Chapter Two

The Metric Spaces

Definition:
Let 𝑋 be a non-empty set and 𝑑: 𝑋 × 𝑋 ⟶ 𝑅 is called the
distance function satisfy the following conditions:
a) d(x,y) ≥ 0 , for all x,y ϵ x
b) d(x,y) = 0 iff x=y
c) d(x,y) = d(y,x) for all x,y ϵ x
d) d(x,y) ≤ d(x,z) + d(z,y) (Triangle inequality) then (x,d) is
called metric space.
Example:
let d:RxR R defined by d(x,y) = |x-y| , for all x,y ϵ R show
that (R,d) is a metric space
Sol:
1) d(x,y) = |x-y| > 0 , for all x,y ϵ R (By def. of absolutely value)
2) d(x,y) = 0
⟺|x-y| = 0
⟺ x-y = 0
⟺x=y
3) d(x,y) = | x-y |
= |- (y-x)|
= |-1| . |y-x|
= |y-x|
= d(y,x)
4) d(x,y) = |x-y|
= |x-z+z-y|
≤ |x-z|+|z-y|
≤ d(x,y) + d(z,y)

1
∴ (R,d) is a metric space
Lemma 2.1 : (Cauchy- Schwarz inequality)
For any real numbers a1, a2, …, an, b1, b2, …, bn we have

(a1 b1 +…+ an bn) ≤ √𝑎12 + ⋯ + 𝑎𝑛2 . √𝑏12 + ⋯ + 𝑏𝑛2

Lemma 2.2 : Minkowski inequality
For any real numbers a1, a2, …, an, b1, b2, …, bn we have
√(𝑎1 + 𝑏1 )2 + (𝑎2 + 𝑏2 )2 + ⋯ + (𝑎𝑛 + 𝑏𝑛 )2

≤ √𝑎12 + ⋯ + 𝑎𝑛2 + √𝑏12 + ⋯ + 𝑏𝑛2

Example :
Let d:RnxRn R defined by
d((x1,x2, …, xn), (y1,y2, … yn)) = √∑𝑛𝑖=1(𝑥𝑖 − 𝑦𝑖 )2
where x= (x1, x2,…, xn) and y=(y1,y2,…, yn)
sol:
1) (𝑥𝑖 − 𝑦𝑖 )2 ≥ 0, ∀ 𝑖 = 1,2, … , 𝑛
𝑛

⟹ ∑(𝑥𝑖 − 𝑦𝑖 )2 ≥ 0
𝑖=1

⟹ √∑(𝑥𝑖 − 𝑦𝑖 )2 ≥ 0
𝑖=1

⟹ 𝑑 (𝑥, 𝑦) ≥ 0, ∀ 𝑥, 𝑦 ∈ 𝑅 𝑛
2) 𝑑 (𝑥, 𝑦) = 0, ∀ 𝑥, 𝑦 ∈ 𝑅 𝑛

2
 𝑛

⟺ √∑(𝑥𝑖 − 𝑦𝑖 )2 = 0
𝑖=1

⟺ ∑(𝑥𝑖 − 𝑦𝑖 )2 = 0
𝑖=1

⟺(xi,yi)2 =0 , ∀ 𝑖 = 1,2, … , 𝑛
⟺xi = yi , ∀ 𝑖 = 1,2, … , 𝑛
⟺ (x1,x2,…, xn) = (y1,y2,…, yn)
⟺x=y
3) d(x,y) = √∑𝑛𝑖=1(𝑥𝑖 − 𝑦𝑖 )2
= √∑𝑛𝑖=1(𝑦𝑖 − 𝑥𝑖 )2
= d(y,x)
4) let x= (x1,…, xn), y= (y1,…, yn), z= (z1,…, zn) ∈ 𝑅
d(x,y) = √∑𝑛𝑖=1(𝑥𝑖 − 𝑦𝑖 )2
= √∑𝑛𝑖=1(𝑥𝑖 − 𝑧𝑖 + 𝑧𝑖 − 𝑦𝑖 )2
≤ √∑𝑛𝑖=1(𝑥𝑖 − 𝑧𝑖 )2 + √∑𝑛𝑖=1(𝑧𝑖 − 𝑦𝑖 )2
[By minkowski inequality]
≤ 𝑑 (𝑥, 𝑧) + 𝑑(𝑧, 𝑦)
∴ 𝑑 𝑖𝑠 𝑚𝑒𝑡𝑟𝑖𝑐 𝑜𝑛 𝑅 𝑛
∴ (𝑅 𝑛 , 𝑑) is a metric space which is called n-dimensional
Euclidean space
Example:
Let d:R2xR2 R defined by d(x,y)= |x1-y1| + |x2-y2|
where x= (x1, x2) and y=(y1,y2)
Sol:
1) ∵ |x1 − y1 | + |x2 − y2 | ≥ 0
∴ 𝑑 (𝑥, 𝑦) ≥ 0

