Metric spaces

Let X be a set.
How to define distance between any two points
Distance must satisfy certain conditions.
Let U be a non-empty set.
Let U be a non-empty set.
Let d : U × U → R.
Let U be a non-empty set.
Let d : U × U → R.
We say that d is a metric (distance) on U if:
(i) d(x, y ) ≥ 0 for all x, y ∈ U.
Let U be a non-empty set.
Let d : U × U → R.
We say that d is a metric (distance) on U if:
(i) d(x, y ) ≥ 0 for all x, y ∈ U.
(ii) d(x, y ) = d(y , x) for all x, y
Let U be a non-empty set.
Let d : U × U → R.
We say that d is a metric (distance) on U if:
(i) d(x, y ) ≥ 0 for all x, y ∈ U.
(ii) d(x, y ) = d(y , x) for all x, y
(iii) d(x, x) = 0 for all x
Let U be a non-empty set.
Let d : U × U → R.
We say that d is a metric (distance) on U if:
(i) d(x, y ) ≥ 0 for all x, y ∈ U.
(ii) d(x, y ) = d(y , x) for all x, y
(iii) d(x, x) = 0 for all x
(iv) d(x, y ) = 0 implies x = y
Let U be a non-empty set.
Let d : U × U → R.
We say that d is a metric (distance) on U if:
(i) d(x, y ) ≥ 0 for all x, y ∈ U.
(ii) d(x, y ) = d(y , x) for all x, y
(iii) d(x, x) = 0 for all x
(iv) d(x, y ) = 0 implies x = y .
(v) d(x, z) ≤ d(x, a) + d(a, z) for all x, z, a ∈ X .
We say that (U, d) is a metric space.
We say that (U, d) is a metric space.
Example 1
Define d : R × R → R by
d(a, b) = |a − b|.
(R, d) is a metric space.
Example 2
Example 2
Another metric on R.
Define ρ(x, y ) = |e x − e y |.
(R, ρ) is a metric space.
Example 3:
Example 3:
Let X be any set.
Define d(x, y ) = 1 if x and y are not equal.
(X , d) is a metric.
d is called discrete metric.
To any set, we can assign a metric.
Example 4:
Example 4:
Let X = R2 .
Take x = (x1 , x2 ) and y = (y1 , y2 ) in R2 .
p
Define d(x, y ) = x1 − y1 )2 + (x2 − y2 )2 .
We shall show that d is a metric on R2 .
d(x, y ) = 0 iff x = y is trivial.
d(x, y ) = d(y , x)
Triangle inequality
Triangle inequality
Follows from Cauchy Schwarz inequality.
x • y ≤ kxkky k.
p
kxk = x12 + x22 .
kx + y k2 = (x + y ) • (x + y ) = kxk2 + 2x • y + ky k2 .
≤ (kxk + ky k)2 .
So, kx + y k ≤ kxk + ky k.
d(x, y ) = kx − y k.
d(x, y ) = kx − y k.
kx − y k = kx − z + z − y k ≤ kx − zk + kz − y k.
This is:
d(x, y ) ≤ d(x, z) + d(z, y ).
Example 5:
Example 5:
Let X = R2
Define kxk1 := max(|x1 |, |x2 |).
Define d(x, y ) := kx − y k1 .
Then (R2 , d) is a metric space.
Example 6
Let X = C [a, b].
C [a, b] contains all continuous real-valued functions
defined on [a, b].
Example 6
Let X = C [a, b].
C [a, b] contains all continuous real-valued functions
defined on [a, b].
Define for f ∈ C [a, b],
qR
b 2
kf k = a f .

Define d2 (f , g ) := kf − g k.
d2 is a metric on X .
Consider C [0, 1].
Consider C [0, 1].
Take f (x) = x and g (x) = x 2
qR
1
d2 (f , g ) = 0 (x − x 2 )2 .
q
1
This is 30 .
Example 7:
Let f ∈ C [a, b].
Then, define kf k∞ := max{|f (t)| : t ∈ [a, b]}.
Define d(f , g ) := kf − g k∞ .
Now d is a metric on C [a, b].
Let X be the set of all bounded sequences.
Let X be the set of all bounded sequences.
X is denoted by l∞ .
Let X be the set of all bounded sequences.
X is denoted by l∞ .
Let a = (an ) be a bounded sequence.
Define kak∞ := sup(|a1 |, . . . , |an |, . . . ).
Define d(a, b) = ka − bk∞ .
d is a metric on l∞ .
Suppose dX is a metric on X and dY is a metric on
Y.
Suppose dX is a metric on X and dY is a metric on
Y.
Define
d : (X × Y ) × (X × Y ) → R

by d((x1 , y1 ), (x2 , y2 )) = dX (x1 , x2 ) + dY (y1 , y2 ).

