Vector-valued operators with singular kernel and Triebel-Lizorkin block spaces with variable exponents (Q272919)

A fundamental part of harmonic analysis consists of results establishing the boundedness of integral operators in relevant function spaces. Such results are often proven first in \(L^p\) spaces, then extended to a much larger class of spaces. For instance, one may want to consider the Morrey spaces used, e.g., in the calculus of variations, and defined by the norms \[ f\mapsto \sup\{\frac{1}{r^{\lambda}} \|1_{B(x,r)}f\|_{p} \;;\; x \in \mathbb{R}^{n}, r>0\} \] for \(\lambda \geq 0\), or their preduals (called block spaces). One often also needs to consider vector-valued generalisations (when functions take values in certain sequence spaces rather than \(\mathbb{C}\)), as well as many other kind of generalisations involving, for instance, variable exponent \(L^p\) spaces. A myriad of such generalisations have been proven by many authors. The paper under review shows that, as far as extensions from \(L^p\) spaces to Morrey and block spaces are concerned, these generalisations follow from a single principle. More precisely, the author establishes theorems that extend boundedness results from a given function space to its Morrey (or block space) counterpart. The framework is as follows. Start with a Banach function space \(X\), and a vector-valued extension \(X(\mathcal{B})\) (where \(\mathcal{B}\) is a family of sequence spaces). Consider a measurable weight \(u: \mathbb{R}^{n} \times (0,\infty) \to (0,\infty)\). One can then consider the corresponding Morrey space \(\mathcal{M}_{u}^{X}(\mathcal{B})\) defined by the norm \[ f\mapsto \sup \{ \frac{1}{u(x,r)} \|1_{B(x,r)}f\|_{X(\mathcal{B})} \;;\; x \in \mathbb{R}^{n}, r>0 \}. \] The main results of the paper under review (Theorems 3.1 and 3.2) show that, for a wide class of operators, boundedness on \(X(\mathcal{B})\) implies boundedness on \(\mathcal{M}_{u}^{X}(\mathcal{B})\) (or the corresponding block space), under mild assumptions on \(u,X,\mathcal{B}\). These results provide a satisfying unification of a range of special cases, and can also be used to approach new problems, such as the study of Triebel-Lizorkin block spaces with variable exponents considered in Section 4. These spaces correspond to the case where \(X\) is a variable exponent \(L^{p(.)}\) space, and \(\mathcal{B}\) is a variable exponent \(\ell^{q(.)}\) space with a variable weight. At this level of generality, some unusual behaviour can be observed, including a failure of the analogue of Fefferman-Stein's vector-valued maximal function inequalities (see [\textit{L. Diening} et al., J. Funct. Anal. 256, No. 6, 1731--1768 (2009; Zbl 1179.46028)]). Nevertheless, the author manages to use his main results to complement the results from [loc. cit.] by proving that the Triebel-Lizorkin block spaces with variable exponents are well defined (in the sense that they are independent of the choice of cut-off functions).