Convergence in Wasserstein Distance for Empirical Measures of Dirichlet Diffusion Processes on Manifolds
arXiv:2005.09290
Abstract
Let be a -dimensional connected compact Riemannian manifold with boundary , let such that is a probability measure, and let be the diffusion process generated by with . Consider the empirical measure under the condition for the diffusion process. If , then for any initial distribution not fully supported on , \begin{align*} &c\sum_{m=1}^\infty \frac{2}{(λ_m-λ_0)^2} \le \liminf_{t\to \infty} \inf_{T\ge t} \Big\{t {\mathbb E}\big[\mathbb W_2(μ_t, μ_0)^2\big|T<τ\big]\Big\} \\ &\le \limsup_{t\to \infty} \sup_{T\ge t} \Big\{ t \mathbb E\big[\mathbb W_2(μ_t, μ_0)^2\big|T<τ\big] \Big\}\le \sum_{m=1}^\infty \frac{2}{(λ_m-λ_0)^2}\end{align*} holds for some constant with when is convex, where for the first Dirichet eigenfunction of , are the Dirichlet eigenvalues of listed in the increasing order counting multiplicities, and the upper bound is finite if and only if . When , decays in the order , while for it behaves like , as .
27 pages