3
 2) d(x,y) = 0
⟺ |x1 − y1 | + |x2 − y2 | = 0
⟺ |x1 − y1 | = 0 𝑎𝑛𝑑 |x2 − y2 | = 0
⟺ x1 = y1 𝑎𝑛𝑑 x2 = y2
⟺x=y
3) d(x,y) =|x1 − y1 | + |x2 − y2 |
= |−(y1 − x1 )| + |−(y2 − x2 )|
= |−1||y1 − x1 | + |−1||y2 − x2 |
= |y1 − x1 | + |y2 − x2 |
= d(y,x)
4) let 𝑧 = (𝑧1 , 𝑧2 ) ∈ 𝑅 2
𝑑 (𝑥, 𝑦) = |𝑥1 − 𝑦1 | + |𝑥2 − 𝑦2 |
= |𝑥1 − 𝑧1 + 𝑧1 − 𝑦1 | + |𝑥2 − 𝑧2 + 𝑧2 − 𝑦2 |
≤ |𝑥1 − 𝑧1 | + |𝑧1 − 𝑦1 | + |𝑥2 − 𝑧2 | + |𝑧2 − 𝑦2 |
≤ (|𝑥1 − 𝑧1 | + |𝑥2 − 𝑧2 |) + (|𝑧1 − 𝑦1 | + |𝑧2 − 𝑦2 |)
≤ 𝑑(𝑥, 𝑧) + 𝑑(𝑧, 𝑦)
∴ 𝑑 𝑖𝑠 𝑎 𝑚𝑒𝑡𝑟𝑖𝑐 𝑜𝑛 𝑅 2
∴ (𝑅 2 , 𝑑)𝑖𝑠 𝑎 𝑚𝑒𝑡𝑟𝑖𝑐 𝑠𝑝𝑎𝑐𝑒
Example:
Let X be a non-empty set defined d: X*X  R by 𝑑 (𝑥, 𝑦) =
0 𝑖𝑓 𝑥 = 𝑦
{
1 𝑖𝑓 𝑥 ≠ 𝑦
Sol:
1) 𝑑 (𝑥, 𝑦) ≥ 0
2) 𝑑 (𝑥, 𝑦) = 0 𝑖𝑓𝑓 𝑥 = 𝑦
3) 𝑑 (𝑥, 𝑦) = 𝑑 (𝑦, 𝑥 )
1 = 1 if 𝑥 ≠ 𝑦
0 = 0 if 𝑥 = 𝑦
4) 𝑑 (𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑧) + 𝑑(𝑧, 𝑦)
4
 1) 𝑖𝑓 𝑥 ≠ 𝑦 ≠ 𝑧
1≤1+1
2) 𝑖𝑓 𝑥 ≠ 𝑦, 𝑥 = 𝑧 & 𝑦 ≠ 𝑧
1≤0+1
3) 𝑖𝑓 𝑥 ≠ 𝑦, 𝑥 ≠ 𝑧 & 𝑦 = 𝑧
1≤1+0
4) 𝑥 = 𝑦 & 𝑥 ≠ 𝑧 ,𝑦 ≠ 𝑧
0<1+1
5) 𝑥=𝑦=𝑧
0≤0+0
Exc. :
Let 𝑐 [𝑎, 𝑏] =
{𝑓: [𝑎, 𝑏] → ℝ 𝑏𝑒 𝑎 𝑐𝑜𝑛𝑡. 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛}, 𝑑𝑒𝑓𝑖𝑛𝑒 𝑑: 𝑐[𝑎, 𝑏] ∗ 𝑐 [𝑎, 𝑏] →
ℝ 𝑎𝑠 𝑑(𝑓, 𝑔) =
𝑏
∫𝑎 |𝑓 (𝑥 ) − 𝑔(𝑥 )|𝑑𝑥 𝑠ℎ𝑜𝑤 𝑡ℎ𝑎𝑡 (𝑐 [𝑎, 𝑏], 𝑑)𝑖𝑠 𝑀. 𝑠.
Exc. :
Let 𝑥 = 𝑅 2 we define 𝑑: 𝑅 2 ∗ 𝑅 2 → 𝑅 by 𝑑 (𝑥, 𝑦) = 𝑚𝑎𝑥. {|𝑥1 −
𝑦1 |, |𝑥2 − 𝑦2 |} (𝑅 2 , 𝑑)𝑎 𝑚𝑒𝑡𝑟𝑖𝑐 𝑠𝑝𝑎𝑐𝑒?

Definition:
Let (x,d) be a metric space and let 𝑥0 ∈ 𝑋, 𝑟 ∈ 𝑅, 𝑟 > 0
𝐵𝑟 (𝑥0 ) = {𝑥 ∈ 𝑋: 𝑑 (𝑥, 𝑥0 ) < 𝑟}
Is called a ball of radius r and center 𝑥0 .
(Neighborhood of 𝑥0 with radius r)
𝐷𝑟 (𝑥0 ) = {𝑥 ∈ 𝑋: 𝑑 (𝑥, 𝑥0 ) ≤ 𝑟} is called disk with radius r and
center 𝑥0 .

5
Example:
Let (R,d) be a metric space where 𝑑 (𝑥, 𝑦) = |𝑥 − 𝑦|, ∀ 𝑥, 𝑦 ∈ 𝑅
𝐵𝑟 (𝑥0 ) = {𝑥 ∈ 𝑅: 𝑑 (𝑥0 , 𝑥 ) < 𝑟}
= {𝑥 ∈ 𝑅: |𝑥 − 𝑥0 | < 𝑟}
= {𝑥 ∈ 𝑅: 𝑥0 − 𝑟 < 𝑥 < 𝑥0 + 𝑟}
= (𝑥0 − 𝑟, 𝑥0 + 𝑟)
𝑥0 − 𝑟 𝑥0 𝑥0 + 𝑟

𝐷𝑟 (𝑥0 ) = {𝑥 ∈ 𝑅: 𝑑(𝑥, 𝑥0 ) ≤ 𝑟}
= {𝑥 ∈ 𝑅: |𝑥 − 𝑥0 | ≤ 𝑟}
= {𝑥 ∈ 𝑅: 𝑥0 − 𝑟 ≤ 𝑥 ≤ 𝑥0 + 𝑟}
= [𝑥0 − 𝑟, 𝑥0 + 𝑟]
𝑥0 − 𝑟 𝑥0 𝑥0 + 𝑟

Example:
Let (𝑅 2 , 𝑑) be a metric space where 𝑑: 𝑅 2 ∗ 𝑅 2 → 𝑅 s.t. d is a
usual distance
𝐵𝑟 (𝑥0 ) = {𝑥 ∈ 𝑅 2 : 𝑑 (𝑥, 𝑥0 ) < 𝑟}
= {𝑥 ∈ 𝑅 2 : √(𝑥 − 𝑥0 )2 + (𝑦 − 𝑦0 )2 < 𝑟}
= {𝑥 ∈ 𝑅 2 : (𝑥 − 𝑥0 )2 + (𝑦 − 𝑦0 )2 < 𝑟 2 }
𝐷𝑟 (𝑥0 ) = {𝑥 ∈ 𝑅 2 : 𝑑(𝑥, 𝑥0 ) ≤ 𝑟}
= {𝑥 ∈ 𝑅 2 : √(𝑥 − 𝑥0 )2 + (𝑦 − 𝑦0 )2 ≤ 𝑟
= {𝑥 ∈ 𝑅 2 : (𝑥 − 𝑥0 )2 + (𝑦 − 𝑦0 )2 ≤ 𝑟
Example:

