11-20【周子锐】五教 5207 “数学优化”系列报告

时间:2024-11-18


报告题目: From Optimization Theory and Algorithms to Practical Solvers: A Case Study with Linear Programming

 

报告人:周子锐


报告时间:1120日(周三)9:45-11:00        

    

报告地点:东区五教5207        

 

摘要

Mathematical optimization is a fundamental aspect of numerous industries, ranging from manufacturing and logistics to finance and artificial intelligence. Consequently, optimization solvers have emerged as indispensable tools in many applications, empowering organizations to make informed decisions and enhance operational efficiency. While optimization theory and algorithmic frameworks establish the foundations, there remain significant challenges in developing optimization solvers that are both efficient and robust in real-world scenarios. This talk focuses on bridging the gap between algorithmic frameworks and practical implementation, sharing computational insights and techniques in the process of crafting highly-efficient optimization solvers. We use simplex method and interior-point method for linear programming as a case study to demonstrate how these textbook algorithms are transformed into high-performance tools suited for practical applications. This talk is particularly relevant for practitioners and researchers aiming to translate research works into applied solutions.


报告人简介

Dr. Zirui Zhou is currently a Principal Researcher with Huawei Technologies Canada. Before joining Huawei Canada, he was an Assistant Professor with the Department of Mathematics at Hong Kong Baptist University. Hereceived the Ph.D. degree in Systems Engineering and Engineering Management from the Chinese University of Hong Kong. His research interests include optimization solvers, convex analysis, numerical optimization, and learning to optimize. Dr. Zhou has four years of experience in solverdevelopment and is a principal contributor to the Optverse AI Solver of Huawei Cloud. His research works on optimization error bound, provable non-convex methods, machine learning enhanced optimization algorithms, and large language models for operations research have been published in top-tier journals and conferences.