Let $\{f_{n}\}$ be a sequence of positive continuous functions on $\mathbb R$; $f_{n}:\mathbb R\to \mathbb R$, for all $n\geq 1$, with the folloing properties:

(1) $\{f_{n}\}$ is uniformly bounded by some constant $C>0$,

(2) $\{f_{n}\}$ is uniformly Lipschitz on $\mathbb R$ (so it is uniformly continuous)

(3) $\{f\,'_{n}\}$ is uniformly bounded (sequence of the derivatives), which I think follows from (2).

Does this sequence converges on $\mathbb R$ (or at least contains a subsequence which converges), uniformly on compact sets, or pointwise, to some continuous function $f$ ? If not, what extra (possible) condition must the sequence have to converge to some continuous function?


By Arzela-Ascoli's theorem your sequence has a uniformly convergent subsequence (the limit of which will, of course, be continuous) on each compact subset. You cannot hope for convergence of the sequence itself, though.

  • $\begingroup$ I just noted that the link I included actually explicitly answers the question $\endgroup$
    – user20266
    Jun 17 '12 at 16:06
  • $\begingroup$ @Berry It does not, this is why I wrote 'on each compact subset'. $\endgroup$
    – user20266
    Jun 17 '12 at 16:22
  • $\begingroup$ In your link, at the middle of the page, it says " Euclidean spaces The Arzelà–Ascoli theorem holds, more generally, if the functions ƒn take values in d-dimensional Euclidean space Rd", does this mean that the result apply in $\mathbb R$, and the convergence is uniform in $\mathbb R$? And, what is the R-valued version of the Arzelà–Ascoli theorem ? $\endgroup$
    – Berry
    Jun 17 '12 at 16:25
  • $\begingroup$ @Berry You can however look at recursively defined subsequences $f^{n}_k:[-n,n]\rightarrow \mathbb{R} $ -- i.e., $f^n_k$ is a subsequence of $f^{n-1}_k$, restricted to $[n,n]$, to arrive a continuous limit function on $\mathbb{R}$. It'd be necessary to check what kind of convergence you'd get, though. Uhm, yes, what's the problem with target $\mathbb{R}^n$? $\endgroup$
    – user20266
    Jun 17 '12 at 16:26
  • $\begingroup$ You are looking at $f_k:\mathbb{R} \rightarrow \mathbb{R}$. That is perfectly fine, if you restrict to compact sets (e.g. closed intervals). The comment on the page just says that the target space may be replaced by a more general space. $\endgroup$
    – user20266
    Jun 17 '12 at 16:32

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.