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

I don't understand the last line of a proof (which is supposed to be obvious...), could you help me?

The context is the following. We have a bounded open set $U$ of $\mathbb{R}^m$, and a $C^\infty$-mapping $$F : \mathbb{R}^m \times \mathbb{R} \times U \to \mathbb{R}^m$$ We also have a sequence $(u_n)_n$ of $W^{1,p}(U)$ ($p > 1$) and $u \in W^{1,p}(U)$ such that $u_n$ converges uniformly to $u$ and $\nabla u_n$ converges weakly in $L^p(U,\mathbb{R}^m)$ to $\nabla u$. Last, we know that $|u| + |\nabla u| < M$ for some $M < +\infty$.

The first thing said is that $F(\nabla u, u_n, .)$ converges uniformly to $F(\nabla u, u, .)$ on $U$. I have the intuition that it's true but I don't know how to prove it formally. Is it just obvious from some sort of theorem about function composition? I suppose I have to use the condition $|u| + |\nabla u| < M$ somewhere...

Secondly, $$\displaystyle\lim\limits_{n\to\infty}\int_U F(\nabla u(x), u_n(x),x)\cdot (\nabla u_n(x) - \nabla u(x)) dx = 0$$(We know that the definite integral exists and is finite for each $n$). They said that "weak convergence is, by its very definition, compatible with linear expressions, and so the limit holds", but I don't understand this statement...

Thank you in advance if you can explain me one of the two affirmations (and sorry for my not perfect english) :). I perhaps have forgotten some hypothesis (even if I hope not), so don't hesitate if you think there is something missing or unclear.

share|cite|improve this question
up vote 3 down vote accepted

To conclude with $F(\nabla u(x),u_n(x),x)\to F(\nabla u(x),u(x),x)$ uniformly, we want $F$ to be uniformly continuous on the set $\{p\in \mathbb R^n: |p|\le M\}\times [-M,M]\times U$. Unfortunately, this set is not compact because of the factor $U$. We can get uniform convergence on any compact subset of $U$, but to get it on all of $U$ an additional hypothesis is needed, such as $F$ being continuous on $\mathbb R^m\times\mathbb R\times \overline{U}$.

The second question is easier to answer. Here is a more general version: if $f_n\to f$ strongly in $L^q(U)$ and $g_n\rightharpoonup g$ weakly in $L^p(U)$ (here $p^{-1}+q^{-1}=1$ as usual), then $\int_U f_ng_n\to \int_U fg$. Proof:

  • $\int_U f(g_n-g)\to 0$ by the definition of weak convergence.
  • $\int_U (f_n-f)g_n \le \|f_n-f\|_q \|g_n\|_p\to 0$, using Hölder's inequality and the fact that weakly convergent sequences are bounded
    • Adding the two, we get $\int_U(f_ng_n-fg)\to 0$. $\quad\Box$

Of course, uniform convergence on a bounded set implies strong convergence in every $L^q$.

share|cite|improve this answer
Oh thank you very very much! I wasn't expecting a so quick and precise answer. You are totally right for the first point, I was wrong with the hypothesis on $F$, it is actually $C^\infty$ on $\mathbb{R}^m \times \mathbb{R} \times \overline{U}$ which gives $F$ uniformly continuous on the set you mention. And your proof is clear for the second question, a reasoning in functional analysis is rarely so clear to me :) – Nicolas Dec 21 '12 at 8:26

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.