Limiting variants of Krasnosel'skiĭ's compact interpolation theorem (Q2253138): Difference between revisions

The authors present the answer, in some cases, to the old problem of compactness of a linear operator \(T\) on Lorentz-Zygmund spaces, regarded as interpolation spaces of a couple of some Lebesgue spaces, if the operator \(T\) is a compact operator on an end-point Lebesgue space. More precisely, they prove the following. Theorem: Let \((R, \mu)\) and \((S, \nu)\) be finite measure spaces, suppose that \(p_{0}, p_{1}, q_{0}, q_{1} \in [1, \infty]\) and that \[ T : L^{p_{0}} (R, \mu) + L^{p_{1}} (R, \mu) \to L^{q_{0}} (S, \nu) + L^{q_{1}} (S, \nu) \] is a linear map such that \[ T : L^{p_{0}} (R, \mu) \to L^{q_{0}} (S, \nu) \] is compact and \[ T : L^{p_{1}} (R, \mu) \to L^{q_{1}} (S, \nu) \] is bounded. If \(1\leq q <\infty \) and \( \alpha_{0} + 1/q > 0 > \alpha_{\infty} + 1/q,\) or \(q =\infty \) and \(\alpha_{0}\geq 0 \geq \alpha_{\infty} ,\) then the map \[ T : (L^{p_{0}} (R), L^{p_{1}} (R))_{0,q,(\alpha_{0} ,\alpha_{\infty} )} \to (L^{q_{0}} (S), L^{q_{1}} (S))_{0,q,(\alpha_{0} ,\alpha_{\infty} )} \] is compact. Here, the Lorentz-Zygmund spaces are spaces close to the Lebesgue spaces, but their definition is too technical to be given in this review. For the details, see the paper. In our opinion, this paper is very interesting mainly for the experts in interpolation theory of operators.