11/10/22 | 4:15pm | E51-145

Peter Frazier

Eleanor and Howard Morgan Professor
Cornell University

Abstract: Universities faced difficult decisions throughout the pandemic. Is it safe to hold classes in person? Should we require masks? What about vaccination? This mirrored tensions throughout society that pitted freedom against safety. At Cornell University, which serves a population of over 30K students and employees, leaders responded to this challenge by turning to data and mathematical models. In fall 2020, when most other universities closed for in-person instruction or reopened with high infection counts, the speaker led a team of Cornell OR faculty and students in using mathematical models to design an asymptomatic screening program testing students multiple times per week. Model-based comparisons to fully virtual instruction convinced university leadership to reopen for in-person classes. The effort was successful, with only 0.5% of students and employees infected over the semester. Building on this success, we continued to use data and mathematical models to support university decisions: policies on masking, vaccination, social distancing, classroom ventilation, and university events; adapting testing frequencies to vaccination and variants; planning hotel capacity for quarantine and isolation housing; and relaxing restrictions as nearly full vaccination and Omicron's reduced virulence allowed a return to normalcy. We offer this experience as an example of how OR/MS professionals can help policy makers to better leverage data and quantitative reasoning.

Bio: Peter Frazier is the Eleanor and Howard Morgan Professor of Operations Research and Information Engineering at Cornell University. He is also a Senior Staff Applied Scientist at Uber. He leads Cornell's COVID-19 Mathematical Modeling Team, which designed Cornell's testing strategy to support safe in-person education during the pandemic. He also works on Bayesian optimization, multi-armed bandits, and incentive design for social learning. At Uber, he managed UberPool's data science group and currently helps to design Uber's pricing systems.