Then d is a metric on X × Y .
Open sets in metric spaces
Open sets in metric spaces
Basic open sets are open balls.
Open balls
Open sets in metric spaces
Basic open sets are open balls.
Open balls
Let a ∈ X and r > 0. Define
Br (a) = {x : d(x, a) < r }.
Open sets in metric spaces
Basic open sets are open balls.
Open balls
Let a ∈ X and r > 0. Define
Br (a) = {x : d(x, a) < r }.
Br (a) is called a open ball with center a and radius
r.
In R, open sets are open intervals. They are
(a − r , a + r ).
Open sets
Open sets
In a metric space X , a set G ⊆ X is called open if
a ∈ G =⇒ Br (a) ⊆ G for some r > 0.
Open sets
In a metric space X , a set G ⊆ X is called open if
a ∈ G =⇒ Br (a) ⊆ G for some r > 0.
Example:
G := {x : x > 0} is open in R.
Let a > 0. Then we can find  > 0 such that
a −  > 0.
(a − , a + ) ⊆ G .
So, G is a open set in R.
Any open ball is an open set
Any open ball is an open set
Consider Br (x0 ).
Any open ball is an open set
Consider Br (x0 ).
Let y ∈ Br (x0 )
Any open ball is an open set
Consider Br (x0 ).
Let y ∈ Br (x0 )
Define r1 = r − d(y , x0 ). Then r1 > 0.
Any open ball is an open set
Consider Br (x0 ).
Let y ∈ Br (x0 )
Define r1 = r − d(y , x0 ). Then r1 > 0.
Claim: Br1 (y ) ⊆ Br (x0 ).
Any open ball is an open set
Consider Br (x0 ).
Let y ∈ Br (x0 )
Define r1 = r − d(y , x0 ). Then r1 > 0.
Claim: Br1 (y ) ⊆ Br (x0 ).
Let p ∈ Br1 (y ).
Any open ball is an open set
Consider Br (x0 ).
Let y ∈ Br (x0 )
Define r1 = r − d(y , x0 ). Then r1 > 0.
Claim: Br1 (y ) ⊆ Br (x0 ).
Let p ∈ Br1 (y ).Then
d(p, x0 ) ≤ d(p, y ) + d(y , x0 ) < r1 + d(y , x0 ) = r .
Any open ball is an open set
Consider Br (x0 ).
Let y ∈ Br (x0 )
Define r1 = r − d(y , x0 ). Then r1 > 0.
Claim: Br1 (y ) ⊆ Br (x0 ).
Let p ∈ Br1 (y ).Then
d(p, x0 ) ≤ d(p, y ) + d(y , x0 ) < r1 + d(y , x0 ) = r .
d(p, x0 ) < r .
Any open ball is an open set
Consider Br (x0 ).
Let y ∈ Br (x0 )
Define r1 = r − d(y , x0 ). Then r1 > 0.
Claim: Br1 (y ) ⊆ Br (x0 ).
Let p ∈ Br1 (y ).Then
d(p, x0 ) ≤ d(p, y ) + d(y , x0 ) < r1 + d(y , x0 ) = r .
d(p, x0 ) < r .Hence, p ∈ Br (x0 ).
Theorem

Let Gα be a open set for each α ∈ A, where A is a

non-empty set. Then, ∪α Gα is open.
Theorem

Let Gα be a open set for each α ∈ A, where A is a

non-empty set. Then, ∪α Gα is open.

Proof
Let x ∈ ∪α Gα .
Theorem

Let Gα be a open set for each α ∈ A, where A is a

non-empty set. Then, ∪α Gα is open.

