Paperlane

no. 13 · Wednesday 11 November 2026

Energy Dissipation Along Trajectories and Wasserstein Gradient Flow of Aggregation-Diffusion Equations

Honglei Chen and Xuewei Wang

math.PR

Abstract

In this work, we established the Wasserstein gradient flow structure of nonlinear Fokker-Planck equations on bounded domains involving both nonlinear diffusion and nonlocal interaction effects. Our approach is based on stochastic analysis. More precisely, we investigated the associated stochastic differential equation (SDE) with normal reflection at the boundary from a trajectorial perspective and derived a semimartingale decomposition of the corresponding energy process. This decomposition leads to a trajectorial energy dissipation identity, which, together with the characterization of the metric derivative, allows us to establish the Wasserstein gradient flow structure of the aggregation-diffusion equation.

fields
Probability, Analysis of PDEs, Dynamical Systems
pdf ↗ page arXiv:2611.14054 ↗