ORC IAP Seminar 2021

1/28/21 - 1/29/21 | 9:00am-5:00pm | Online

Policymaking in Operations Research 

If you have any questions, please contact us via email: orc_iapcoordinators@mit.edu.

Dates: Thursday, January 28th - Friday, January 29th

Place: Online (via Zoom)

If you are not already on an email list that is receiving IAP seminar announcements, you may opt-in via the link below. You can add yourselves to this email list to receive the seminar announcements, updates, and Zoom links. You can also remove yourself from the list at any time using the same link.

http://mailman.mit.edu/mailman/listinfo/or_remote_iapseminar (Zoom links will be provided the day before and morning of the seminar)


Schedule Day 1

Talk 1 - January 28th 9:00am-10:30am

Mohammad Fazel Zarandi

Senior Lecturer, MIT Sloan School of Management

Prescriptive Machine Learning for Public Policy: The Case of Immigration Enforcement

Historically, machine learning applications have focused on making predictions by discovering complex patterns in data. In this talk, we demonstrate that successful implementations of data-driven policies entail going beyond machine learning predictions by carefully connecting such predictions to decisions. Specifically, we show that a crucial component of the prediction-decision interaction is accounting for both observed and unobserved factors in the data-generating process. We present a framework that embeds machine learning predictions into a quasi-experimental design to allow for an unbiased evaluation of an algorithm's recommendations. We apply this framework to design a national immigration policy that prevents crime, using federal government administrative data comprised of millions of observations. After accounting for unobserved confounders, we show that our algorithm reduces severe crimes by 26% and successfully decreases re-offending rates for immigrants in both federal and state prison systems in the United States.



Mohammad Fazel‐Zarandi is a Senior Lecturer at the MIT Sloan School of Management. His research interests are in the areas of data-driven decision making and machine learning, with applications in public policy. His current research focuses on the use of analytics to analyze and improve decision-making in the public policy domain. Some of the topics he studies include immigration policy, criminal justice, and public health.

Mohammad's research has been published in flagship journals such as Operations Research and Management Science, and has been reported by the Los Angeles Times, the Boston Globe, the New York Times, the Washington Post, Bloomberg, and many other major media outlets.

Before joining MIT, Mohammad was a Postdoctoral Associate and Lecturer at the Yale School of Management. He received his PhD from the Rotman School of Management, University of Toronto in operations management and his MS in industrial engineering, also from the University of Toronto.


Talk 2 - January 28th 11:00am-12:30pm

Jonathan Mattingly

James B. Duke Distinguished Professor of Mathematics, Duke

Sampling to Understand Gerrymandering and Influence Public Policy

Understanding and identifying Gerrymandering requires understanding the interaction of the geopolitical structure of the electorate with the legal and policy principles of redistricting. While it is tempting to postulate overarching ideas of “fairness” based on idealized principles, we have preferred to understand what the properties of a generic or typical redistricting possess and use them to critique proposed redistricting. Our formulation of this problem involves sampling a high dimensional probability distribution and leads to an operational definition of “fairness”.

This approach has been gaining traction and has been used by a few groups in the advisory and judicial setting around redistricting. The speaker and his collaborators have been involved in several court cases around political and racial gerrymandering. I have testified in two court cases: Common Cause v. Rucho (that went to the US Supreme Court) and Common Cause v. Lewis. This work was also submitted in a tired case in NC and as part of pretrial motions in a Pennsylvania case. The result of the NC cases was the redrawing of the NC State Legislative district maps and NC congressional maps for the 2020 elections.

On the mathematical side, we have been driven to understand interesting sampling methods and develop new algorithms around sampling and optimization. On the policy side, there are many interesting questions ranging from how to encode the law mathematically to how best to explain and distill the results for the courts and political stakeholders.

I will endeavor to give both insight into the mathematical questions and the policy story the methods tell.



Jonathan Christopher Mattingly grew up in Charlotte, NC where he attended Irwin Ave elementary and Charlotte Country Day. He graduated from the NC School of Science and Mathematics and received a BS is Applied Mathematics with a concentration in physics from Yale University. After two years abroad with a year spent at ENS Lyon studying nonlinear and statistical physics on a Rotary Fellowship, he returned to the US to attend Princeton University where he obtained a PhD in Applied and Computational Mathematics in 1998. After 4 years as a Szego assistant professor at Stanford University and a year as a member of the IAS in Princeton, he moved to Duke in 2003. He is currently a Professor of Mathematics and of Statistical Science.

His expertise is in the longtime behavior of stochastic system including randomly forced fluid dynamics, turbulence, stochastic algorithms used in molecular dynamics and Bayesian sampling, and stochasticity in biochemical networks.

Since 2013 he has also been working to understand and quantify gerrymandering and its interaction of a region's geopolitical landscape. This has lead him to testify in a number of court cases including in North Carolina, which led to the NC congressional and both NC legislative maps being deemed unconstitutional and replaced for the 2020 elections.

