Approximate Wasserstein attraction flows for dynamic mass transport over networks

This paper presents a Wasserstein attraction approach for solving dynamic mass transport problems over networks. In the problem networks, we start with distribution set of nodes that needs to be “transported” target accounting network topology. We exploit specific structure problem, characterized by computation implicit gradient steps, and formulate an based on discretized flows. As result, our proposed algorithm relies iterative constrained barycenters. show how method finds approximate solutions taking into account topology network, capacity communication channels, individual nodes. Finally, performance this applied large-scale water transportation