On a Muckenhoupt-type condition for Morrey spaces (Q1949858): Difference between revisions

This paper considers the description of the admissible weights for Morrey spaces. It is known that the classical Muckenhoupt class \(A_p\) may not be an appropriate class of weights ensuring the boundedness of maximal and/or singular integrals on weighted Morrey spaces. The description of the corresponding admissible weights \(w\) for Morrey spaces is still an open problem. Let \(\Omega\subset \mathbb{R}^n\), \(p\in[1,\infty)\), \(\lambda\in[0,1]\), and let \(L^{p,\lambda}(\Omega, w)\) denote the weighted Morrey space normed by \[ \|f\|_{p,\lambda;w}:=\sup_{x\in \Omega; r>0}\left(\frac1{|B(x,r)|^\lambda}\int_{B(x,r)}|f(y)|^p w(y)\,dy\right)^{1/p}, \] where \(f\) is assumed to be continuous beyond \(\Omega\) as zero whenever necessary. Obviously, \(L^{p,\lambda}(\Omega, 1)\) is just the classical Morrey space \(L^{p,\lambda}(\Omega)\). Instead of the priori assumption \(w, w^{1-p'}\in L^1_{\text{loc}}(\mathbb{R}^n)\) in the definition of \(A_p\), in this paper the author first gives some \(\lambda\)-dependent a priori assumptions on the admissible weights \(w\) for Morrey spaces as follows: \[ \chi_Bw^{1/p}\in L^{p,\lambda}(\Omega) \quad\text{and}\quad \chi_Bw^{-1/(\lambda+p-1)}\in L^{p,\lambda}(\Omega,w)\quad \text{for all balls}\;B, \] which reverts to \(w, w^{1-p'}\in L^1_{\text{loc}}(\mathbb{R}^n)\) when \(\lambda=0\). A weight function \(w\) is called a \((p,\lambda)\)-admissible weight if it satisfies such a priori assumptions. The author then defines the class \(\mathcal{A}_{p,\lambda}\) as the set of all \((p,\lambda)\)-admissible weights \(w\) such that \[ \sup_B \frac{\|\chi_B\|_{p,\lambda;w}}{\|\chi_B\|_{p,\lambda;w^\ast}} \left(\int_B w^{-\frac1{\lambda+p-1}}\,dy\right)<\infty,\quad w^\ast:=w^{-\frac{1-\lambda}{\lambda+p-1}}, \] where the supremum is taken with respect to all balls \(B\). This class \(\mathcal{A}_{p,\lambda}\) coincides with \(A_p\) when \(\lambda=0\). In the one-dimensional case, the author proves that the condition \(w\in \mathcal{A}_{p,\lambda}\) is necessary for the boundedness of the Hilbert transform in the weighted Morrey space \(L^{p,\lambda}(\mathbb{R}, w)\).