I'm having trouble filling in the details of a proof in Lions' book Quelques méthodes de résolution des problèmes aux limites non linéaires. He constructs approximate solutions to the wave equation on a bounded open set $\Omega\subset\mathbb{R}^n$ which satisfy the following conditions: \begin{align} u_n \rightarrow u \qquad&\text{weak-star in $L^\infty(0,T;H_0^1(\Omega)\cap L^p(\Omega))$} \\ u_n' \rightarrow u' \qquad&\text{weak-star in $L^\infty(0,T;L^2(\Omega))$} \\ \end{align} where $u_n(0)\rightarrow u_0$ strongly in $H_0^1(\Omega)\cap L^p(\Omega))$. (I have omitted some details to shorten the setup.)

I am having trouble with the claim on p. 14 that $u_n(0) \rightarrow u(0)$ weakly in $L^2(\Omega)$. Since, in particular, $u_n(0)\rightarrow u_0$ strongly in $L^2(\Omega)$, it follows that $u(0)=u_0$.


First note that by interpolation, $u_n, u \in C([0,T];L^2(\Omega))$ for all $n\in\mathbb{N}$. Thus we interpret $u(0)$ as an $L^2$ limit, same for $u_n(0)$.

Let $\phi \in L^2(\Omega)$ be an arbitrary function and denote $$ f_n(t) = \int_\Omega \phi (u_n-u) \, dx $$ and $\psi_m(t) = m*\chi_{[0,1/m]}(t)$. Then $$ |f_n(0)| \leq \left|f_n(0) - \int_0^T \psi_m \, f_n(t) \, dt \right| + \left|\int_0^T \psi_m \, f_n(t) \, dt \right|. $$ For fixed $n$, the first term can be made small by the Lebesgue differentiation theorem (and noting the continuity of f). For fixed $m$, the second term can be made small by the weak convergence statement. But I couldn't get a diagonal sequence to conclude $f_n(0) \rightarrow 0$ as $n\rightarrow\infty$.

  • $\begingroup$ Don't you have $u_n\to u$ in $L^\infty(0;T, L^2(\Omega))$ strongly? $\endgroup$ – daw Nov 25 '16 at 16:57
  • $\begingroup$ By Aubin-Lions? Yes. $\endgroup$ – dls Nov 25 '16 at 18:06
  • $\begingroup$ Then the second integral can be made small for large $n$ independent of $m$ $\endgroup$ – daw Nov 25 '16 at 19:22

By Aubin--Lions you have that $u_n\to u$ strongly in $L^\infty(0,T;L^2(\Omega))$, hence the second integral can be made arbitrarily small for large $n$ independent of $m$, as $f_n\to0$ in $L^\infty(0,T;L^2(\Omega))$.


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.