Title: Ranking and selection: parallel implementation, VRT combination, and adaptive sampling
Speaker: Prof. LUO Jun, Associate Professor, Antai College of Economics and Management at Shanghai Jiao Tong University
Time: 9:00am, May 10, 2021
Venue: Room 109, Building 12, Wushan Campus
Introduction to the speaker:
LUO Jun is a tenured Associate Professor of Management Science in Antai College of Economics and Management at Shanghai Jiao Tong University. He received his PhD degree in Industrial Engineering and Logistics Management (IELM) at Hong Kong University of Science and Technology (HKUST) in 2013 and a B.S. degree in Statistics at Nanjing University in 2009. He is the principle investigator of a number of research funds, including the Key Program of National Natural Science Foundation of China (NSFC), NSFC for Excellent Young Scientists, NSFC for Young Scientists, and Chenguang Program of Shanghai Education Commission. His research interests include stochastic models, simulation optimization, and business analytics, with their applications in service operations management, healthcare management, risk management, and logistics management. His work has been published in journals such as Operations Research, INFORMS Journal on Computing, Naval Research Logistics and so on. He also received several awards, including the Excellent Research Award in Social Science from Ministry of Education of China (Second Prize), the Excellent Young Scholar Award from the Stochastic Service and Operations Management (SSOM) Chapter of the Operations Research Society of China (ORSC), the Teaching and Education Award (Third Prize) and the Excellent Teacher Award (Second Prize) from SJTU. Currently, he serves as the Council Member of the SSOM Chapter at ORSC, the Financial Engineering and Risk Management Chapter at ORSC, and the Management and Decision Science Chapter at Chinese Academy of Management.
Abstract:
In the past several decades, many ranking-and-selection (R&S) procedures have been proposed in order to select the best system design with the largest (or smallest) mean performance measure from a finite number of alternatives. In this talk, we will briefly discuss about three recent developments on R&S. First, with the rapid growth of computing technology, parallel computing are ubiquitous and accessible for ordinary users. How to implement existing fully sequential procedures in parallel to solve large-scale problems is an important topic. Second, various variance reduction techniques (VRTs) have been developed to fast the simulation process. How to combine VRTs into R&S procedure designs to improve the efficiency is another interesting question. Third, as samples are sequentially generated in computers, it is attractive to dynamically determine which alternative to be simulated next in order to achieve high efficiency. Then, how to design procedures with adaptive sampling rules is also worth exploring.
