ⊆ U .
Hence U is open.
(10) U = {(x , y) ∈ R2 : −1 < x + y < 1 } is an open subset of (R2 , d) where d is the
Euclidean metric.
(1, 0) ∈ U. So U 6= ∅.
Let p = (p1 , p2 ) ∈ U .
So, −1 < p1 + p2 < 1.
=⇒ 0 < p1 + p2 + 1, 0 < 1 − p1 − p2 .
p1 + p 2 + 1 1 − p1 − p2
Choose r = min , . Thus r > 0.
2 2
p1 + p2 + 1 1 − p 1 − p2
This implies r ≤ , and r ≤ .
2 2
That means 2r ≤ p1 + p2 + 1 and 2r ≤ 1 − p1 − p2 .
So, −1 ≤ p1 + p2 − 2r and p1 + p2 + 2r ≤ 1. (∗)
Claim: B(p, r) ⊆ U .
Let (x1 , x2 ) ∈ B(p, r).
=⇒ d (x1 , x2 ), (p1 , p2 < R.
p
=⇒ (x1 − p1 )2 + (x2 − p2 )2 < r.
=⇒ |x1 − p1 | < r, |x2 − p2 | < r.
=⇒ −r < x1 − p1 < r, −r < x2 − p2 < r.
=⇒ p1 − r < x1 < r + p1 , p2 − r < x2 < r + p2 .
=⇒ p1 − r < x1 < r + p1 , p2 − r < x2 < r + p2 1.
=⇒ p1 + p2 − 2r < x1 + x2 < p1 + p2 + 2r.
From (∗) we have, −1 ≤ p1 + p2 − 2r < x1 + x2 < p1 + p2 + 2r ≤ 1.
=⇒ −1 < x1 + x2 < 1.
74 CHAPTER 1. UNIT I: METRIC SPACES

Hence (x1 , x2 ) ∈ U .
Thus B(p, r) ⊆ U .
Hence U is open.

(11) Let (X, d1 ) and (Y, d2 ) be metric spaces. Consider the metric d : (X × Y ) × (X ×
Y ) −→ R defined by d((x1 , y1 ), (x2 , y2 )) = max {d1 (x1 , x2 ), d2 (y1 , y2 )}. Let p ∈
X, q ∈ Y and r, s > 0. Show that B(p, r) × B(q, s) is an open set in (X × Y, d).
(p, q) ∈ Bd1 (p, r) × Bd2 (q, s) =⇒ Bd1 (p, r) × Bd2 (q, s) 6= ∅.
Let (x, y) ∈ Bd1 (p, r) × Bd2 (q, s).
So, x ∈ B(p, r) and y ∈ B(q, s).
Since Bd1 (p, r) and Bd2 (q, s) are open subsets of X, Y respectively, there exist δ1 , δ2 >
0 such that Bd1 (x, δ1 ) ⊆ Bd1 (p, r) and Bd2 (y, δ2 ) ⊆ Bd2 (q, s).
Choose δ = min{δ1 , δ2 }.
Bd1 (x, δ) ⊆ Bd1 (x, δ1 ) ⊆ Bd1 (p, r) and Bd2 (y, δ) ⊆ Bd2 (y, δ2 ) ⊆ Bd2 (q, s). (∗)
Claim: Bd (x, y), δ ⊆ Bd1 (p, r) × Bd2 (q, s).

Let (x1 , y1 ) ∈ Bd (x, y), δ .

d (x1 , y1 ), (x, y) < δ.
max{d1 (x1 , x), d2 (y1 , y)} < δ.
This implies d1 (x1 , x) < δ, d2 (y1 , y) < δ.
∴ x1 ∈ Bd1 (x, δ), y1 ∈ Bd2 (y, δ).
=⇒ x1 ∈ Bd1 (p, r) and y1 ∈ Bd2 (q, s). (from (∗))
Therefore(x1 , y1 ) ∈ B
d1 (p, r) × B d2 (q, s).
Hence Bd (x, y), δ ⊆ Bd1 (p, r) × Bd2 (q, s).

(12) Let (X, d1 ), (Y, d2 ) be metric spaces. Consider the product space
X × Y with the
product metric d : (X×Y )×(X×Y ) −→ R, d (x1 , y1 ), (x2 , y2 ) = max{d1 (x1 , y1 ), (x2 , y2 )}.
Let U, V be open subsets of X, Y respectively. Show that U × V is an open subset
of X × Y.
If U × V = ∅ then U × V is open.
Suppose U × V 6= ∅.
Let (p, q) ∈ U × V.
So, p ∈ U and q ∈ V .
Since U and V are open subsets of X, Y respectively, there exist δ1 , δ2 > 0 such that
Bd1 (p, δ1 ) ⊆ U and Bd2 (q, δ2 ) ⊆ V.
Choose δ = min{δ1 , δ2 }.
Bd1 (p, δ) ⊆Bd1 (p, δ1 ) ⊆ U and Bd2 (q, δ) ⊆ Bd2 (q, δ2 ) ⊆ V. (∗)
Claim: Bd (p, q), δ ⊆ U × V .

Let (x, y) ∈ Bd (p, q), δ .

d (x, y), (p, q) < δ.
max{d1 (x, p), d2 (y, q)} < δ.
This implies d1 (x, p) < δ, d2 (y, q) < δ.
75

∴ x ∈ Bd1 (p, δ), y ∈ Bd2 (q, δ).

=⇒ x ∈ U and y ∈ V. (from (∗))
Therefore(x, y) ∈ U × V.
Hence Bd (p, q), δ ⊆ U × V .

(13) Let (X, k k) be a normed linear space and A 6= ∅, A ⊆ X. Show that if U 6= ∅, U ⊆ X

is an open set then U + A is open.
First we will prove that p + U is open ∀p ∈ A.
Let p ∈ A.
Consider p + U .
Let x ∈ p + U . So, x = p + u for some u ∈ U.
Since U is open, ∃ r > 0 such that B(u, r) ⊆ U.
Claim B(x, r) ⊆ p + U .
Let y ∈ B(x, r).
So, y ∈ B(p + u, r).
But, B(p + u, r) = p + u + rB(0, 1) = p + (u + rB(0, 1)) = p + B(u, r) ⊆ p + U =⇒
y ∈ p + U =⇒ B(x, r) ⊆ p + U .
∴ p + U is open.
S
Now A + U = p∈A (p + U ).
p + U is open for every p ∈ A. S
Since arbitrary union of open sets is open, p∈A (p + U ) is open =⇒ A + U is open.
76 CHAPTER 1. UNIT I: METRIC SPACES

(13) B1 (0, 1) is open in (C [0 , 1 ], k k∞ ). (14) B∞ (0, 1)is not open in (C [0 , 1 ], k k1 ).

First we will show that We show that for every r > 0, there
kf k1 ≤ kf k∞ for all f ∈ C[0, 1]. exists a function f ∈ B1 (0, r),
Z 1 such that f ∈ / B∞ (0, 1) and hence we
kf k1 = |f (t)|dt will have for every r > 0, B1 (0, r) 6⊆
0 B∞ (0, 1). Let r > 0.
Z 1
By Archimedean Property, there ex-
≤ sup{|f (t)| : t ∈ [0, 1]}dt ists
0
Z 1 n0 ∈ N such that n10 < r.
= sup{|f (t)| : t ∈ [0, 1]} dt Define f(: [0, 1] −→ R, as follows:
0
Z 1 1 − n0 x if 0 ≤ x ≤ n10
f (x) = .
= kf k∞ dt 0 if n10 < x ≤ 1
0
0 ≤ x ≤ n10 =⇒ 0 ≤ n0 x ≤ 1.
= kf k∞ ∗ 1
So, −1 ≤ −n0 x ≤ 0.
= kf k∞ Adding 1 on both sides we get,
0 ≤ 1 − n0 x ≤ 1.
Hence kf k1 ≤ kf k∞ for all f ∈
Thus, for 0 ≤ x ≤ n10 , 0 ≤ f (x) ≤ 1
C [0 , 1 ].
Now we will prove that and for n10 ≤ x ≤ 1, f (x) = 0.
B1 (0, 1) is open in (C [0 , 1 ], k k∞ ). So, f (x) ≤ 1 for all x ∈ [0, 1].
Let f ∈ B1 (0, 1). Also, f (0) = 1.
This implies kf − 0k1 < 1, Hence sup{f (x) : x ∈ [0, 1]} = 1.
that is kf k1 < 1. =⇒ kf k∞ = 1.
So 0 < 1 − kf k1 . That is, kf − 0k∞ = 1.
Take δ = 1 − kf k1 . So, f ∈/ B∞ (0, 1).
Claim: B∞ (f, δ) ⊆ B1 (0, 1). But We will show that f ∈ B1 (0, r).
Let g ∈ B∞ (f, δ). Z 1
kg − f k∞ < δ. kf − 0k1 = |f (t)|dt.
Since kg − f k1 ≤ kg − f k∞ < δ, 0
1
Z Z 1
kg − f k1 < δ. n0
= |f (t)|dt + |f (t)|dt
By Triangle Inequality, 0 1
n0
kgk1 ≤ kg − f k1 + kf k1 ≤ kg − f k∞ + Z 1 Z 1
n0

kf k1 < δ + kf k1 = 1. = 1 − n0 t dt + 0 dt
Hence kgk1 < 1. 0 1
n0

So, g ∈ B1 (0, 1). h n0 t2 i n10 h i1

Hence B∞ (f, δ) ⊆ B1 (0, 1). = x− + 0 1.
2 0 n0
Hence B1 (0, 1) is open in
1 n0
(C[0, 1], k k∞ ). = − 2 − 0 + 0.
n0 2n0
1 1
= < < r.
2n0 n0
Hence f ∈ B1 (0, r) but f ∈/ B∞ (0, 1).
So there does not exist r > 0 such that
B1 (0, r)
77

1.8 Open sets in Subspaces

Theorem 1.8.1 Let (X, d) be a metric space and Y be a non-empty subset of X. Prove
that a subset G of Y is open in the subspace (Y, d) if and only if G = V ∩ Y where V is an
open set in (X, d).

Proof: First we show thatTif BY (p, r) is an open ball with center p ∈ Y and r > 0 in (Y, d)
then BY (p, r) = BX (p, r) Y where BX (p, r) is an open ball in (X, d) with center p and
radius r.

BY (p, r) = {x ∈ Y : d(x, p) < r}

\
= {x ∈ X : d(x, p) < r} Y
\
= BX (p, r) Y.

