报告题目:Near optimal controls for partially observed stochastic linear quadratic problems
报 告 人:熊捷 教授(南方科技大学)
报告时间:2025年8月11日(星期一)10:00—11:00
报告地点:数学科学学院114(小报告厅)
校内联系人:李娜 教授 联系方式:84708354
报告摘要:In this talk, we consider a stochastic linear quadratic control problem with partial observation, and the aim is to find $\epsilon$-optimal controls in the weak formulation. The main challenge arises not only from the presence of the control in the diffusion term of the state equation but also from the circular dependence between the control process and the filtration generated by the observation. Furthermore, the observation process contains an unbounded drift term, which prevents us from directly applying Girsanov theorem. We address these difficulties by restricting the control to a smaller domain, which enables us to apply the conditional version of the Girsanov theorem and thereby break the circular dependence. Subsequently, we study the restricted problem using a non-standard variation method. As a result, the $\epsilon$-optimal controls are derived by passing to the limit over an approximate sequence. This talk is based on a paper jointly with Sun, Wen and Xu.。
报告人简介:熊捷,1983年在北京大学数学系获得学士学位,1986年获得北京大学概率统计系硕士学位,1992年在美国卡罗莱纳大学教堂山分校统计系获得博士学位。
1992年-2014年间,熊捷执教于美国田纳西大学数学系,在此期间,于2002-2003年在加拿大被授予随机过程与滤波Canada Research Chair;于2003-2004年获得洪堡奖学金并赴德国进行合作研究。2014年正式入职澳门大学数学系,并于2017年底作为讲席教授加入南方科技大学数学系。
他的研究领域包括:随机滤波和控制,数理金融,数理生物,测度值随机过程以及随机偏微分方程。