2
$\begingroup$

Let $\{f_n\} \subset C([0,1])$ be a sequence of functions with $f_n(x) = x^n$ for $n = 1, 2, \ldots$

  1. Prove that $\{f_n\}$ does not have any uniformly convergent subsequence.

  2. By (1) to prove that the closed unit ball in $C([0,1])$ is not compact.

By Arzela-Ascoli Theorem, I think $\{f_n\}$ seems to contain a uniformly convergent subsequence since $\{f_n\}$ is equicontinuous and pointwise bounded. Is it right?

$\endgroup$
  • $\begingroup$ How do you know this set is equicontinuous? $\endgroup$ – Alex S Jul 9 '15 at 5:29
  • $\begingroup$ What metric are you using? The sup metric, $L^2$,...? $\endgroup$ – user217285 Jul 9 '15 at 5:47
4
$\begingroup$

Note that any subsequence of $(f_n)_{n\in\mathbb N}$ converges pointwise the function $$f(x)\equiv\begin{cases}0&\text{if $x\in[0,1)$,}\\1&\text{if $x=1$.}\end{cases}$$ Hence, if uniform convergence occurred, then the uniform limit would have to agree with the pointwise limit. However, the pointwise limit $f$ is not continuous. Since the uniform limit of continuous functions is continuous, it follows that no subsequence can converge uniformly.

Now, $\|f_n\|=\sup_{x\in[0,1]}|x^n|=1$ for each $n\in\mathbb N$, so that the sequence $(f_n)_{n\in\mathbb N}$ is contained in the unit ball of $C([0,1])$ (when this space is endowed with the uniform norm). If the unit ball were compact, there would exist a uniformly convergent subsequence, which, as one has seen, is not possible.


The Arzelà–Ascoli theorem doesn't work here because the sequence is not equicontinuous at $x=1$. To see this, pick any $\varepsilon\in(0,1)$. If the sequence were equicontinuous, there would exist some $\delta>0$ such that if $y\in(\max\{1-\delta,0\},1]$, then $$|f_n(1)-f_n(y)|<\varepsilon\quad\forall n\in\mathbb N.$$ But this is impossible, as $$|f_n(1)-f_n(y)|=|1-y^n|\to 1\quad\text{as $n\to\infty$}$$ whenever $y\in(0,1)$, so $|f_n(1)-f_n(y)|$ eventually exceeds $\varepsilon<1$.


It may be worth keeping in mind that, in general, if $(X,\|\cdot\|)$ is any infinite-dimensional normed vector space, then its closed unit ball is never compact with respect to the norm topology.

$\endgroup$

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.