[2 marks]
Let G be an open subset of Y .
∴ ∀y
[ ∈ G ∃ ry > 0 such that BY (y, ry ) ⊆ G
∴ BY (y, ry ) ⊆ G (∗)
y∈G

[ {y} ⊆ B[
Now Y (y, ry )
∴ {y} ⊆ BY (y, ry )
y∈G y∈G
[
∴G⊆ BY (y, ry ) (∗∗)
y∈G
∴ from
[(∗) and (∗∗) we get,
G= BY (y, ry ) [2 marks]
y∈G
[ \
∴G= BX (y, ry ) Y (as shown above)
y∈G
[ \
∴G= BX (y, ry ) Y
y∈G
[
Let V = BX (y, ry ) [1 mark]
y∈G
Now open balls are open sets.
∴ BX (y, ry ) is an open set in (X, d) for every y ∈ G
Since
[ arbitrary union of open sets is open,
∴ BX (y, ry ) is open in X.
y∈G T
∴ V is open in X and G = VT Y [1 mark]
Conversely, suppose G = V Y where V is an open subset of X
TST G is open in Y .
If G is empty then it is open.
78 CHAPTER 1. UNIT I: METRIC SPACES

T
Let y ∈ G ∴y ∈V Y.
∴ y ∈ V and V is open in X.
∴ ∃ r > 0 such
T that BXT(y, r) ⊆ V
∴ BX (y, r) Y ⊆V Y
∴ BY (y, r) ⊆ G
∴ G is open in Y .
Examples 1.8.2
√ √
(1) Show that A = {x ∈ Q : − 2 < x < 2} is open in the subspace Q of R with
usual distance.
(2) Show that S = {x ∈ Q : 3 < x2 < 5} is open in the subspace Q of R with usual
metric.
(3) Let Y = {(x, y) ∈ R2 : x ≥ 0, y ≥ 0} in R2 w.r.t. the Euclidean distance. Let
A = {(x, y) ∈ Y : 0 ≤ x < 1, 0 ≤ y < 1}. Is A open in Y ?
(4) Let X = R, Y = Z. Which subsets of Y are open in Y ?.
(5) Let (X, d) be a metric space and Y be an open subset of X. Then prove that A ⊆ Y
is open in Y if and only if A is open in X.
(6) Let A = [0, 1) ⊆ R have the induced usual metric from R. Find BA (0, r) for any
r > 0? (BA (x, r) is the open ball in A centred at x and radius r with respect to the
induced metric.)
(7) Let A =

Solution
√ √ √ √
(1) A = {x ∈ Q : − 2 < x < 2} = (− 2, 2) ∩ Q. √ √
Since open intervals are open sets in (R, usual), (− 2, 2)) is an open set in
(R, usual)
√ √
∴ (− 2, 2)) ∩ Q is an open set in the subspace Q of (R, usual)
(G ⊆ Y is open in the subspace (Y, d) of (X, d) if and only if G = V ∩ Y for some
open subset V of X )
Hence A is open in the subspace Q.
√ √
= {x ∈ Q : √
(2) S √ 3 < x2 < √5} =⇒√S = S = {x√∈ Q √ : 3 < x < 5} ∪ S = {x ∈ Q :
− 5 <√ x < −√ 3} = (− 5, √− √3) ∩ Q ∪ ( √ 3, 5)√∩ Q. √ √
S = (− 5, − 3) ∩ Q ∪ [ 3, 5] ∩ Q. = (− 5, − 3) ∪ ( 3, 5) ∩ Q. [1M]
Since √
open√intervals
√ are
√ open sets in (R, usual) and union of open sets is open,
∴ (−√5, √3)) ∪ (√3, √5)) is an open set in (R, usual).
∴ (− 5, 3)) ∪ ( 3, 5)) ∩ Q is an open set in the subspace Q of (R, usual).

(3) We can write A as A = (0, 1) × (0, 1) ∩ Y.
Since (0, 1) × (0, 1) is open in R2 w.r.t. the Euclidean space, A is open in Y w.r.t.
the induced Euclidean space.
79

(4) Let X = R, Y = Z. Which subsets of Y are open in Y ?.

Let A ⊆ Z such that A is open in Z.
So, A is written as A = V ∩ Z where V is an open subset of R w.r.t the usual
distance.
This means V is a union of finite or countable family of pairwise disjoint open
intervals.
So, A is finite or countable.

(5) Let (X, d) be a metric space and Y be an open subset of X. Then prove that A ⊆ Y
is open in Y if and only if A is open in X.
Suppose A is open in X. Since A ⊆ Y, we can write A = A ∩ Y.
Since A is open in X, A ∩ Y is open in Y. Thus A is open in Y.
Conversely suppose A is open in Y .
We want to show that A is open in X.
There exists U ⊆ X such that U is open in X and A = U ∩ Y.
Now, U, Y are open in X. Hence A is open in X. (finite intersection of open sets is
open)

(6) Let A = [0, 1) ⊆ R have the induced usual metric from R. Find BA (0, r) for any
r > 0? (BA (x, r) is the open ball in A centred at x and radius r with respect to the
induced metric.)

BA (x, r) = B(x, r) ∩ [0, 1)

= (0 − r, 0 + r) ∩ [0, 1)
= (−r, r) ∩ [0, 1)
= [0, r)

2
(7) Let A= {(x, y) : x ≥ 0, y ≥ 0} have the induced Euclidean metric from R . Draw
B(A,d) (0, 0), 1 .
80 CHAPTER 1. UNIT I: METRIC SPACES

x ≥ 0, y ≥ 0

B(A,d) (0, 0), 1

(0, 0)

1.9 Structure of an open set in R

Theorem 1.9.1 let (X, d) be a metric space. A subset G of X is open if and only if G is a
union of open balls.

Proof:
Let G ⊆ X be such that G is open.
If G = ∅ then G is empty union of open balls.
If G 6= ∅, then for each x ∈ G, ∃ rx > 0 such that B(x, rx ) ⊆ G.
Since x ∈ B(x, r), {x} ⊆ B(x, rx ).
=⇒ {x}
[ ⊆ B(x,[ rx ) ⊆ G.
=⇒ {x} ⊆ B(x, rx ) ⊆ G.
x∈G [ x∈G
=⇒ G ⊆ B(x, rx ) ⊆ G.
x∈G
[
=⇒ G ⊆ B(x, rx ).
x∈G
Conversely, suppose
[ G is a union of open balls.
This means G = B(x, r)
Since each open ball is an open set, hence G is an arbitrary union of open sets.
We will show that arbitrary union of open sets is open.
Since arbitrary union of open sets is open, ∴ G is open.
81

Theorem 1.9.2 A nonempty open set in (R, d) (where d is the usual distance) is an union
of finite or countable family of pairwise disjoint open intervals.

Proof:
Let U 6= ∅, U ⊆ R be open.
Let x ∈ U .
∃ rx > 0 such that B(x, rx ) ⊆ U .
Since B(x, rx ) = (x − rx , x + rx ), so B(x, rx ) is an interval.
Let Jx denote
[ the union of all open intervals that contain x and are contained in U .
i.e. Jx = Iα
α∈Ω
where Iα is an open interval such that x ∈ Iα and Iα ⊆ U ∀α ∈ Ω.
Since open intervals are open sets and arbitrary union of open sets is open, Jx is an open
set.
We will show that Jx is an open interva.
Consider a, b, c ∈ R be such that a < c < b, and a, b ∈ Jx .
We want to show that c ∈ Jx .
Since a, b ∈ Jx , a ∈ Iα and b ∈ Iβ for some α, β ∈ Ω.
Case 1: x ≤ a.
∴ x ≤ a < c < b.
x, b ∈ Iβ and Iβ is an interval,
∴ [x, b] ⊆ Iβ =⇒ c ∈ Iβ ⊆ Jx .
∴ c ∈ Jx .
Case 2: a < x.
So, there are subcases such as a < b < x or a < x < b.

Case 2a If a < b < x then Case 2a If a < x < b then we have follow-
a, x ∈ Iα =⇒ [a, x] ⊆ Iα ⊆ Jx . ing cases.
∴ a < c < b < x =⇒ c ∈ Ix ⊆ Jx . If x ≤ c then a < x ≤ c < b =⇒ c ∈
Iβ ⊆ Jx .
If c < x then a < c < x =⇒ c ∈ Iα ⊆
Jx .

Thus in any case a < c < b, a, b ∈ Jx =⇒ c ∈ Jx . Hence Jx is an interval.

Also Jx is open as arbitrary union of open sets is open.
Thus Jx is open and is an interval, that is Jx is an open interval.
Now we wil show that Jx , Jy are either disjoint or equal ∀x, y ∈ U.
Let x, y ∈ U.
S Jx ∩ Jy 6= ∅.
Suppose
∴ Jx Jy is an open interval containing both x, y. S
Since JSx is the union of all open intervals containing x and contained in U, Jx Jy ⊆ Jx .
=⇒ Jx Jy = Jx . (∗)
Similarly,
S Jy is the union of all open intervals containing y and contained in U.
∴ Jx Jy ⊆ Jy .
82 CHAPTER 1. UNIT I: METRIC SPACES

S
∴ Jx Jy = Jy (∗∗)
∴ from (∗) and (∗∗), we get, Jx = Jy
Since U = ∪x∈U {x} ⊆ ∪x∈U {Jx } ⊆ U
∴ U = ∪x∈U {Jx }
Also, since Jx , Jy are either equal or disjoint, we can write U as a union of a family
{Jxi : i ∈ I} of pairwise disjoint open intervals.
Thus ∴ U = ∪i∈I Jxi .
Choose one rational number say ri ∈ Jxi for each i ∈ I.
Define f : I −→ Q such that f (i) = ri .
So f is injective.
Since Q is countable, ∴ ∃ a bijective function g : Q −→ N.
So the composite function g ◦ f : I −→ N is an injective function.
Hence I is countable.
Thus U can be expressed as a countable union of open intervals.

1.9.1 Translation invariance of the metric induced by norm

Theorem 1.9.3 Let (X, k k) be a normed linear space. Let x ∈ X and r > 0. Then
B(x, r) = x + rB(0, 1)
We want to show that B(x, r) = x + rB(0, 1), x ∈ X, r > 0
Let p ∈ B(x, r).
∴ d(x, p) < r
∴ kx − pk < r
1
∴ kx − pk < 1
r
1
∴ k (x − p)k < 1
r
1 1
∴ k (x) − (p)k < 1
r r
1 1
∴k (p) − (x) − 0k < 1
r r
1 1
∴ (p) − (x) ∈ B(0, 1)
r r
1 1
∴ (p) − (x) = y for some y ∈ B(0, 1)
r r
1 1
∴ (p) − (x)) = y for some y ∈ B(0, 1)
r r
∴ p − x = ry where y ∈ B(0, 1)
∴ p = x + ry for some y ∈ B(0, 1)
∴ p ∈ x + rB(0, 1)
Thus B(x, r) ⊂ x + rB(0, 1) (∗)
Now suppose p ∈ x + rB(0, 1)
∴ p = x + ry for some y ∈ B(0, 1)
∴ p − x = ry where y ∈ B(0, 1)
∴ kp − xk = kryk = |r|kyk = rkyk < r ∗ 1 since y ∈ B(0, 1)
∴ kp − xk < r =⇒ p ∈ B(x, r)
83

∴ x + rB(0, 1) ⊆ B(x, r) (∗∗)

From (∗) and (∗∗) we get, x + rB(0, 1) = B(x, r).
Theorem 1.9.4 Let (X, k k) be a normed linear space and A 6= ∅, A ⊆ X. If U 6= ∅, U ⊆ X
is an open set then U + A is open.
Proof: First we will prove that p + U is open ∀p ∈ A.
Let p ∈ A.
Consider p + U .
Let x ∈ p + U . So, x = p + u for some u ∈ U.
Since U is open, ∃ r > 0 such that B(u, r) ⊆ U.
Claim B(x, r) ⊆ p + U .
Let y ∈ B(x, r).
So, y ∈ B(p + u, r)
But, B(p + u, r) = p + u + rB(0, 1) = p + (u + rB(0, 1)) = p + B(u, r) ⊆ p + U =⇒ y ∈
p + U =⇒ B(x, r) ⊆ p + U . S
∴ p + U is open. Now A + U = p∈A (p + U )
p + U is open for every p ∈ A S
Since arbitrary union of open sets is open, p∈A (p + U ) is open =⇒ A + U is open.

1.10 Interior of a set

Definition 1.10.1 Let (X, d) be a metric space and A ⊆ X. We say that x ∈ A is an
interior point of A if there exists r > 0 such that B(x, r) ⊆ A. The set of all interior points
of A is denoted by A◦ .

Remark 1.10.2 A◦ = {x ∈ A : ∃ r > 0 such that B(x, r) ⊆ A}

Theorem 1.10.3 Let (X, d) be a metric space and A, B ⊆ X. Then following results hold.
(i) A ⊆ B =⇒ A◦ ⊆ B ◦ .

(ii) (A ∩ B)◦ = A◦ ∩ B ◦ .

