Converse of the extreme value theorem

Theorem 1(Extreme value theorem).

Let be a metric space. Then is compact if and only if every continuous real valued function on takes a maximum and a minimum value.

We have already shown the forward implication here.

Proof of
Assume every real valued function on takes a minimum and maximum value. To show that is compact, it is necessary and sufficient to show that it is complete and totally bounded.

Argue by contradiction to show is totally bounded. If is not totally bounded, there exists and a countably infinite subset of which we may enumerate as , for which the collection of open balls is disjoint. For each natural number , define by

Define the function by

Since each is continuous and vanishes outside and the collection is disjoint, is well-defined and continuous. But for each natural number , , and hence is unbounded above. This is a contradiction.

It remains to show that is complete. Let be a Cauchy sequence in . For each , we infer form the triangle inequality that is a Cauchy sequence in , which converges since is complete. Thus, we can define by

is continuous: . By assumption, there is a point at which takes a minimum value. Since is Cauchy, the infimum of on is . Therefore and hence . Thus is complete.

Theorem 2(Lebesgue covering Lemma).

Let be an open cover of a compact metric space . Then there exists , called a Lebesgue number for the cover, such that for each , the ball is contained in some member of the cover.

Proof
Assume there is no such Lebesgue number. Then for every , there exists such that is not contained in any . Let be the sequence of such points generated by the sequence . Since is compact, there exists a subsequence of which converges, say to a point . Now there exists some for which . Since is open, there exists an open ball . We may choose an index such that and . Then, , which is a contradiction.