Proof
Let x ∈ ∪α Gα .
So, x ∈ Gβ for some β ∈ A.
Theorem

Let Gα be a open set for each α ∈ A, where A is a

non-empty set. Then, ∪α Gα is open.

Proof
Let x ∈ ∪α Gα .
So, x ∈ Gβ for some β ∈ A.
So, Br (x) ⊆ Gβ for some r > 0.
Proof is complete.
Closed sets
Let (X , d) be a metric space. We say that F ⊆ X is
closed if F c is open. That is, X r F is open.
Closed sets
Let (X , d) be a metric space. We say that F ⊆ X is
closed if F c is open. That is, X r F is open.
Example
N is a closed set in R.
Closed sets
Let (X , d) be a metric space. We say that F ⊆ X is
closed if F c is open. That is, X r F is open.
Example
N is a closed set in R.
Nc = (x : x < 0) ∪ (0, 1) ∪ (1, 2) ∪ · · ·
Closed sets
Let (X , d) be a metric space. We say that F ⊆ X is
closed if F c is open. That is, X r F is open.
Example
N is a closed set in R.
Nc = (x : x < 0) ∪ (0, 1) ∪ (1, 2) ∪ · · ·
So, Nc is open.
Closed sets
Let (X , d) be a metric space. We say that F ⊆ X is
closed if F c is open. That is, X r F is open.
Example
N is a closed set in R.
Nc = (x : x < 0) ∪ (0, 1) ∪ (1, 2) ∪ · · ·
So, Nc is open.
So, N is a closed subset of R.
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
Proof:
Let a ∈ G . Then, a ∈ Gi for all i = 1 : m.
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
Proof:
Let a ∈ G . Then, a ∈ Gi for all i = 1 : m.
There exist r1 , . . . , rm > 0 such that
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
Proof:
Let a ∈ G . Then, a ∈ Gi for all i = 1 : m.
There exist r1 , . . . , rm > 0 such that
Bri (a) ⊆ Gi for all i.
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
Proof:
Let a ∈ G . Then, a ∈ Gi for all i = 1 : m.
There exist r1 , . . . , rm > 0 such that
Bri (a) ⊆ Gi for all i.
Let r := min(r1 , . . . , rm ).
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
Proof:
Let a ∈ G . Then, a ∈ Gi for all i = 1 : m.
There exist r1 , . . . , rm > 0 such that
Bri (a) ⊆ Gi for all i.
Let r := min(r1 , . . . , rm ).
Then, Br (a) ⊆ Gi for all i.
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
Proof:
Let a ∈ G . Then, a ∈ Gi for all i = 1 : m.
There exist r1 , . . . , rm > 0 such that
Bri (a) ⊆ Gi for all i.
Let r := min(r1 , . . . , rm ).
Then, Br (a) ⊆ Gi for all i.
So, Br (a) ⊆ G .
In a metric space, if G1 , . . . , Gm are open sets, then
G := G1 ∩ · · · ∩ Gm is open.
Proof:
Let a ∈ G . Then, a ∈ Gi for all i = 1 : m.
There exist r1 , . . . , rm > 0 such that
Bri (a) ⊆ Gi for all i.
Let r := min(r1 , . . . , rm ).
Then, Br (a) ⊆ Gi for all i.
So, Br (a) ⊆ G .
Let X be a metric space. Then, X itself is open.
Let X be a metric space. Then, X itself is open.
Also, X is closed.
Let X be a metric space. Then, X itself is open.
Also, X is closed.
Show that F := [0, 1] is a closed set in R.
Let X be a metric space. Then, X itself is open.
Also, X is closed.
Show that F := [0, 1] is a closed set in R.
F c = {x : x < 0} ∪ {x : x > 1}.
Let X be a metric space. Then, X itself is open.
Also, X is closed.
Show that F := [0, 1] is a closed set in R.
F c = {x : x < 0} ∪ {x : x > 1}.
So, F c is closed.
Sequences in metric spaces
