Boundedness properties in a family of weighted Morrey spaces with emphasis on power weights (Q2192351)

For \(p\in(0, \infty)\), \(\lambda_1, \lambda_2\in\mathbb R\), consider the Morrey space \(\mathcal L^{p,\lambda}(\omega)\) \(\lambda=(\lambda_1, \lambda_2)\), formed by the collections of all measurable functions \(f\) such that \[\|f\|_{\mathcal L^{p,\lambda}(\omega)}=\sup_{x\in\mathbb R^n, r>0}\left(\frac1{r^{\lambda_1}\omega(B(x,r))^{\lambda_2/n}} \int_{B(x,r)}|f|^p\omega\right)^{1/p}<\infty.\] This general formulation encompasses three interesting special cases of weighted Morrey spaces that have been considered in the literature: \par 1) the spaces introduced by \textit{Y. Komori} and \textit{S. Shirai} [Math. Nachr. 282, No. 2, 219--231 (2009; Zbl 1160.42008)] when \(\lambda_1=0\) and \(\lambda_2\in (0, n)\); \par 2) the spaces introduced by \textit{N. Samko} [J. Math. Anal. Appl. 350, No. 1, 56--72 (2009; Zbl 1155.42005)] when \(\lambda_1\in(0, n)\) and \(\lambda_2=0\); \par 3) the spaces considered by \textit{J. Poelhuis} and \textit{A. Torchinsky} [Trans. Am. Math. Soc. 367, No. 11, 7957--7998 (2015; Zbl 1361.42015)] when \(\lambda_1<0\) and \(\lambda_2=n\). Let \(p\in(1, \infty)\), \(\omega\in A_p\), \(\lambda_2\in[0, n]\) with \(\lambda_1\sigma^\prime_\omega+\lambda_2>0\), where \(\sigma=\sup\{\sigma: \omega\in RH_\sigma\}\) and \(RH_\sigma\) is the reverse Hölder class. In this article, the authors show that the Hardy-Littlewood maximal operator is bounded on \(\mathcal L^{p,\lambda}(|x|^\alpha \omega)\) for \(\lambda_1\sigma^\prime_\omega+\lambda_2<n\) and \(\alpha\) with \[ 0\le \alpha< \left\{\begin{array}{ll}\frac n{n-\lambda_2}(\lambda_1+\lambda_2\theta), & \lambda_2<n \\ \infty, & \lambda_2\ge n.\end{array}\right. \] Here \(\theta\) satisfies that for all \(0<r<R\), \[\frac{\omega(B(0, r))}{\omega(B(0, R))}\le c\big(\frac rR\big)^{n\theta}.\] Moreover, the authors use extrapolation techniques to obtain abstract results in the class of weighted Morrey spaces: let \(p_0\in[1, \infty)\) and \(\mathcal F\) be a collection of non-negative measurable pairs of functions. Assume that for every \((f,g)\in \mathcal F\) and every \(u\in A_{p_0},\) one has \[\|g\|_{L^{p_0}(u)}\le C \|f\|_{L^{p_0}(u)},\] where \(C\) does not depend on the pair \((f,g)\) and it depends on \(u\) only in terms of \([u]_{A_{p_0}}\). Then for \(p\in(1, \infty)\) and \(\omega\in A_p\), it holds \[\|g\|_{\mathcal L^{p, \lambda}(|x|^\alpha \omega)}\le C \|f\|_{\mathcal L^{p, \lambda}(|x|^\alpha \omega)}\] for \(\lambda_2\in[0, n]\), \(0<\lambda_1\sigma^\prime_\omega+\lambda_2<n\) and \(\alpha\) as stated above. Special cases are also studied.