6
$\begingroup$

Let $X$ be a compact metric space, and let $Y$ be another metric space.

I am looking for examples of, and especially references to, theorems that give conditions under which any continuous mapping $f:X\rightarrow Y$ can be arbitrarily well-approximated in the uniform (supremum) distance by Lipschitz mappings $g:X\rightarrow Y$.

If $Y=\mathbb{R}$, then this is answered positively by the Stone-Weierstrass Theorem. (Note that real-valued Lipschitz functions always separate points in a metric space.)

By applying this to the coordinate functions, we get a positive answer as well in the case where $Y=\mathbb{R}^n$ or even some infinite-dimensional Banach spaces like $\ell^p$.

On the other hand, it is easy to come up with pairs of spaces with many continuous maps between them but no non-trivial Lipschitz maps. For example, take $X=[0,1]\subseteq\mathbb{R}$ and $Y=\mathbb{R}$, but where $Y$ is equipped with the metric $$ d_Y(p,q) = |p-q|^{1/2}.$$ In this case there are of course many continuous maps from $X$ to $Y$, but every Lipschitz map must be constant.

Are there any references to general theorems that say something positive in the case where, for example, $Y$ is not a Banach space?

$\endgroup$
3
  • 1
    $\begingroup$ @PiotrHajlasz: Anonymity has been something that's been discussed since the earliest days of MO. See for instance meta.mathoverflow.net/questions/1113/… and meta.mathoverflow.net/questions/2574/anonymity-policy. I would say the consensus was that we encourage real names for users who are comfortable with that, but if they are not (and there are many possible reasons for this), we respect that choice. $\endgroup$ Apr 10, 2018 at 21:41
  • $\begingroup$ @NateEldredge You are right. I deleted my comment. $\endgroup$ Apr 10, 2018 at 21:47
  • $\begingroup$ Has the OP found an answer to this old post? $\endgroup$
    – ABIM
    Oct 21, 2021 at 12:14

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.