Localization and dimension free estimates for maximal functions (Q2437044)

\textit{A. Naor} and \textit{T. Tao} [J. Funct. Anal. 259, No. 3, 731--779 (2010; Zbl 1196.42018)] recently introduced a new class of measures with a so-called micro-doubling property and proved a localization principle for the associated maximal operators. This proof by Naor and Tao is probabilistic and relies on random martingales and Dood-type maximal inequalities. Consequently, Naor and Tao observed that in order to extend Stein's and Strömberg's '\(n\log n\)' bound to general measure spaces it suffices to assume that dilations by the factor \(\displaystyle \left(1+\frac{1}{n}\right)\) preserve essentially the volumes of balls and that the measures of intersecting balls with the same radius are comparable. The aim of the present paper is twofold. First, the two authors prove a new localization principle that considers localization not only in time but also in space. The proof is geometrical and is based on covering lemmas and selection processes. This principle may be used to prove the maximal theorem presented by Naor and Tao. Second, the authors restrict themselves to the Euclidean case and prove that a uniform condition for micro-doubling measures provides dimension free bounds for their maximal functions in all \(\displaystyle \mathbb{L}^p\), for \(p>1\). This is achieved by means of a new technique that allows differentiation through dimensions. In the last section, the authors provide some examples of doubling measures with associated maximal operators that fail to have uniform bounds.