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 $(g_n)$ be a sequence of twice differentiable functions defined on $[0,1]$, and assume that $g_n(0)=g_n'(0)$ for all $n$. Suppose also that $|g_n'(x)|\leq 1$ for all $n\in\mathbb{N}$ and all $x\in[0,1]$. Prove that there is a subsequence of $(g_n)$ converging uniformly on $[0,1]$.

This problem is screaming at me Arzelà-Ascoli Theorem. However, I am neither sure how to show that $(g_n)$ is uniformly bounded nor how to show that $(g_n)$ is equicontinuous.

Any help would be appreciated.

Edit: $(g_n)$ is uniformly bounded since $$|g_n(x)|=\left|g(0)+\int_0^xg_n'(x)dx\right| \leq |g_n(0)|+\int_0^x|g_n'(x)|dx \leq |g_n(0)|+ \int_0^x dx=1+x\leq2.$$

share|cite|improve this question
You made $g(0)$ disappear in there ; your upper bound should become $2$ instead of $1$. The reason for this is that the sequence $g_n(x) = 1+x$ satisfies the hypothesis, therefore I know your bound is wrong. =) – Patrick Da Silva Feb 21 '12 at 7:25
you are correct i was thinking $g_n(0)=0$ which is not true – Galois Feb 21 '12 at 7:33
Lol even you? XD I mean, you're asking the question! You should've at least read it right. =P This is most often the reason why intelligent people give wrong answers ; they read the question wrong. I am fortunate to have the luck to read things well at this time of the day (2:34 a.m.), so I understand your badluck. ^^ – Patrick Da Silva Feb 21 '12 at 7:34
up vote 3 down vote accepted

For equicontinuity, you have that for some $\zeta \in [0,1]$, $$ |g_n(x) - g_n(y)| = |g'_n(\zeta)||x-y| \le |x-y| $$ by Taylor's theorem, so that $\{ g_n \}$ are all Lipschitz with constant $1$, hence equicontinuous. I don't know why you need twice differentiability ; $C^1$ is enough here.

For uniformly bounded, $$ |g_n(x)| = |g_n(x) - g_n(0) + g_n(0)| = |g_n'(\zeta)x + g_n'(0)| \le |g'_n(\zeta)||x|+|g'_n(0)| \le 1+x. $$ again by Taylor's theorem and since $|g'_n(x)| \le 1$, plus with $x \ge 0$ I can remove the $| \cdot |$.

Hope that helps,

share|cite|improve this answer

The fact that the sequence $(g_n)$ is equicontinuous follows immediately from the mean value theorem, and the assumption $|g_n'(x)|\leq 1$. Since $|g_n(x)-g_n(y)|=|g_n'(\xi)|\cdot|x-y|\leq |x-y|$ (where $\xi$ is some element in $(x,y)$).

share|cite|improve this answer

Uniformly bounded: For any $n$, we have $$|g_n(x)|=\left|g_n(0)+\int_0^x g_n'(t)dt\right|\leq |g_n(0)|+\int_0^x|g_n'(t)|dt\leq1+\int_0^xdt = 1+x$$ establishing that the sequence is uniformly bounded.

Equicontinuous: Let $\epsilon>0$ and set $\delta=\epsilon$. If $|x-y|<\delta$, then $$|g_n(x)-g_n(y)|=\left|\int_x^yg_n'(t)dt\right|\leq \left|\int_x^y|g'(t)|dt\right|\leq \left|\int_x^ydt\right|=|x-y|<\delta=\epsilon$$ thus the sequence $(g_n)$ is equicontinuous.

share|cite|improve this answer
thanks i just realized that and apparently eddited my post as you were answering me – Galois Feb 21 '12 at 7:19
@Galois accidentally pressed enter too soon, which is why my answer was incomplete. I touched it up now and included equicontinuity. – Alex Becker Feb 21 '12 at 7:22
@PatrickDaSilva I forgot my apostrophes. – Alex Becker Feb 21 '12 at 7:25
But still, your proof is wrong ; the bound is incorrect. You need to compute $$ |g_n(x)| = |g_n(x) - g_n(0) + g_n(0)| = |\int_0^x g_n(t) \, dt + g_n(0)| $$ which leaves you with the correct upper bound $1+x$ when you replace $g_n(0)$ by $g_n'(0)$ which is bounded by $1$. – Patrick Da Silva Feb 21 '12 at 7:28
@PatrickDaSilva You're quite right, I misread the OP as saying $g_n(0)=0$. I'll fix that. – Alex Becker Feb 21 '12 at 7:32

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.