Convergence in Wasserstein Distance for Empirical Measures of Dirichlet Diffusion Processes on Manifolds

preprint2022arXivOpen access

Abstract

Let MM be a dd-dimensional connected compact Riemannian manifold with boundary ∂M\partial M, let V∈C2(M)V\in C^2(M) such that μ(dx):=eV(x)dxμ({\rm d} x):={\rm e}^{V(x)}{\rm d} x is a probability measure, and let XtX_t be the diffusion process generated by L:=Δ+∇VL:=Δ+\nabla V with τ:=inf⁡{t≥0:Xt∈∂M}τ:=\inf\{t\ge 0: X_t\in\partial M\}. Consider the empirical measure μt:=1t∫0tδXsdsμ_t:=\frac 1 t \int_0^t δ_{X_s}{\rm d} s under the condition t<τt<τ for the diffusion process. If d≤3d\le 3, then for any initial distribution not fully supported on ∂M\partial M, \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 c∈(0,1]c\in (0,1] with c=1c=1 when ∂M\partial M is convex, where μ0:=ϕ02μμ_0:= ϕ_0^2μ for the first Dirichet eigenfunction ϕ0ϕ_0 of LL, {λm}m≥0\{λ_m\}_{m\ge 0} are the Dirichlet eigenvalues of −L-L listed in the increasing order counting multiplicities, and the upper bound is finite if and only if d≤3d\le 3. When d=4d=4, sup⁡T≥tE[W2(μt,μ0)2∣T<τ]\sup_{T\ge t} \mathbb E\big[\mathbb W_2(μ_t, μ_0)^2\big|T<τ\big] decays in the order t−1log⁡tt^{-1}\log t, while for d≥5d\ge 5 it behaves like t−2d−2t^{-\frac 2 {d-2}}, as t→∞t\to\infty.

Explore this paper’s authors, topics and related work.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Reviews 0

Write a reviewWrite

No reviews yet.

Discussion 0

Add a commentWrite

No comments yet.