(iii) A◦ ∪ B ◦ ⊆ (A ∪ B)◦ .

Proof:
(i) A ⊆ B =⇒ A◦ ⊆ B ◦
Let x ∈ A◦ =⇒ there is an open ball B(x, r) ⊆ A
B(x, r) ⊆ A =⇒ x ∈ B (as A ⊆ B) and B(x, r) ⊆ A ⊆ B.
A ⊆ B =⇒ A◦ ⊆ B ◦ .

(ii) (A ∩ B)◦ = A◦ ∩ B ◦
(A ∩ B) ⊆ A =⇒ (A ∩ B)◦ ⊆ A◦
(A ∩ B) ⊆ B =⇒ (A ∩ B)◦ ⊆ B ◦
Hence (A ∩ B)◦ ⊆ A◦ ∩ B ◦ (∗)
Conversely, x ∈ A ∩ B =⇒ x ∈ A and x ∈ B ◦ .
◦ ◦ ◦
84 CHAPTER 1. UNIT I: METRIC SPACES

=⇒ x ∈ A, ∃ an open ball B(x, r1 ) ⊆ A and x ∈ B,

∃ an open ball B(x, r2 ) ⊆ B. x ∈ A◦ ∩ B ◦ =⇒ x ∈ A ∩ B and if r = min{r1 , r2 }
then B(x, r) ⊆ B(x, r1 ) ⊆ A B(x, r) ⊆ B(x, r2 ) ⊆ B.
=⇒ x ∈ A ∩ B and ∃ an open ball B(x, r) ⊆ A ∩ B.
=⇒ x ∈ (A ∩ B)◦ .
Hence A◦ ∩ B ◦ ⊆ (A ∩ B)◦ (∗∗).
From (∗) and (∗∗) we have, A ∩ B ⊆ (A ∩ B)◦
◦ ◦

(iii) A◦ ∪ B ◦ ⊆ (A ∪ B)◦ and the inequality may be strict.

A ⊆ A ∪ B =⇒ A◦ ⊆ (A ∪ B)◦
B ⊆ A ∪ B =⇒ B ◦ ⊆ (A ∪ B)◦
Thus A◦ ∪ B ◦ ⊆ (A ∪ B)◦ .
If we consider the metric space (R, d) where d is the usual distance and A = Q, B =
R \ Q then A◦ = ∅, B ◦ = ∅
Thus A◦ ∪ B ◦ = ∅ but A ∪ B = R =⇒ (A ∪ B)◦ = R.
(If x ∈ Q, B(x, r) = (x − r, x + r) * Q for (x − r, x + r) has irrational numbers.
∴ x ∈ Q◦ and Q◦ = ∅. Similarly (R \ Q)◦ = ∅.

Remark 1.10.4 If we consider the metric space (R, d) where d is the usual distance and
A = Q, B = R \ Q then A◦ = ∅, B ◦ = ∅
Thus A◦ ∪ B ◦ = ∅ but A ∪ B = R =⇒ (A ∪ B)◦ = R.
(If x ∈ Q, B(x, r) = (x − r, x + r) * Q for (x − r, x + r) has irrational numbers. ∴ x ∈ Q◦
and Q◦ = ∅. Similarly (R \ Q)◦ = ∅.

Theorem 1.10.5 Let (X, d) be a metric space and A ⊆ X. Show that

(i) A◦ is an open set and is the largest open set contained in A .

(ii) A is open if and only if A = A◦ .

Proof:
Let (X, d) be a metric space and A ⊆ X.

(i) Since A◦ = {x ∈ A : ∃ r > 0 such that B(x, r) ⊆ A}, A◦ ⊆ A (∗)

◦ ◦
If A = ∅ then A is open. (since empty set is an open set).
Suppose A◦ 6= ∅.
Let x ∈ A◦ .
∴ ∃ r > 0 such that B(x, r) ⊆ A.
Claim: B(x, r) ⊆ A◦ .
Let y ∈ B(x, r) ∴y∈A
Since open balls are open sets, ∃ δ > 0 such that B(y, δ) ⊆ B(x, r) (take δ =
r − d(x, y))
As B(x, r) ⊆ A, B(y, δ) ⊆ A and hence y ∈ A◦ .
Thus B(x, r) ⊆ A◦ .
∴ ∀ x ∈ A◦ , ∃ r > 0 such that B(x, r) ⊆ A◦ .
85

Hence A◦ is an open set. (∗∗)

Now, we will show that A◦ is the largest open set containded in A.
By (∗) and (∗∗), A◦ is an open set contained in A.
Suppose G is an open set contained in A.
If G = ∅ then G ⊆ A◦ .
Suppose G 6= ∅.
Let x ∈ G.
Since G is open, ∃ r > 0 such that B(x, r) ⊆ G.
Now G ⊆ A.
So, B(x, r) ⊆ A and hence x ∈ A◦ .
Thus G ⊆ A◦ .
Hence A◦ is the largest open set contained in A.
(II) We show that A is open if and only if A = A◦ .
Suppose A is open.
This means A is open and A ⊆ A.
But A◦ is the largest open set contained in A.
Therefore, A ⊆ A◦ .
Also from the definition of A◦ , A◦ ⊆ A.
Hence A◦ = A.
Conversely if A◦ = A then since A◦ is open, so A is open.
[
Theorem 1.10.6 Let (X, d) be a metric space and A ⊆ X. Then A◦ = G.
G open,G⊆A

Proof: Let x ∈ A◦ . So, x ∈ A and there exists r > 0 such that B(x, r) ⊆ A.
Let G = B(x, r).
Then G ⊆ A, G is open[ and x ∈ G. (as x ∈ B(x, r) and B(x, r) ⊆ G)
This implies x ∈ G.
G open,G⊆A
[
◦
Therefore, A ⊆ G.
G open,G⊆A
[
Now, let x ∈ G.
G open,G⊆A
So, x ∈ G for some G, G ⊆ A and G is open.
So, x ∈ G for some G ⊆ X such that

1.11 Equivalent metric spaces

Definition 1.11.1 Two metrics d1 and d2 on a nonempty set X are said to be equivalent
if the open subsets of X with respect to d1 and d2 are same. In that case, the metric spaces
(X, d1 ) and (X, d2 ) are said to be equivalent metric spaces.

Examples 1.11.2

(1) (N, d) and (N, d1 ) are equivalent metric spaces where d is the usual distance and d1
is the discrete metric.
86 CHAPTER 1. UNIT I: METRIC SPACES

(2) Let X be a finite set and d1 and d2 be any two metrics on X. Then d1 and d2 are
equivalent.
1
(3) Let d and d1 be metrics on N defined by d(m, n) = |m−n| and d1 (m.n) = 1+
m+n
for all m, n ∈ N. Show that d and d1 are equivalent metrics on N.

Solution:

(1) (N, d) and (N, d1 ) are equivalent metric spaces where d is the usual distance and d1
is the discrete metric.
First we will show that {x} is open in (N, d) as well as in (N, d1 ).

Bd (x, 1)in (N, d) = {y ∈ N : d(y, x) < 1}. Bd1 (x, 1)in (N, d1 ) = {y ∈ N : d1 (y, x) < 1}.
= {y ∈ N : |y − x| < 1}. = {y ∈ N : d1 (y, x) = 0} (as d1 (x
= {y ∈ N : y ∈ (x − 1, x + 1)}. = {y ∈ N : y = x}.
= {x} (as x, y ∈ N). = {x}.

Since open balls are open sets, Since open balls are open sets,
{x} is open in (N, d). {x} is open in (N, d1 ).
[
If G ⊆ X, G 6= ∅, then we can write G = {x}.
x∈G
Since arbitrary union of open sets is open, G is open in both (N, d) and (N, d1 ).
Thus every subset of X is open in both (N, d) and (N, d1 ).
Hence a subset of X is open in (N, d) if and only if it is open in (N, d1 ).
Therefore, d and d1 are equivalent metrics on N.

(2) Let X be a finite set and d1 and d2 be any two metrics on X. Then d1 and d2 are
equivalent.

If X = {x} then the open sets in (X, d1 ) and (X, d2 ) are ∅, X.

Hence the two spaces have same open subsets. So, they are equivalent.
If |X| > 1, then we show {x} is open in (X, d1 ) as well as (X, d2 ) for each x ∈ X.
Let r1 = min{d1 (x, y) : y ∈ X, y 6= x}.
r2 = min{d2 (x, y) : y ∈ X, y 6= x}.
Then r1 , r2 > 0.
Bd1 (x, r1 ) = {x} in (X, d1 ) and Bd2 (x, r1 ) = {x} in (X, d2 ).
Since open balls are open [sets, {x} is open in (X, d1 ) and (X, d2 ).
Let G ⊆ X. Then G = {x}.
x∈G
This means G is a union of open sets in (X, d1 ) as well as (X, d2 ).
Therefore G is open in (X, d1 ) and (X, d2 ).
87

Hence every subset of X is open in (X, d1 ) and (X, d2 ).

This means the collection of open sets in (X, d1 ) and (X, d2 ) is same.
Hence d1 and d2 are equivalent metrics on X.

Let d and d1 be metrics on N defined by d(m, n) = |m − n| and d1 (m.n) =

(3) 
1 + 1 if m 6= n
m+n . Show that d and d1 are equivalent metrics on N.
0 if m = n
First we will show that {x} is open in (N, d) as well as in (N, d1 ).

Theorem 1.11.3 Two metrics d1 and d2 on X are equivalent if and only for each x ∈ X,
for every r > 0, the open ball Bd2 (x, r) contains an open ball Bd1 (x, r1 ) for some r1 > 0 and
the open ball Bd1 (x, r) contains an open ball Bd2 (x, r2 ) for some r2 > 0. (for every r > 0,
there exist r1 , r2 > 0 such that Bd1 (x, r1 ) ⊆ Bd2 (x, r) and Bd2 (x, r2 ) ⊆ Bd1 (x, r) for each
x ∈ X.)

Proof: Let x ∈ X and r > 0.

Consider Bd2 (x, r). Consider Bd1 (x, r).

Since open balls are open sets, Since open balls are open sets,
Bd2 (x, r) is an open set in (X, d2 ). Bd1 (x, r) is an open set in (X, d1 ).
As d1 and d2 are equivalent metrics on X, As d1 and d2 are equivalent metrics on X,
Bd2 (x, r) is open set in (X, d1 ). Bd1 (x, r) is an open set in (X, d2 ).
Now, x ∈ Bd2 (x, r) and Bd2 (x, r) is Now, x ∈ Bd1 (x, r) and Bd1 (x, r) is
open in (X, d1 ). open in (X, d2 ).
So, there exists r1 > 0 such that So, there exists r2 > 0 such that
Bd1 (x, r1 ) ⊆ Bd2 (x, r). Bd2 (x, r2 ) ⊆ Bd1 (x, r).
88 CHAPTER 1. UNIT I: METRIC SPACES

Examples 1.11.4

R2 induced by k k1 , k k2 and k k∞ respec-

(1) Let d1 , d2 , d∞ be three metrics defined on p
tively, where kxk1 = |x1 | + |x2 |, kxk2 = x21 + x22 , kxk∞ = max{|x1 |, |x2 |} ∀x =
(x1 , x2 ) ∈ R2 . Let r > 0 Prove the following:

(i) d1 and d2 are equivalent metrics on R2 .

(ii) d1 and d∞ are equivalent metrics on R2 .
(iii) d2 and d∞ are equivalent metrics on R2 .

(2) Let (X, d) be a metric space. Consider the metric space d1 : X ×X −→ R, d1 (x, y) =
d(x, y)
. Show that d and d1 are equivalent metrics on X.
1 + d(x, y)
(3) Let (X, d) be a metric space. Consider the metric space d1 : X ×X −→ R, d1 (x, y) =
min{1, d(x, y)}. Show that d and d1 are equivalent metrics on X.

