So I'm working on a problem with the following statement:

Problem Statement: Assume metric space $C$ is complete and has this property:

If $(x_n)_{n \ge 0}$ is a sequence in $C$ and there exists $\varepsilon > 0$ such that $d(x_i,x_j) > \varepsilon$ for all $i \neq j$, then the sequence $(x_n)$ is not bounded.

Prove that if $A \subset C$ bounded and closed, then $A$ is compact.

Here is what I have done so far/my thoughts...

Proof: From the statement, we know that $C$ is complete thus this means that every Cauchy sequence converges to a point in metric space $C$. Now, compactness means that a subset $A$ in metric space $C$ is compact when the set $A$ is contained in the union of a collection of open subsets of $C$, then $A$ is contained in a finite number of these subsets.

From the property given, we know that the metric space $C$ contains a divergent (non-convergent) sequence that is not bounded. Hence, a subset $A$ of $C$ must be select points which are bounded thus we let $\varepsilon > 0$ such that $d(x,y) < \varepsilon$ $\forall$ $x,y \in A\subset C$. Since $A$ is bounded and closed, we select $x < y$ such that A is monotonically increasing, thus $A$ is a convergent sequence because it is a bounded monotonically increasing sequence. Thus, since $A$ contains a convergent subsequence it is compact. Thus, metric space $C$ is compact.

To me, I feel like I'm in the right direction but am struggling to have that closing logic for this proof. I appreciate the help!

  • 4
    $\begingroup$ $x<y$ doesn't mean anything in a general metric space. $\endgroup$ – aschepler Nov 11 '17 at 22:57
  • $\begingroup$ @aschepler can you elaborate on that? Is that from the definition of a metric space? $\endgroup$ – Inti Nov 11 '17 at 23:02
  • $\begingroup$ @Inti There doesn't have to be an order defined on the metric space. What if for example $C = \mathbb{R}^2$ with the Euclidean metric? $\endgroup$ – Demophilus Nov 11 '17 at 23:38
  • $\begingroup$ You assumed x_n in C and then out of the blue, you are talking about unrelated x_i and x_j. $\endgroup$ – William Elliot Nov 12 '17 at 4:06

You lost me when you say $x<y$: as pointed out in the comment, it is not clear what the ordering here is referring to. In a general metric space, there is no such ordering.

To solve your question one can do the following:

Since $A$ is closed and $C$ is complete, $A$ is also complete. From the answer to this question, it suffices to show that $A$ is totally bounded. That is, for all $\epsilon >0$, there are finitely many open balls of radius $\epsilon$ which covers $A$.

This can be shown easily: Fix an $\epsilon$. If $A$ is not covered by any finite number of $\epsilon$-open balls, then there is a sequence $x_1, x_2, x_3\cdots$ in $A$ so that $d(x_i, x_j) >\epsilon$ (Try to construct this sequence inductively, start with any point $x_1\in A$). But then from the condition on $C$, this sequence is unbounded and this contradicts the fact that $A$ is bounded. Hence $A$ is totally bounded and thus is compact.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.