hw12_solutions
hw12_solutions
Exercises:
§27, 28, 31
1. Let X be a compact topological space and (Y, d) a metric space. Let C(X, Y ) denote the set of all
continuous functions f : X → Y .
(a) For f, g ∈ C(X, Y ), show that h : X → R defined by h(x) := d(f (x), g(x)) is continuous.
(b) Show that D(f, g) := supx∈X d(f (x), g(x)) exists and defines a metric on C(X, Y ).
(c) Let ϕ : X → X be a continuous function. Show that the map Φ : C(X, Y ) → C(X, Y ) defined by
Φ(f ) := f ◦ ϕ is uniformly continuous with respect to the metric D.
2. Let (X, d) be a metric space. For x ∈ X and nonempty A ⊂ X, recall that d(x, A) := inf a∈A d(x, a).
(a) Show that d(x, A) = 0 if and only if x ∈ A.
(b) Suppose A ⊂ V for A compact and V open. Show that there exists > 0 so that
[
Bd (a, ) ⊂ V.
a∈A
4. Let X be a normal topological space and let A, B ⊂ X be disjoint closed subsets of X. Show that
there are open subsets U, V ⊂ X satisfying A ⊂ U , B ⊂ V , and U ∩ V = ∅.
5. Let X be a normal topological space. We say A ⊂ X is a Gδ set if it is a countable intersection of open
sets. Show that A ⊂ X is a closed Gδ set if and only if there exists a continuous function f : X → [0, 1]
with f (x) = 0 for all x ∈ A and f (x) > 0 for all x 6∈ A. [Hint: use Urysohn’s Lemma.]
6*. For d ∈ N and a = 0, 1, . . . , d − 1 define
Ud,a := {dn + a | n ∈ Z} ⊂ Z.
In this exercise you will use topology to show that there are infinitely many prime numbers.
(a) Show that the collection B := {Ud,a | d ∈ N, a = 0, 1, . . . , d − 1} forms a basis for a topology on
Z.
(b) Show that Ud,a is clopen in this topology.
(c) Show that if U ⊂ Z is nonempty and open in this topology, then U is infinite.
(d) Let P ⊂ N be the subset of prime numbers. Consider
[
A := Up,0 .
p∈P
Solutions:
1. (a) Observe that h is the composition of the functions f × g : X → Y × Y and d : Y × Y → R. The
former is continuous by Exercise 1 on Homework 6, and the latter is continuous by Exercise 4 on
Homework 7. Hence h is continuous.
(b) By the previous part, x 7→ d(f (x), g(x)) is continuous. Since X is compact, the extreme value
theorem implies the supremum D(f, g) is achieved, and in particular exists. Now, D(f, g) ≥ 0
since d(f (x), g(x)) ≥ 0 for all x ∈ X. If f = g, then d(f (x), g(x)) = 0 for all x ∈ X and hence
D(f, g) = 0. Conversely, if D(f, g) = 0, then d(f (x), g(x)) = 0 for all x ∈ X, which means
f (x) = g(x) for all x ∈ X. That is, f = g. The symmetry D(f, g) = D(g, f ) follows from the
corresponding symmetry of d. Finally, for f, g, h ∈ C(X, Y ) and each x ∈ X we have
d(f (x), h(x)) ≤ d(f (x), g(x)) + d(g(x), h(x)) ≤ D(f, g) + D(g, h).
Taking a supremum over x ∈ X on the left yields D(f, h) ≤ D(f, g) + D(g, h). Hence D is a
metric.
(c) Let > 0. Set δ = , and suppose f, g ∈ C(X, Y ) satisfy D(f, g) < δ = , then d(f (x), g(x)) ≤
D(f, g) for all x ∈ X. In particular, this holds for all x ∈ ϕ(X): d(f (ϕ(x)), g(ϕ(x))) ≤ D(f, g)
for all x ∈ X. Therefore
D(Φ(f ), Φ(g)) = sup d(f ◦ ϕ(x), g ◦ ϕ(x)) ≤ D(f, g) < δ =
x∈X
Indeed, otherwise Bd (a, ) ∩ (X \ V ) 6= ∅ for some a ∈ A. Thus d(a, y) < for some y ∈ X \ V
and therefore
f (a) = d(a, X \ V ) ≤ d(a, y) < = f (a0 ),
contradicting f (a0 ) being the smallest element.
3. If f (x) = f (y), then
d(x, y) = d(f (x), f (y)) = 0,
so that x = y. Thus f is injective. We also note that f is continuous according to the -δ definition of
continuity for metric spaces by simply choosing δ = .
Suppose, towards a contradiction, that f is not surjective. Then there exists y ∈ X \ f (X). Since
X is compact and f is continuous, f (X) is compact and in particular closed (since metric spaces are
Hausdorff). Consequently, X \ f (X) is open and so there exists > 0 so that Bd (y, ) ⊂ X \ f (X).
For each n ∈ N, define xn := f ◦ · · · ◦ f (y). Thus (xn )n∈N ⊂ X is a sequence and for m < n we have
| {z }
n times
d(xn , xm ) = d(f (xn−1 ), f (xm−1 )) = d(xn−1 , xm−1 ) = · · · = d(xn−m+1 , x1 ) = d(f (xn−m ), f (y)) = d(xn−m , y).
