Orateur
Anastasia Hraivoronska
Description
We introduce a fully discrete version of the Jordan–Kinderlehrer–Otto (JKO) scheme for gradient flows subject to hard constraints. We discuss the construction of the scheme and establish its convergence toward the corresponding evolution equations. The fully discrete formulation naturally leads to computational methods based on discrete optimal transport, for which we use the network simplex algorithm. We illustrate the approach through numerical simulations and applications to crowd motion, tumor growth, and specific cross-diffusion systems.