Math 313: Lecture 13
M. K. Vemuri
Compactness (review)
Definition
By an open cover of a set E in a metric space X we S
mean a
collection {Gα } of open subsets of X such that E ⊆ α Gα .
Definition
A subset K of a metric space X is said to be compact if every
open cover of K contains a finite subcover.
More explicitly, the requirement is that given an open cover
{Gα }α∈I of K , there exist finitely many indices α1 , . . . , αn ∈ I such
that
K ⊆ Gα1 ∪ · · · ∪ Gαn .
Compactness (contd)
In order to prove that “K is compact”, one has to show that for
every open cover {Gα }α∈I of K , there exists a finite set
{α1 , . . . , αn }⊆ I such that
K ⊆ Gα1 ∪ · · · ∪ Gαn .
On the other hand, in order to prove that “E is not compact”, it
suffices to find one open cover {Gα }α∈I of E which does not have
a finite subcover.
Compactness (contd)
Theorem
Compact subsets of metric spaces are bounded.
Proof.
Let K be aScompact subset of the metric space X . Let p ∈ X .
Then X = r ∈(0,∞) Nr (p), so {Nr (p)}r ∈(0,∞) is an open cover of
K . Since K is compact, there exist r1 , · · · rk ∈ (0, ∞) such that
k
[
K⊆ Nrj .
j=1
Put r = max{r1 , . . . , rk }. Then K ⊆ Nr (p), so K is bounded.
Compactness (contd)
Theorem
Compact subsets of metric spaces are closed.
Proof.
Let E be a subset of the metric space X , and assume that E is not
closed. We will construct an open cover of E which does not have
a finite subcover. Since E is not closed, there exists p ∈ E ′ \ E .
Let Gn = {x ∈ X | d(p, x) > 1/n}, n ∈ N. If x ∈ Gn , and
r = d(p, x) − 1/n, then r > 0, S and Nr (x) ⊆ Gn . Therefore each Gn
is open. Also, E ⊆ X \ {p} = n∈N Gn . Therefore {Gn }n∈N is an
open cover of E . We claim that it has no finite subcover. Indeed,
if M is a finite subset of N, and m = max M, then 1/m > 0 and
since p is aSlimit point of E , there exists q ∈ N1/m ∩ E . This
q∈ / Gm = n∈M Gn , so {Gn }n∈M is not a cover of E .
Compactness (contd)
Theorem
Let K be a compact subset of X . If F is a closed set in X and
F ⊆ K , then F is compact.
Proof.
Let Σ be an open cover of F . This means each element of Σ is an
open set and F ⊆ U∈Σ U. Since F is closed, F c is open. Let
S
Ω = Σ ∪ {F c }. Then Ω is an open cover of K . Since K is
compact, there is a finite subcollection Φ of Ω which covers K and
hence F . Let Ψ = Φ \ {F c }. Then Ψ ⊆ Σ and covers F .
Therefore F is compact.
Corollary
If F is closed and K is compact then F ∩ K is compact.
The Heine-Borel theorem in R
Theorem
Let c, d ∈ R. In the metric space R with the usual metric, the set
[c, d] is compact.
Proof. Suppose {Gα }α∈I is an open cover of [c, d]. Let
( )
[
A= x ∈ [c, d] ∃M ⊆ I , |M| < ∞, [c, x] ⊆ Gα .
α∈M
We need to show that d ∈ A.
Since [c, c] = {c} ⊆ Gα for some α, the set A is not empty. Also
d is an upper bound of A. Therefore a = sup A exists. It is clear
that a ∈ [c, d].
The Heine-Borel theorem in R (contd)
First, we claim that a ∈ A. Indeed, there exists β ∈ I such that
a ∈ Gβ . Since Gβ is open, there exists r > 0 such that
(a − r , a + r ) = Nr (a) ⊆ Gβ . Since a − r is not an upper bound of
A, there exists x ∈ A such that a − r < x. Since S x ∈ A, there
exists M ⊆ I such that |M| < ∞ and [c, x] ⊆ α∈M Gα . Put
M ′ = M ∪ {β}. Then M ′ ⊆ I , |M ′ | < ∞ and
[
[c, a] ⊆ [c, x] ∪ (a − r , a + r ) ⊆ Gα .
α∈M ′
Therefore a ∈ A, and the claim is proved.
The Heine-Borel theorem in R (contd)
Next, we claim that a = d. Suppose not, i.e. S a < d. Since a ∈ A,
there is a finite set M ⊆ I such that [c, a] ⊆ α∈M Gα . In
particular, a ∈ Gβ for some β ∈ M. Since Gβ is open, there exists
r > 0 such that (a − r , a + r ) = Nr (a) ⊆ Gβ . Let
b = max{a + r /2, d}. Then b > a and
[
[0, b] ⊆ Gα ,
α∈M
so b ∈ A. This contradicts the fact that a = sup A, and proves the
claim. The two claims together imply that d ∈ A, and the theorem
is proved.
The Heine-Borel theorem in R (contd)
Theorem (Heine-Borel)
If K is a closed bounded set in the metric space R, then K is
compact.
Proof.
Since K is bounded, there exist q, M ∈ R such that K ⊆ NM (q).
Therefore K ⊆ [q − M, q + M]. Since K is closed and
[q − M, q + M] is compact, it follows that K is compact.
The Bolzano-Weierstrass theorem in R
Theorem
Let K be a compact subset of a metric space X . If E is an infinite
subset of K then E has a limit point in K .
Proof.
If no point of K were a limit point of E , then each q ∈ K would
have a neighborhood Vq which contains at most one point of E
(namely, q, if q ∈ E ). It is clear that no finite subcollection of
{Vq } can cover E ; and the same is true of K , since E ⊆ K . This
contradicts the compactness of K .
The Bolzano-Weierstrass theorem in R (contd)
Theorem
Every bounded infinite set of real numbers has a limit point.
Proof.
Let E be a bounded infinite subset of R. Then E ⊆ [c, d] for some
c, d ∈ R. Since [c, d] is compact, it follows that E has a limit
point in [c, d].