INFORMS 2021 Program Book

INFORMS Anaheim 2021

MD02

L. Allison Jones-Farmer, Miami University, Information Systems & Analytics, Farmer School of, Oxford, OH, 45056, United States

Monday Keynote 03 CC Ballroom C /Virtual Theater 3

Omega Rho Distinguished Lecturer: A Journey through Public Sector Operations Research Informs Special Session: Keynote Informs Special Session Session 1 - A Journey through Public Sector Operations Research Laura Albert, University of Wisconsin-Madison, Madison, WI, 53562, United States Societally important problems have driven the theory and application of operations research since its origins in World War II. Recent events have highlighted the enormous number of challenges that require expertise from operations research and analytics. The operations research community has a long history of stepping up to address challenging problems in the public sector through modeling, computation, and data analytics that has influenced policy and impacted practice. This has been a central theme of my academic career, which has focused on security, emergency response, public safety, and risk management. This talk discusses several research problems, focusing on how operations research has made a difference, and offers a blueprint for how the operations research community can tackle future challenges, impact society, and broadcast our message to the world. Keynote: A Dynamic Queueing Road Map from Communication Systems to Resource Sharing Services Keynote Session 1 - A Dynamic Queueing Road Map from Communication Systems to Resource Sharing Services William A. Massey, Professor, Princeton University, ORFE Department, Sherrerd Hall, Princeton, NJ, 08544, United States The field of operations research applies mathematics to the creation of quantitative languages designed for strategic decision making. Queueing theory was invented just over a century ago to design efficiency into communication systems. In the 21st century, it plays this same role in the design of resourcesharing services.Rates for customer service demand can easily be dependent on the time of day, week, or seasonal effects. Hence dynamic rate queues are more realistic stochastic models than their traditional constant rate counterparts. Moreover, since they are not amenable to classical steady state analysis techniques, dynamic rate problems lead to greater mathematical challenges. Along with many collaborators, this talk covers a personal research journey to develop a dynamic rate queueing theory. We also show how the guideposts for our path evolved from communication systems to resource-sharing services. MD01 CC Ballroom A / Virtual Theater 1 Hybrid Panel Discussion on Editor’s Perspective in Publishing Data Science-Focused Papers Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Xiaowei Yue, Virginia Tech, Blacksburg, VA, 24061, United States Co-Chair: Raed Al Kontar, University of Michigan, Ann Arbor, MI, 48109-2117, United States 1 - Panelist Yu Ding, Texas A&M University, Dept Industrial & Systems Engineering, College Station, TX, 77843-3131, United States 2 - Panelist Jing Li, Georgia Institute of Technology, School Of Computing Informatics & Decision Sy Po B, Tempe, AZ, 85287-8809, United States 3 - Panelist Ramaswamy Ramesh, SUNY Buffalo, East Amherst, NY, 14051- 1687, United States 4 - Panelist Monday Keynote 04 CC Ballroom D /Virtual Theater 4 Monday, 2:45PM 4:15PM

MD02 CC Ballroom B / Virtual Theater 2 Hybrid Recent Advances in Planning and Scheduling Under Uncertainty Sponsored: OPT/Optimization Under Uncertainty Sponsored Session Chair: Karmel S. Shehadeh, Lehigh University, Bethlehem, PA, 18015- 1518, United States 1 - Optimized Scenario Reduction: Solving Large-scale Stochastic Programs with Quality Guarantees Wei Zhang, Faculty of Business, The Hong Kong Polytechnic University, MN037, Hong Kong, Hong Kong, Alexandre Jacquillat, Kai Wang, Shuaian Wang Stochastic programming involves large-scale optimization with exponentially many scenarios. We propose an optimization-based scenario reduction approach to generate high-quality solutions and tight lower bounds by only solving small- scale instances. First, we design a scenario subset selection model that minimizes the recourse approximation error over a pool of solutions. We provide theoretical results to support our formulation, and a tailored heuristic algorithm to solve it. Second, we propose a scenario assortment optimization approach that generates a lower bound—hence, a solution quality guarantee—by relaxing nonanticipativity constraints across scenario bundles. We formulate an optimization model to maximize this lower bound, and design exact row-generation and column- generation algorithms to solve it. 2 - Strategic Idling in Appointment Systems with Sequential Servers You Hui Goh, Nanyang Technological University, Singapore, Singapore, Zhenzhen Yan This paper studies an appointment scheduling problem with two sequential servers from a distributionally robust optimization (DRO) perspective. Conventionally, schedules are optimized to minimize the expected total cost including customers’ waiting costs and servers’ overtime costs. Yet, the schedule obtained can lead to imbalanced waiting times in two servers, concentrating on the downstream server. To ensure a balanced waiting time in two servers without rescheduling patients, we adopt an idea in the queueing literature to strategically idle the upstream server. We propose a DRO model to calculate the optimal strategic idling (SI) policy considering the correlations in service times. 3 - Distributionally Robust Home Service Routing and Appointment Scheduling with Random Travel and Service Times Man Yiu Tsang, Lehigh University, Bethlehem, PA, United States, Karmel S. Shehadeh We study an integrated routing and appointment scheduling (RAS) problem arising from home service practice. Given a set of customers within a region that an operator needs to serve, we seek for the operator’s route and time schedule. The travel time and service time of each customer are random with unknown distributions. Only a possibly small set of historical data is available. To address this, we propose and analyze two distributionally robust home service RAS (DHRAS) models that search for decisions to minimize the worst-case expectation of operational costs over distributions residing within an ambiguity set. We use a moment based ambiguity set and a 1-Wasserstein distance based ambiguity set. We derive equivalent MILP reformulations of both models. In an extensive numerical experiment, we investigate the proposed models’ performances and derive insights into DHRAS. 4 - Presenter Chrysanthos Gounaris, Carnegie Mellon University, Pittsburgh, PA, United States 5 - Valid Inequalities for Approximating the Robust Surgery Scheduling Problem Ankit Bansal, University of Minnesota, Minneapolis, MN, United States, Jean-Philippe P. Richard, Bjorn Berg, Yu-Li Huang An approximation of the two-stage robust optimization surgery-to-OR allocation problem is presented. The second-stage problem is linearly relaxed and three types of valid inequalities which approximate the integer hull are derived. The resulting linear relaxation of the second stage problem is then dualized and integrated into the first-stage problem. A column-generation based approach is used to solve the resulting MILP, yielding an approximation of the problem. Data from an academic medical center is used to compare the computational performance of the approximate approach and its solution quality with the only known exact approach in the literature. Managerial insights are discussed.

73