6
Let (𝑅 𝑛 , 𝑑) be a metric space where 𝑑: 𝑅 𝑛 ∗ 𝑅 𝑛 → 𝑅 s.t. d is a
usual distance on 𝑅 𝑛
𝐵𝑟 (𝑥0 ) = {𝑥 ∈ 𝑅 𝑛 : 𝑑(𝑥, 𝑥0 ) < 𝑟}
= {𝑥 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 ) ∈
𝑅 𝑛 : √(𝑥1 − 𝑥02 )2 + (𝑥2 − 𝑥02 )2 + ⋯ + (𝑥𝑛 − 𝑥0𝑛 )2 < 𝑟
= {(𝑥1 , 𝑥2 , … , 𝑥𝑛 ) ∈ 𝑅 𝑛 : (𝑥1 − 𝑥01 )2 + ⋯ + (𝑥𝑛 − 𝑥0𝑛 )2 < 𝑟 2 }
Where 𝑥0 = (𝑥01 , 𝑥02 , … , 𝑥0𝑛 )
𝐷𝑟 (𝑥0 ) = (𝐸𝑥𝑐. )
Define:
Let (x,d) be a metric space and 𝐴 ⊆ 𝑋, an element 𝑃 ∈ 𝐴 is called
interier point if ∃𝐵𝑟 (𝑝)𝑠. 𝑡. 𝐵𝑟 (𝑝) ⊆ 𝐴, and all interier points of A
denoted by 𝐴0
Ex. : let (𝑅, 𝑑) be a metric space where 𝐴 = (0,1), 𝐵 = [−1,1], 𝑐 =
𝑧
Find 𝐴0 , 𝐵0 , 𝐶 0
Sol: 𝐴0 = (0,1), 𝐵0 = (−1,1)𝑎𝑛𝑑 𝐶 0 = 𝜙
𝐴0 ∀𝑥 ∈ 𝐴 → (𝑥−∈, 𝑥+∈) ⊆ 𝐴
𝐵0 ∀ 𝑥 ∈ 𝐵 → ∃∈> 0 𝑠. 𝑡. (𝑥−∈, 𝑥+∈) ⊆ 𝐵
1 ∈ 𝐵, ∄ ∈> 0 𝑠. 𝑡. (1−∈, 1+∈) ⊆ 𝐵
−1 ∈ 𝐵, ∄ ∈> 0 𝑠. 𝑡. (−1−∈, −1+∈) ⊆ 𝐵
𝐶 0 , ∀ 𝑥 ∈ 𝐶 → ∃∈> 0 𝑠. 𝑡. (𝑥−∈, 𝑥+∈) ⊈ 𝐶
Definition:
Let (x,d) be a metric space and 𝐴 ⊆ 𝑋, A is called an open set if
∀ 𝑃 ∈ 𝐴 there exists 𝑟 > 0 (𝑟 ∈ 𝑅) such that 𝐵𝑟 (𝑝) ⊆ 𝐴.

7
i.e. A is open set iff 𝐴0 = 𝐴.
Ex. : let (R,d) be a metric space, which of the following sets is
open: 𝐴 = (0,1)𝑖𝑠 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡 𝐴0 = (0,1) = 𝐴.
Theorem 2.1:
Every ball (neighborhood) is an open set proof: 𝐵𝑟 (𝑥0 ) =
{𝑥 ∈ 𝑋: 𝑑 (𝑦, 𝑥0 ) < 𝑟}
Let 𝑦 ∈ 𝐵𝑟 (𝑥0 ) → 𝑑 (𝑦, 𝑥0 ) = 𝑟1 < 𝑟 take ∈= 𝑟 − 𝑟1 > 0
T.P. 𝐵∈ (𝑦) ⊆ 𝐵𝑟 (𝑥0 )
Let 𝑧 ∈ 𝐵∈ (𝑦) 𝑇. 𝑃. 𝑧 ∈ 𝐵𝑟 (𝑥0 )
𝑑 (𝑧, 𝑦) <∈ 𝑇. 𝑃. 𝑑(𝑧, 𝑥0 ) < 𝑟
𝑑 (𝑧, 𝑥0 ) ≤ 𝑑 (𝑧, 𝑦) + 𝑑 (𝑦, 𝑥0 )
< ∈ +𝑟1
< 𝑟 − 𝑟1 + 𝑟1
<𝑟
∴ 𝑑 (𝑧, 𝑥0 ) < 𝑟 → 𝑧 ∈ 𝐵𝑟 (𝑥0 )
∴ 𝐵∈ (𝑦) ⊆ 𝐵𝑟 (𝑥0 )
Hence every point of 𝐵𝑟 (𝑥0 )𝑖𝑠 𝑎𝑛 𝑖𝑛𝑡𝑒𝑟𝑖𝑒𝑟 𝑝𝑜𝑖𝑛𝑡.
∴ 𝐵𝑟 (𝑥0 )𝑖𝑠 𝑎𝑛 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡.
Remark: every open interval in R is an open set
Ex. : (𝑎, ∞), (−∞, 𝑎), (𝑎, 𝑐) ore open sets.
Sol: ∀ 𝑏 ≠ 𝑎, ∃𝑑 = |𝑏 − 𝑎|
s.t. (𝑏−∈, 𝑏+∈) ⊂ (𝑎, ∞)
∴ (𝑎, ∞)0 = (𝑎, ∞)

