Winter term 2017/18 Lecture Optimal Transport and Wasserstein Gradient Flows
Description: The optimal transport problem was already formulated by Gaspard Monge in the 18 century. It deals with the relocation of an initial distribution of mass to a final distribution, such that the cost of transport is minimal. The formulation of this problem was generalized by Kantorovich in 1942. Besides the original applications in economy, new connections to Problems in geometry, probability theory, and analysis emerged. In particular, in the recent decades a strong connection between partial differential equations, that describe diffusion processes, could be made. These diffusion problems can be formulated as gradient flows of the system's entropy and the so-called Wasserstein distance.
In this module, we introduce the problem of optimal transport, discuss basic results and applications: Monge- and Kantorovich formulation, existence of optimal transport plans, dual formulation, dynamical formulation, diffusion equations as Wasserstein gradient flows.