He is the recipient of a Sloan Fellowship and a PECASE CAREER award. He is also a fellow of the IMS and the AMS. He was awarded the Defender of Freedom award by Common Cause for his work on Quantifying Gerrymandering.


Talk 3 - January 28th 1:30pm-3:00pm

Lawrence Wein

Jeffrey S. Skoll Professor of Management Science, Stanford University

Solving Violent Crimes

We present research on three approaches to solving homicides or sexual assaults: ballistic imaging, DNA matching, and investigative genetic genealogy. We use data from Stockton CA (ballistic imaging), Detroit and San Francisco (sexual assault kits), and the DNA Doe Project (genetic genealogy) to guide the development of new models to address these problems, with the goal of maximizing violent crime solving subject to capacity constraints.



Wein received his PhD in operations research from Stanford in 1988 and has taught the core MBA course in operations management throughout his entire career, both at MIT’s Sloan School of Management from 1988 to 2002, where he was the DEC Leaders for Manufacturing Professor of Management Science, and at Stanford since 2002, where he is currently The Jeffrey S. Skoll Professor of Management Science. He also is a senior fellow at Stanford’s Center for International Security and Cooperation.

In manufacturing, health care, and homeland security, he has published widely and impacted practice. His HIV work on drug-switching policies led to a successful multicenter clinical trial. His smallpox work influenced the George W. Bush administration’s post-attack vaccination policy; his anthrax work led to plans in Washington, D.C., to use postal workers to distribute antibiotics after a large attack; and his testimony before a congressional committee on his biometric analysis of the US-VISIT Program was instrumental in the switch from a two-finger to a ten-finger system. He has won several research awards and was editor-in-chief of Operations Research from 2000 to 2005. He is a member of the National Academy of Engineering.


Talk 4 - January 28th 3:30-5:00pm

Alexandre Jacquillat

Assistant Professor of Operations Research and Statistics, MIT

Michael Lingzhi Li, Agni Orfanoudaki, and Holly Wiberg

PhD Candidates at Operations Research Center, MIT

From Predictions to Prescriptions: A Data-Driven Response to COVID-19

The COVID-19 pandemic has created unprecedented challenges worldwide. Strained healthcare providers make difficult decisions on patient triage, treatment and care management on a daily basis. Policy makers have imposed social distancing measures to slow the disease, at a steep economic price. We design analytical tools to support these decisions and combat the pandemic. Specifically, we propose a comprehensive data-driven approach to understand the clinical characteristics of COVID-19, predict its mortality, forecast its evolution, and ultimately alleviate its impact. By leveraging cohort-level clinical data, patient-level hospital data, and census-level epidemiological data, we develop an integrated four-step approach, combining descriptive, predictive and prescriptive analytics. First, we aggregate hundreds of clinical studies into the most comprehensive database on COVID-19 to paint a new macroscopic picture of the disease. Second, we build personalized calculators to predict the risk of infection and mortality as a function of demographics, symptoms, comorbidities, and lab values. Third, we develop a novel epidemiological model to project the pandemic’s spread and inform social distancing policies. Fourth, we propose an optimization model to reallocate ventilators and alleviate shortages. Our results have been used at the clinical level by several hospitals to triage patients, guide care management, plan ICU capacity, and re-distribute ventilators. At the policy level, they are currently supporting safe back-to-work policies at a major institution and equitable vaccine distribution planning at a major pharmaceutical company, and have been integrated into the US Center for Disease Control’s pandemic forecast.



Alexandre Jacquillat is an Assistant Professor of Operations Research and Statistics at the MIT Sloan School of Management. His research focuses on data-driven decision-making, spanning stochastic optimization, large-scale optimization, mechanism design and field experimentation.

His primary focus is on problems of scheduling, operations and pricing in the transportation sector—with a particular interest in air traffic management and on-demand mobility.

Alexandre is the recipient of several research awards, including the 2017 Best Paper Award from the INFORMS Transportation Science and Logistics Society, the 2015 George B. Dantzig Dissertation Award from INFORMS, the 2015 Dissertation Award from the INFORMS Transportation and Logistics Society, the Milton Pikarsky Memorial Award from the Council of University Transportation Centers, and the L.E. Rivot Medal from the French Academy of Science.


Schedule Day 2

Talk 5 - January 29th 9:00am-10:30am

Swati Gupta

Assistant Professor and Fouts Family Early Career Professor, Georgia Tech


Mitigating the Impact of Bias in Selection Algorithms


The introduction of automation into the hiring process has put a spotlight on a persistent problem: discrimination in hiring on the basis of protected-class status. Left unchecked, algorithmic applicant-screening can exacerbate pre-existing societal inequalities and even introduce new sources of bias; if designed with bias-mitigation in mind, however, automated methods have the potential to produce fairer decisions than non-automated methods. In this work, we focus on selection algorithms used in the hiring process (e.g., resume-filtering algorithms) given access to a "biased evaluation metric". That is, we assume that the method for numerically scoring applications is inaccurate in a way that adversely impacts certain demographic groups.

