来源:统计学院

12月10日 | 余长君:A Hybrid Direct Method for Solving Optimal Control Problem

来源:统计学院发布时间:2022-12-07浏览次数:106

时   间:2022年12月10日 15:30-16:30

地   点:理科大楼 A1514

             腾讯会议ID:273-893-586

报告人:余长君  副教授

邀请人:吴贤毅  教授

摘   要:

Control parameterizationand direct collocation methods are known as the direct method for solvingoptimal control problems. The methods used for numerical integration anddifferentiation form different types of collocation methods. However, there isnot too much discussion on the numerical method for integration and differentiationwhen control parameterization method is used. Hence, it is difficult to make afair comparison between this two methods in terms of the solution accuracy aswell as the computational efficiency. In this talk, we propose a new hybridcomputational scheme for solving optimal control problems. The discretizationstrategy is adopted from control parameterization where only the controlfunction is discretized, however, the numerical integration and differentiationmethods is in line with those in existing direct collocation methods. Variousbenchmark examples are solved using the proposed method and conventional directcollocation methods, numerical results show that our new method is highlyeffective.

报告人简介:

余长君,副教授、博士生导师,上海大学理学院副院长。2005年本科毕业于电子科技大学应用数学学院;2012年于澳大利亚 Curtin University获得博士学位;2012 - 2016 任澳大利亚 Curtin University 数学系Research Fellow; 2016年进入上海大学理学院数学系工作至今。余长君博士主要从事非线性最优化与最优控制理论、算法及其在信息和工程中的应用的研究,在IEEE Transactions on Automatic Control,Automatica等国际著名学术期刊发表论文40余篇,在Springer出版专著一部。主持了多项国家级科研项目;余长君博士于2015年入选上海市青年东方学者计划;2018年获得上海大学蔡冠深优秀青年教师奖;自2019年1月担任SCI期刊《Journal of Industrial andManagement Optimization》AssociateEditor;目前担任中国运筹学会理事,科普工作委员会副主任,数学规划分会青年理事,中国工业与应用数学学会竞赛工作委员会委员,上海市运筹学会常务副理事长兼秘书长。