講座題目：Online Resource Allocation with Limited Flexibility 2017-06-02

題目：Online Resource Allocation with Limited Flexibility

報告人：香港科技大學Xuan Wang助理教授

時間： 2017年6月8日星期四上午10：00-12:00

地點：bwin必贏唯一官網(wǎng)7002室

Abstract：

We consider a class of online resource allocation problems in whichthere are several types of resources with limited initial inventory and severaldemand classes. The resources are flexible in that each type of resources canserve more than one demand class. In this talk, we focus on a special class ofstructures with limited flexibility, the long chain design, which has been animportant concept in the design of sparse flexible processes. We study the longchain design in an online stochastic environment where the requests are drawnrepeatedly and independently from a known probability distribution over thedifferent demand classes. Also, the decision on how to address each requestmust be made immediately upon its arrival. We show the effectiveness of thelong chain design in mitigating supply-demand mismatch under a simple myopiconline allocation policy. In particular, we provide an upper bound on the expectedtotal number of lost sales that is irrespective of how large the market sizeis.

Biography:

Xuan Wang is an Assistant Professor in Operations Management at theSchool of Business and Management, The Hong Kong University of Science andTechnology. Her main research interests lie in the broad area ofdecision-making under uncertainty with applications in supply chainoptimization and revenue management. Xuan holds a Ph.D. in OperationsManagement from the Stern School of Business at New York University, and aBachelor in Operations Research and Industrial Engineering from TsinghuaUniversity, China.