# Presenter : Adam Cheol Woo Kim / Postdoctoral Researcher / Harvard University

Title: Machine Learning for Optimization and Control Problems

Abstract: In many practical settings, similar optimization and control problems frequently arise and must be solved repeatedly. We propose novel methods that leverage patterns from pre-solved instances using machine learning, resulting in significantly faster solutions once training is complete. This talk is divided into three parts, each addressing a different class of optimization and control problems. First, we introduce a machine learning approach to the optimal control of multiclass fluid queueing networks. We demonstrate that optimal control policies can be learned using a decision tree algorithm, Optimal Classification Trees with Hyperplane Splits. The second part presents a machine learning framework for solving two-stage linear adaptive robust optimization problems with binary here-and-now decisions and polyhedral uncertainty sets. Finally, we discuss a prescriptive machine learning approach to mixed integer convex optimization. We test our approach on various synthetic and real-world problems. Using the proposed methods, we can obtain high-quality solutions to a broad range of large-scale optimization and control problems in real-time – within milliseconds.

 

Bio: I am a postdoctoral researcher at Harvard University, advised by Professor Milind Tambe. I completed my PhD in Operations Research at MIT, where I was advised by Professor Dimitris Bertsimas. I obtained my undergraduate degree from Seoul National University. My PhD research focused on developing machine learning techniques to solve optimization and control problems. I am broadly interested in AI for decision-making, with applications in healthcare and public policy. My long-term goal is to make a real-world impact by enabling fast, effective, and socially beneficial decisions.” 

We look forward to meeting you.

Thank you.