INFORMS 2021 Program Book

INFORMS Anaheim 2021

ME04

3 - Interview Hoarding

3 - COVID-19: Prediction, Prevalence and the Operations of Vaccine Allocation Georgia Perakis, Massachusetts Institute of Technology, Belmont, MA, 02478-1706, United States Abstract not available 4 - Unmasking Human Trafficking Risk in Commercial Sex Supply Chains with Machine Learning Pia Ramchandani, University of Pennsylvania, Philadelphia, PA, 19104, United States, Hamsa Bastani, Emily Wyatt Abstract not available

Vikram Manjunath, Thayer Morrill Many centralized matching markets are preceded by interviews between the participants. We study the impact on the final match of an increase in the number of interviews for one side of the market. Our motivation is the match between residents and hospitals where, due to the COVID-19 pandemic, interviews for the 2020-21 season of the National Residency Matching Program were switched to a virtual format. This drastically reduced the cost to applicants of accepting interview invitations. However, the reduction in cost was not symmetric since applicants, not programs, previously bore most of the costs of in-person interviews. We show that, starting from a situation where the final matching is stable, if doctors can accept more interviews, but the hospitals do not increase the number of interviews they offer, then no doctor is better off and many doctors are potentially harmed. This adverse consequence is the result of what we call interview hoarding. We prove this analytically and characterize optimal mitigation strategies for special cases. We use simulations to extend these insights to more general settings. 4 - The College Portfolio Problem Ran Shorrer, Penn State University, State College, PA, 16802, United Statesm, S. Nageeb Ali A college applicant faces the following risky choice: she applies to a portfolio of colleges while being uncertain about which school would admit her. Admissions decisions are correlated insofar as being rejected by a lower ranked school may imply that she is rejected by a higher ranked school. We show that solutions to this decision problem involve applying to a combination of reach, match, and safety schools. When application costs decrease, a college applicant broadens the range of schools to which she applies by including both those that are more selective and those that are safer options. Hybrid JFIG Paper Competition Sponsored: Junior Faculty Interest Group Sponsored Session Chair: Alice E. Smith, Auburn University, Auburn, AL, 36849, United States Co-Chair: Dorit Simona Hochbaum, University of California-Berkeley, Berkeley, CA, 94720-1777, United States Co-Chair: Manish Bansal, Virginia Tech., Blacksburg, VA, 24061-1019, United States ME05 CC Ballroom E / Virtual Theater 5 Hybrid PSOR Best Paper Award Sponsored: Public Sector OR Sponsored Session Chair: Justin J. Boutilier, University of Wisconsin-Madison, Madison, WI, 53706-1603, United States Co-Chair: Somya Singhvi, Massachusetts Institue of Technology, Cambridge, MA, 02139, United States Co-Chair: Yanchong (Karen) Zheng, Massachusetts Institute of Technology, Cambridge, MA, 02142-1508, United States 1 - Off-Grid Lighting Business Models to Serve the Poor: Evidence From a Structural Model and Field Experiments in Rwanda Serguei Netessine, The Wharton School, Philadelphia, PA, United States, Bhavani Shanker Uppari, Ioana Popescu, Rowan P. Clarke A significant proportion of the world’s population does not have access to grid- based electricity. Rechargeable lamp-based technology is becoming prominent as an alternative off-grid lighting model in developing countries. We explore, in close collaboration with Nuru Energy in Rwanda, the consumer behavior and the operational inefficiencies that result under this model. Specifically, we are interested in measuring the impact of inconvenience along with the impact of liquidity constraints on lamp usage, and evaluating the efficacy of strategies that address these factors. Our undertaking has implications for both firm-level operational decisions and government-level policy decisions. 2 - Deploying a Reinforcement Learning System for COVID-19 Testing at the Greek Border Kimon Drakopoulos, University of Southern California, Los Angeles, CA, 90305-1028, United States Abstract not available ME04 CC Ballroom D / Virtual Theater 4

ME06 CC Room 303A In Person: Operations Research & Vulnerable Populations General Session

Chair: Jiayi Lin, College Station, TX, 77845, United States 1 - Designing Policies for Allocating Housing to Persons Experiencing Homelessness Bill Tang, University of Southern California, Los Angeles, CA, United States, Phebe Vayanos, Cagil Kocyigit We study the problem of allocating scarce housing resources of different types to individuals experiencing homelessness based on their observed covariates. Our goal is to leverage administrative data collected in deployment to design an online policy that maximizes mean outcomes while satisfying budget requirements. We propose a policy in which an individual receives the resource maximizing the difference between their mean treatment outcomes and the resource bid price, or roughly the opportunity cost of using a resource. Our approach has nice asymptotic guarantees and is easily interpretable. We evaluate it on synthetic and real-world Homeless Management Information System data to illustrate practical usage of our methodology. 2 - Client-Volunteer Relationships and Satisfaction in a Non-Profit Organization: The Case of Meals on Wheels Atlanta Shikha Safaya, Georgia Institute of Technology, Atlanta, GA, United States, Basak Kalkanci, Ravi Subramanian We partner with Meals on Wheels Atlanta, a non-profit organization providing meals and personal interactions to seniors who have limited mobility and are food insecure. We examine how key factors related to service design, including service frequency and duration, contribute to satisfaction of seniors and volunteers and ensure high service quality and sustained volunteer engagement. 3 - Targeted Mass Screening under Limited Testing Capacity with Application to Covid-19 Jiayi Lin, Texas A&M University, College Station, TX, United States, Hrayer Aprahamian Mass screening is an essential tool that arises in various settings, e.g., the ongoing COVID-19 pandemic. The objective is to classify subjects as positive or negative for an infectious disease as efficiently and accurately as possible. Under limited testing capacity, administrators must target those among the population who need to be screened the most. This work aims to address this decision problem by taking advantage of population-level risk information in order to identify the optimal subset of subjects to screen. We consider two models: (i) individual testing, and (ii) group testing. We solve the resulting optimization problems to global optimality through a parameterized reformulation scheme. Our case study on real COVID-19 risk data reveals substantial benefits over conventional methods, highlighting the importance of data-driven informed policies. 4 - Designing School Choice for Diversity in the San Francisco Unified School District Katherine L. Mentzer, Stanford University, Stanford, CA, 94305, United States, Irene Yuan Lo, Itai Ashlagi, Maxwell Allman Prompted by a redesign of the San Francisco Unified School District (SFUSD) school choice system, we explore how choice mechanisms affect tradeoffs between choice, diversity, and other school district goals. We used simulations combining zone optimization with choice to propose new assignment policies. We found that zones must be designed with choice, as choice can lead to resegregation of diverse zones. However, well-designed zones combined with minority reserves could attain SFUSD diversity goals, as well as other district objectives such as predictability and proximity. In SFUSD, traditional school choice tools such as priorities can also attain diversity goals and provide choice, at the expense of predictability and proximity. Based on our findings, we recommended a policy of medium zones and reserves that was approved by the SFUSD Board of Education.

88