The back-and-forth method for Wasserstein gradient flows
Numerical Analysis
2020-11-17 v1 Numerical Analysis
Analysis of PDEs
Abstract
We present a method to efficiently compute Wasserstein gradient flows. Our approach is based on a generalization of the back-and-forth method (BFM) introduced by Jacobs and L\'eger to solve optimal transport problems. We evolve the gradient flow by solving the dual problem to the JKO scheme. In general, the dual problem is much better behaved than the primal problem. This allows us to efficiently run large-scale simulations for a large class of internal energies including singular and non-convex energies.
Keywords
Cite
@article{arxiv.2011.08151,
title = {The back-and-forth method for Wasserstein gradient flows},
author = {Matt Jacobs and Wonjun Lee and Flavien Léger},
journal= {arXiv preprint arXiv:2011.08151},
year = {2020}
}