Global heat kernel estimates for symmetric Markov processes dominated by stable-like processes in exterior \(C^{1,\eta}\) open sets (Q494703)

This is a rather technical paper about bilateral estimates on transition densities (heat kernels) of a class of symmetric, pure jump Markov processes. As a matter of fact, besides a few well-known examples, it is in general quite difficult to extract a closed, exact form for these transition densities, and hence sharp estimates are always welcome, as actually pointed out by the quoted literature. The class of Markov processes involved is defined starting with the Hunt processes \(Y\) associated to the regular Dirichlet forms established through a family of symmetric, measurable, wisely chosen functions \(J(x, y)\), and then by taking the process \(Y^D\) which is just \(Y\) killed upon exiting an open set \(D\). The aim of the paper -- also extending previous results of the same author -- is to establish suitable, two-sided estimates for the transition density \(p_D(t, x, y)\) of \(Y_D\) when \(D\) is an exterior \(C^{1,\eta}\) open set. It turns out (in the main results summarized in Theorem 1.2) that the upper and lower bounds are implemented by the kernels of stable-like processes with an impressive panoply of parameters. A few corollaries conclude the paper, also adding an estimate on the Green function \(G_D(x, y)\) of \(Y^D\).