Speaker
Description
There are several variants of the optimal matching on the unit cube differing by boundary conditions, e.g.\ periodic boundary conditions, free transport cost over the boundary or the classical Wasserstein distance on the cube. We show that for the matching problem w.r.t.\ squared distance cost between uniform points and the Lebesgue measure all these problems lead to the same asymptotic cost. More generally, for the transport problem between two random measures, for any p-th power of the cost function all but the classical Wasserstein distance lead to the same asymptotic constant.
This asymptotic constant can then be taken as a cost function for transport problems between point processes revealing some structural properties of point processes.
In this talk, we explain the equivalence of the asymptotic transport cost and review some of the recent results on point processes.
(Based on joint work with Michael Goldman, Thomas Leblé, Dario Trevisan, and Arthur Wachtel).