LU - Department of Mathematics

Dr H. Gebran

M2201 Exercises of chapter 3 – Statements and solutions Fall 2020-2021

1. Show that a finite union of compact subspaces of a space X is compact.

Solution. It is enough to prove this for two subspaces. Let K1 and K2 be two compact
subspaces of X. Let (Oλ )λ∈L be a covering of K1 ∪K2 by open subsets of X. Then (Oλ )λ∈L is a
covering of K1 by open subsets of X. Since K1 is compact, (Oλ )λ∈L has a finite subcollection
(Oλ )λ∈L1 that covers K1 . Similarly, (Oλ )λ∈L is a covering of K2 by open subsets of X.
Compactness of K2 implies that (Oλ )λ∈L has a finite subcollection (Oλ )λ∈L2 that covers K2 .
Then (Oλ )λ∈L1 ∪L2 is a finite collection of (Oλ )λ∈L that covers K1 ∪ K2 .

2. Let (K
Tn ) be a decreasing sequence of nonempty compact sets in a metric space X. Let
K= ∞ n=1 Kn . Show that if U is an open set containing K, then U contains Kn for all n large
enough.

Solution. It is enough to prove that U contains some Km (because the sequence is

decreasing). If not, then Kn ∩ (X − U ) 6= ∅ for all n ∈ N∗ . Thus, we get a decreasing
sequence of nonempty closed sets in the compact space K1 . Their intersection should be
therefore nonempty. But
∞
\

Kn ∩ (X − U ) = K ∩ (X − U ) = ∅
n=1

since K ⊂ U , a contradiction.

3. Let X and Y be two metric spaces. Show that if Y is a compact, then the projection π1 :
X × Y → X is a closed map.

Solution. Let A be a closed subset of X × Y . We have to show that π1 (A) is closed, i.e,
equal to its closure. Let x ∈ π1 (A). By Proposition 1.8., there exists a sequence (xn ) of π1 (A)
that converges to x. xn ∈ π1 (A) means that xn = π1 (xn , yn ) where (xn , yn ) ∈ A ⊂ X × Y .
Compactness of Y - which is equivalent to sequential compactness of Y - implies that (yn ) has
a subsequence (ynk ) that converges to some y ∈ Y . It follows that (xnk , ynk ) converges to
(x, y). Since A is closed, (x, y) ∈ A. This means that x = π1 (x, y) ∈ π1 (A).

4. Let f : X → Y be a map between metric spaces. Show that if Y is compact and the graph of
f , G = {(x, f (x)); x ∈ X} is closed in X × Y , then f is continuous.

Solution. We will use Proposition 2.1.
Let (xn ) be a sequence of
X that converges to x.
Since Y is compact the sequence f (xn ) has a subsequence

f (xnk ) that converges to some
y ∈ Y . It follows that the sequence xnk , f (xnk ) converges to (x, y). Since the graph

of f is
closed, it follows that (x, y) ∈ G(f ). This means that y = f (x). Therefore f (xnk ) converges
to f (x). By Proposition 2.1 f is continuous at x.

5. Let A, B ⊂ Rn . We set A + B = {x + y; x ∈ A, y ∈ B}.

1
 (a) Show that if A and B are compact, then A + B is compact.
(b) Show that if A is closed and B is compact then A + B is closed.

Solution. (a) First method. Note that A + B is the image of A × B under the continuous
map (x, y) 7→ x + y. If A and B are compact, then A × B is compact and so its image under
a continuous map is compact as well. Therefore A + B is compact.

Second method. We show that A+B is sequentially compact. Let (xn ) be a sequence of A+B.
We can write xn = an + bn where an ∈ A and bn ∈ B. Compactness of A implies that (an )
has a subsequence (ank ) that converges to some a ∈ A. Compactness of B implies that (bnk )
has a subsequence still denoted by (bnk ) that converges to some b ∈ B. It follows that (xnk )
converges to a + b ∈ A + B.