Sequences in metric spaces
A sequence is a function from f : N → X .
{f (n) : n ∈ N} is a sequence in X .
Sequences in metric spaces
A sequence is a function from f : N → X .
{f (n) : n ∈ N} is a sequence in X .
Usually denoted by (fn ).
Sequences in metric spaces
A sequence is a function from f : N → X .
{f (n) : n ∈ N} is a sequence in X .
Usually denoted by (fn ).
We say that a sequence (xn ) in X converges to a if
for each  > 0, there exists n0 ∈ N such that
n ≥ n0 =⇒ d(xn , a) < .
Sequences in metric spaces
A sequence is a function from f : N → X .
{f (n) : n ∈ N} is a sequence in X .
Usually denoted by (fn ).
We say that a sequence (xn ) in X converges to a if
for each  > 0, there exists n0 ∈ N such that
n ≥ n0 =⇒ d(xn , a) < .
Example
Example
(1/n) is a sequence in R.
This sequence converges to 0.
Example
(1/n) is a sequence in R.
This sequence converges to 0.
This follows from Archimedian principle:
Example
(1/n) is a sequence in R.
This sequence converges to 0.
This follows from Archimedian principle:
If x ∈ R then there exists n0 ∈ N such that x < n0 .
Example
(1/n) is a sequence in R.
This sequence converges to 0.
This follows from Archimedian principle:
If x ∈ R then there exists n0 ∈ N such that x < n0 .
1
Let  > 0. Then, there exists n0 such that  < n0 .
1
So, n0 < .
Example
(1/n) is a sequence in R.
This sequence converges to 0.
This follows from Archimedian principle:
If x ∈ R then there exists n0 ∈ N such that x < n0 .
1
Let  > 0. Then, there exists n0 such that  < n0 .
1
So, n0 < .
1 1 1
If n > n0 , then n < n0 . So, n <  if n ≥ n0 .
In R2 consider the sequence {( n1 , 1)}.
In R2 consider the sequence {( n1 , 1)}.
This sequence converges to (0, 1) in the standard
metric.
In R2 consider the sequence {( n1 , 1)}.
This sequence converges to (0, 1) in the standard
metric.
Consider ( n1 ) in R where d is the discrete metric.
In R2 consider the sequence {( n1 , 1)}.
This sequence converges to (0, 1) in the standard
metric.
Consider ( n1 ) in R where d is the discrete metric.
Then, (1/n) is not convergent.
In R2 consider the sequence {( n1 , 1)}.
This sequence converges to (0, 1) in the standard
metric.
Consider ( n1 ) in R where d is the discrete metric.
Then, (1/n) is not convergent.
Suppose this sequence converges to a.
In R2 consider the sequence {( n1 , 1)}.
This sequence converges to (0, 1) in the standard
metric.
Consider ( n1 ) in R where d is the discrete metric.
Then, (1/n) is not convergent.
Suppose this sequence converges to a.
Let  = 1/2.
In R2 consider the sequence {( n1 , 1)}.
This sequence converges to (0, 1) in the standard
metric.
Consider ( n1 ) in R where d is the discrete metric.
Then, (1/n) is not convergent.
Suppose this sequence converges to a.
Let  = 1/2.
There is n0 such that
n ≥ n0 =⇒ d( n1 , a) < 21 .
There is n0 such that
n ≥ n0 =⇒ d( n1 , a) < 21 .
So, d( n10 , a) = 0.
There is n0 such that
n ≥ n0 =⇒ d( n1 , a) < 21 .
So, d( n10 , a) = 0.
1
n0 = a.
1
Similarly, n0 +1 = a.
There is n0 such that
n ≥ n0 =⇒ d( n1 , a) < 21 .
So, d( n10 , a) = 0.
1
n0 = a.
Similarly, n01+1 = a.
This is a contradiction.
Theorem