8
∴ (𝑎, ∞)𝑖𝑠 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡.
Ex. : is 𝐴 = [𝑎, 𝑏)𝑜𝑝𝑒𝑛 𝑠𝑒𝑡
1
Sol: for all 𝑥 ∈ (𝑎, 𝑏) → ∃𝑛 𝑠. 𝑡. <∈ (𝑥−∈, 𝑥+∈) ⊆ (𝑎, 𝑏)
𝑛

but for all ball (𝑎−∈, 𝑎+∈) ⊈ [𝑎, 𝑏)

∴ 𝐴0 = (𝑎, 𝑏)
∴ 𝐴0 ≠ 𝐴
∴ 𝐴 𝑖𝑠 𝑛𝑜𝑡 𝑜𝑝𝑒𝑛.
Ex. :
𝐻 = {(𝑥, 𝑦) ∈ 𝑅 2 : 𝑥 ∈ 𝑅, 𝑦 ≥ 0}.
Is H open set in 𝑅 2 ?
Sol : for all (𝑥, 𝑦) ∈ 𝐻 𝑠. 𝑡. 𝑦 > 0 → ∃𝐵𝑟 (𝑥, 𝑦)𝑠. 𝑡. 𝐵𝑟 ((𝑥, 𝑦)) ⊆ 𝐻
but if 𝑦 = 0 and 𝑥 ∈ 𝑅
→ 𝐵𝑟 ((𝑥, 𝑦)) ⊈ 𝐻
Exc.: show that 𝑘 = {(𝑥, 𝑦) ∈ 𝑅 2 : 𝑥 ∈ 𝑅, 𝑦 > 0} is open subset of
𝑅2 .
Ex.: the set of rational is not open set since any interval in Q with
𝑝
center ∈ 𝑄 doesn't contain rationales only (by the density of
𝑞
irrational)
Theorem 2.2: For any collection {𝐺𝑖 }𝑖𝐺𝐼 of open sets then
𝑈𝑖∈𝐼 𝐺𝑖 𝑖𝑠 𝑜𝑝𝑒𝑛.
Proof: let 𝑥 ∈ 𝑈𝑖∈𝐼 𝐺𝑖
→ 𝑥 ∈ 𝐺𝑘 , 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 𝑘 ∈ 𝐼
Since 𝐺𝑘 𝑖𝑠 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡

9
∴ 𝑥 𝑖𝑠 𝑎𝑛 𝑖𝑛𝑡𝑒𝑟𝑖𝑒𝑟 𝑝𝑜𝑖𝑛𝑡 𝑜𝑓 𝐺𝑘
∴ 𝑖. 𝑒. ∃𝐵𝑟 (𝑥 )𝑠. 𝑡. 𝐵𝑟 (𝑥 ) ⊆ 𝐺𝑘
∴ 𝐵𝑟 (𝑥 ) ⊆ 𝑈𝑖∈𝐼 𝐺𝑖 𝑖𝑠 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡
Theorem 2-3:
The intersection of a finite number of open set is open.
Proof: let 𝑢1 , 𝑢2 , … , 𝑢𝑛 be a set of finite number of open set.
T.P. ⋂𝑛𝑖=1 𝑢𝑖 𝑖𝑠 𝑜𝑝𝑒𝑛
Let 𝑥 ∈ ⋂𝑛𝑖=1 𝑢𝑖
⟹ 𝑥 ∈ 𝑢𝑖 , ∀ 𝑖 = 1,2, … , 𝑛
∵ 𝑢𝑖 𝑖𝑠 𝑜𝑝𝑒𝑛 , ∀ 𝑖 = 1,2, … , 𝑛
∴ ∃𝑟𝑖 > 0 𝑠. 𝑡. 𝐵𝑟𝑖 (𝑥 ) ⊆ 𝑢𝑖 , ∀𝑖 = 1,2, … , 𝑛
Take 𝑟 = min. {𝑟1 , 𝑟2 , … , 𝑟𝑛 }
𝑛

𝐵𝑟 (𝑥 ) ⊆ ⋂ 𝑢𝑖
𝑖=1
𝑛

∴ ⋂ 𝑢𝑖 𝑖𝑠 𝑜𝑝𝑒𝑛.
𝑖=1

Remark: the intersection of infinite number of open set needn't be

open, as the following example show:
Ex. : let (𝑅, 𝑑) be a metric space
−1 1
∀𝑛 ∈ 𝑁, 𝑙𝑒𝑡 𝐴𝑛 = ( , )
𝑛 𝑛

⋂ 𝐴𝑛 = {0}
𝑛

10
By Arch. Property.
1
If ∃0 ≠ 𝑥, 𝑥 > 0, ∃𝑘 ∈ 𝑁 𝑠. 𝑡. < 𝑥
𝑘
−1 1
∴𝑥∉( , )
𝑘 𝑘

∴ 𝑥 ∉ ⋂ 𝐴𝑛
𝑛

By arch. Property.
1
If 0 ≠ 𝑥, 𝑥 < 0 ⟹ −𝑥 > 0 ⟹ ∃𝑡 ∈ 𝑁 𝑠. 𝑡. < −𝑥
𝑡
−1
⟹ >𝑥
𝑡
−1 1
∴𝑥∉( , )
𝑡 𝑡

⟹ 𝑥 ∉ ⋂ 𝐴𝑛
𝑛

{0} is not open since ∀∈> 0

𝐵∈ (0) = (−∈, ∈) ⊄ {0}
Proposition 2.4: let (x,d) be a metric space and 𝐴 ⊆ 𝑋 then A is
open iff A is a union of balls.
Proof: (⟹)𝑠𝑢𝑝𝑝𝑜𝑠𝑒 𝑡ℎ𝑎𝑡 𝐴 𝑖𝑠 𝑎𝑛 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡.
⟹ ∀𝑥 ∈ 𝐴, ∃𝑟𝑥 > 0 𝑠. 𝑡. 𝐵𝑟𝑥 (𝑥 ) ⊆ 𝐴