(4) Let d1 and d2 be metrics on a non-empty set X such that there exist k1 , k2 > 0 such
that k1 d1 (x, y) ≤ d2 (x, y) ≤ k2 d1 (x, y), ∀x, y ∈ X.

(5) Let X = C[0, 1] and d1 be metrics on N defined by d(m, n) = |m − n| and d1 (m.n) =

1
1+ for all m, n ∈ N. Show that d and d1 are equivalent metrics on N.
m+n
Solution:

Examples 1.11.5

R2 induced by k k1 , k k2 and k k∞ respec-

(1) Let d1 , d2 , d∞ be three metrics defined on p
tively, where kxk1 = |x1 | + |x2 |, kxk2 = x21 + x22 , kxk∞ = max{|x1 |, |x2 |} ∀x =
2
√ 1 , x2 ) ∈ R . Let r > 0. We know2 that d∞ (x, y) ≤ d2 (x, y) ≤ d1 (x, y) ≤
(x
2d2 (x, y) ≤ 2d∞ (x, y) for all x, y ∈ R .
Prove the following:
2
(i) To prove that d1 and d2 are equivalent
√ metrics on R .
We know that d2 (x, y) ≤ d1 (x, y) ≤ 2d2 (x, y) (∗).

We show that ∃ r1 > 0 such that We show that ∃ r2 > 0 such that
Bd1 (x, r1 ) ⊆ Bd2 (x, r) Bd2 (x, r2 ) ⊆ Bd1 (x,
√ r)
Since d2 (x, y) ≤ d1 (x, y), we choose Since d1 (x, y) ≤ 2d2 (x, y), we have
r1 = r. d1 (x, y)
√ ≤ d2 (x, y). (∗∗)
Let y ∈ Bd1 (x, r). 2
This menas d1 (y, x) < r. r
Choose r2 = √ .
From (∗), d2 (x, y) ≤ d1 (x, y) 2
Hence d1 (y, x) < r =⇒ d2 (y, x) < r. Let y ∈ Bd2 (x, r2 ).
r
Thus y ∈ Bd2 (x, r). This menas d2 (y, x) < √ .
2
Hence Bd1 (x, r1 ) ⊆ Bd2 (x, r).
89

d1 (x, y) That is d1 (y, x) < r.

From (∗∗), √ ≤ d2 (x, y).
2 Thus y ∈ Bd1 (x, r).
r d1 (y, x)
Hence d2 (y, x) < √ =⇒ √ < Hence Bd2 (x, r2 ) ⊆ Bd1 (x, r).
2 2
r
√ .
2

(ii) To prove that d1 and d∞ are equivalent metrics on R2 .

Let r > 0.
We know that d∞ (x, y) ≤ d1 (x, y) ≤ 2d∞ (x, y). (∗)

We show that ∃ r1 > 0 such that We show that ∃ r2 > 0 such that
Bd1 (x, r1 ) ⊆ Bd∞ (x, r) Bd∞ (x, r2 ) ⊆ Bd1 (x, r)
Since d∞ (x, y) ≤ d1 (x, y), we choose Since d1 (x, y) ≤ 2d∞ (x, y), we have
r1 = r. d1 (x, y)
≤ d∞ (x, y). (∗∗)
Let y ∈ Bd1 (x, r). 2
r
This menas d1 (y, x) < r. Choose r2 = .
2
From (∗), d∞ (x, y) ≤ d1 (x, y) Let y ∈ Bd∞ (x, r2 ).
Hence d1 (y, x) < r =⇒ d∞ (y, x) < r. r
This menas d∞ (y, x) < r2 = .
Thus y ∈ Bd∞ (x, r). 2
Hence Bd1 (x, r1 ) ⊆ Bd∞ (x, r). d1 (x, y)
From (∗∗), ≤ d∞ (x, y).
2
r d1 (y, x) r
Hence d∞ (y, x) < =⇒ < .
2 2 2
That is d∞ (y, x) < r.
Thus y ∈ Bd∞ (x, r).
Hence Bd∞ (x, r2 ) ⊆ Bd1 (x, r).

(iii) d2 and d∞ are equivalent metrics on R2 .

Let r > 0.
We know that d∞ (x, y) ≤ d2 (x, y) ≤ 2d∞ (x, y). (∗)

We show that ∃ r0 > 0 such that We show that ∃ r00 > 0 such that
Bd2 (x, r0 ) ⊆ Bd∞ (x, r) Bd∞ (x, r00 ) ⊆ Bd2 (x, r)
Since d∞ (x, y) ≤ d2 (x, y), we choose Since d2 (x, y) ≤ 2d∞ (x, y), we have
r0 = r. d2 (x, y)
≤ d∞ (x, y). (∗∗)
Let y ∈ Bd2 (x, r0 ). 2
r
This menas d2 (y, x) < r0 = r. Choose r00 = .
2
From (∗), d∞ (x, y) ≤ d2 (x, y) Let y ∈ Bd∞ (x, r00 ).
Hence d2 (y, x) < r =⇒ d∞ (y, x) < r. r
This menas d∞ (y, x) < r2 = .
Thus y ∈ Bd∞ (x, r). 2
Hence Bd2 (x, r0 ) ⊆ Bd∞ (x, r). d2 (x, y)
From (∗∗), ≤ d∞ (x, y).
2
90 CHAPTER 1. UNIT I: METRIC SPACES

r d2 (y, x) r
Hence d∞ (y, x) < =⇒ < . Thus y ∈ Bd∞ (x, r).
2 2 2 Hence Bd∞ (x, r2 ) ⊆ Bd2 (x, r).
That is d∞ (y, x) < r.

(2) Let (X, d) be a metric space. Consider the metric space d1 : X ×X −→ R, d1 (x, y) =
d(x, y)
. Show that d and d1 are equivalent metrics on X.
1 + d(x, y)
First we show that d1 (x, y) ≤ d(x, y) for all x, y ∈ X.

0 ≤ d(x, y) < 1 + d(x, y) ∀ x, y ∈ X.

Hence d1 (x, y) ≤ d(x, y) for all x, y ∈ X.

To prove that d1 and d are equivalent metrics on X.
Let r > 0.

We show that ∃ r1 > 0 such that We show that ∃ r2 > 0 such that
Bd1 (x, r1 ) ⊆ Bd (x, r) Bd (x, r2 ) ⊆ Bd1 (x, r)
r Since d1 (x, y) ≤ d(x, y), we choose
Choose r1 = .
1+r r2 = r.
Let y ∈ Bd1 (x, r1 ).
r Let y ∈ Bd (x, r2 ).
This menas d1 (y, x) < . This menas d(y, x) < r2 = r.
1+r
From (∗), d1 (x, y) ≤ d(x, y)
d(y, x) r Hence d(y, x) < r =⇒ d1 (y, x) < r.
≤ .
1 + d(y, x) 1+r Thus y ∈ Bd1 (x, r).
Hence d(y, x)(1 + r) < r(1 + d(y, x)). Hence Bd (x, r1 ) ⊆ Bd1 (x, r).
=⇒ d(y, x) < r.
Thus y ∈ Bd (x, r).

Hence Bd1 (x, r1 ) ⊆ Bd (x, r).

(3) Let (X, d) be a metric space. Consider the metric space d1 : X ×X −→ R, d1 (x, y) =
min{1, d(x, y)}. Show that d and d1 are equivalent metrics on X.
Since d1 (x, y) = min{1, d(x, y)}, d1 (x, y) ≤ d(x, y) for all x, y ∈ X. (∗)
To prove that d1 and d are equivalent metrics on X.
Let r > 0.

We show that ∃ r1 > 0 such that We show that ∃ r2 > 0 such that
Bd1 (x, r1 ) ⊆ Bd (x, r). Bd (x, r2 ) ⊆ Bd1 (x, r)
Choose r1 = min{1, r}. Since d1 (x, y) ≤ d(x, y), we choose
So, r1 ≤ 1, r1 ≤ r. Let y ∈ Bd1 (x, r1 ). r2 = r.
This means d1 (y, x) < r1 . Let y ∈ Bd (x, r2 ).
As r1 ≤ 1, d1 (y, x) < r1 ≤ 1. This menas d(y, x) < r2 = r.
This implies d1 (y, x) < 1. From (∗), d1 (x, y) ≤ d(x, y)
So, d1 (y, x) 6= 1. Hence d(y, x) < r =⇒ d1 (y, x) < r.
But d1 (y, x) = 1 or d(y, x). Thus y ∈ Bd1 (x, r).
Hence d1 (y, x) = d(y, x). Hence Bd (x, r1 ) ⊆ Bd1 (x, r).
Then d1 (y, x) < r1 =⇒ d(y, x) < r1 .
But r1 ≤ r. So, d(y, x) < r.
Hence y ∈ Bd (x, r).
Thus Bd1 (x, r1 ) ⊆ Bd (x, r).

(4) Let d1 and d2 be metrics on a non-empty set X such that there exist k1 , k2 > 0 such
that k1 d1 (x, y) ≤ d2 (x, y) ≤ k2 d1 (x, y), ∀x, y ∈ X.
To prove that d1 and d2 are equivalent metrics on X.
Let r > 0.
Given k1 d1 (x, y) ≤ d2 (x, y) ≤ k2 d1 (x, y) ≤ for all x, y ∈ X.
d2 (x, y)
That is ≤ d1 (x, y) and k1 d1 (x, y) ≤ d2 (x, y) for all x, y ∈ X. (as k1 , k2 >
k2
0.) (∗)

We show that ∃ r1 > 0 such that We show that ∃ r2 > 0 such that
Bd1 (x, r1 ) ⊆ Bd2 (x, r) Bd2 (x, r2 ) ⊆ Bd1 (x, r)
d2 (x, y) From (*), k1 d1 (x, y) ≤ d2 (x, y).
Since ≤ d1 (x, y), we choose
k2 Choose r2 = k1 r.
r
r1 = > 0. Let y ∈ Bd2 (x, r2 ).
k2 This menas d2 (y, x) < r2 = k1 r.
Let y ∈ Bd1 (x, r1 ).
r From (∗∗), k1 d1 (x, y) ≤ d2 (x, y).
This menas d1 (y, x) < r1 = . Hence d2 (y, x) < k1 r =⇒ k1 d1 (y, x) <
k2
d2 (x, y) k1 r.
From (∗), ≤ d1 (x, y) That is d1 (y, x) < r.
k2
r d2 (y, x) r Thus y ∈ Bd1 (x, r).
Hence d1 (y, x) < =⇒ < . Hence Bd2 (x, r2 ) ⊆ Bd1 (x, r).
k2 k2 k2
So, d2 (y, x) < r.
Thus y ∈ Bd2 (x, r).
Hence Bd1 (x, r1 ) ⊆ Bd2 (x, r).
92 CHAPTER 1. UNIT I: METRIC SPACES

(5) Show that k k1 and k k∞ are not equivalent metrics on C [0 , 1 ].

