I am faced with the following problem:

Let $X$ be a totally bounded metric space. If $f$ is a uniformly continuous mapping from $X$ to a metric space $Y$, show that $f(X)$ is totally bounded. Is the same true if $f$ is only required to be continuous?

For the first part -i.e., showing that $f(X)$ is totally bounded, so far I have the following:

Suppose that $X$ is totally bounded. Then $\forall \epsilon > 0$, $X$ can be covered by a finite number of open balls of radius $\epsilon$, thus also making $X$ immediately compact. Since $f$ is uniformly continuous, it is also continuous. Therefore, $f(X)$ is compact.

Now, I know that it is true that if a metric space $X$ is compact, it is sequentially compact, and if a metric space $X$ is sequentially compact, it is complete and totally bounded. However, $f(X)$ is not necessarily a metric space is it?

And although as of yet, I have not used the fact that the continuity is uniform, I believe it is important, and that the second part of the problem is therefore false. What I'm thinking is that the uniform continuity has something to do with making $f(X)$ totally bounded in addition to just being compact, but I'm missing a piece of the puzzle here.

Edit: But then, there is also a theorem that says that a continuous mapping from a compact metric space into a metric space is uniformly continuous, so isn't $f$ uniformly continuous regardless of whether we explicitly say so? Still, though, I am not sure I'm correct then in assuming this would guarantee that $f(X)$ is totally bounded, because although a metric space $X$ is totally bounded if and only if it is compact, I'm not convinced that $f(X)$ is a metric space. I know that $Y$ is, but $f(X)$ is just a subset of a metric space. How do I know/show that it is a subspace??

Could somebody please help me out?

Thank you

  • $\begingroup$ A metric space is compact iff it is totally bounded and complete. see maths.kisogo.com/… $\endgroup$ – Alec Teal Dec 7 '15 at 13:51
  • $\begingroup$ @AlecTeal, so $f(X)$ is a metric space? I don't care how stupid that sounds - humor me. $\endgroup$ – ALannister Dec 7 '15 at 13:53
  • $\begingroup$ You're wrong in saying "therefore it is compact" you require completeness for it to be compact. See the link I gave for exact conditions $\endgroup$ – Alec Teal Dec 7 '15 at 13:55
  • $\begingroup$ @AlecTeal, I thought there was a theorem that said that the image of any compact metric space was also compact. Or are you saying that I need to show that $X$ is complete first? $\endgroup$ – ALannister Dec 7 '15 at 13:56
  • $\begingroup$ The metric subspace $B_1(0)$ of $\mathbb{R}^2$ is totally bounded (ball of radius 1 at the origin IS the entire space) but not complete, (consider sequence $1-\frac{1}{n}$) $\endgroup$ – Alec Teal Dec 7 '15 at 13:59

First $X$ is not necessarily compact. For example $X=(0,1) \subset \mathbb R$ is a totally bounded space but is not compact.

To prove that $f(X)$ is totally bounded, you have to prove that for any $\epsilon > 0$, $f(X)$ is contained in the union of a finite number of balls of radius $\epsilon$.

So take $\epsilon > 0$. As $f$ is uniformly continuous, you can find $\delta > 0$ such that $d_X(x,y) < \delta$ implies $d_Y(f(x),f(y)) < \epsilon$. As $X$ is supposed to be totally bounded, you can find a finite number of balls $B_X(a_1, \delta), \dots, B_X(a_n, \delta)$ such that $$\displaystyle X \subset \cup_{i=1}^n B_X(a_i,\delta)$$ but then $$\displaystyle f(X) \subset \cup_{i=1}^n B_Y(f(a_i),\epsilon).$$


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.