Let F ⊆ X . Suppose F is closed. Let (fn ) be a

sequence such that fn ∈ F for all n. If fn → f , then
f ∈ F.
Theorem

Let F ⊆ X . Suppose F is closed. Let (fn ) be a

sequence such that fn ∈ F for all n. If fn → f , then
f ∈ F.

Proof
Suppose f ∈ F c .
Theorem

Let F ⊆ X . Suppose F is closed. Let (fn ) be a

sequence such that fn ∈ F for all n. If fn → f , then
f ∈ F.

Proof
Suppose f ∈ F c .
F c is open.
Theorem

Let F ⊆ X . Suppose F is closed. Let (fn ) be a

sequence such that fn ∈ F for all n. If fn → f , then
f ∈ F.

Proof
Suppose f ∈ F c .
F c is open.
Let r > 0 be such that Br (f ) ⊆ F c .
Let r > 0 be such that Br (f ) ⊆ F c .
Let  = 2r .
Let r > 0 be such that Br (f ) ⊆ F c .
Let  = 2r .
There exists n0 such that
n ≥ n0 ⇒ d(fn , f ) < 2r .
Let r > 0 be such that Br (f ) ⊆ F c .
Let  = 2r .
There exists n0 such that
n ≥ n0 ⇒ d(fn , f ) < 2r .
This implies fn0 ∈ Br (f ).
Let r > 0 be such that Br (f ) ⊆ F c .
Let  = 2r .
There exists n0 such that
n ≥ n0 ⇒ d(fn , f ) < 2r .
This implies fn0 ∈ Br (f ).
So, fn0 ∈ F c .
Let r > 0 be such that Br (f ) ⊆ F c .
Let  = 2r .
There exists n0 such that
n ≥ n0 ⇒ d(fn , f ) < 2r .
This implies fn0 ∈ Br (f ).
So, fn0 ∈ F c .
This contradicts our assumption fn ∈ F for each n.
Continuous functions:
Continuous functions:
Let (X , dX ) and (Y , dY ) be two metric spaces.
Continuous functions:
Let (X , dX ) and (Y , dY ) be two metric spaces.
Let f : X → Y .
Continuous functions:
Let (X , dX ) and (Y , dY ) be two metric spaces.
Let f : X → Y .
Let a ∈ X .
Continuous functions:
Let (X , dX ) and (Y , dY ) be two metric spaces.
Let f : X → Y .
Let a ∈ X .
We say that f is continuous at a if for each  > 0
there exists δ such that
dX (x, a) < δ =⇒ dY (f (x), f (a)) < .
Continuous functions:
Let (X , dX ) and (Y , dY ) be two metric spaces.
Let f : X → Y .
Let a ∈ X .
We say that f is continuous at a if for each  > 0
there exists δ such that
dX (x, a) < δ =⇒ dY (f (x), f (a)) < .
Continuity at a point can be characterised through
sequences
f : X → y is continuous at a if and only if

xn → a =⇒ f (xn ) → f (a).
Continuity at a point can be characterised through
sequences
f : X → y is continuous at a if and only if

xn → a =⇒ f (xn ) → f (a).

