Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $x_{n}$ be a sequence of continuous functions uniformly convergent to the function $x$, the domain of all functions $x_{n}$ be an interval $[a,b]$ and let $g_{n}(t):=f(t,x_{n}(t))$ and $g(t):=f(t,x(t))$, where $f$ is real-valued function with open domain.

Question: What assumptions about $f$ should be made to get the uniform convergence $g_{n}(t)$ to $g(t)$ on the interval $[a,b]$?

I suspect that continuity of $f$ is not sufficient, probably one can assume the uniform continuity of $f$, but I can not prove any of these claims.
I would be very grateful for any hints.

share|cite|improve this question

I think that the necessary uniformly continuity of $f$ in such a case is obtained automatically on our set of interest(it is my point 2 following from Cantor's theorem as stated in Lang's Real and Functional Analysis).

Below I sketch how continuity of $f$ is sufficient:

  1. $C=\{(t,x(t)):t\in[a,b]\}$ is compact (for the continuity of $x$), and it is contained in the domain of definition of $f$ (under your assumption);
  2. $f$ is continuos so for Cantor's theorem it is uniformly continuous relatively at $C,$ this means that for any $\varepsilon$ there exists $\delta$ such that $$\|(t',x')-(t'',x'')\|<\delta\textrm{ and }(t',x')\in C\Rightarrow |f(t'',x'')-f(x',t')|<\varepsilon.$$
  3. $\{x_n\}_{n\in\mathbb{N}}$ is uniformly convergent towards $x,$ that means for any $r$ there exists $\nu\in\mathbb{N}$ s.t. $\|x_n-x\|<r$ for any $n>\nu.$

From 1.,2., and 3. we conclude easily the uniform convergence of $\{f\circ(\textrm{id},x_n)\}_{n\in\mathbb{N}}$ towards $f\circ(\textrm{id},x)$ on $[a,b].$

share|cite|improve this answer
Thank you Giuseppe for your answer. – John Apr 29 '12 at 10:05

I think you may need a Hölder or Lipschitz uniform continuity condition of some order $\alpha > 0$ to get this. But when you do it comes out so easily that you may wonder if uniform continuity of $f$ -on this open domain you mentioned - wouldn't be enough.. I'll have to look for counterexamples. I don't think there are necessary conditions on $f$, because you might 'tweak' an example of $x_n$'s given a particular $f$ where say $g_n(t)$ was identically constant or something like that. This is why I say your question 'what conditions should be made about $f'$ is somewhat misleading, really it's 'what statements I can prove about f will guarantee this result'. If I can find some counterexamples where uniform continuity is not enough, etc. I will work on it.(later that AM) I see that if $x_n\to x$ uniformly and if $f$ is merely continuous, then we have pointwise convergence of $g_n\to g$ (all continuous too). But this can allow for some absurdities(maybe not if your paths were in $\mathbb R^2$ or less to begin with).

share|cite|improve this answer
Thank you James. May I ask if you found an counterexample where uniform continuity is not enough? For me it's hard to imagine that such a counterexample exists... – John Apr 29 '12 at 10:07

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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