∴ ⋃ 𝐵𝑟𝑥 (𝑥 ) ⊆ 𝐴
𝑥∈𝐴

(⟸)𝑙𝑒𝑡 𝐴 = ⋃ 𝐵𝑖 , 𝐵𝑖 𝑎𝑟𝑒 𝑏𝑎𝑙𝑙𝑠

𝑖∈𝐴

11
∵ 𝑒𝑣𝑒𝑟𝑦 𝑏𝑎𝑙𝑙 𝑖𝑠 𝑎𝑛 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡

⟹ ⋃ 𝐵𝑖 𝑖𝑠 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡. [𝑡ℎ𝑒𝑜𝑟𝑜𝑚 2.2]

𝑖∈𝐴

Def. : Two metrics d and 𝑑1 on the some set X are said to be

equivalent, if every open set in (𝑥, 𝑑) is open in (𝑥, 𝑑1 ).
Ex. : let (𝑥, 𝑑) be a metric space and P be a function on 𝑋 ∗ 𝑋,
defined by 𝑝(𝑥, 𝑦) = 𝑚𝑖𝑛. {1, 𝑑 (𝑥, 𝑦)}, ∀𝑥, 𝑦 ∈ 𝑋
Sol:
1) ∵ 𝑑 (𝑥, 𝑦) ≥ 0
∴ 𝑚𝑖𝑛. {1, 𝑑(𝑥, 𝑦)} ≥ 0
⟹ 𝑃(𝑥, 𝑦) ≥ 0
2) 𝑃(𝑥, 𝑦) = 0
⟺ 𝑚𝑖𝑛. {1, 𝑑(𝑥, 𝑦)} = 0
⟺ 𝑑 (𝑥, 𝑦) = 0
⟺𝑥=𝑦
3) 𝑃(𝑥, 𝑦) = min{1, d(x, y)}
= min{1, 𝑑 (𝑦, 𝑥 )}
= 𝑃(𝑦, 𝑥 ), ∀𝑥, 𝑦 ∈ 𝑋
4) Let 𝑥, 𝑦, 𝑧 ∈ 𝑋, If at least one of say 𝑑(𝑥, 𝑦) ≥ 1
Then 𝑃(𝑥, 𝑦) = min{1, 𝑑 (𝑥, 𝑦)} = 1
∴ 𝑃(𝑥, 𝑦) + 𝑃(𝑦, 𝑧) ≥ 1 ≥ 𝑃(𝑥, 𝑧)
Also in case 𝑑 (𝑥, 𝑦) < 1 𝑎𝑛𝑑 𝑑 (𝑦, 𝑧) < 1
𝑃(𝑥, 𝑦) = min{𝑑(𝑥, 𝑦), 1} = 𝑑 (𝑥, 𝑦)
𝑃(𝑦, 𝑧) = min{𝑑(𝑦, 𝑧), 1} = 𝑑(𝑦, 𝑧)
∴ 𝑃(𝑥, 𝑦) + 𝑃(𝑦, 𝑧) = 𝑑 (𝑥, 𝑦) + 𝑑 (𝑦, 𝑧)
≥ 𝑑 (𝑥, 𝑧)𝑏𝑦 𝑡𝑟𝑖𝑎𝑛𝑔𝑙𝑒 𝑖𝑛𝑒𝑞𝑢𝑎𝑙𝑖𝑡𝑦 ≥ 𝑃(𝑥, 𝑧)

12
∴ 𝑃(𝑥, 𝑧) ≤ 𝑃(𝑥, 𝑦) + (𝑦, 𝑧).
∴ 𝑃 𝑖𝑠 𝑎 𝑚𝑒𝑡𝑟𝑖𝑐 𝑜𝑛 𝑋 𝑎𝑛𝑑 (𝑋, 𝑃) 𝑖𝑠 𝑎 𝑚𝑒𝑡𝑟𝑖𝑐 𝑠𝑝𝑎𝑐𝑒.
Now to show that P is equivalent to d.
T.P. every open set in (X,P) is open in (x,d)
Let G be any open subset of X in (X,P)
Let 𝑥 ∈ 𝐺 ⟹ ∃𝑎𝑛 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡
{𝑦 ∈ 𝑋: 𝑃(𝑥, 𝑦) < 𝑟} ⊆ 𝐺
∴ 𝑃(𝑥, 𝑦) ≤ 𝑑 (𝑥, 𝑦), ∀ 𝑥, 𝑦 ∈ 𝑋
∴ {𝑦 ∈ 𝑋 ∷ 𝑑 (𝑥, 𝑦) < 𝑟} ⊆ {𝑦 ∈ 𝑋: 𝑃(𝑥, 𝑦) < 𝑟} ⊆ 𝐺
∴ 𝐺 𝑖𝑠 𝑜𝑝𝑒𝑛 𝑖𝑛 (𝑥, 𝑑)
Hence every open set in (X,P) is open set in (x,d)
Next, let H be an open set in (x,d)⟹ ∀ 𝑥 ∈ 𝐻, ∃ 𝑎𝑛𝑑 𝑏𝑎𝑙𝑙
𝑦 ∈ 𝑋: 𝑑(𝑥, 𝑦) < 𝑟} ⊆ 𝐻
Let 𝑟1 = min{1, 𝑟} , 𝑠𝑜 𝑟1 ≤ 𝑟, 𝑡ℎ𝑒𝑛
{𝑦 ∈ 𝑋: 𝑃(𝑥, 𝑦) < 𝑟1 } ⊆ {𝑦 ∈ 𝑋: 𝑑 (𝑥, 𝑦) < 𝑟} ⊆ 𝐻.
∴ 𝐻 𝑖𝑠 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡 𝑖𝑛 (𝑋, 𝑃)
∴ 𝐸𝑣𝑒𝑟𝑦 𝑜𝑝𝑒𝑛 𝑠𝑒𝑡 𝑖𝑛 (𝑥, 𝑑)𝑖𝑠 𝑜𝑝𝑒𝑛 𝑖𝑛 (𝑋, 𝑃)
Hence d and P are equivalent metrics.
Ex. : Let (x,d) be a metric space, and let
𝑑(𝑥,𝑦)
𝑑 𝑥 (𝑥, 𝑦) = , ∀ 𝑥, 𝑦 ∈ 𝑋
1+𝑑(𝑥,𝑦)

Show that 𝑑 ∗ is a metric on X equivalent to d.

