English Version

首页» 学术科研» 讲座信息

[2016-5-23]Dynamic Optimization for Modern Service and E-commerce Systems




Dynamic Optimization for Modern Service and E-

commerce Systems


报告人  王欣上 博士


主持人: 侍乐媛 教授






In this talk, we consider two dynamic optimization problems arising in modern service and e-commerce systems.


In the first part of the talk, we study a personalized recommendation system based on sending mobile push messages. In recent years, e-commerce companies are seeing an increasing amount of transactions completed via mobile platforms, such as apps in iOS and Android systems. In China, the e-commerce market share of a mobile app developed by Alibaba Group, which has been installed on several hundred million devices, is rapidly replacing that of traditional e-commerce markets hosted on webpages. We study the problem of managing the allocation of push messages sent to users by this app which recommends products tailored to every user. The model deviates from canonical revenue management models as the app can dynamically sequence the users to receive messages.


In the second part of the talk, we study web and mobile applications that are used to schedule advance service, from medical appointments to restaurant reservations. We model these problems as online weighted bipartite matching problems with non-stationary arrivals. We give the first online, data-driven algorithms with performance guarantees for these problems. We show that the average performance of our algorithms is bounded by 1/2 times that of an optimal offline algorithm. Our algorithms can also be applied to a number of related problems, including display-ad allocation problems and revenue-management problems for opaque products. We test the empirical performance of our algorithms against several well-known heuristics by using appointment-scheduling data from a department within a major academic hospital system in New York City. The results show that the algorithms exhibit the best performance among all the tested policies. In particular, our algorithms are 21% more effective than the actual scheduling strategy used in the hospital system according to our performance metric.





Xinshang Wang is a PhD. candidate in Operations Research at the Department of Industrial Engineering and Operations Research at Columbia University. His research spans several areas of stochastic and combinatorial optimization for modern service applications, including data-driven optimization under uncertain and dependent demands, and modeling of customer choice behavior for resource allocation problems. Applications of his interest include, but are not limited to, revenue management, healthcare operations and supply-chain management. Previously, Xinshang earned a B.S. degree in Physics from Peking University.