Proof
Suppose f is continuous at a.
Let xn → a.
We need to show that f (xn ) → f (a).
Let xn → a.
We need to show that f (xn ) → f (a).
Let  > 0.
Let xn → a.
We need to show that f (xn ) → f (a).
Let  > 0.
Then there exists δ > 0 such that
dX (x, a) < δ =⇒ dY (f (x), f (a) < .
Let xn → a.
We need to show that f (xn ) → f (a).
Let  > 0.
Then there exists δ > 0 such that
dX (x, a) < δ =⇒ dY (f (x), f (a) < .
There is a n0 such that
n ≥ n0 =⇒ dX (xn , a) < δ.
Let xn → a.
We need to show that f (xn ) → f (a).
Let  > 0.
Then there exists δ > 0 such that
dX (x, a) < δ =⇒ dY (f (x), f (a) < .
There is a n0 such that
n ≥ n0 =⇒ dX (xn , a) < δ.
So, n ≥ n0 =⇒ dY (f (xn ), f (a)) < .
Let xn → a.
We need to show that f (xn ) → f (a).
Let  > 0.
Then there exists δ > 0 such that
dX (x, a) < δ =⇒ dY (f (x), f (a) < .
There is a n0 such that
n ≥ n0 =⇒ dX (xn , a) < δ.
So, n ≥ n0 =⇒ dY (f (xn ), f (a)) < .
So, f (xn ) → f (a).
Converse:
Converse:
Suppose xn → a =⇒ f (xn ) → f (a).
Converse:
Suppose xn → a =⇒ f (xn ) → f (a).
Suppose f is not continuous at a.
Converse:
Suppose xn → a =⇒ f (xn ) → f (a).
Suppose f is not continuous at a.
There exists  > 0 with the following property:
Converse:
Suppose xn → a =⇒ f (xn ) → f (a).
Suppose f is not continuous at a.
There exists  > 0 with the following property:
dX (x1 , a) < 1 =⇒ dY (f (x1 ), f (a)) ≥ .
dX (x2 , a) < 1/2 =⇒ dY (f (x2 ), f (a)) ≥ .
Converse:
Suppose xn → a =⇒ f (xn ) → f (a).
Suppose f is not continuous at a.
There exists  > 0 with the following property:
dX (x1 , a) < 1 =⇒ dY (f (x1 ), f (a)) ≥ .
dX (x2 , a) < 1/2 =⇒ dY (f (x2 ), f (a)) ≥ .
In general,
dX (xn , a) < 1/n =⇒ dY (f (xn ), f (a)) ≥ .
Clearly xn → a but f (xn ) does not converge to f (a).
Clearly xn → a but f (xn ) does not converge to f (a).
We will do the following problem:
Clearly xn → a but f (xn ) does not converge to f (a).
We will do the following problem:
If f : X → R is continuous at a, and if f (a) > 0,
then there exists δ > 0 such that
x ∈ Bδ (a) =⇒ f (x) > 0.
Suppose for each δ, there exists y ∈ Bδ (a) such that
f (y ) < 0.
Suppose for each δ, there exists y ∈ Bδ (a) such that
f (y ) < 0.
Let yn ∈ B1/n (a) and f (yn ) < 0.
Suppose for each δ, there exists y ∈ Bδ (a) such that
f (y ) < 0.
Let yn ∈ B1/n (a) and f (yn ) < 0.
Now yn → a and f (yn ) converges to f (a).
f (a) ≤ 0 (Why?)
Suppose for each δ, there exists y ∈ Bδ (a) such that
f (y ) < 0.
Let yn ∈ B1/n (a) and f (yn ) < 0.
Now yn → a and f (yn ) converges to f (a).
f (a) ≤ 0 (Why?)
This contradicts our assumption.
Lemma

If (xn ) is a sequence in R, and if xn < 0, xn → x,

then x ≤ 0.
Lemma

If (xn ) is a sequence in R, and if xn < 0, xn → x,

then x ≤ 0.

Proof:
Lemma

If (xn ) is a sequence in R, and if xn < 0, xn → x,

then x ≤ 0.

Proof:
Suppose x > 0.
Lemma

If (xn ) is a sequence in R, and if xn < 0, xn → x,

then x ≤ 0.

Proof:
Suppose x > 0.
Let  > 0 be such that x −  > 0.
Lemma

If (xn ) is a sequence in R, and if xn < 0, xn → x,

then x ≤ 0.

Proof:
Suppose x > 0.
Let  > 0 be such that x −  > 0.
There is n0 such that
n ≥ n0 =⇒ |xn − x| < .
n ≥ n0 =⇒ xn ∈ (x − , x + ).
n ≥ n0 =⇒ xn ∈ (x − , x + ).
xn0 > x −  > 0.
n ≥ n0 =⇒ xn ∈ (x − , x + ).
xn0 > x −  > 0.
Contradicts that xn < 0 for all n.
We say that f : X → Y is a continuous function if
f is continuous at every point.
We say that f : X → Y is a continuous function if
f is continuous at every point.

Theorem

f is continuous iff inverse image of every open set is

open.
We say that f : X → Y is a continuous function if
f is continuous at every point.

Theorem

f is continuous iff inverse image of every open set is

open.

Proof:
We say that f : X → Y is a continuous function if
f is continuous at every point.

Theorem

f is continuous iff inverse image of every open set is

open.