Sol. : First to show 𝑑 ∗ is a metric on X

13
 1) ∵ 𝑑 (𝑥, 𝑦) ≥ 0 ⟹ 𝑑 ∗ (𝑥, 𝑦) ≥ 0
2) 𝑑 ∗ (𝑥, 𝑦) = 0
⟺ 𝑑 (𝑥, 𝑦) = 0
⟺ 𝑥 = 𝑦 𝑠𝑖𝑛𝑐𝑒 𝑑 𝑖𝑠 𝑚𝑒𝑡𝑟𝑖𝑐 𝑜𝑛 𝑋
𝑑(𝑥,𝑦) 𝑑(𝑦,𝑥)
3) 𝑑 ∗ (𝑥, 𝑦) = = = 𝑑 ∗ (𝑦, 𝑦)
1+𝑑(𝑥,𝑦) 1+𝑑(𝑦,𝑥)

4) For all 𝑥, 𝑦, 𝑧 ∈ 𝑋, 𝑤𝑒 ℎ𝑎𝑣𝑒:

𝑑 (𝑥, 𝑦) 𝑑 (𝑦, 𝑧)
𝑑 ∗ (𝑥, 𝑦) + 𝑑 ∗ (𝑦, 𝑧) = +
1 + 𝑑(𝑥, 𝑦) 1 + 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)
≥
1 + 𝑑(𝑥, 𝑦) + 𝑑 (𝑦, 𝑧)
1
≥1−
1 + 𝑑 (𝑥, 𝑦) + 𝑑(𝑦, 𝑧)
∵ 𝑑 𝑖𝑠 𝑎 𝑚𝑒𝑡𝑟𝑖𝑐 𝑜𝑛 𝑋
∴ 𝑑 (𝑥, 𝑦) + 𝑑(𝑦, 𝑧) ≥ 𝑑 (𝑥, 𝑧)
⟹ 1 + 𝑑 (𝑥, 𝑦) + 𝑑 (𝑦, 𝑧) ≥ 1 + 𝑑 (𝑥, 𝑧)
1 1
⟹ ≤
1 + 𝑑 (𝑥, 𝑦) + 𝑑 (𝑦, 𝑧) 1 + 𝑑 (𝑥, 𝑧)
1
⟹1−
1 + 𝑑(𝑥, 𝑦) + 𝑑 (𝑦, 𝑧)
1
≥1−
1 + 𝑑 (𝑥, 𝑧)
∗(
≥ 𝑑 𝑥, 𝑧)
∴ 𝑑 ∗ (𝑥, 𝑧) ≤ 𝑑∗ (𝑥, 𝑦) + 𝑑 ∗ (𝑦, 𝑧)
Exc. : now, to show d and 𝑑 ∗ are equivalent.
Let 𝐵𝑅 (𝑥 ), 𝑟 > 0, 𝑏𝑒 𝑎𝑛𝑦 𝑑 − 𝑜𝑝𝑒𝑛 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 𝐵𝑃 (𝑥 )
𝑟
Be 𝑑 ∗ − 𝑜𝑝𝑒𝑛 ball where 𝑃 =
1+𝑟

14
T.P. 𝐵𝑃 (𝑥 ) ⊆ 𝐵 (𝑥, 𝑟)
Let 𝑦 ∈ 𝐵𝑃 (𝑥 ) ⟹ 𝑑 ∗ (𝑥, 𝑦) < 𝑃
𝑑(𝑥,𝑦) 𝑟
⟹ ,
1+𝑑(𝑥,𝑦) 1+𝑟
⟹ 𝑑 (𝑥, 𝑦) + 𝑟𝑑 (𝑥, 𝑦) < 𝑟 + 𝑟𝑑 (𝑥, 𝑦)
⟹ 𝑑 (𝑥, 𝑦) < 𝑟
⟹ 𝑦 ∈ 𝐵𝑟 (𝑥 )
∴ 𝐵𝑃 (𝑥 ) ⊆ 𝐵𝑟 (𝑥 )
Next, let 𝐵𝑃 (𝑥 ), 𝑃 > 0 𝑏𝑒 𝑑 ∗ − 𝑜𝑝𝑒𝑛
∵ 𝑑 ∗ (𝑥, 𝑦) ≤ 1, ∀𝑥, 𝑦 ∈ 𝑋
We take 0 < 𝑃 < 1
𝑃
Let 𝐵𝑟 (𝑥 )𝑏𝑒 𝑑 − 𝑜𝑝𝑒𝑛, 𝑟 =
1−𝑃

T.P. 𝐵𝑟 (𝑥 ) ⊆ 𝐵𝑃 (𝑥 )
Let 𝑦 ∈ 𝐵𝑟 (𝑥 ) ⟹ 𝑑 (𝑥, 𝑦) < 𝑟
𝑑∗ (𝑥, 𝑦) 𝑃 ∗(
𝑑 (𝑥, 𝑦)
⟹ < [𝑑 𝑥, 𝑦 ) = ⟹ 𝑑(𝑥, 𝑦)
1 − 𝑑∗ (𝑥, 𝑦) 1 − 𝑃 1 + 𝑑(𝑥, 𝑦)
𝑑 ∗ (𝑥, 𝑦)
=
1 − 𝑑 ∗ (𝑥, 𝑦)
⟹ 𝑑 ∗ (𝑥, 𝑦) − 𝑃𝑑∗ (𝑥, 𝑦) < 𝑃 − 𝑃𝑑 ∗ (𝑥, 𝑦)
⟹ 𝑑 ∗ (𝑥, 𝑦) < 𝑃
⟹ 𝑦 ∈ 𝐵𝑃 (𝑥 )
∴ 𝐵𝑟 (𝑥 ) ⊆ 𝐵𝑃 (𝑥 )
∴ 𝐸𝑣𝑒𝑟𝑦 𝑑 − 𝑜𝑝𝑒𝑛 𝑖𝑠 𝑎 𝑑∗ − 𝑜𝑝𝑒𝑛 𝑎𝑛𝑑 𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑒𝑙𝑦.
Definition: Let (𝑋. 𝑑) be a metric space, a point 𝑃 ∈ 𝑋 is said to
be a limit point of a set 𝐴 ⊂ 𝐶, if every ball (nbd.) of 𝑃 contains a
point of 𝑃.