We will show that B∞ (0, 1) is not open in (C [0 , 1 ], k k1 ).
We show that for every r > 0, there exists a function f ∈ B1 (0, r), such that
f∈/ B∞ (0, 1) and hence we will have for every r > 0, B1 (0, r) 6⊆ B∞ (0, 1).
Let r > 0.
By Archimedean Property, there exists n0 ∈ N such that n10 < r.
(
1 − n0 x if 0 ≤ x ≤ n10
Define f : [0, 1] −→ R, as follows: f (x) = .
0 if n10 < x ≤ 1
0 ≤ x ≤ n10 =⇒ 0 ≤ n0 x ≤ 1.
So, −1 ≤ −n0 x ≤ 0.
Adding 1 on both sides we get,
0 ≤ 1 − n0 x ≤ 1.
Thus, for 0 ≤ x ≤ n10 , 0 ≤ f (x) ≤ 1
and for n10 ≤ x ≤ 1, f (x) = 0.
So, f (x) ≤ 1 for all x ∈ [0, 1]. Also, f (0) = 1. Hence sup{f (x) : x ∈ [0, 1]} = 1.
This implies kf k∞ = 1.
That is, kf − 0k∞ = 1. So, f ∈ / B∞ (0, 1).
But We will show that f ∈ B1 (0, r).
Z 1 Z 1 Z 1 Z 1 Z 1
n0 n0

kf − 0k1 = |f (t)|dt. = |f (t)|dt + |f (t)|dt = 1 − n0 t dt + 0 dt.
1 1
0 0 n
0 n0
0
h n0 t2 i n10 h i1
= x− + 0 1.
2 0 n0
1 n0
= − − 0 + 0.
n0 2n20
1 1
= < < r.
2n0 n0
Hence f ∈ B1 (0, r) but f ∈/ B∞ (0, 1).
So there does not exist r > 0 such that B1 (0, r) ⊆ B∞ (0, 1).
Hence B∞ (0, 1) is not open in (C [0 , 1 ], k k1 ).
Since open balls are open sets, B∞ (0, 1) is an open set in (C [0 , 1 ], k k∞ ).

Thus B∞ (0, 1) is an open set in (C [0 , 1 ], k k∞ ) but it is not an open set in (C [0 , 1 ], k k1 ).

Thus the two metric spaces do not have same open sets.

Hence k k1 and k k∞ are not equivalent metrics on C [0 , 1 ].

1.12 Distance of a point from a set in a metric space

Definition 1.12.1 Distance of a point from a set: Let A be a nonempty subset of a
metric space (X, d) and let x ∈ X. The distance of x from A is denoted by x(x, A) and is
given by
d(x, A) = inf{d(x, a) : a ∈ A}.
93

Theorem 1.12.2 Let A be a nonempty subset of a metric space (X, d). Then for x, y ∈
X, |d(x, A) − d(y, A)| ≤ d(x, y)..

Proof:
Let x, y ∈ X and a ∈ A.
By Triangle Inequality, d(x, a) ≤ d(x, y) + d(y, a). (∗)
This implies inf{d(x, a) : a ∈ A} ≤ inf{d(y, a) : a ∈ A} + inf{d(x, y) : a ∈ A}.
Now, inf{d(x, a) : a ∈ A} = d(x, A),
inf{d(y, a) : a ∈ A} = d(y, A) and
inf{d(x, y) : a ∈ A}d(x, y).
So, d(x, A) ≤ d(x, y) + d(y, A). (I)
This implies d(x, A) − d(y, A) ≤ d(x, y).
If we interchange x and y in this inequality, we get d(y, A) − d(x, A) ≤ d(x, y) (II).
Hence from (I) and (II), |d(x, A) − d(y, A)| ≤ d(x, y) for all x, y ∈ X.

Theorem 1.12.3 Let (X, d) be a metric space and A ⊆ X. Then A = {x ∈ X : d(x, A) =

Proof: Since d(x, a) ≥ 0 for all a ∈ A, 0 is a lower bound of {d(x, a) : a ∈ A}.

d(x, A) = 0 if and only if ∀ r > 0, there exists ar ∈ A such that 0 ≤ d(x, ar ) < 0 + r.
if and only if ∀ r > 0, there exists ar ∈ A such that 0 ≤ d(x, ar ) < r.
if and only if ∀ r > 0, there exists ar ∈ A such that ar ∈ B(x, r).
if and only if ∀ r > 0, B(x, r) ∩ A 6= ∅.
if and only if x ∈ A.

Hence A = {x ∈ X : d(x, A) = 0}.

Examples 1.12.4

(1) Let A = (0, 1) ⊂ R with usual distance. Find dA (x) for x ∈ [−1, 2]. Draw the graph
of the function dA .
(2) Let x ∈ R with usual metric. What is dQ (x)?
(3) Let p = (a, b) ∈ R2 with Euclidean metric and A be the x− axis. What is dA (p)?
(4) What is dA (p) where A = {(x, y) ∈ R2 : x2 +y 2 = 1} and p = (a, b). Find an explicit
expression.

(1) Let A = (0, 1) ⊂ R with usual distance. Find dA (x) for x ∈ [−1, 2]. Draw the map
of the function dA .
Let x ∈ [−1, 2].

Case 1: −1 ≤ x ≤ 0.
d(−1, A) = |0 − (−1)| = 1, d(−0.5, A) = 0.5. and so on. We can see that
d(x, A) = −x for all x ∈ [−1, 0].
94 CHAPTER 1. UNIT I: METRIC SPACES

Case 2: 0 < x < 1. That is x ∈ A. Hence dA (x) = 0 for all x ∈ (0, 1).

Case 3: 1 ≤ x ≤ 2.
d(1, A) = |1 − (1)| = 0, d(1.5, A) = 0.5. and so on. We can see that d(x, A) =
x − 1 for all x ∈ [1, 2].

y = −x y =x−1

−1 y=0 1 2
0

−1

(2) Let x ∈ R with usual metric. What is dQ (x)?

We know that Q = R.
Therefore x ∈ R =⇒ x ∈ Q.
Hence by theorem 1.12.3, d(x, Q) = 0. That is, dQ (x) = 0.

(3) Let p = (a, b) ∈ R2 with Euclidean metric and A be the x− axis. What is dA (p)?
dA (p) = |b|.

(4) What is dA (p) where A = {(x, y) ∈ R2 : x2 +y 2 = 1} and p = (a, b). Find an explicit
expression.
dA (p) = a2 + b2 − 1.

1.13 Distance between two sets in a metric space

Definition 1.13.1 Let A and B be two nonempty subsets of a metric space (X, d). The
distance d(A, B) between them is given by d(A, B) = inf{d(a, b) : a ∈ A, b ∈ B}.

Examples 1.13.2
1
(i) Let A = {n + n
: n ∈ N}. Find d(A, N).

(ii) Find the distance d(A, B) between A and B where

(1) A = Q, B = R \ Q.
95

(2) A = Q and B is any nonempty subset of R.

(3) A is the rectangular hperbola xy = 1 and B is the union of axes xy = 0.

(iii) Let A be a subset in a metric space (X, d). Show that the set {x ∈ X : dA (x) < }
is open for any > 0.

Solution

1
(i) Let A = {n + n
: n ∈ N}. Find d(A, N).

d(A, N) = inf{d(a, m) : a ∈ A, m ∈ N}.

1
= inf d n + , m : n, m ∈ N
n

1 1
d n + , n : n ∈ N ⊆ d n + , m : n, m ∈ N
n n

1 1
n + − n : n ∈ N ⊆ d n + , m : n, m ∈ N
n n

1 1
: n ∈ N ⊆ d n + , m : n, m ∈ N
n n

1 1
inf d n + , m : n, m ∈ N ≤ inf :n∈N
n n
So, d(A, N) ≤ 0
Also d(A, N) ≥ 0
∴ d(A, N) = 0

(ii) Find the distance d(A, B) between A and B where

(1) A = Q, B = R \ Q.

d(A, B) = inf{d(a, b) : a ∈ A, b ∈ B}.

= inf{d(a, b) : a ∈ Q, b ∈ R \ Q}.

We will show that d(A, B) = 0.

Since d(a, b) ≥ 0 for all a ∈ A and b ∈ B, 0 is a lower bound of {d(a, b) : a ∈
A, b ∈ B}.
Let > 0.
Let p ∈ Q.
Consider (p − , p + ).
There exists q ∈ R \ Q such that q ∈ (p, , p + ). (from Density Theorem)
This means there exists q ∈ R \ Q such that d(p, q) < .
Thus there exist p ∈ Q and q ∈ R \ Q such that 0 ≤ d(p, q) < 0 + .
Hence inf{d(a, b) : a ∈ Q, b ∈ R \ Q} = 0.
96 CHAPTER 1. UNIT I: METRIC SPACES

(2) A = Q and B is any nonempty subset of R.

d(A, B) = inf{d(a, b) : a ∈ A, b ∈ B}.

= inf{d(a, b) : a ∈ Q, b ∈ B}.

Case 1 B ∩ Q 6= ∅.
Let x ∈ B ∩ Q.
Therefore d(x, x) ∈ {d(a, b) : a ∈ Q, b ∈ B}.
That is 0 ∈ {d(a, b) : a ∈ Q, b ∈ B}.
This means {d(a, b) : a ∈ Q, b ∈ B} contains its lower bound 0.
Hence 0 = inf{d(a, b) : a ∈ Q, b ∈ B}.
Case 2 B ∩ Q = ∅.
Since B 6= ∅, there exists y ∈ R \ Q such that y ∈ B.
Since Q = R, therefore y ∈ R =⇒ there exists a sequence of rationals say
xn −→ y.
Now xn −→ y =⇒ d(xn , y) −→ 0.
We will show that 0 = inf{d(a, b) : a ∈ Q, b ∈ B}.
Clearly 0 is a lower bound of {d(a, b) : a ∈ Q, b ∈ B}.
We will show that for every > 0, there exists a ∈ Q, b ∈ B such that
0 ≤ d(a, b) < 0 + .
Let > 0.
Since xn −→ y, there exists n0 ∈ N such that d(xn , y) < for all n ≥ n0 .
In particular, d(xn0 , y) < .
Now, (xn ) is a sequence in Q =⇒ xn0 ∈ Q.
Hence d(xn0 , y) ∈ {d(a, b) : a ∈ Q, b ∈ B} and d(xn0 , y) < .
So, there exists xn0 ∈ Q, y ∈ B such that d(xn0 , y) < 0 + .
Hence 0 = inf{d(a, b) : a ∈ Q, b ∈ B}.
(3) A is the rectangular hyperbola xy = 1 and B is the union of axes xy = 0.
A = {(x, y) ∈ R2 : xy = 1}, B = {(x, 0) : x ∈ R} ∪ {(0, y) : y ∈ R}.
As 0 ≤ d(a, b) for all a ∈ A and b ∈ B, 0 ≤ inf{d(a, b) : a ∈ A, b ∈ B}.
show that inf{d(a, b) : a ∈ A,
We will b ∈ B}≤ 0.
1 1
Since n, ∈ A and (n, 0) ∈ B, d n, , (n, 0) ∈ {d(a, b) : a ∈ A, b ∈
n n
B}.
1
This means ∈ {d(a, b) : a ∈ A, b ∈ B} for all n ∈ N.
n
Claim: inf{d(a, b) : a ∈ A, b ∈ B} = 0.
Let > 0.
1
By Archimedean property, there exists n0 ∈ N such that < .
n0
1 1
That is 0 < < 0 + and ∈ {d(a, b) : a ∈ A, b ∈ B}.
n0 n0
Hence inf{d(a, b) : a ∈ A, b ∈ B} = 0.

(iii) Let A be a subset in a metric space (X, d).

Show that the set {x ∈ X : dA (x) < } is open for any > 0.
Let > 0 Let x ∈ {x ∈ X : dA (x) < }.
Therefore dA (x) < .
That is inf{d(x, a) : a ∈ A} < .
There exists a ∈ A such that d(x, a) < .
Let δ = − d(x, a). So, δ > 0.
Claim: B(x, δ) ⊆ {x ∈ X : dA (x) < }.
Let y ∈ B(x, δ). This means d(y, x) < δ.
We show that dA (y) < .
Since a ∈ A, dA (y) ≤ d(y, a).
By Triangle Inequality, d(y, a) ≤ d(y, x) + d(x, a).
So, dA (y) ≤ d(y, x) + d(x, a) < δ + d(x, a) = − d(x, a) + d(x, a) = .
So, y ∈ {x ∈ X : dA (x) < }.
This implies B(x, δ) ⊆ {x ∈ X : dA (x) < }.
Hence {x ∈ X : dA (x) < } is open.

