Bounding the Local Dimension of the Convolution of Measures
Kevin G. Hare, Joaquin G. Prandi
TL;DR
This work investigates the local dimension of the convolution of two finite measures on groups and the real line. The central contribution is a general upper bound: if the upper local dimension of $oldsymbol{ u}$ satisfies $ar{ ext{dim}}_{loc}oldsymbol{ u}(x) leq \lambda$ on the interior of its support, then for any regular measure $oldsymbol{ ho}$, $ar{ ext{dim}}_{loc}(oldsymbol{ u}*oldsymbol{ ho})(z) leq \lambda$ for all $z$ in the interior of the product of supports; this is established via a group-theoretic argument using sets $M_z$, $N_z$ and translation maps. The paper extends the result to the real line with gap conditions on the supports, analyzes local dimensions at special boundary points, and discusses sharpness and limitations through concrete examples. These results advance understanding of how convolution re-distributes mass and how interior points versus boundary points control local dimensions, with implications for multifractal analysis and potential generalizations to higher dimensions. The findings also raise open questions about weaker hypotheses for lower bounds and extensions to $ ext{R}^n$ where gap geometry becomes more complex.
Abstract
We study the local dimension of the convolution of two measures. We give conditions for bounding the local dimension of the convolution on the basis of the local dimension of one of them. Moreover, we give a formula for the local dimension of some special points in the support of the convolution.