(b) Let x belong to the closure of A + B. Then, there exists a sequence (xn ) of A + B that
converges to x. We can write xn = an + bn where an ∈ A and bn ∈ B. Compactness of B
implies that (bn ) has a subsequence (bnk ) that converges to some b ∈ B. Since (xnk ) converges
to x, the sequence (ank ) converges to a := x − b. Since A is closed, a ∈ A. Therefore,
x = a + b ∈ A + B.

6. Let X be a compact metric space. Let (xn ) ⊂ X be a sequence with a unique cluster point a.

(a) Show that the sequence (xn ) converges to a.

(b) Show by an example that if the compactness assumption is removed, the conclusion may
fail.

Solution. (a) Suppose that (xn ) does not converge to a. Then

¬ (∀ε > 0 ∃n0 ∈ N ∀n > n0 d(xn , a) < ε) .

Therefore,
∃ε0 ∀m ∈ N ∃n > m d(xn , a) ≥ ε0 .
Step 1. Take m = 1 in the above condition. Accordingly, there is n1 > 1 such that
d(xn1 , a) ≥ ε0 .

Step 2. Take m = n1 in the above condition. Accordingly, there is n2 > n1 such that
d(xn2 , a) ≥ ε0 .

Continuing in this way, we construct a subsequence (xnk ) of (xn ) such that

d(xnk , a) ≥ ε0 ∀k ∈ N∗ .

Compactness of X implies that (xnk ) has a convergent subsequence (xnk` ). By assumption

this subsequence converges to a. But this contradicts the inequality

d(xnk` , a) ≥ ε0 ∀` ∈ N∗ .

(b) Consider the real sequence 1, 12 , 2, 13 , 3, . . . , n1 , n, . . .. Zero is the unique cluster point of this
sequence. However the sequence does not converge. Note that he sequence does not live in
compact space.

7. Let X be a metric space. For A, B ⊂ X, A, B 6= ∅, we set

d(A, B) = inf d(x, B) = inf inf d(x, y).

x∈A x∈A y∈B

2
(a) Show that if A is compact and B is closed with A ∩ B = ∅, then d(A, B) > 0. Does the
result hold if A is not assumed compact?
S
(b) Show that {x ∈ X; d(x, A) < r} = a∈A B(a, r).
(c) Suppose that A is compact and U is an open set containing A. Show that there exists
ε > 0 such that {x ∈ X; d(x, A) < ε} ⊂ U . Does the result hold if A is not assumed
compact?

Solution. (a) First method. We know that the function x 7→ d(x, B) is continuous.
Therefore it has a minimum value on the compact set A. This means that there exists a ∈ A
such that d(a, B) = inf x∈A d(x, B) = d(A, B). Since A and B are disjoint and a ∈ A, then
a∈/ B. Since B is closed d(a, B) > 0. Thus, d(A, B) > 0.

Second method. Suppose that d(A, B) = 0. By a property of the infimum, there exist
a sequence (an ) of A and a sequence (bn ) of B such that d(an , bn ) → 0. Compactness of
A implies that (an ) has subsequence (ank ) that converges to some a ∈ A. The condition
d(ank , bnk ) → 0 implies that (bnk ) converges to a as well. Since B is closed, we get a ∈ B.
Therefore, a ∈ A ∩ B, a contradiction.
Without the assumption of compactness, the result need not hold. Take for example in R2 ,
 
1
A = (x, y); y ≥ > 0 and B = {(x, y); y ≤ 0} .
x
Then both A and B are closed but d(A, B) = 0.

(b) ⊃ If d(x, a) < r for some a ∈ A, then d(x, A) < r. ⊂ Conversely, let d(x, A) < r and
let ε = r − d(x, A). By a property of the infimum, there exists a ∈ A such that d(x, a) <
d(x, A) + ε = r.

