英文版
您当前的位置: 新闻中心 »学术报告

3-31 清华大学工业工程系张玉利博士学术讲座:Moment-based Robust Optimization and its Applications

  • 发布日期:2017-03-30 01:50:00

题目:Moment-based Robust Optimization and its Applications

主讲人:张玉利 博士 (清华大学工业工程系)

时间:2017年3月31日(周五)下午15:30-17:00

地点:主楼216

主讲人介绍:

    Dr Yuli Zhang received his B.S. degree from the Department of Automation, Wuhan University, Wuhan, China, and his M.S., and Ph.D. degrees from the Department of Automation, Tsinghua University, Beijing, China. He is currently a Postdoc Research Fellow with Department of Industrial Engineering, Tsinghua University, Beijing, China. His research interest focuses on operations research and stochastic/robust optimization in the fields of inventory management, production scheduling and intelligent transportation systems. He has published in various journals, such as Production and Operations Management, Transportation Research Part B, European Journal of Operational Research and IEEE Systems, Man, and Cybernetics Part A. His work was supported by the National Natural Science Foundation of China and Postdoctoral Science Foundation of China.

内容介绍:

    Uncertainty is an inevitable element in real-world systems and has a significant impact on the system performance. In the last ten years, robust optimization approaches has been widely adopted as a powerful tool to model uncertain system parameters with inexact distributions and seek optimal robust decisions. In this talk, we first review the moment-based robust optimization (M-RO) models, which only require the support set and moments of the uncertain system parameters and evaluate a feasible solution by its worst-case average performance over all possible distributions with given support set and moments. Then, we give several recent applications of M-RO in the fields of inventory management, machine scheduling and intelligent routing services. Although M-RO models are hard multi-level min-max optimization problems in general, we show how to equivalently reformulate these problems into tractable conic optimization problems. Furthermore, by exploiting the structural properties of the resulted reformulations, we design exact (or inexact) algorithms for problems with uncorrelated (or correlated) uncertain system parameters. We also discuss the managerial insight given by the M-RO models and the computational efficiency of the proposed algorithms.


 

(承办:技术经济与战略管理系,科研与学术交流中心)