Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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 am trying to comprehend the proof of a theorem in the calculus of variations (sketch: the functional $\int\limits_\Omega f(Du_j(x))~dx$ is weak*-sequentially semicontinuous if and only if $f$ is quasiconvex), but I am stuck on a (probably) simple argument which is not explained there.

Let $f:\mathbb{R}^{m} \rightarrow \mathbb{R}$ be continuous and let $u_j \rightarrow u \text{ weak}^*$ in $L^\infty(\Omega,\mathbb{R}^m)$, $\Omega$ a bounded Lipschitz-domain. Does the following hold: $f(u_j) \rightarrow f(u)$ $\text{weak}^*$ $L^\infty(\Omega)$?

Here I identfied $L^{\infty}(\Omega,\mathbb{R}^m)$ with the dual of $L^{1}(\Omega,\mathbb{R}^m)$.

Edit: I added "$\text{weak}^*$" in the last line.

share|cite|improve this question
If $u_j \in L^\infty(\Omega,\mathbb{R}^m)^*$, then $u_j : L^\infty(\Omega,\mathbb{R}^m) \to \mathbb{R}$. What does $f(u_j)$ mean? – copper.hat Nov 26 '12 at 19:16
$u_j \in L^\infty(\Omega,\mathbb{R}^m)$ - I changed the question to make this clear, too. – Ole Burghardt Nov 27 '12 at 12:09
Have you tried with $\Omega:=(0,1)$, $u_j(x):=\sin(j\pi x)$ and $f(x):=|x|$? – Davide Giraudo Nov 27 '12 at 19:52
Thanks for your suggestion, Davide. Actually it is a counterexample, see copper.hat's answer. – Ole Burghardt Dec 1 '12 at 0:03
up vote 0 down vote accepted

I think this is a counterexample. It hinges on my understanding of $f(u_k)$.

Suppose $\Omega=[0,1]$, $m=2$. Let $f(x) = \sqrt{x_1^2+x_2^2}$.

Let $v_k(t) = (\cos 2 \pi k t, \sin 2 \pi k t)$, and $u_k(x) = \int_{[0,1]} v_k(t)^T x(t) dt$, with $x \in L^1([0,1], \mathbb{R}^2)$.

I am presuming that $f(u_k)$ is defined as $f(u_k)(\eta) = \int_{[0,1]} f(v_k(t)) \eta(t) dt$, with $\eta \in L^1([0,1], \mathbb{R})$. Then $f(u_k)(\eta) = \int_{[0,1]} \eta(t) dt$, for all $k$, and $f(0) = 0$.

Then we have $u_k(x) \to 0$ for all $x$, but $f(u_k)(\eta) = \int_{[0,1]} \eta(t) \neq f(0)(\eta) = 0$.

share|cite|improve this answer
Thank you. So this is not why the argument in the proof holds. But I am confident of finding out how it does. – Ole Burghardt Nov 30 '12 at 23:58

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.