{"id":27737,"date":"2026-04-23T17:13:04","date_gmt":"2026-04-23T09:13:04","guid":{"rendered":"http:\/\/sem.tongji.edu.cn\/semen\/?p=27737"},"modified":"2026-04-23T17:13:04","modified_gmt":"2026-04-23T09:13:04","slug":"an-milp-based-solution-scheme-for-factored-markov-decision-processes","status":"publish","type":"post","link":"https:\/\/sem.tongji.edu.cn\/semen\/27737.html","title":{"rendered":"An MILP-Based Solution Scheme for Factored Markov Decision Processes"},"content":{"rendered":"<p><strong>SPEAKER: <\/strong>\u4f59\u6587\u5fe0 \u526f\u6559\u6388, \u9999\u6e2f\u7406\u5de5\u5927\u5b66<\/p>\n<p><strong>TIME\/DATE: <\/strong>2026.4.27\u00a0 10\uff1a00<\/p>\n<p><strong>CLASSROOM: <\/strong>A308<\/p>\n<p><strong><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-27738\" src=\"https:\/\/sem.tongji.edu.cn\/semen\/wp-content\/uploads\/sites\/3\/2026\/04\/word-image-2.jpeg\" alt=\"\" width=\"200\" height=\"249\" srcset=\"https:\/\/sem.tongji.edu.cn\/semen\/wp-content\/uploads\/sites\/3\/2026\/04\/word-image-2.jpeg 319w, https:\/\/sem.tongji.edu.cn\/semen\/wp-content\/uploads\/sites\/3\/2026\/04\/word-image-2-241x300.jpeg 241w\" sizes=\"auto, (max-width: 200px) 100vw, 200px\" \/><\/strong><\/p>\n<p><strong>ABSTRACT:\u00a0<\/strong><\/p>\n<p>Factored Markov decision processes (MDPs) are a prominent paradigm within the artificial intelligence community for modeling and solving large-scale MDPs whose rewards and dynamics decompose into smaller, loosely interacting components. Through the use of value function approximations, dynamic Bayesian networks and context-specific independence, factored MDPs can achieve an exponential reduction in the state space of an MDP and thus scale to problem sizes that are beyond the reach of classical MDP algorithms. However, factored MDPs are typically solved using custom-designed algorithms that can require meticulous implementations and considerable fine-tuning. In this paper, we propose a mathematical programming approach to solving factored MDPs. Unlike existing solution schemes, our approach leverages off-the-shelf solvers, which enables streamlined implementation and maintenance. Moreover, we exploit the factored structure in both state and action spaces, and we employ feature representations to unify existing methods while taking advantage of context-specific independence in problem classes that other approaches cannot solve efficiently. To further enhance scalability, we introduce a feature learning scheme that automatically identifies informative features and a dynamic basis approximation scheme that adaptively refines our value function approximations. Our numerical experiments demonstrate the potential of our approach.<\/p>\n<p><strong>GUEST BIO:<\/strong><\/p>\n<p>Dr. Man-Chung Yue is currently an Associate Professor at the Department of Applied Mathematics of The Hong Kong Polytechnic University. He received his B.Sc. degree in Mathematics and Ph.D. degree in Systems Engineering and Engineering Management, both from The Chinese University of Hong Kong. He also worked at The University of Hong Kong as an Assistant Professor and Imperial College London as a Research Associate. His research focuses on continuous optimization and its interplay with decision-making under uncertainty, signal processing, machine learning, and operations research.<\/p>\n<p>&nbsp;<\/p>\n<p>&nbsp;<\/p>\n","protected":false},"excerpt":{"rendered":"<p>SPEAKER: \u4f59\u6587\u5fe0 \u526f\u6559\u6388, \u9999\u6e2f\u7406\u5de5\u5927\u5b66 TIME\/DATE: 2026.4.27\u00a0 10\uff1a00 CLASSROOM: A308 ABSTRACT:\u00a0 Factored Markov decision processes (MDPs) are a prominent paradigm within the artificial intelligence community for modeling and solving large-scale MDPs whose rewards and dynamics decompose into smaller, loosely interacting components. Through the use of value function approximations, dynamic Bayesian networks and context-specific independence, factored MDPs [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":27739,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[59],"tags":[],"class_list":["post-27737","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-lectures-seminars"],"_links":{"self":[{"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/posts\/27737","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/comments?post=27737"}],"version-history":[{"count":1,"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/posts\/27737\/revisions"}],"predecessor-version":[{"id":27740,"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/posts\/27737\/revisions\/27740"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/media\/27739"}],"wp:attachment":[{"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/media?parent=27737"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/categories?post=27737"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/sem.tongji.edu.cn\/semen\/wp-json\/wp\/v2\/tags?post=27737"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}