Date(s) - February 17, 2023
10:40 am - 11:30 am
Categories No Categories
Title : Multi-Armed Bandits with Endogenous Learning Curves: An Application to Split Liver Transplantation (with Savannah Tang, Andrew Li and Sridhar Tayur)
Abstract : Proficiency in many sophisticated tasks is attained through experience-based learning, in other words, learning by doing. For example, transplant surgeons need to practice difficult surgeries to master the skills required, call center staff need to handle customer calls to improve their ability to resolve customer issues, and new franchisees learn to operate smoothly over time. This experience-based learning may affect other stakeholders, for example, patients eligible for transplant surgeries. Such a situation illustrates the classical exploration versus exploitation trade-off: A central planner may want to identify and develop surgeons with high aptitudes, while ensuring that patients still have excellent outcomes and equitable access to organs. We formulate a multi-armed bandit (MAB) model, in which parametric learning curves are embedded in the reward functions to capture experience-based learning. In addition, our model includes provisions ensuring that the choices of arms are subject to fairness constraints (ensuring equity), incorporates queueing dynamics (to capture waiting time dynamics), and arm dependence (to capture learning across similar surgeries). To solve our MAB problem we propose the L-UCB, FL-UCB, and QFL-UCB algorithms, all variants of the upper confidence bound (UCB) algorithm that attain O(log t) regret on problems enhanced with experience-based learning, fairness concerns, queueing dynamics, and arm dependence. We demonstrate our model and algorithms on the split liver transplantation (SLT) allocation problem, showing that our algorithms have superior numerical performance compared to standard bandit algorithms in a setting where experience-based learning and fairness exist. From a methodological point of view, our proposed MAB model and algorithms are generic and have broad application prospects. From an application standpoint, our algorithms could be applied to help evaluate potential strategies to increase the proliferation of SLT and other technically-difficult medical procedures.
Bio : Alan Scheller-Wolf is the Richard M. Cyert Professor of Operations Management at the Tepper School of Business. He has previously served as Senior Associate Dean of Research at Tepper, and the head of the doctoral program. He received his PhD from the IE/OR department of Columbia University in 1996, having completed his doctorate under the advising of Karl Sigman. He has a Bachelor of Science in Mathematics and Computational Science, and a Bachelor of Arts in Art History, from Stanford University. Prior to his time at Columbia, Alan served for 2 1/2 years in the Peace Corps as a mathematics teacher, in Serowe, Botswana.
Alan’s research interests include inventory theory (especially ATO systems, systems with capacities, alternate supply options and/or perishable products), healthcare (organ transplantation, blood supply, treatment for opioid addiction disorder) energy, service systems, computer science, stochastic processes and queueing theory. He has served on the editorial boards of Management Science, Operations Research, M&SOM, and QUESTA. He has completed consulting projects with Amazon, Caterpillar, John Deere, The American Red Cross, and The Vera Institute of Justice. He currently teaches courses in Quality and Sustainable Operations.