On-diagonal lower bounds for heat kernels and Markov chains (Q1365165): Difference between revisions

The authors establish estimates for on-diagonal lower bound for heat kernels on Riemannian manifolds of two kinds: (i) a suplower bound (i.e., a lower bound for \(\sup_{x\in M}p_t(x,x)\) where \(p_t(x,y)\) is the heat kernel on a manifold \(M\)); (ii) a pointwise lower bound (i.e., a lower bound for \(p_t(x,x)\) for a fixed given point \(x\in M\)). There are also two kinds of assumptions about the geometry of the manifold (considered alternatively): (i) an anti-isoperimetric (or anti--Faber-Krahn-type) inequality; (ii) an upper bound for the volume growth function or a doubling volume property. The authors' approach to lower bounds of heat kernels is independent of known ones for upper bounds and it works even if the Harnack principle is unknown or false. The approach involves, in particular, random walks on graphs and other probabilistic ideas.