Proof:
Let f be continuous.
Let H be open set in Y .
Let H be open set in Y .
Claim: f −1 (H) is open in X .
Let H be open set in Y .
Claim: f −1 (H) is open in X .
Let c ∈ f −1 (H)
Let H be open set in Y .
Claim: f −1 (H) is open in X .
Let c ∈ f −1 (H)
f (c) ∈ H.
Let H be open set in Y .
Claim: f −1 (H) is open in X .
Let c ∈ f −1 (H)
f (c) ∈ H.
There exists  > 0 such that
B (f (c)) ⊆ H.
f is continuous at c.
f is continuous at c.
So, there is a δ > 0 such that
x ∈ Bδ (c) =⇒ f (x) ∈ B (f (c)).
f is continuous at c.
So, there is a δ > 0 such that
x ∈ Bδ (c) =⇒ f (x) ∈ B (f (c)).
This means x ∈ Bδ (c) =⇒ x ∈ f −1 (B (f (c))).
f is continuous at c.
So, there is a δ > 0 such that
x ∈ Bδ (c) =⇒ f (x) ∈ B (f (c)).
This means x ∈ Bδ (c) =⇒ x ∈ f −1 (B (f (c))).
So, Bδ (c) ⊆ f −1 (B (f (c))).
f is continuous at c.
So, there is a δ > 0 such that
x ∈ Bδ (c) =⇒ f (x) ∈ B (f (c)).
This means x ∈ Bδ (c) =⇒ x ∈ f −1 (B (f (c))).
So, Bδ (c) ⊆ f −1 (B (f (c))).Since B (f (c)) ⊆ H,
f −1 (B (f (c))) ⊆ f −1 (H).
Bδ (c) ⊆ f −1 (H).
Inverse image of open set is open.
Converse:
Converse:
Suppose inverse image of every open set is open.
Converse:
Suppose inverse image of every open set is open.
Claim: f is continuous.
Converse:
Suppose inverse image of every open set is open.
Claim: f is continuous.
Let c ∈ X and  > 0.
Converse:
Suppose inverse image of every open set is open.
Claim: f is continuous.
Let c ∈ X and  > 0.
B (f (c)) is open in Y .
Converse:
Suppose inverse image of every open set is open.
Claim: f is continuous.
Let c ∈ X and  > 0.
B (f (c)) is open in Y .
f −1 (B (f (c))) is open in X .
Converse:
Suppose inverse image of every open set is open.
Claim: f is continuous.
Let c ∈ X and  > 0.
B (f (c)) is open in Y .
f −1 (B (f (c))) is open in X .
c ∈ f −1 (B (f (c))).
There exists δ > 0 such that
Bδ (c) ⊆ f −1 (B (f (c))).
There exists δ > 0 such that
Bδ (c) ⊆ f −1 (B (f (c))).
This means
dX (x, c) < δ =⇒ f (x) ∈ B (f (c)).
There exists δ > 0 such that
Bδ (c) ⊆ f −1 (B (f (c))).
This means
dX (x, c) < δ =⇒ f (x) ∈ B (f (c)).
dX (x, c) < δ =⇒ dY (f (x), f (c)) < .
More on open sets
More on open sets
Let (X , d) be a m.space.
More on open sets
Let (X , d) be a m.space.
Suppose A ⊂ X .
More on open sets
Let (X , d) be a m.space.
Suppose A ⊂ X .
Now, define d 0 (a, b) := d(a, b) for all a, b ∈ A.
More on open sets
Let (X , d) be a m.space.
Suppose A ⊂ X .
Now, define d 0 (a, b) := d(a, b) for all a, b ∈ A.
Clearly, d 0 is a metric on A.
More on open sets
Let (X , d) be a m.space.
Suppose A ⊂ X .
Now, define d 0 (a, b) := d(a, b) for all a, b ∈ A.
Clearly, d 0 is a metric on A.
Question: What are the open sets in the m.space
(A, d 0 )?
Let U be an open ball in (A, d 0 ).
Let U be an open ball in (A, d 0 ).
Then, there exist x ∈ X and  > 0 such that
U := {x ∈ A : d 0 (x, a) < }.
Let U be an open ball in (A, d 0 ).
Then, there exist x ∈ X and  > 0 such that
U := {x ∈ A : d 0 (x, a) < }.This is
U = A ∩ {x : d(x, a) < }.
Let U be an open ball in (A, d 0 ).
Then, there exist x ∈ X and  > 0 such that
U := {x ∈ A : d 0 (x, a) < }.This is
U = A ∩ {x : d(x, a) < }.
So, U = A ∩ G where G is an open set in (X , d).
Theorem

If U is an open set in (A, d 0 ), then U = X ∩ G

where G is an open set in (X , d).
Theorem

If U is an open set in (A, d 0 ), then U = X ∩ G

where G is an open set in (X , d).

