Title: Sequencing and Scheduling Arrivals to a Stochastic Service Delivery System
One of the central problems in service delivery management is the design of appointment system for customers. A good appointment system will help to balance the conflicting goals of efficient utilization of available capacity, with minimal waiting time experienced by the customers. This is a central problem in the development of ship berthing schedules at container ports, gate assignment for airplanes at the air ports, patients' appoint time in clinics, and booking of Operating Rooms for surgery operations in hospitals etc.
In the design of the appointment system, the central questions are: (1) the sequence of customer arrivals to the system, and (2) the schedules (i.e. timings) of arrivals. There have been a surge of research activities in this area, due to the heightened interest in the area of service management. Both problems pose significant challenges to the research community. Very few results are known of the sequencing problem, as the mathematics appears to be intractable. Even the simplest case involving deterministic service duration, but fixed appointment slots, is already NP-hard. There have been more successess dealing with the scheduling problem, which have been studied using the tools of stochastic programming, local search, and recently discrete convex analysis.
In this paper, we address these problems using the moments approach - we assume that the service duration of each customer is not completely specified, except the means and covariances information are known. Under the worst case model with moments restrictions, it turns out that the development of the appointment times (including sequencing and scheduling decisions) can be formulated as a mixed integer conic programming problem. This approach builds on recent theoretical works of Burer (2008), and Natarajan, Teo and Zheng (2009). We present preliminary computational results on the problem of appointment system design using this approach, and provide a glimpse into the structure of the optimal policies, for different classes of appointment system.
Biosketch of Prof. Teo Chung-Piaw：
Dr. Chung-Piaw Teo is a Professor with the Department of Decision Sciences, and concurrently Vice-Dean (Research) and Director of the PhD Program in the NUS Business School.
He graduated from MIT with a PhD in Operations Research, and has taught in NUS (Singapore) and Sungkyunkwan University (Korea). He was a fellow with the Singapore-MIT Alliance Program, a research associate with the Logistics Institute – Asia Pacific (TLI-AP), an Eschbach Scholar with Northwestern University (US), and a Distinguished Visiting Professor in YuanZe University (Taiwan). He is currently an associate editor with journals such as Operations Research, Management Science, IIE Transactions, and Flexible Services and Manufacturing Journal etc. His research works have appeared in journals such as Operations Research, Management Science, Journal of Economic Theory, SIAM Review, Mathematics of Operations Research, Mathematical Programming, SIAM Journal on Optimization, SIAM Journal on Computing, etc.