Convergence to global equilibrium for Fokker-Planck equations on a graph and Talagrand-type inequalities

TL;DR

This paper investigates the exponential convergence of Fokker-Planck equations on finite graphs to a global equilibrium, establishing bounds based on spectral gap and Sobolev constants, and deriving Talagrand-type inequalities.

Contribution

It provides the first analysis of convergence rates for Fokker-Planck equations on graphs, linking spectral properties to convergence speed and inequalities.

Findings

Convergence to equilibrium is exponential with rate bounds.

Decay rates relate to spectral gap and Sobolev constants.

Established Talagrand-type inequalities for graph-based probability measures.

Abstract

In recent work, Chow, Huang, Li and Zhou introduced the study of Fokker-Planck equations for a free energy function defined on a finite graph. When $N\ge 2$ is the number of vertices of the graph, they show that the corresponding Fokker-Planck equation is a system of $N$ nonlinear ordinary differential equations defined on a Riemannian manifold of probability distributions. The different choices for inner products on the space of probability distributions result in different Fokker-Planck equations for the same process. Each of these Fokker-Planck equations has a unique global equilibrium, which is a Gibbs distribution. In this paper we study the {\em speed of convergence} towards global equilibrium for the solution of these Fokker-Planck equations on a graph, and prove that the convergence is indeed exponential. The rate as measured by the decay of the $L_2$ norm can be bound in terms of the spectral gap of the Laplacian of the graph, and as measured by the decay of (relative) entropy be bound using the modified logarithmic Sobolev constant of the graph. With the convergence result, we also prove two Talagrand-type inequalities relating relative entropy and Wasserstein metric, based on two different metrics introduced in [CHLZ] The first one is a local inequality, while the second is a global inequality with respect to the "lower bound metric" from [CHLZ].