15
i.e. 𝑃 is a limit point of 𝐴 if ∀𝐵𝑟 (𝑃) then [𝐵𝑟 (𝑃) − {𝑃}] ∩ 𝐴 ≠ 𝜙
Definition: The set of all points of a set 𝐴 ⊂ 𝑋 is known as the
derived set and is denoted by 𝐴′ .
Example: let (ℝ, 𝑑) be a metric space and 𝐴 = (0,1)
Solution: 0 1
{𝑃}
∀𝑃 ∈ (0,1) ⟹ ∀𝐵𝜖 (𝑃); 𝜖 > 0
[𝐵𝜖 (𝑃) − {𝑃}] ∩ (0,1) ≠ 𝜙
∀𝐵𝑟 (𝑃), 𝑟 > 0
[𝐵𝑟 (𝑃) − {𝑃}] ∩ (0,1) ≠ 𝜙
∀𝜖 > 0, ∀𝐵𝜖 (0)
0 1
[𝐵𝜖 (0) − {0}] ∩ (0,1) ≠ 𝜙
[𝐵𝜖 (1) − {1}] ∩ (0,1) ≠ 𝜙, 𝐵𝜖 (0)
0 1
If 𝑃 < 0 ⟹ ∃𝜖 > 0 𝑠. 𝑡. 𝐵𝜖 (𝑃)
[𝐵𝜖 (𝑃) − {𝑃}] ∩ (0,1) = 𝜙 𝑃

If 𝑃 > 1 ⟹ ∃𝜖 > 0 𝑠. 𝑡. 𝐵𝜖 (𝑃)

0 1
[𝐵𝜖 (𝑃) − {𝑃}] ∩ (0,1) = 𝜙
𝑃
∴ 𝐴′ = [0,1]
Example: let (ℝ, 𝑑) be a metric space and 𝐴 = [2,5]
Solution:
∀𝑃 ∈ [2,5] ⟹ ∀𝐵𝜖 (𝑃) 2 {𝑃} 5

[𝐵𝜖 (𝑃) − {𝑃}] ∩ [2,5] ≠ 𝜙

𝑟 > 0, ∀𝐵𝑟 (𝑃)
[𝐵𝑟 (𝑃) − {𝑃}] ∩ [2,5] ≠ 𝜙

16
∀𝑃 < 2 ⟹ ∃𝜖 > 0 𝑠. 𝑡. 𝐵𝜖 (𝑃)
[𝐵𝜖 (𝑃) − {𝑃}] ∩ [2,5] = 𝜙
∀𝑃 > 5 ⟹ ∃𝜖 > 0 𝑠. 𝑡. 𝐵𝜖 (𝑃) {𝑃} 2 5 {𝑃}

[𝐵𝜖 (𝑃) − {𝑃}] ∩ (0,1) = 𝜙

∴ 𝐴′ = [2,5]
Example: let (ℝ, 𝑑) be a metric space and 𝐴 = ℤ
Solution:
∀𝑃 ∈ ℤ ⟹ ∃𝜖 > 0 𝑠. 𝑡.
−1 0 1 2
[𝐵𝜖 (𝑃) − {𝑃}] ∩ ℤ = 𝜙
∀𝑃 ∈ ℝ − ℤ ⟹ ∃𝜖 > 0 𝑠. 𝑡.
[𝐵𝜖 (𝑃) − {𝑃}] ∩ ℤ = 𝜙 −1 0 1 2

∴ 𝐸′ = 𝜙
Example: let (ℝ, 𝑑) be a metric space and 𝐴 = ℚ
Solution:
∀𝑃 ∈ ℚ, ∃𝜖 > 0 𝑠. 𝑡.
[𝐵𝜖 (𝑃) − {𝑃}] ∩ ℚ ≠ 𝜙
∀𝑃 ∈ ℚ , ∀𝐵𝑟 (𝑃) 𝑠. 𝑡.
[𝐵𝑟 (𝑃) − {𝑃}] ∩ ℚ ≠ 𝜙
∀𝑃 ∈ ℝ − ℚ , ∀𝜖 > 0
[𝐵𝜖 (𝑃) − {𝑃}] ∩ ℚ ≠ 𝜙
∀𝑃 ∈ ℝ − ℚ , ∀𝐵𝑟 (𝑃)
[𝐵𝑟 (𝑃) − {𝑃}] ∩ ℚ ≠ 𝜙
∴ 𝐴′ = ℝ

17
Exercise: let (ℝ, 𝑑) be a metric space find the derived set of
1
𝐴 = {1,2, … ,10}, 𝐵 = ℕ, 𝐶 = { : 𝑛 ∈ ℕ} , 𝐷 = [−4,2)
𝑛

