I have a question concerning the sequential weak-$*$ compactness in a Bochner space. Let $H$ be some non-separable Hilbert space. Consider the Bochner space $L^{\infty}(0,T;H)$. It is known that this space is isometrically isomorph to the dual of $L^1(0,T;H)$. Consequently, the Banach-Alaoglu theorem applies and yields the weak-$*$ compactness of a unit ball in $L^{\infty}(0,T;H)$. Now, since $L^1(0,T;H)$ is not separable (as $H$ is not), one cannot use the sequential version of the Banach-Alaoglu theorem in this case. But it seems that the conclusion must still be true in this case, i.e., a ball in $L^{\infty}(0,T;H)$ should be sequentially weak-$*$ compact. Here is my argumentation. Consider $L^2(0,T;H)$. This space is reflexive, so that Eberlein–Šmulian theorem applies. This result ensures in particular that each $L^{\infty}(0,T;H)$-bounded sequence $x_n$ has a subsequence $x_{n_k}$ which weakly converges to some $x\in L^2(0,T;H)$. But $L^2(0,T;H)$ is dense in $L^1(0,T;H)$, and the sequence is uniformly bounded in $(L^1(0,T;H))^*$. Therefore, due to the Banach-Steinhaus theorem the limit $x$ should actually belong to $(L^1(0,T;H))^*$ as well. Is this a correct argumentation? If yes, could somebody please recommend a suitable citation? I searched, e.g., in the book on vector measures by Diestel and Uhl, but could not find such a result.

Thanks a lot in advance!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.