Since xn−m = f (xn−m−1 ) ∈ f (X), the above distance is at least . Thus d(xn , xm ) > for all distinct
n, m ∈ N. Now, since X is compact, the sequence (xn )n∈N necessarily has a convergent subsequence
(xnk )k∈N , say converging to x ∈ X. Thus there exists K ∈ N so that k ≥ K implies
d(xnk , x) < .
2
But then for k, ` ≥ K we have
d(xnk , xn` ) ≤ d(xnk , x) + d(x, xn` ) < + = ,
2 2
contradicting d(xn , xm ) ≥ for all n, m ∈ N. Thus f must be surjective, and therefore bijective.
It remains to show f −1 : X → X is continuous. But if f (x) = a and f (y) = b then
d(f −1 (a), f −1 (a)) = d(x, y) = d(f (x), f (y)) = d(a, b).
Thus f −1 is an isometry and is therefore continuous by the same argument as above.
4. Since X is normal, there exists disjoint open sets U1 , V1 ⊂ X with A ⊂ U1 and B ⊂ V1 . By a proposition
from the lecture on §31, there exists a neighborhood U of A satisfying U ⊂ U1 . Set V := V1 , and note
that V ⊂ X \ U1 , which is a closed set. Hence V ⊂ X \ U1 ⊂ X \ U . Consequently, U ∩ V = ∅.
5. (=⇒): Suppose A is a closed Gδ set. Let {Un : n ∈ N} be open subsets of X with
\
Un = A.
n∈N
Using Urysohn’s Lemma, for each n ∈ N we can find a continuous function fn : X → [0, 1] with fn |A ≡ 0
and fn |X\Un ≡ 1. Define f : X → [0, 1] by
∞
X
f (x) = 2−n fn (x).
n=1
so f is well-defined. Also, f (x) = 0 for all x ∈ A since fn (x) = 0. On the other hand, if x 6∈ A then
there must be some n ∈ N such that x 6∈ Un . Consequently, fn (x) = 1 and so f (x) ≥ 2−n > 0. It
remains to show that f is continuous. We will show that the partial sums
N
X
SN (x) := 2−n fn (x)
n=1
converge uniformly to f . Since each SN is continuous (as a finite sum of continuous functions), the
Uniform Limit Theorem will imply f is continuous. Let > 0. Let N0 ∈ N be large enough so that
2−N0 < . Then for all N ≥ N0
X ∞
X
2−n = 2−N 2−n = 2−N < .
n=N +1 n=1
Thus
sup |f (x) − SN (x)| ≤
x∈X
6*. (a) First note that U1,0 = Z, and so every element of Z is contained in a set in the collection B. Next,
suppose x ∈ Ud,a ∩ Ue,b . Then x = dn + a = em + b for some n, m ∈ Z. Let f be the least common
multiple of d and e and let c be the remainder one gets after dividing x by f : x = f m + c for
some multiple m ∈ Z. Then x ∈ Uf,c , and in fact Uf,c = {x + f n : n ∈ Z}. Observe that f n is a
multiple of both d and e since f is the least common multiple, and therefore x + f n ∈ Ud,a ∩ Ue,b
for all n ∈ Z. Hence Uf,c ⊂ Ud,a ∩ Ue,b , and therefore B is a basis for a topology on Z.
(b) Ud,a is certainly open. To see that it is closed, observe that
[
Z \ Ud,a = Ud,b .
b6=a
Indeed, x ∈ Ud,b if and only if x = dn+b for some n ∈ Z. So if x ∈ Ud,a ∩Ud,b , then dn+b = dm+a
for some n, m ∈ Z, and therefore a − b = d(n − m). This implies a − b is divisible by d, but
a − b ∈ {−(d − 1), . . . , −1, 0, 1, . . . , d − 1} and so this is only possible if a = b. This shows Ud,a is
disjoint from Ud,b for all b 6= a and so
[
Ud,b ⊂ Z \ Ud,a .
b6=a
Conversely, if x ∈ Z \ Ud,a , let b ∈ {0, . . . , d − 1} be the remainder one gets after dividing x by d:
x = dn + b for some n ∈ N. So x ∈ Ud,b , and since x 6∈ Ud,a , we must have b 6= a. This yields the
other inclusion and establishes the desired equality.
(c) Suppose U ⊂ Z is nonempty and open. Let x ∈ U . Since B is a basis, there exists d ∈ N and
a ∈ {0, 1, . . . , d − 1} such that x ∈ Ud,a ⊂ U . Define f : N → U by f (n) := dn + a. This is
injective, and hence U is infinite.
(d) If x ∈ Z \ A, this means x 6∈ Up,0 for any prime number p. Consequently, x is not divisible by
p for any prime p. There are only two numbers not divisible by any prime number: ±1. Thus
Z \ A = {−1, 1}, which is of course finite.
(e) Suppose, towards a contradiction that P is finite. Then A from the previous part is a finite union
of closed sets (recall that Up,0 is closed by part (b)), and thus is closed. Therefore its complement
Z \ A is open. However, in the previous part we showed Z \ A is finite and nonempty, which
contradicts part (c). Thus P must be infinite.