0
$\begingroup$

Let $\Omega$ be a bounded domain of $\mathbb{R}^n$ and $u_n \colon [0,T] \to L^2(\Omega)$ a sequence of functions such that $||u_n(t)||_{L^2(\Omega)} \leq C$ for every $n$ and for every $t \in [0,T]$.

We know that $u_n \rightharpoonup^* u$ in $L^\infty(0,T;L^2(\Omega))$ to some function $u \in L^\infty(0,T;L^2(\Omega))$, i.e. $$ \int_0^T \int_\Omega u_n(t,x) v(t,x) \,dxdt \to \int_0^T \int_\Omega u(t,x) v(t,x) \,dxdt \quad \text{for all } v \in L^1(0,T;L^2(\Omega)). $$ At the same time, for each fixed $t \in [0,T]$, we have $u_n(t) \rightharpoonup w^t$ in $L^2(\Omega)$ to some function $w^t \in L^2(\Omega)$, i.e. $$ \int_\Omega u_n(t,x)v(x)\,dx \to \int_\Omega w^t(x)v(x)\,dx \quad \text{for all } v \in L^2(\Omega). $$

Is it true that $u(t) = w^t$ for almost every $t \in [0,T]$?

$\endgroup$
0
$\begingroup$

Define the sequence $f_n(t)=\sin(n \pi t)$. Then it holds $f_n \rightharpoonup^*0$ in $L^\infty(0,1)^*$. The sequence converges pointwise to zero only on a set of zero measure. Now multiply $f_n$ with a not-null $L^2(\Omega)$ function $g$. Then $f_ng$ converges weak star to $0$. It converges to zero only on a set of zero measure.

$\endgroup$
  • $\begingroup$ Thanks @daw So loosely speaking, taking $v \in L^2(\Omega)$, what prevents me from doing something like $\lim_n \int_0^T \int_\Omega u_n v \,dx\,dt= \int_0^T \lim_n \int_\Omega u_n(t) v\,dx\,dt = \int_0^T \int_\Omega w^t v \,dx\,dt$ is the fact that the sequence $F_n(t) = \int_\Omega u_n(t) v \,dx$ has no pointwise limit right? So in general I need some stronger control on $\{F_n\}_n$... $\endgroup$ – Mauro Feb 6 '18 at 8:54
  • $\begingroup$ Yes, weakly converging sequences do not need to converge pointwise. $\endgroup$ – daw Feb 6 '18 at 9:19

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.