Compound Poisson particle approximation for McKean-Vlasov

主题:Compound Poisson particle approximation for McKean-Vlasov

主讲人:张希承

时间:2023-10-21 09:00:00

地点:松江校区2号学院楼331室

组织单位:理学院统计系

主讲人中文简介:

张希承是北京理工大学数学与统计学院党委书记、教授、博士生导师,教育部新世纪优秀人才(2010)、德国洪堡学者(2006),2016年入选国家级人才。目前担任多个知名学术期刊的副主编和编委,主要从事随机分析及其相关领域问题研究,研究兴趣主要集中于非光滑随机流、随机微分方程、Levy过程与非局部算子分析、热核估计、动力学方程以及与随机方程的联系等方面。到目前为止发表学术论文一百多篇。曾主持国家自然科学基金重点项目、面上项目等多项。

活动内容摘要:

We propose a full discretization scheme for McKean-Vlasov SDEs driven by Brownian motions or alpha-stable processes, in terms of the compound Poisson particle approximations. The advantage of the scheme is that it simultaneously discretizes the time and space variables, and the approximation processes can be considered as a Markov chain with values in lattice. In particular, we show the propagation of chaos under quite weak assumptions on the coecients, where the coecients can be polynomial growth. Moreover, we study a functional CLT for the approximation of ODEs and the convergence of invariant measures for SDEs. Applications to 2D-NSE and SQG equation are also discussed.

主持人:闫理坦、胡良剑、张振中

撰写:李学元