Paley type inequality on the Hardy type space in the Dunkl setting (Q707962)

The \textit{Dunkl transform} \({\mathcal F}_k\) on \({\mathbb R}^d\) is a generalization of the Fourier transform, which is \({\mathcal F}_0\). It is defined by \[ {\mathcal F}_k (f) (x) = \int_{{\mathbb R}^d} E_k (-ix, y)\, f (y)\,d\mu_k (y), \] where the \textit{Dunkl kernel} \(E_k : {\mathbb R}^d \times {\mathbb R}^d \to {\mathbb C}\) is the unique analytic solution, satisfying \(E_k (0, y) = 1\), of some e.d.p. associated to some \textit{reflection group} \(G\) of \({\mathbb R}^d\) generated by a finite \textit{root system} \(R \subseteq {\mathbb R} \setminus \{0\}\), and where \(k : R \to {\mathbb C}\) is a function constant on each orbit under the action of \(G\) (the \textit{multiplicity function}) and where \(d\mu_k (y) = w_k (y)\,dy\), the weight \(w_k\) depending on \(R\) and \(k\). In the paper under review, the author defines, for \(0 < p \leq 1\), an \textit{atomic Hardy space} \({\mathcal H}_{p, k}\) as a linear subspace of the tempered distributions and proves the following \textit{Paley type inequality}: \[ \int_{{\mathbb R}^d} | {\mathcal F}_k (f) (x) |^p \|x \|^{(2\gamma + d)(p - 1)}\, d\mu_k (x) \leq C\, \|f\|_{{\mathcal H}_{p, k}}^p \leqno (1) \] for every \(f \in {\mathcal H}_{p, k}\), where \(\gamma = \sum_{\alpha \in R_+} k (\alpha)\). In particular, for \(p = 1\), that gives a Hardy type inequality: \(\int_{{\mathbb R}^d} {| {\mathcal F}_k (f) (x) | \over \| x \|^{2\gamma + d}}\, d\mu_k (x) \leq C\, \|f\|_{{\mathcal H}_{1, k}}\) for \(f \in {\mathcal H}_{1, k}\). Actually, the author proves, using Marcinkiewicz's interpolation theorem, and the Plan\-cherel theorem for Dunkl transform proved by \textit{M. F. E. de Jeu} [Invent. Math. 113, No. 1, 147--162 (1993; Zbl 0789.33007)] that \((1)\) also holds for \(1 < p \leq 2\) for every \(f \in L^p ({\mathbb R}^d, \mu_k)\).