Zaiwen Wen

Zaiwen Wen is a tenured professor at the Beijing International Center for Mathematical Research, Peking University, and a Boya Distinguished Professor at Peking University. He also serves as the Chair of the Department of Industrial Engineering and Management at the College of Engineering. In 2016, he was awarded the China Youth Science and Technology Award; he was recognized as a National Leading Talent in Scientific and Technological Innovation under the “Ten Thousand Talents Program” in 2020, and he was selected as a Changjiang Distinguished Professor by the Ministry of Education in 2024. He is currently an editorial board member of journals such as Journal of Scientific Computing, Journal of the Operations Research Society of China, Journal of Computational Mathematics, CSIAM Transactions on Applied Mathematics and a technical editor of Mathematical Programming Computation (MPC). He has published more than 80 papers in leading academic journals such as Mathematical Programming, the SIAM and IEEE series. He has developed software packages including ARNT, SSNSDP, and MCPG. He has also authored the textbook “Optimization: Modeling, Algorithms, and Theory”, which has been adopted by over 100 universities, including PKU, Tsinghua and Fudan, across fields such as mathematics, statistics, data science, and artificial intelligence, with over 25,000 copies printed.

Title of talk: A Monte Carlo Policy Gradient Method with Local Search for Binary Optimization

Binary integer programming problems are ubiquitous in many practical applications, including the MaxCut and cheeger cut problem, the MIMO detection and MaxSAT, etc. They are NP-hard due to the combinatorial structure. In this talk, we present a policy gradient method using deep Monte Carlo local search to ensure sufficient exploration in discrete spaces. The local search method is proved to improve the quality of integer solutions and the policy gradient descent converges to stationary points in expectation. Numerical results show that this framework provides near-optimal solutions efficiently for quite a few binary optimization problems.