Maximal Averages on the Affine Group $G_n$ and applications

Abstract

Let \(G_n=\mathbb R^n\rtimes\mathbb R_+\) be equipped with the left Haar measure \( d\mu(x,y)=\frac{dx\,dy}{y^{n+1}}. \) We study maximal averages associated with three basic motions on \(G_n\): horizontal translations, vertical dilations, and fixed hyperbolic geodesics in the upper half-space model. The translation maximal operator is the Euclidean Hardy--Littlewood maximal operator on each horizontal slice. The Haar-compatible dilation maximal operator is of weak type \((1,1)\) and bounded on \(L^p(G_n)\) for \(1<p\le\infty\), but it is not strongly bounded on \(L^1(G_n)\). By contrast, the unweighted Lebesgue dilation average is unbounded on every finite \(L^p(G_n)\) and is not of weak type \((1,1)\). For fixed hyperbolic geodesic averages, the large-time part is strongly bounded on \(L^1(G_n)\) because of modular exponential decay. The small-time part is a finite-type parabolic maximal problem. Using the corresponding local finite-type \(L\log\log L\) endpoint estimate for the geodesic slice, we prove \( \mathcal M_{\gamma_\omega}:L\log\log L(G_n) \longrightarrow L^{1,\infty}(G_n) \) in weak Orlicz form, together with the strong \(L^p(G_n)\) bounds for \(1<p\le\infty\). We also show that the strong \(L^1\) endpoint fails. Finally, we record a discrete random-walk maximal inequality whose sufficient condition is expressed through the modular drift $$ \rho_p(\sigma)=\int_{G_n}y(h)^{n/p}\,d\sigma(h), $$ where \(\sigma\) is the probability measure defining the right random walk.