学术活动
学术活动名称:The metric nature of Hamilton-Jacobi equation in space of probability measures
2014-06-30
来源:科技处 点击次数:主讲人:Jin Feng 教授(Department of Mathematics, University of Kansas,Lawrence, KS 66049)
时 间:2014年6月30日(周一)10:00-11:00
地 点:必赢76net线路官网北二区教学楼517教室
摘 要:In this part of the talk, I will use metric space analysis to formulate an abstract
well-posedness theory for Hamilton-Jacobi equations in metric space with geodesic
property.Then, I will apply these results to an example in the Wasserstein space. The Wasserstein space example is associated with
variational characterization of compressible Euler equation. The problem had been open for some time. A key, and new, step
is realization of metric nature of the problem. This lead to the use of geometric tangent cones, instead of topological linear tangent
spaces, to define Hamilton-Jacobi operator. Elements in the geometric tangent cone can be identified with bi-variate probability measures. Some subtle convex analysis techniques in such context is needed and I will go over them.
Probabilistic coupling technique plays an important role in this regard.This is a joint work with Luigi Ambrosio in Scuola Normale
Superiori di Pisa, Italy.