(c) Since A ⊂ U , we have A ∩ (X − U ) = ∅. By (a), d(A, X − U ) > 0 because X − U is closed

and A is compact. Now take 0 < ε < d(A, X − U ). We claim that {x ∈ X; d(x, A) < ε} ⊂ U .
Indeed, let d(x, A) < ε. Then, d(x, A) < d(A, X − U ). If x ∈
/ U , then, x ∈ X − U and so

d(x, A) ≥ inf d(z, A) = d(X − U, A) = d(A, X − U ).

z∈X−U

Then, d(x, A) < d(x, A), a contradiction. Therefore, x ∈ U .

The result need not hold if A is not assumed to be compact. For example take A = [0, 1[ and
U =] − ∞, 1[. Then {x; d(x, A) < ε} =] − ε, 1 + ε[ and this set is not contained in U no
matter how ε is small.

3
8. Let f : X → Y be a continuous map between two metric spaces. We say that f proper if the
inverse image under f of a compact set is a compact set.

(a) Show that a proper map is closed.

(b) Equip Rn and Rm with the Euclidean norm. Let f : Rn → Rm be a continuous map.
Show that the following conditions are equivalent.
(i) f is proper.
(ii) The inverse image under f of a bounded set is a bounded set.
(iii) lim ||f (x)|| = +∞.
||x||→∞
(c) Give an example of a proper map and an example of a continuous map which is not proper.

Solution. (a) Let A ⊂ X be closed. We will prove that f (A) = f (A). Let y ∈
f (A). By Proposition 1.8., there is a sequence (xn ) of A such that f (xn ) → y. Let
K := {f (x1 ), f (x2 ), . . .} ∪ {y}. We know from the lectures that K is compact. But f (xn ) ∈
K. Therefore, xn ∈ f −1 (K). The properness assumption implies that f −1 (K) is compact.
Therefore (xn ) has a subsequence (xnk ) that
converges to some x. Since A
is closed x ∈ A.
Continuity of f at x implies that f (xnk ) converges to f (x). But f (xnk ) converges to y.
Therefore, y = f (x) ∈ f (A).
(b) (i) ⇒ (ii). Let B ⊂ Rm be bounded. Then B is compact since it is closed and bounded
in Rm . Properness of f implies that f −1 (B) is compact and therefore bounded it follows that
f −1 (B) is bounded as well.
(ii) ⇒ (i). Let K ⊂ Rm be compact. Then, it is closed and bounded. Continuity of f implies
that f −1 (K) is closed in Rn . Our assumption implies that f −1 (K) is bounded. Being closed
and bounded in Rn , f −1 (K) is compact.
(ii) ⇒ (iii). Let M > 0 be given. Since f −1 (B 0 (0, M )) is bounded, it is contained in some ball
B 0 (0, R) of Rn . By contraposition, if ||x|| > R, then, x ∈
/ f −1 (B 0 (0, M )) and so ||f (x)|| > M .
This precisely means that lim ||f (x)|| = +∞.
||x||→∞

(iii) ⇒ (ii). You should be able to write the proof.

(c) Any polynomial is proper. The map f : Rn → R given by f (x) = e||x|| is proper. Any
homeomorphism is proper. The map f : Rn → R given by f (x) = e−||x|| is continuous but not
proper. The map f : R → R given by f (x) = sin x is continuous but not proper. The map
1
f : R → R given by f (x) = 1+x 2 is continuous but not proper.

9. Show that the functions x 7→ x2 and x 7→ sin x2 are not uniformly continuous on R.

Solution. You should have proved the following characterization of uniform continuity. Let
f : X → Y be a map between two metric spaces. Then, f is uniformly continuous if and only
if for any any two sequences (xn ) and (yn ) of X, we have
d(xn , yn ) → 0 ⇒ d(f (xn ), f (yn ) → 0.
Let f (x) = x2 . Take xn = n and yn = n + n1 . Then, |xn − yn | = n1 → 0. However,

1 2
 
2 1
|f (xn ) − f (yn )| = n − n + =2+ 2
n n
which does not converge to 0.
Let f (x) = sin x2 . Take xn = π2 + nπ and yn = π2 + (n + 1)π. You should be able to
p p

check that |xn − yn | → 0 but |f (xn ) − f (yn )| = 2. Otherwise, you know what to do.