Proof
Theorem

If U is an open set in (A, d 0 ), then U = X ∩ G

where G is an open set in (X , d).

Proof
Let u ∈ U.
Theorem

If U is an open set in (A, d 0 ), then U = X ∩ G

where G is an open set in (X , d).

Proof
Let u ∈ U.
Then there exists u > 0 such that
Theorem

If U is an open set in (A, d 0 ), then U = X ∩ G

where G is an open set in (X , d).

Proof
Let u ∈ U.
Then there exists u > 0 such that
{x ∈ A : d 0 (x, u) < u } ⊆ U.
Theorem

If U is an open set in (A, d 0 ), then U = X ∩ G

where G is an open set in (X , d).

Proof
Let u ∈ U.
Then there exists u > 0 such that
{x ∈ A : d 0 (x, u) < u } ⊆ U.
We call this open ball by Bu .
∪{Bu : u ∈ U} = U.
∪{Bu : u ∈ U} = U.
Corresponding to Bu we define
∪{Bu : u ∈ U} = U.
Corresponding to Bu we define
Bu0 = {x ∈ X : d(x, u) < u }.
∪{Bu : u ∈ U} = U.
Corresponding to Bu we define
Bu0 = {x ∈ X : d(x, u) < u }.
Then, Bu = A ∩ Bu0 .
∪{Bu : u ∈ U} = U.
Corresponding to Bu we define
Bu0 = {x ∈ X : d(x, u) < u }.
Then, Bu = A ∩ Bu0 .
∪Bu = U = ∪A ∩ Bu0 = A ∩ (∪Bu0 ).
∪{Bu : u ∈ U} = U.
Corresponding to Bu we define
Bu0 = {x ∈ X : d(x, u) < u }.
Then, Bu = A ∩ Bu0 .
∪Bu = U = ∪A ∩ Bu0 = A ∩ (∪Bu0 ).
UBu0 is an open set in X .
∪{Bu : u ∈ U} = U.
Corresponding to Bu we define
Bu0 = {x ∈ X : d(x, u) < u }.
Then, Bu = A ∩ Bu0 .
∪Bu = U = ∪A ∩ Bu0 = A ∩ (∪Bu0 ).
UBu0 is an open set in X .
So, U = A ∩ G where G is open in X .
Converse:

Theorem

Suppose G is an open set in X . Then, A ∩ G is an

open set in (A, d 0 ).
Converse:

Theorem

Suppose G is an open set in X . Then, A ∩ G is an

open set in (A, d 0 ).

Proof
Converse:

Theorem

Suppose G is an open set in X . Then, A ∩ G is an

open set in (A, d 0 ).

Proof
Let x0 ∈ A ∩ G .
Converse:

Theorem

Suppose G is an open set in X . Then, A ∩ G is an

open set in (A, d 0 ).

Proof
Let x0 ∈ A ∩ G .
There exists r > 0 such that
{x ∈ X : d(x, x0 ) < r } ⊆ G .
So, {x ∈ A : d 0 (x, x0 ) < r } ⊆ G .
So, {x ∈ A : d 0 (x, x0 ) < r } ⊆ G .
Note that {x ∈ A : d 0 (x, x0 ) < r } ⊆ A.
So, {x ∈ A : d 0 (x, x0 ) < r } ⊆ G .
Note that {x ∈ A : d 0 (x, x0 ) < r } ⊆ A.
So, {x ∈ A : d 0 (x, x0 ) < r } ⊆ A ∩ G .
So, {x ∈ A : d 0 (x, x0 ) < r } ⊆ G .
Note that {x ∈ A : d 0 (x, x0 ) < r } ⊆ A.
So, {x ∈ A : d 0 (x, x0 ) < r } ⊆ A ∩ G .
Hence A ∩ G is open.
Let (X , d) be a metric space. Let A ⊆ X . Then Ā
is the intersection of closed sets that contain A.
Let (X , d) be a metric space. Let A ⊆ X . Then Ā
is the intersection of closed sets that contain A.

Problem: Show that x ∈ Ā iff there exists B (x)

such that B (x) ∩ (A r {x}) is non-empty.