1
$\begingroup$

The following Theorem was described in a text I was reading as a compactness result. The proof is probably too advanced for me but I was just wondering with respect to what topology we have compactness or relatively compactness here. The first one that came to my mind is the product topology $X^{(0,T)}$ where $X$ is endowed with the weak convergence topology. However in this topology, convergence of the $f_{j_k}$ would mean $$\langle \psi,f_{j_k}(t)-f(t)\rangle\to 0\text{ when }k\to\infty\text{ for all }\psi\in X'\text{(dual of $X$) and all }t\in[0,T]$$ so it seems to be another topology that we consider.

Following is the theorem I mention...

Theorem : Let $X$ be a reflexive and separable Banach space and $T\in\mathbb{R}_{\ge 0}$. Let $(f_j)_{ j\in\mathbb{N}}$ be a sequence of functions in $L^q((0,T),X)$ with $1 \le q < +\infty$. Let us also assume that there exists a map $m ∈ L_q (0,T)$ such that $\|f_j(t)\|_X\le m(t)$ for almost all $t ∈ [0,T]$. Then there exist a subsequence $(f_{j_k} )_{k\in\mathbb{N}}$ and a function $f\in L^q((0, T ), X)$ such that $$\lim\limits_{k\to\infty}\int_0^T\langle\phi(t),f_{j_k}(t,\cdot)-f(t,\cdot)\rangle dt=0$$ for all $\phi ∈ L^{q'} ((0, T ), X' )$, with $q'$ the conjugate exponent of $q$, and $$\int_{t_1}^{t_2}f_{j_k}(t)dt\stackrel{}{\rightharpoonup}\int_{t_1}^{t_2}f(t)dt,\quad\text{ when }k\to\infty\text{ for all } t_1,t_2\in[0,T]$$ where "$\stackrel{}{\rightharpoonup}$" denotes the weak convergence in $X$ and the integrals are in the sense of Bochner.

$\endgroup$
8
  • 3
    $\begingroup$ I don't know if that's the case here, but it's not uncommon to use the words "compactness result" for something that just feels like some kind of compactness (some sequence has a subsequence such that ...), without necessarily being (obviously) equivalent to a statement: the subset $A$ of the topological space $\mathcal T$ is compact. $\endgroup$ Commented Apr 22, 2021 at 17:38
  • 6
    $\begingroup$ For this to make sense, shouldn't $\phi$ take values in $X'$, not $X$? In that case, this looks like weak sequential compactness if $L^{q'}((0,T),X')$ is the dual of $L^q((0,T),X)$, which I think is true under some conditions that I forget. $\endgroup$ Commented Apr 22, 2021 at 17:47
  • $\begingroup$ Yes, I edited thank you $\endgroup$
    – edamondo
    Commented Apr 22, 2021 at 18:01
  • 1
    $\begingroup$ @edamondo: A reference for the duality in case that $X$ is reflexive is, for instance, Corollary 1.3.22 in "Hytönen, van Neerven, Veraar, Weis: Analysis in Banach Spaces, Volume I (2016)". In fact, what one needs is not necessarily reflexivity of $X$, but the weaker property that the dual space $X'$ have the Radon-Nikodým property; see [op. cit., Def 1.3.9, Thms 1.3.10 and 1.3.26, and Def 1.3.27]. $\endgroup$ Commented Apr 23, 2021 at 0:55
  • 3
    $\begingroup$ @edamondo: In addition to Jochen's hint, the classic "Vector Measures" by Diestel and Uhl is a source to look at. $\endgroup$ Commented Apr 23, 2021 at 19:15

0

You must log in to answer this question.