1.14 Diameter of a set in a metric space

Definition 1.14.1 Let A be a nonempty subset of a metric space (X, d). The diameter
diam(A) of A is defined by diam(A) = sup{d(x, y) : x, y ∈ A}

Theorem 1.14.2 Let (X, d) be a metric space and A be a nonempty subset of X. Then
diam A = diamA.

Since A ⊆ A, a, b ∈ A =⇒ a, b ∈ A.
Hence d(a, b) ≤ diam(A) ∀a, b ∈ A.
So, sup{d(a, b) : a, b ∈ A} ≤ diam (A).
Hence diam (A) ≤ diam (A) (1) [2 marks]
We show that diam A < diam (A) + ∀ > 0.
Now given > 0 ∃ a, b ∈ A such that
d(a, b) > diam (A) − 2 (by definition of sup) (∗) [1 mark]
Since a, b ∈ A, we have B 4 (a) ∩ A 6= ∅ and B 4 (b) ∩ A 6= ∅ and [1 mark]
Let a0 ∈ B 4 (a) ∩ A and b0 ∈ B 4 (b) ∩ B

So, a0 , b0 ∈ A and d(a0 , a) < 4 and d(b0 , b) < 4 .

We have diam (A) − 2 < d(a, b) (from (∗))
So, diam A − 2 < d(a, b) ≤ d(a, a0 ) + d(a0 , b0 ) + d(b0 , b) < 4 + d(a0 , b0 ) + 4 .

Hence diam (A) − 2 < d(a0 , b0 ) + 2

Now, a0 , b0 ∈ A =⇒ d(a0 , b0 ) ≤ diam (A).
Hence diam (A) − 2 < diam A + 2 .
Thus diam (A) < diam (A) + ∀ > 0.
Hence diam (A) ≤ diam (A) (2) [2 marks]
Thus diam (A) = diam (A) (from (1) and (2))

Examples 1.14.3
98 CHAPTER 1. UNIT I: METRIC SPACES

Definition 1.14.4 Bounded Set: A subset A of a metric space is said to be bounded if

it is empty or its diameter diam(A) of A is finite where diam A = sup{d(x, y) : x, y ∈ A}.

(1) Show that the diameter diam(B(x, r)) ≤ 2r and that the strict inequality can occur.

Solution
1 Consider (R, d) where d is the discrete metric.
Let p ∈ R. So B(p, 1) = {x ∈ R : d(x, p) < 1} = {x ∈ R : d(x, p) = 0} = {p}.
Let x, y ∈ B(p, 1) = {p}. So x = p, y = p. This implies d(x, y) = 0.
Hence sup{d(x, y) : x, y ∈ B(p, 1)} = 0 =⇒ diam B(p, 1) < 1 < 2 = 2 ∗ r.

1.15 Closed balls in a metric space

Definition 1.15.1 Let (X, d) be a metric space. Let x ∈ X and r > 0. The subset
Bd [x, r] = {y ∈ X : d(x, y) ≤ r} is called the closed ball centred at x with radius r with
respect to the metric d.

1.16 Closed Sets in a metric space

Definition 1.16.1 Let (X, d) be a metric space and F ⊆ X.F is said to be closed if its
complement X \ F is open.

Remarks 1.16.2

(i) ∅, X are both open and closed in any metric space. (From definition of open sets,
∅, X are open. Since X \ ∅ = X is open, ∅ is closed. Similarly X \ X = ∅ is open,
X is closed. )

(ii) In (R, d) where d usual, [a, b], [a, ∞), (−∞, b] are closed. (a, b), (a, b], [a, b) are not
closed.

Examples 1.16.3

(i) Any finite subset of a metric space is closed.

(ii) Every subset of a discrete metric space is both open and closed.

(iii) Every subset of a finite metric space is is both open and closed.

(iv) Every subset of Z is is both open and closed w.r.t. the induced usual metric from
R.

(v) S = {(x, y) ∈ R2 : x2 + y 2 = 1} is a closed subset of R2 , where the distance being

Euclidean.
99

√ √
(vi) Show that A = {x ∈ Q : − 2 < x < 2} is both open and closed in the subspace
Q of R with usual distance.
(vii) Show that S = {x ∈ Q : 3 < x2 < 5} is both open and closed in the subspace Q of
R with usual metric.

Solution
(i) Any finite subset of a metric space is closed.
Let (X, d) be a metric space and F = {x1 , x2 , · · · , xn } be a finite subset of X.
Consider X \ F.
If X \ F = ∅ then X \ F is open and hence F is closed.
Suppose X \ F 6= ∅.
Let x ∈ X \ F.
This implies x 6= x1 , x 6= x2 , · · · x 6= xn . So d(x, x1 ) > 0, d(x, x2 ) > 0 · · · , d(x, xn ) >
0.
Choose r = min{d(x, xi ) : 1 ≤ i ≤ n}. So r > 0.
Claim B(x, r) ⊆ X \ F.
Let y ∈ B(x, r). So, d(y, x) ≤ r < d(xi , x) for all i, 1 ≤ i ≤ n.
This implies y 6= xi , for all i, 1 ≤ i ≤ n.
Hence y ∈/ {x1 , x2 , · · · , xn }.
This means y ∈ / F . Hence y ∈ X \ F.
Thus B(x, r) ⊆ X \ F. Hence X \ F is open and F is closed.
(ii) Every subset of a discrete metric space is is both open and closed.
Let (X, d) be a discrete metric space. First we show that every subset of a discrete
metric space is open.
Let p ∈ X.

B(p, 1) = {x ∈ X : d(x, p) < 1}.

= {x ∈ X : d(x, p) = 0}.
= {x ∈ X : x = p}.
= {p}.

Let G ⊆ X.
If G = ∅ then G is open.
S G 6= ∅.
So, suppose
∴ G = Sp∈G {p}
∴ G = p∈G B(p, 1)
Since open balls are open sets, we have expressed G as a union of open sets.
G is open in (X, d) as arbitrary union of open sets is open.

Now we show that G is closed.

As G ⊆ X, X \ G ⊆ X.
Since every subset of X is open, X \ G is open and hence G is closed.
100 CHAPTER 1. UNIT I: METRIC SPACES

(iii) Every subset of a finite metric space is both open and closed.
Let (X, d) be a finite metric space.
So, X = {a1 , a2 , · · · , an }.
We show that every singleton set is open.
Let p ∈ X.
This implies p = ai for some i, 1 ≤ i ≤ n.
Let r = min{d(p, aj ) : 1 ≤ j ≤ n, j 6= i.
Since j 6= i, aj 6= ai =⇒ aj 6= p.
=⇒ d(p, aj ) > 0 ∀ j, j 6= i, 1 ≤ j ≤ n.
Therefore r > 0.
Claim: B(p, r) = {p}.
Let x ∈ B(p, r). So d(x, p) < r.
But r ≤ d(p, aj ), for all 1 ≤ j ≤ n, j 6= i.
This implies d(x, p) < d(aj , p), for all 1 ≤ j ≤ n, j 6= i.
Hence x 6= aj for all j, 1 ≤ j ≤ n, j 6= i.
Thus we have, x ∈ {a1 , a2 , · · · , ai = p, · · · , an } and x 6= aj for all j, 1 ≤ j ≤ n, j 6= i.
Hence x = ai . That is x = p.
Hence B(p, r) ⊆ {p}.
As d(p, p) = 0 < r, p ∈ B(p, r). So {p} ⊆ B(p, r).
Hence B(p, r) = {p}.

Let G ⊆ X.
If G = ∅ then G is open.
S G 6= ∅.
So, suppose
∴ G = Sp∈G {p}
∴ G = p∈G Bd (p, 1)
Since open balls are open sets,
we have expressed G as a union of open sets.
G is open in (X, d)
(as arbitrary union of open sets is open)

(iv) Every subset of Z is both open and closed w.r.t. the induced usual metric from R.
Let p ∈ Z.

BZ (p, 1) = {x ∈ Z : d(x, p) < 1}.

= {x ∈ Z : |x − p| < 1}.
= {x ∈ Z : x ∈ (p − 1, p + 1)}.
= Z ∩ (p − 1, p + 1).
= {p}. (as p ∈ Z.)

Let G ⊆ Z. If G = ∅ then G is open.

S G 6= ∅.
So, suppose
∴ G = Sp∈G {p}
∴ G = p∈G BZ (p, 1)
101

Since open balls are open sets,

we have expressed G as a union of open sets.
G is open in (X, d) as arbitrary union of open sets is open.
Now we show that G is closed.
As G ⊆ X, X \ G ⊆ X.
Since every subset of X is open, X \ G is open and hence G is closed.

(v) S = {(x, y) ∈ R2 : x2 + y 2 = 1} is a closed subset of R2 , where the distance being

Euclidean.

(We will have to give another proof using the definition of closed sets)
S = {(x, y) ∈ R2 : x2 + y 2 = 1}
Let p = (p1 , p2 ) ∈ S.
So, ∃ a sequence (xn , yn ) in S such that (xn , yn ) −→ (p1 , p2 ) in R2 , distance being
Euclidean.
Hence x2n + yn2 = 1 ∀n ∈ N and (xn , yn ) −→ (p1 , p2 ) in R2 , distance being
Euclidean.
Now, (xn , yn ) −→ (p1 , p2 ) in R2 , distance being Euclidean =⇒ xn −→ p1 , yn −→ p2
in R w.r.t. the usual distance.
So, x2n −→ p21 and yn2 −→ p22 in R (distance being usual).
Hence x2n + yn2 −→ p21 + p22 in R (distance being usual).
But x2n + yn2 = 1 ∀n ∈ N
=⇒ x2n + yn2 −→ 1
=⇒ p21 + p22 = 1 (as a convergent sequence converges to unique limit).
Hence (p1 , p2 ) ∈ S.
Thus S ⊆ S.
Hence S = S (as S ⊆ S).
Hence S is closed.

1.16.1 Closed Sets in a sub space

Theorem 1.16.4 Let (X, d) be a metric space and Y be a non-empty subset of X. A subset
F of Y is closed in the subspace (Y, d) if and only if F = K ∩ Y where K is a closed set in
(X, d).

Proof: Suppose F is closed in Y.

Therefore, Y \ F is open in Y .
102 CHAPTER 1. UNIT I: METRIC SPACES

Therefore, Y \ F = V ∩ Y for some V ⊂ X, V open in X.

F = Y \ (V ∩ Y )
= Y ∩ (V ∩ Y )c
= Y ∩ (V c ∪ Y c )
= (Y ∩ V c ) ∪ (Y ∩ Y c )
=Y ∩Vc

Since V is open in X, V c is closed in X.

Put V c = K
Therefore F = K ∩ Y where K is closed in X.
Conversely, F = K ∩ Y where K is closed in X.
We show that F is closed in Y .
Consider Y \ F .

Y \ F = Y \ (K ∩ Y )
c
=Y ∩ K ∩Y
= Y ∩ (K c ∪ Y c )
= (Y ∩ K c ) ∪ (Y ∩ Y c )
= (Y ∩ K c ) ∪ ∅
= Y ∩ Kc

Now, K is closed in X. So, K c is open in X.

Hence, Y ∩ K c is open in Y .
Thus, Y \ F is open in Y .
Hence, F is closed in Y .

Examples 1.16.5
√ √
(I) Show that A = {x ∈ Q : − 2 < x < 2} is both open and closed in the subspace
Q of R with usual distance.