We analyze the classical online secretary algorithms under two models of bias or inaccuracy in evaluations: (i) first, we assume that the candidates belong to disjoint groups (e.g., race, gender, nationality, age), with unknown true utility Z, and “observed” utility Z/\beta for some unknown \beta that is group-dependent, (ii) second, we propose a “poset” model of bias, wherein certain pairs of candidates can be declared incomparable. We show that in the biased setting, group-agnostic algorithms for online secretary problem are suboptimal, often causing starvation of jobs for groups with \beta>1. We bring in techniques from matroid secretary literature and order theory to develop group-aware algorithms that are able to achieve certain “fair” properties, while obtaining near-optimal competitive ratios for maximizing true utility of hired candidates in a variety of adversarial and stochastic settings. Keeping in mind the requirements of U.S. anti-discrimination law, however, certain group-aware interventions can be construed as illegal, and we will conclude the talk by partially addressing tensions with the law and ways to argue legal feasibility of our proposed interventions. This talk is based on work with Jad Salem and Deven R. Desai.



Dr. Swati Gupta is an Assistant Professor and Fouts Family Early Career Professor in the H. Milton Stewart School of Industrial & Systems Engineering at Georgia Institute of Technology. She received a Ph.D. in Operations Research from MIT in 2017 and a joint Bachelors and Masters in CS from IIT, Delhi in 2011. Dr. Gupta’s research interests are in optimization, machine learning and algorithmic fairness. Her work spans various application domains such as revenue management, energy and quantum computation. She received the NSF CISE Research Initiation Initiative (CRII) Award in 2019. She was also awarded the prestigious Simons-Berkeley Research Fellowship in 2017-2018, where she was selected as the Microsoft Research Fellow in 2018. Dr. Gupta received the Google Women in Engineering Award in India in 2011. Dr. Gupta’s research is partially funded by the NSF and DARPA.


Talk 6 - January 29th 11:00am-12:30pm

Edward Kaplan

William N. and Marie A. Beach Professor of Operations Research, Yale

Adventures in Policy Modeling

Policy modeling refers to the application of operations research and allied quantitative methods to model policy problems. The idea is to help make better decisions when data are uncertain or unavailable, the time allowed for analysis is very short relative to the natural time scale of the problem, or other resources seemingly necessary for analysis are constrained. Looking back over the past 30 years or so, this talk reviews some personal adventures in policy modeling including HIV prevention via needle exchange, smallpox bioterror preparedness, counterterrorism and intelligence operations, and monitoring and control of local coronavirus outbreaks.



Professor Kaplan's research has been reported on the front pages of the New York Times and the Jerusalem Post, editorialized in the Wall Street Journal, recognized by the New York Times Magazine's Year in Ideas, and discussed in many other major media outlets. The author of more than 125 research articles, Professor Kaplan received both the Lanchester Prize and the Edelman Award, two top honors in the operations research field, among many other awards. An elected member of both the National Academy of Engineering and the Institute of Medicine of the US National Academies, he has also twice received the prestigious Lady Davis Visiting Professorship at the Hebrew University of Jerusalem, where he has investigated AIDS policy issues facing the State of Israel. Kaplan’s current research focuses on the application of operations research to problems in counterterrorism and homeland security.

In 2014, he was elected to the presidency of the Institute for Operations Research and the Management Sciences (INFORMS), the world’s largest society of operations research and analytics academics and professionals.  He will serve as President-Elect in 2015, President in 2016, and Past-President in 2017.


Talk 7 - January 29th 1:30pm-3:00pm

Susan Martonosi

Professor of Mathematics, Harvey Mudd College

Balancing Cost and Profit in the Public Sector Vaccine Market: Case Studies of DTaP and COVID-19

Vaccine markets in the United States are vulnerable to the development of monopolies due to few manufacturers and high research and development costs. This work addresses how the government can ensure the cost-effective procurement of pediatric vaccines and the new COVID-19 vaccine from private manufacturers. The Centers for Disease Control and Prevention’s (CDC) significant patronage of vaccines affords them leverage in negotiating public-sector prices that prevent the formation of monopolies, but existing vaccine pricing literature excludes the CDC as a rational player. We combine optimization and game theoretic techniques to address cost-effective immunization. We present two case studies, one of the pediatric diphtheria-tetanus-pertussis vaccine (DTaP) and one of the recently developed COVID-19 vaccine.



Susan E. Martonosi, PhD, Professor of Mathematics, focuses her research on the application of operations research and analytics methodology to problems in the public sector, including homeland security, humanitarian logistics, and public policy. Her work has included probabilistic models to guide aviation security policy related to passenger and cargo screening and shipping container screening policy; game theory, social networks analysis and graph theory to solve problems in resource allocation and terrorist network disruption; epidemiological techniques coupled with optimization models for the efficient allocation of interventions against malaria; and game theory models for negotiating pediatric vaccine prices in the public sector. She is currently engaging students in research in professional basketball analytics.

Martonosi has conducted research with the RAND Corporation, examining the feasibility of screening options for shipping containers at US ports. She also served as a high school mathematics teacher with the Peace Corps in the Republic of Guinea.


Event Time: 

2021 - 09:00