From Arzelà-Ascoli to Riesz-Kolmogorov (Q739431)

From MaRDI portal
scientific article
Language Label Description Also known as
English
From Arzelà-Ascoli to Riesz-Kolmogorov
scientific article

    Statements

    From Arzelà-Ascoli to Riesz-Kolmogorov (English)
    0 references
    0 references
    0 references
    18 August 2016
    0 references
    The study of compactness properties of subsets of Banach function spaces is a classical topic in functional analysis and lies at the center of some fundamental results in mathematical analysis. Let us mention for example the Dunford-Pettis theorem that relates uniform integrability and relative weak compactness for subsets of \(L^1\)-spaces. Extensions of this kind of results for \(L^p\)-spaces and in general for Banach function spaces, and also for spaces of vector-valued functions, have been exhaustively studied since this original result appeared, and is still a current topic. In general, the relation among limit properties involving integration and compactness for subsets of functions is central in mathematical analysis, both from the theoretical point of view and by its applications. In this paper, the authors center their attention on giving necessary and sufficient conditions for a subset of a Banach function space defined over a metric measure space to be totally bounded. They show a description of relatively compact sets using some previous results obtained by using the Arzelà-Ascoli theorem. Let \((X, \rho, \mu)\) be a metric measure space equipped with a metric \(\rho\) and a Borel regular measure \(\mu\). The measure \(\mu\) is said to be continuous with respect to the metric \(\rho\) if, for all \(x \in X\) and \(r > 0\), the following condition holds: \[ \lim_{y \to x} \mu(B(x, r) \Delta B(y, r)) = 0. \] The main result of the paper is given in Theorem 3.1, which states the following. Assume that \((X, \rho, \mu)\) is a proper metric measure space with continuous measure. Let \(E =E(X, \mu)\) be a Banach function space. If the family \(\mathcal F \subset E\) satisfies the following conditions: {\parindent=0.7cm\begin{itemize}\item[(a)] \(\mathcal F\) is bounded in \(E\), i.e., \(\sup_{f \in \mathcal F} \| f\|_E < \infty,\) \item[(b)] \(\lim_{r \to 0} \, \sup_{f \in \mathcal F} \big\| \frac{\int_{B(\cdot,r)} f d\mu}{\mu(B(\cdot,r))} - f \big\|_E =0,\) \item[(c)] for some \(x_0 \in X\), \(\lim_{R \to \infty} \sup_{f \in \mathcal F} \big\| f \chi_{X \setminus B(x_0,R)} \big\|_E=0\), \end{itemize}} then the family \(\mathcal F\) is totally bounded in \(E\). A sort of converse is given in Theorem 3.2, which however needs some extra limit properties and the boundedness of the associated maximal operator. Using these results, the authors provide some compactness criteria for several function spaces via Hausdorff's criterion -- a set is compact if and only if it is complete and totally bounded. This allows to obtain also a Kolmogorov-type compactness criterion for special Banach function spaces; for example, Theorem 4.2 provides such a result for variable exponent \(L^p\)-spaces. A partial converse of this result is given in Theorem 4.5, which is not correct as stated in the original paper. The reader can find a discussion and the correction in the note by the authors published in [ibid. 149, 177--179 (2017; Zbl 1362.46022)].
    0 references
    0 references
    Banach function spaces
    0 references
    metric measure spaces
    0 references
    Riesz-Kolmogorov theorem
    0 references
    grand variable exponent Lebesgue spaces
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references