(II) Show that S = {x ∈ Q : 3 < x2 < 5} is both open and closed in the subspace Q of
R with usual metric.

open subset V of X )
Hence A is open in the subspace
√ √ Q. √ √ √ √
We√can√also write A as [− 2, 2] ∩ Q as − 2, 2 ∈ / Q =⇒ A = (− 2, 2) ∩ Q =
[− 2, √ 2] ∩√
Q. √ √
Since [− 2, 2] is closed in R, [− 2, 2] ∩ Q is closed in the subspace Q with usual
metric induced from R.
√ √ √
(2) S = √{x ∈ Q : 3√< x2√< 5} =⇒ S √ = {x√ ∈ Q : − 5 < x < − 3} ∪ {x ∈ Q : 3<
x < 5} √= (−√ 5, − 3) ∩√ Q ∪√ ( 3, 5) ∩ Q. √ √ √ √
S = (− 5, − 3) ∩ Q ∪ [ 3, 5] ∩ Q. = (− 5, − 3) ∪ ( 3, 5) ∩ Q. [1M]
Since open intervals are open sets in (R, usual) and union of open sets is open,
h √ √ i h√ √ i √ √ √ √
We can also write S as − 5, − 3 ∪ 3, 5 ∩Q as − 3, 3, − 5, 5 ∈ / Q =⇒
√ √ √ √ √ √ √ √
S = (− √5, −√ 3) ∩ Q ∪ ( 3, 5) √∩ Q
√ = [− 5, − 3] ∪ [ 3, 5] ∩ Q.
Since [− 2, 2] is closed in R, [− 2, 2] ∩ Q is closed in the subspace Q with usual
metric induced from R.
√ √ √ √
∴ (−√5, √3)) ∪ (√3, √5)) is an open set in (R, usual).
∴ (− 5, 3)) ∪ ( 3, 5)) ∩ Q is an open set in the subspace Q of (R, usual).

1.16.2 Examples of closed sets in various

metric spaces.

1.17 Limit Point of a set

Theorem 1.17.1 A subset F of a metric space (X, d), the following statements are equiv-
alent:
(I) F is closed

(II) F contains all its limit points.

Proof: Suppose F ⊆ X is closed.

TST F contains all its limit points.
i.e. TST D(F ) ⊆ F where D(F ) = set of all limit points of F
If D(F ) = ∅, D(F ) ⊆ F.
Suppose D(F ) 6= ∅.
Let if possible ∃ x ∈ D(F ) such that x ∈
/ F.
Then x ∈ X \ F, which is an open set.
∴ ∃ r > 0 such that B(x, r) ⊂ X \ F.
∴ B(x, r) ∩ F = ∅
since x ∈
/ F, ∴ F \ {x} = F .

∴ B(x, r) ∩ F \ {x} = ∅
∴x∈/ D(F ), which is not true.
104 CHAPTER 1. UNIT I: METRIC SPACES

∴ our assumption that x ∈

/ F is wrong.
∴ x ∈ D(F ) =⇒ x ∈ F ∴ D(F ) ⊆ F [4 marks]

Conversely, suppose D(F ) ⊆ F.

we show that X \ F is open (i.e. F is closed.)
If X \ F = ∅, X \ F is open.
Suppose X \ F 6= ∅, x ∈ X \ F .
Then x ∈/ D(F ) (For x ∈ D(F ) =⇒ x ∈ F )
∴ ∃ r > 0 such that B(x, r) ∩ F \ {x} = ∅.
B(x, r) ∩ F = ∅ (since x ∈ X \ F, ∴ F \ {x} = F )
∴ B(x, r) ⊆ X \ F .
∴ X \ F is open and F is closed.

Theorem 1.17.2 Let (X, d) be a metric space and S ⊆ X. Show D(S) is a closed subset
of X where D(S) denotes the set of all limit points of S.

We show that D(S) is a closed set.

Since a set is closed if and only if it contains all its limit points, we show that D D(S) ⊆
D(S). [1M]
Let p ∈ D D(S)
TST p ∈ D(S)
i.e. TST for every r > 0, B(p, r) contains infinitely many points of S.
Let r > 0
Since p ∈ D D(S) ∴ for this r > 0, B(p, r) ∩ D(S) \ {p} 6= ∅. [1 mark]

∴ ∃ q ∈ B(p, r) ∩ D(S) \ {p} [1M]
∴ q ∈ B(p, r), q ∈ D(S) and q 6= p. [1 mark]
Since B(p, r) is an open ball and open balls are open sets, ∴ B(p, r) is an open set.
Since q ∈ B(p, r), ∴ ∃ δ > 0 such that B(q, δ) ⊆ B(p, r). [1 mark]
Now, q ∈ D(S) =⇒ B(q, δ) contains infinitely many points of S. [1 mark]
Since B(q, δ) ⊆ B(p, r) ∴ B(p, r) contains infinitely many points of S. [1 mark]
∴ for every r > 0, B(p, r) contains infinitely many points of S. [1 mark]
∴ p ∈ D(S).

Thus D D(S) ⊆ D(S)
Hence D(S) is a closed set.

Theorem 1.17.3 . Every infinite bounded subset of R has a limit point.

Let A be an infinite bounded subset of R.

Since A is bounded, there exists M ∈ R+ such that |a| ≤ M ∀a ∈ A
So, −M ≤ a ≤ M ∀a ∈ A
Let E = {x ∈ R : x ≤ for infinitely many a ∈ A}.
∴ −M ∈ E.
So, E 6= ∅.
We will show that E is also bounded above by M .
105

Let x ∈ E.
So, there exists infinitely many a ∈ A such that x ≤ a.
Now, a ≤ M ∀a ∈ A
∴ x ≤ M ∀x ∈ E.
Thus E is bounded above.
So, E 6= ∅, E ⊂ R and E is bounded above.
∴ by order completeness of R, sup E exists.
Let L = sup E.
Claim: L is a limit point of E.
Let > 0 .
Since L = sup E, ∴ for this > 0 ∃ x ∈ E such that L − < x ≤ L (∗)
Since x ∈ E, ∴ ∃ infinitely many a ∈ A such that x ≤ a.
∴ L−<x≤a for infinitely many a ∈ A
∴ L − < a for infinitely many a ∈ A
∴ L < a + for infinitely many a ∈ A (∗∗)
0
Infinitely many from these a s are a < L + , if not then a − < L for finitely many
0
a s such that L − < a.
So, except for these finitely many a0 s, infinitely many ∴ from (∗) and (∗∗), M − <
x ≤ a < M − for infinitely may a ∈ A

1.18 Closure of a set

Theorem 1.18.1 Let (X, d) be a metric space and A ⊆ X. Then

(I) A is a closed set.

(II) A is closed if and only if A = A.

(III) A is the smallest closed set containing A.

(I) We show that D(D(A)) ⊆ D(A)

If D(D(A)) = ∅ then D(D(A)) ⊆ D(A)
Suppose D(D(A)) 6= ∅.
Let x ∈ D(D(A)).
We will show that x ∈ D(A). That is to show that x is a limit point of A.
Let r > 0.
We show that B(x, r) ∩ A \ {x} = 6 ∅.
Since x ∈ D(D(A)), B(x, r) ∩ D(A) \ {x} = 6 ∅.
Let y ∈ B(x, r) ∩ D(A) \ {x}
Therefore, y ∈ B(x, r), y ∈ D(A), y 6= x.
Since open balls are open sets, B(x, r) is an open set.
So, ∃ δ > 0 such that B(x, δ) ⊆ B(x, r) (we may take δ = r − d(x, y))
Since y ∈ D(A), B(y, δ) contains infinitely many points of A.
Therefore ∃ z ∈ B(y, δ) such that z ∈ A, z 6= x, z 6= y.
z ∈ B(y, δ) ⊆ B(x, r) =⇒ z ∈ B(x, r).
106 CHAPTER 1. UNIT I: METRIC SPACES

Thus z ∈ B(x, r) ∩ A.
Also, z 6= x =⇒ z ∈ B(x, r) ∩ A \ {x}.
Hence B(x, r) ∩ A \ {x} =
6 ∅.
Hence x ∈ D(A).
Thus, D(D(A)) ⊆ D(A).
Hence D(A) is closed.

(II) To show that A is closed if and only if A = A.

Suppose A is closed.
Therefore D(A) ⊆ A.
Hence A ∪ D(A) ⊆ A.
Therefore, A ⊆ A. · · · · · · (I)
Also, A = A ∪ D(A) =⇒ A ⊆ A · · · · · · (II)
Therefore, A = A.
Conversely, suppose A = A =⇒ D(A) ⊆ A =⇒ A is closed.

(III) A is the smallest closed set containing A.

A = A ∪ D(A)

D(A) = D A ∪ D(A) = D(A) ∪ D D(A) ⊆ D(A) ∪ D(A)

(as D D(A) ⊆ D(A))
Hence, D(A) ⊆ D(A).
Now, (A) = A ∪ D(A) ⊆ A ∪ D(A) ⊆ A (as D(A) ⊆ A)
But A ⊆ (A)
So, A = (A)
Therefore, A is a closed set. [3 M]
Also, A ⊆ A =⇒ A is a closed set contaitning A. [1 M]
Let F be any closed set containting A.
Then A ⊆ F =⇒ A ⊆ F = F (as F is closed.) [2M]
Therefore, A

Theorem 1.18.2 (X, d) is a metric space and A, B ⊆ X. Then

(I) If A ⊆ B then A ⊆ B.

(II) A ∩ B ⊆ A ∩ B and equality may not hold.

(III) A ∪ B = A ∪ B.

(I) To prove that If A ⊆ B then A ⊆ B

x ∈ A =⇒ ∀ r > 0, B(x, r) ∩ A 6= ∅
=⇒ ∀ r > 0, B(x, r) ∩ B 6= ∅ (As A ⊆ B, B(x, r) ∩ A ⊆ B(x, r) ∩ B
=⇒ x ∈ B
Hence A ⊆ B =⇒ A ⊆ B
107

(II) To prove that A ∩ B ⊆ A ∩ B and equality may not hold.

A∩B ⊆A A∩B ⊆B
A∩B ⊆A A∩B ⊆B
A∩B ⊆A∩B

Consider (R, d) where d is the usual distance in R, A = Q, B = R \ Q.

Then A = R, B = R.
If x ∈ R, for any r > 0, B(x, r) = (x − r, x + r).
∴ ∃ s ∈ Q, t ∈ R \ Q such that x − r < s < x + r, x − r < t < x + r (by density
theorem).
∴ s ∈ B(x, r) ∩ Q, t ∈ B(x, r) ∩ R \ Q.
Thus B(x, r) ∩ Q 6= ∅ & B(x, r) ∩ R \ Q 6= ∅.
∴x∈Q & x∈R\Q
Hence R ⊆ Q ⊆ R & R ⊆ R \ Q ⊆ R.
Thus Q = R & R \ Q = R
A ∩ B = Q ∩ (R \ Q) = ∅ = ∅
A ∩ B = R ∩ R = R.
A∩B (A∩B
(III) To prove that A ∪ B = A ∪ B.

A⊆A∪B B ⊆A∪B
A⊆A∪B B ⊆A∪B
A∪B ⊆A∪B (∗)

x∈
/ A ∪ B =⇒ x ∈
/A & x∈ / B.
=⇒ ∃ r1 > 0 such that B(x, r1 ) ∩ A = ∅ and ∃ r2 > 0 such that B(x, r2 ) ∩ B = ∅

Let r = min{r1 , r2 }, ∴ r > 0, and B(x, r) ⊆ B(x, ri ) for i = 1, 2

Then B(x, r) ∩ A = ∅, B(x, r) ∩ B = ∅.
=⇒ B(x, r) ∩ (A ∪ B) = ∅.
=⇒ x ∈/ A ∪ B.
Hence x∈ / A ∪ B =⇒ x ∈ / A ∪ B.
The contrapositive statement gives:
x ∈ A ∪ B =⇒ x ∈ A ∪ B.
Thus A ∪ B ⊆ A ∪ B (∗∗).
From (∗) and (∗∗)
A ∪ B = A ∪ B. is the smallest closed set containing A.

