活动时间：2023-06-27 14:00

活动地点：二号学院楼331

主讲人：李运章

主讲人中文简介：

李运章，复旦大学智能复杂体系实验室青年副研究员。2020年博士毕业于复旦大学数学科学学院。2020年至2022年在复旦大学从事博士后研究工作，期间被聘为香港中文大学名誉博士后。主要研究领域为随机系统的最优控制问题的高阶精度数值算法，相关成果发表于SIAM J. Control. Optim., SIAM J. Sci. Comput., SIAM J. Financial Math., ESAIM: M2AN等知名学术期刊。入选上海市晨光计划，国家博士后创新人才支持计划，上海市“超级博士后”激励计划。主持中国博士后科学基金面上资助。

活动内容摘要：

In this talk, we propose a time discretization scheme for stochastic control problem under partial observation, which allows one to solve the problem in a discrete induction algorithm with finitely many steps. With the help of the compactification method and the BMO martingale theory, we prove the convergence of the time discretization. Using the dynamic programming principle for the discrete filter process, we present an implementable scheme by making the discrete signal and observation take values in a countable set. Numerical experiment with the linear quadratic structure is performed to validate the theoretical results.

主持人：田琳琳