Random martingales and localization of maximal inequalities (Q982510): Difference between revisions

A metric measure space \((X,d, \mu)\) is a separable metric space \((X,d)\), equipped with a Radon measure \(\mu\). The Hardy-Littlewood maximal function is defined by \[ Mf(x) \equiv \sup_{r>0} \frac{1}{\mu (B(x,r))} \int_{B(x,r)} | f(y) | d\mu, \] where \(B(x,r) = \{ y \in X: d(x,y) \leq r \}.\) In most cases of interest, it is difficult to compute the weak \((1,1)\) operator norm \(\| M \|_{L^1(X) \to L^{1, \infty}(X)} \) exactly; a notable exception to this statement is the real line \({\mathbb R}\), equipped with the usual metric and Lebesgue measure [see \textit{A. D. Melas}, Ann. Math. (2) 157, No. 2, 647--688 (2003; Zbl 1055.42013)]. Many studies have been made to estimate the asymptotic behavior of the various operator norms of maximal functions. The authors introduce a variant of maximal function and \textit{microdoubling condition}. For \(R \subseteq (0, \infty )\), let \[ M_R f(x) \equiv \sup_{r \in R} \frac{1}{\mu (B(x,r))} \int_{B(x,r)} | f(y) | d\mu. \] We say that \((X,d, \mu)\) is \(n\)-microdoubling if there exists a constant \(K\) such that \[ \mu \left( B \left( x, \left( 1 + \frac{1}{n} \right) r \right) \right) \leq K \mu (B(x,r)). \] They prove that \(M_R\) has the following localization property: \[ \| M_R \|_{L^1(X) \to L^{1, \infty}(X)} \approx \sup_{r>0} \| M_{R \cap [r, nr]} \|_{L^1(X) \to L^{1, \infty}(X)} \] if \((X,d, \mu)\) is \(n\)-microdoubling for some \(n>1\). As a consequence of this result they obtain the next result. If \((X,d, \mu )\) is Ahlfors-David \(n\)-regular where \(n \geq 2\), then \[ \| M \|_{L^1(X) \to L^{1, \infty}(X)} = O(n \log n). \] This generalizes a result of \textit{E. M. Stein} and \textit{J.-O. Strömberg} [Ark. Mat. 21, 259-269 (1983; Zbl 0537.42018)].