On the conformal walk dimension: Quasisymmetric uniformization for symmetric diffusions

TL;DR

This paper introduces the conformal walk dimension, linking elliptic and parabolic inequalities, and characterizes when it is finite and equal to two for certain symmetric diffusion spaces, with applications to fractals.

Contribution

It defines the conformal walk dimension for symmetric Dirichlet spaces and establishes conditions for its finiteness and attainment, especially on self-similar fractals.

Findings

Conformal walk dimension is two for spaces with metric doubling and elliptic Harnack inequality.

Necessary conditions are provided for the attainment of the infimum of the conformal walk dimension.

The infimum is not attained for Vicsek set and higher-dimensional Sierpiński gaskets, but is for the 2D Sierpiński gasket.

Abstract

We introduce the notion of conformal walk dimension, which serves as a bridge between elliptic and parabolic Harnack inequalities. The importance of this notion is due to the fact that, for a given strongly local, regular symmetric Dirichlet space in which every metric ball has compact closure (MMD space), the finiteness of the conformal walk dimension characterizes the conjunction of the metric doubling property and the elliptic Harnack inequality. Roughly speaking, the conformal walk dimension of an MMD space is defined as the infimum over all possible values of the walk dimension with which the parabolic Harnack inequality can be made to hold by a time change of the associated diffusion and by a quasisymmetric change of the metric. We show that the conformal walk dimension of any MMD space satisfying the metric doubling property and the elliptic Harnack inequality is two, and provide a necessary condition for a pair of such changes to attain the infimum defining the conformal walk dimension when it is attained by the original pair. We also prove a necessary condition for the existence of such a pair attaining the infimum in the setting of a self-similar Dirichlet form on a self-similar set, and apply it to show that the infimum fails to be attained for the Vicsek set and the $N$-dimensional Sierpi\'{n}ski gasket with $N\geq 3$, in contrast to the attainment for the two-dimensional Sierpi\'{n}ski gasket due to Kigami [Math. Ann. 340 (2008), no. 4, 781--804].