Theorem 1.18.3 Let A be a subset of a metric space (X, d). Then

(I) (X \ A) = X \ A◦
(II) (X \ A)◦ = X \ (A)
108 CHAPTER 1. UNIT I: METRIC SPACES

(I) To prove that X \ A0 = X \ A 3 marks

x ∈ X \ A0 if and only if x ∈
/ A0
. if and only if @ r > 0 such thatTB(x, r) ⊆ A
. if and only if ∀r > 0 B(x, r) X \ A 6= ∅
. if and only if x ∈ X \ A
. if and only if x ∈ X \ A

(II) X \ Ā = (X \ A)0 3 marks

x ∈ X \ Ā if and only if x ∈/ Ā T
. if and only if ∃ r > 0 such that B(x, r) A = ∅
. if and only if ∃ r > 0 B(x, r) ⊆ X \ A
. if and only if x ∈ X \ A◦
. if and only if x ∈ (X \ A)◦ .

1.19 Dense Sets

Prove that a subset A of a metric space (X, d) is dense in X if and only if G ∩ A 6= ∅ for
each non-empty open subset G of X.
Suppose A is a dense subset of X and G ⊆ X, G open and G 6= ∅.
To show G ∩ A 6= ∅.
G 6= ∅.
∃ x ∈ G. [1M]
As G is open, x ∈ G, ∃ an open ball B(x, r) ⊆ G. [1M]
x ∈ X = A.
So, B(x, r) ∩ A 6= ∅. [1M]
Hence B(x, r) ⊆ G =⇒ G ∩ A 6= ∅.
Conversely, suppose for each non-empty open subset G of X, G ∩ A 6= ∅.
We show A = X.
Let x ∈ X.
Then ∀r > 0, B(x, r) is a non-empty open set. [1M]
Hence ∀r > 0, B(x, r) ∩ A 6= ∅. [1M]
∴ x ∈ A. [1M]
X ⊆ A ⊆ X.
∴ A = X.
Hence A is dense in X.
1.20 boundary of a set

Riemannian Geometry Study Guide
No ratings yet
Riemannian Geometry Study Guide
123 pages
Advanced Ordinary Differential Equations: Third Edition Athanassios G. Kartsatos
No ratings yet
Advanced Ordinary Differential Equations: Third Edition Athanassios G. Kartsatos
232 pages
My Project PDF
100% (1)
My Project PDF
27 pages
Metric Spaces and Complex Analysis
No ratings yet
Metric Spaces and Complex Analysis
161 pages
255tutnotes Analysis
No ratings yet
255tutnotes Analysis
100 pages
Algebraic Topology I (MAT431) : Lecture Notes
No ratings yet
Algebraic Topology I (MAT431) : Lecture Notes
82 pages
Script Convex Analysis
No ratings yet
Script Convex Analysis
167 pages
MT Notes 2024-4
No ratings yet
MT Notes 2024-4
96 pages
Lecture Notes On Electrodynamics 118120
No ratings yet
Lecture Notes On Electrodynamics 118120
189 pages
A YT FunctAnalysis Vol 1 146 JM 01
No ratings yet
A YT FunctAnalysis Vol 1 146 JM 01
386 pages
476278MP
No ratings yet
476278MP
147 pages
Notes On Complex Analysis (Utopia)
No ratings yet
Notes On Complex Analysis (Utopia)
108 pages
Abstract Algebra Essentials
No ratings yet
Abstract Algebra Essentials
230 pages
000 VC Chapters en
No ratings yet
000 VC Chapters en
154 pages
General Relativity Lecture Notes
No ratings yet
General Relativity Lecture Notes
64 pages
CA Lectures
No ratings yet
CA Lectures
196 pages
Complete
No ratings yet
Complete
216 pages
Lecture Notes
No ratings yet
Lecture Notes
118 pages
Notes Topology Ngalla
No ratings yet
Notes Topology Ngalla
96 pages
P352 - Notes
No ratings yet
P352 - Notes
171 pages
Lattice Polytopes
100% (1)
Lattice Polytopes
241 pages
Notes Pde Ngalla
No ratings yet
Notes Pde Ngalla
130 pages
Convex Analysis and Optimization Guide
No ratings yet
Convex Analysis and Optimization Guide
70 pages
Abs Alg
No ratings yet
Abs Alg
401 pages
Idolikecfd Vol1 2ed v2p2
No ratings yet
Idolikecfd Vol1 2ed v2p2
299 pages
Melhor Material
No ratings yet
Melhor Material
182 pages
Mathematical Methods II Lecture Notes: Manuel Mañas Baena Luis Martínez Alonso
No ratings yet
Mathematical Methods II Lecture Notes: Manuel Mañas Baena Luis Martínez Alonso
249 pages
Functions of Several Real Variables
No ratings yet
Functions of Several Real Variables
269 pages
Mechanics Book Fall 2014
No ratings yet
Mechanics Book Fall 2014
206 pages
CMU Prob-Grad-Notes - Tomasz Tkocz
No ratings yet
CMU Prob-Grad-Notes - Tomasz Tkocz
226 pages
Infinite Dimensional Manifolds
No ratings yet
Infinite Dimensional Manifolds
171 pages
Ele Teo Rel
No ratings yet
Ele Teo Rel
152 pages
FNF Notes
No ratings yet
FNF Notes
74 pages
7490 Dmth516 Topology II
No ratings yet
7490 Dmth516 Topology II
90 pages
Polycopie de Cours Berhail Amel Sobolev
No ratings yet
Polycopie de Cours Berhail Amel Sobolev
112 pages
tm2016 17
No ratings yet
tm2016 17
181 pages
Filters and Ultrafilters in Real Analysis
No ratings yet
Filters and Ultrafilters in Real Analysis
31 pages
Notes On MATH 441
No ratings yet
Notes On MATH 441
122 pages
MO Apuntes Anuales
No ratings yet
MO Apuntes Anuales
200 pages
Lectures
No ratings yet
Lectures
289 pages
Topology. Course Notes - Gary Gruenhage
No ratings yet
Topology. Course Notes - Gary Gruenhage
244 pages
Differential Geometry Notes
100% (1)
Differential Geometry Notes
179 pages
(Rothe) Topic From Relativity (2010)
No ratings yet
(Rothe) Topic From Relativity (2010)
107 pages
Topology: Hu NH Quang Vũ
No ratings yet
Topology: Hu NH Quang Vũ
281 pages
KTRM Analytic Mechanics
100% (1)
KTRM Analytic Mechanics
144 pages
Groups PDF
100% (1)
Groups PDF
789 pages
Classical Field Theory
No ratings yet
Classical Field Theory
180 pages
Analysis 3
No ratings yet
Analysis 3
127 pages
Mechanics Dec 05
No ratings yet
Mechanics Dec 05
309 pages
Differential Topology Basics
No ratings yet
Differential Topology Basics
74 pages
Rings and Modules
100% (2)
Rings and Modules
141 pages
DiBenedetto 2002 Real Analysis
0% (1)
DiBenedetto 2002 Real Analysis
505 pages
Metric Topo
No ratings yet
Metric Topo
149 pages
Solutions Ch1-5
No ratings yet
Solutions Ch1-5
72 pages
Ph106bc: Electrodynamics: Sunil Golwala
100% (2)
Ph106bc: Electrodynamics: Sunil Golwala
960 pages
M. Sc. I Sem. II Real Analysis All
No ratings yet
M. Sc. I Sem. II Real Analysis All
175 pages
Diff Geom
No ratings yet
Diff Geom
487 pages
Qamrul Hasan Ansari, Metric Spaces - Including Fixed Point Theory and Set-Valued Maps
100% (2)
Qamrul Hasan Ansari, Metric Spaces - Including Fixed Point Theory and Set-Valued Maps
209 pages
Ph.D. Qualifying Exam, Real Analysis Fall 2009,: J J 1 N 1 N N 1 N
No ratings yet
Ph.D. Qualifying Exam, Real Analysis Fall 2009,: J J 1 N 1 N N 1 N
2 pages
Introduction To Real Analysis Fall 2014 Lecture Notes: Vern I. Paulsen November 6, 2014
No ratings yet
Introduction To Real Analysis Fall 2014 Lecture Notes: Vern I. Paulsen November 6, 2014
81 pages
BTech ECE 2018 RP PDF
No ratings yet
BTech ECE 2018 RP PDF
188 pages
Linear Algebra Course Guide
No ratings yet
Linear Algebra Course Guide
2 pages
Radius of Curvature
100% (1)
Radius of Curvature
11 pages
Lipschitz Cont Slides
No ratings yet
Lipschitz Cont Slides
11 pages
MCQ Na A
0% (1)
MCQ Na A
9 pages
XI Real Numbers-1
No ratings yet
XI Real Numbers-1
9 pages
Maths-Crash Course-1
No ratings yet
Maths-Crash Course-1
5 pages
Continuous-Time Signals: David W. Graham EE 327
No ratings yet
Continuous-Time Signals: David W. Graham EE 327
18 pages
Module 1 Lesson 1
No ratings yet
Module 1 Lesson 1
17 pages
AA HL - Questionbank - Calculus - Easy
No ratings yet
AA HL - Questionbank - Calculus - Easy
74 pages
Two-Dimensional Digital Image Correlation For In-Plane Displacement and Strain Measurement A Review
No ratings yet
Two-Dimensional Digital Image Correlation For In-Plane Displacement and Strain Measurement A Review
18 pages
PDF HMH Texas Algebra 2 Timothy D. Kanold Download
100% (5)
PDF HMH Texas Algebra 2 Timothy D. Kanold Download
53 pages
Monday 24 May 2021: Further Pure Mathematics
No ratings yet
Monday 24 May 2021: Further Pure Mathematics
6 pages
Exercise 3 - Attempt Review-10of10 PDF
No ratings yet
Exercise 3 - Attempt Review-10of10 PDF
3 pages
William Arveson - An Invitation To C - Algebras (1976) (978!1!4612-6371-5)
100% (5)
William Arveson - An Invitation To C - Algebras (1976) (978!1!4612-6371-5)
117 pages
A Collection of Dice Problems: With Solutions and Useful Appendices
No ratings yet
A Collection of Dice Problems: With Solutions and Useful Appendices
95 pages
Trigonometric Functions in Real-Life Applications
No ratings yet
Trigonometric Functions in Real-Life Applications
11 pages
Mat201 Partial Differential Equations and Complex Analysis, December
No ratings yet
Mat201 Partial Differential Equations and Complex Analysis, December
3 pages
Transportation Problems
No ratings yet
Transportation Problems
64 pages
Triangular Fuzzy Matrices: Dr. V. Ramadoss, M.SC., M.Phil., PH.D., & S. Keerthana
No ratings yet
Triangular Fuzzy Matrices: Dr. V. Ramadoss, M.SC., M.Phil., PH.D., & S. Keerthana
4 pages
MATLAB Impulse Response Guide
No ratings yet
MATLAB Impulse Response Guide
5 pages
CST Blueprint - Algebra 2
100% (5)
CST Blueprint - Algebra 2
3 pages
History & Applications of Calculus of Variations
No ratings yet
History & Applications of Calculus of Variations
26 pages
Vectoranalysis
No ratings yet
Vectoranalysis
466 pages
Curved Beams
No ratings yet
Curved Beams
19 pages
Analysis of Stress and Strain PDF
No ratings yet
Analysis of Stress and Strain PDF
102 pages
Press - 09 Tanzania
No ratings yet
Press - 09 Tanzania
9 pages
DIP Lecture 06
No ratings yet
DIP Lecture 06
26 pages
Portfolio Theory Lecture Notes
100% (1)
Portfolio Theory Lecture Notes
2 pages
S3 Maths Annual Scheme
No ratings yet
S3 Maths Annual Scheme
8 pages