Definition: A subset 𝐴 of a metric space (𝑋, 𝑑) is said to be closed

if 𝐴 contains all of its limit points.
i.e. 𝐴 ⊆ 𝑋 is closed iff 𝐴′ ⊆ 𝐴.
Example: let (ℝ, 𝑑) be a metric space and 𝐴 = (0, ,1)
∵ 𝐴′ = [0,1] and 𝐴′ ⊈ 𝐴
∴ 𝐴 = (0,1) is not closed.
Example: let (ℝ, 𝑑) be a metric space and 𝐴 = [2,7]
∵ 𝐴′ = [2,7] and 𝐴′ ⊆ 𝐴
∴ 𝐴 = [2,7] is closed.
Example: let (ℝ2 , 𝑑) be a metric space and
𝐻 = {(𝑥, 𝑦) ∈ ℝ2 : 𝑦 ≥ 0}
Solution:
𝐻 ′ = {(𝑥, 𝑦) ∈ ℝ2 : 𝑦 ≥ 0} (check)
∴ 𝐻′ ⊆ 𝐻
∴ 𝐻 is closed
Example: let (ℝ2 , 𝑑) be a metric space and
𝐾 = {(𝑥, 𝑦) ∈ ℝ2 : 𝑦 > 0}
𝐾 ′ = {(𝑥, 𝑦) ∈ ℝ2 : 𝑦 ≥ 0}
∴ 𝐾′ ⊈ 𝐾
∴ 𝐾 is not closed
Theorem 2.5: In a metric space a set 𝐸 is closed if and only if its
complement is open.

18
Proof: Suppose that 𝐸 is closed set
T.P. 𝐸 𝑐 is open
Let 𝑥 ∈ 𝐸 𝑐 ⟹ 𝑥 ∉ 𝐸
∵ 𝐸 is closed
∴ 𝑥 is not a limit point of 𝐸
⟹ ∃𝑟 > 0, 𝑠. 𝑡. 𝐵𝑟 (𝑥 ) ∩ 𝐸 = 𝜙
⟹ 𝑥 ∈ 𝐵𝑟 (𝑥 ) ⊆ 𝐸 𝑐
⟹ 𝐸 𝑐 is open
Suppose that 𝐸 𝑐 is open
T.P. 𝐸 is closed
Let 𝑥 be a limit point of 𝐸
∴ ∀𝐵𝑟 (𝑥) s.t. 𝐵𝑟 (𝑥 ) ∩ 𝐸 ≠ 𝜙
∴ 𝐵𝑟 (𝑥) ⊈ 𝐸 𝑐
∵ 𝐸 𝑐 is open
∴ 𝑥 ∉ 𝐸𝑐
∴𝑥∈𝐸
∴ 𝐸 is closed
Theorem 2.6: for any collection {𝐸𝑖 }𝑖∈𝐼 of closed sets, then ⋂𝑖∈𝐼 𝐸𝑖
is closed.
Proof: let {𝐸𝑖 }𝑖∈𝐼 is closed
T.P. ⋂𝑖∈𝐼 𝐸𝑖 is closed
𝑐

(⋂ 𝐸𝑖 ) = ⋃ 𝐸𝑖𝑐
𝑖∈𝐼 𝑖∈𝐼

19
∵ 𝐸𝑖𝑐 is open set, ∀𝑖 ∈ 𝐼 [Theorem 2.5]
By theorem 2.2 we get
⋃𝑖∈𝐼 𝐸𝑖𝑐 is open
Hence ⋂𝑖∈𝐼 𝐸𝑖 is closed [Theorem 2.5]
Theorem 2.7: for any finite collection 𝐸1 , 𝐸2 , … , 𝐸𝑛 of closed sets
then ⋃𝑛𝑖=1 𝐸𝑖 is closed.
Proof: suppose that 𝐸1 , 𝐸2 , … , 𝐸𝑛 are closed sets T.P. ⋃𝑛𝑖=1 𝐸𝑖 is
closed.
Since (⋃𝑛𝑖∈𝐼 𝐸𝑖 )𝑐 = ⋂𝑛𝑖∈𝐼 𝐸𝑖𝑐
𝐸𝑖𝑐 is open set 𝑖 = 1, … , 𝑛 [Theorem 2.5]

By theorem 2.3 we get

⋂𝑛𝑖∈𝐼 𝐸𝑖𝑐 is open
Hence ⋃𝑛𝑖∈𝐼 𝐸𝑖 is closed.

Definition: In a metric space (𝑋, 𝑑) the closure of a set 𝐸 is

denoted by 𝐸̅ or 𝑐𝑙(𝐸) which is defined by
𝐸̅ = 𝐸 ∪ 𝐸 ′
Example: let (ℝ, 𝑑) be a metric space, let 𝐸 = (0,1), 𝐴 = ℤ.
Then 𝐸̅ = 𝐸 ∪ 𝐸 ′ = (0,1) ∪ [0,1] = [0,1]
𝐴̅ = 𝐴 ∪ 𝐴′ = ℤ ∪ 𝜙 = ℤ

Theorem 2.8: if (𝑋, 𝑑) is a metric space and 𝐸 ⊂ 𝑋 then:

a) 𝐸̅ is closed,
b) 𝐸 = 𝐸̅ iff 𝐸 is closed,
c) 𝐸̅ ⊆ 𝐹 for every closed set 𝐹 ⊆ 𝑋 such that 𝐸 ⊂ 𝐹.
Proof:

20
a) If 𝑃 ∈ 𝑋 and 𝑃 ∉ 𝐸̅
⟹ 𝑃 ∉ 𝐸 and 𝑃 ∉ 𝐸 ′
∴ (𝐸̅ )𝑐 is open
∴ 𝐸̅ is closed [Theorem 2.5]
b) Suppose that 𝐸 = 𝐸̅
By (a) we get 𝐸 is closed
Suppose that 𝐸 is closed
⟹ 𝐸′ ⊂ 𝐸
⟹ 𝐸 = 𝐸 ∪ 𝐸′
⟹ 𝐸 = 𝐸̅
c) Suppose that 𝐸 ⊆ 𝐹 and 𝐹 is closed
T.P. 𝐸̅ ⊆ 𝐹
∵ 𝐹 is closed
∴ 𝐹 ′ ⊆ 𝐹 and 𝐸 ⊆ 𝐹
⟹ 𝐸′ ⊆ 𝐹
⟹ 𝐸 ∪ 𝐸′ ⊆ 𝐹
⟹ 𝐸̅ ⊆ 𝐹

21