INFORMS 2021 Program Book

INFORMS Anaheim 2021

SD06

Sunday, 2:45PM 4:15PM

SD04 CC Ballroom D / Virtual Theater 4 Hybrid MSOM Student Paper Competition II Sponsored: Manufacturing and Service Operations Management Sponsored Session Chair: Vishal Agrawal, Georgetown University, Washington, 20057, United States Co-Chair: Dragos Florin Ciocan, INSEAD, Fontainebleau, France Co-Chair: Yanchong (Karen) Zheng, Massachusetts Institute of Technology, Cambridge, MA, 02142-1508, United States 1 - Online Policies for Efficient Volunteer Crowdsourcing Scott Rodilitz, Yae University, New Haven, CT, 90278, United States, Vahideh Manshadi Nonprofit crowdsourcing platforms encourage volunteers to complete tasks by using nudging mechanisms to notify a subset of volunteers with the hope that at least one of them responds positively. However, since excessive notifications may reduce volunteer engagement, the platform faces a trade-off between notifying more volunteers for the current task and saving them for future ones. Motivated by these applications, we introduce the online volunteer notification problem and develop an online randomized policy that achieves constant-factor guarantees. Further we demonstrate the effectiveness of our policy by testing it on data from a volunteer-based food recovery platform. 2 - Contextual Learning with Online Convex Optimization: Theory and Application to Chronic Diseases Esmaeil Keyvanshokooh, University of Michigan, Ann Arbor, MI, 77807, United States, Mohammad Zhalechian, Cong Shi, Mark P. VanOyen, Pooyan Kazemian We formulate a new contextual multi-armed bandit model under a two- dimensional control with a nested structure, where each arm (treatment) has a control (dosage) that affects the arm’s performance. Reward (disease progression) is binary and is modeled as the outcome of a logistic random variable that depends on the chosen arm and a convex function of the corresponding control. We develop a joint contextual bandit learning and stochastic gradient descent algorithm, that integrates the strength of contextual bandit learning with online convex optimization. We prove a sub-linear regret, which is provably tight up to a logarithmic factor. We illustrate the effectiveness of our methodology by using case data on patients with type 2 diabetes. 3 - Distributionally Robust Batch Contextual Bandits Nian Si, Stanford University, Stanford, CA, 94305, United States, Fan Zhang, Zhengyuan Zhou, Jose Blanchet Policy learning using historical observational data is an important problem that has found widespread applications. However, existing literature rests on the crucial assumption that the future environment where the learned policy will be deployed is the same as the past environment that has generated the data—an assumption that is often false or too coarse an approximation.In this paper, we lift this assumption and aim to learn a distributionally robust policy with incomplete (bandit) observational data. We propose a novel learning algorithm that is able to learn a robust policy to adversarial perturbations and unknown covariate shifts. We first present a policy evaluation procedure in the ambiguous environment and then give a performance guarantee based on the theory of uniform convergence. SD06 CC Room 303A In Person: Causality, Machine Learning, and Optimization General Session Chair: Nur Kaynar, University of California-Los Angeles, Los Angeles, CA, 90049-5554, United States 1 - Exact Logit-Based Product Design Irem Akcakus, Anderson School of Management, University of California-Los Angeles, Los Angeles, CA, United States, Velibor Misic The share-of-choice product design problem is to find the product that maximizes market share arising from a collection of customer segments. When customers follow a logit model of choice, the market share is given by a weighted sum of logistic probabilities, leading to a challenging problem to solve: one must optimize an objective function that is neither convex nor concave, over an exponentially- sized set of attribute combinations. We develop an exact methodology for solving this problem based on modern integer, convex and conic optimization by showing that the resulting problem can be reformulated as a mixed-integer convex program, which can be further reformulated using conic constraints. Using synthetic problem instances and instances derived from real conjoint data sets, we show that our approach can solve large instances in operationally feasible time frames.

SD01 CC Ballroom A / Virtual Theater 1 Hybrid AAS Award Session Sponsored: Aviation Applications Sponsored Session Chair: Alessandro Bombelli, Delft University of Technology, Delft, 2612 GR, Netherlands Co-Chair: Alexandre Jacquillat, MIT Sloan School of Management, Cambridge, MA, 2142, United States 1 - Modeling of Supply-Demand Interactions in the Optimization of Air Transport Networks Sebastian Birolini, University of Bergamo, Dalmine (BG), 24044, Italy 2 - Modeling and Control of Queuing Networks: Applications to Airport Surface Operations Sandeep Badrinath, Massachusetts Institute of Technology, London, United States 3 - A Stochastic Integer Programming Approach to Air Traffic Scheduling and Operations Alexandre Jacquillat, MI T. Sloan School of Management, Cambridge, MA, 2142, United States SD02 CC Ballroom B / Virtual Theater 2 Hybrid HAS Graduating PhD Job Flash Session Sponsored: Health Applications Society Sponsored Session Chair: Sait Tunc, Virginia Tech, Blacksburg, VA, 24061, United States Co-Chair: Pengyi Shi, Purdue University, West Lafayette, 47907, United States Hybrid APS Panel Discussion Sponsored: Applied Probability Society Sponsored Session Chair: Shane Henderson, Cornell University, Ithaca, NY, 14853, United States 1 - Panelist Jose Blanchet, Columbia University, Dallas, United States 2 - Panelist Shane Henderson, Cornell University, School of ORIE, Rhodes Hall, Cornell University, Ithaca, NY, 14853, United States 3 - Panelist Jim Dai, Cornell University & CUHK-Shenzhen, Ithaca, NY, 14853, United States 4 - Panelist Amy R. Ward, The University of Chicago Booth School of Business, Chicago, IL, 60637-1610, United States 5 - Panelist Devavrat Shah, Massachusetts Institute of Technology,, Cambridge, MA, 02139-4301, United States SD03 CC Ballroom C / Virtual Theater 3

33