2024 Harvey J Greenberg Research Award
Winners: Selvaprabu Nadarajah (University of Illinois Chicago) and Andre Cire (University of Toronto) for their paper "Self-adapting network relaxations for weakly coupled Markov decision processes." Management Science (2024).
The paper "Self-adapting Network Relaxations for Weakly Coupled Markov Decision Processes" by Nadarajah and Cire (2024) introduces Feasibility Network Relaxations (FNR), a novel approach to approximating weakly coupled Markov decision processes by employing network representations to fully capture combinatorial constraints. Theoretical and empirical analyses demonstrate that FNR outperforms Lagrangian methods in terms of model size and solution quality, with significant improvements
observed in practical applications such as inventory management and telecommunications maintenance.
Honorable Mention: Pierre Bonami, Andrea Lodi, and Giulia Zarpellon. "A classifier to decide on the linearization of mixed-integer quadratic problems in CPLEX." Operations Research 70, no. 6 (2022): 3303-3320.
This paper presents an end-to-end methodology for integrating a learning algorithm into the IBM CPLEX solver, which has been incorporated into version 12.10 (Fall 2019) and has achieved a twofold speed-up for mixed-integer quadratic programming (MIQP) problems. This advancement marks a significant step towards the practical application of machine learning techniques in optimization.
Committee: Jeff Linderoth (University of Wisconsin Madison), Zeyu Liu (West Virginia University), Esra Buyuktahtakin Toy (Virginia Tech), Hamidreza Validi (Texas Tech), and Willem-Jan Van Hoeve, Chair (Carnegie Mellon)
2023 Harvey J Greenberg Research Award
Winners: Wei Zhang (Tsinghua), Alexandre Jacquillat (MIT), Kai Wang (Tsinghua), Shuaian Wang (Hong Kong Polytechnic) for their paper "Routing Optimization with Vehicle-Customer Coordination," Management Science, 2023.
Ride sharing and dial-a-ride systems are impacting lives daily. This paper expands upon the literature on dial-a-ride problems by considering the vehicle-customer coordination and by being able to support real-time operations in large-scale systems. The paper provides a deep and rigorous study of several variants considering floating targets (single-stop and multi-stop variants, and a dial-a-ride problem in its deterministic and online setting). The article sets-up theoretical and methodological foundations for further research involving routing applications with floating targets.
Runner-up: Li Chen (NUS), Long He (George Washington), Yangfang (Helen) Zhou (Singapore Management), for their paper "An Exponential Cone Programming Approach for Managing Electric Vehicle Charging", Operations Research, 2023.
Committee: David Bergman (Connecticut), Akshay Gupte (Edinburgh, in lieu of Cynthia Rudin), Ivana Ljubic, Chair (ESSEC Paris)
2022 Harvey J Greenberg Research Award
Winners: Zeyu Liu and Anahita Khojandi and Xueping Li and Akram Mohammed and Robert L Davis (Tennessee), and Rishikesan Kamaleswaran (Emory) for their paper "A Machine Learning–Enabled Partially Observable Markov Decision Process Framework for Early Sepsis Prediction," INFORMS Journal on Computing, published online March 22,2022.
Sepsis can be triggered by the body's extreme response to an infection and can be life-threatening. Existing sepsis prediction algorithms suffer from high false-alarm rates. The authors present an integrated machine learning (ML) and partially observable Markov decision process framework to address this issue. This approach is calibrated and tested using physiological data collected from bedside monitors. The framework reduces false-alarm rates and improves sepsis prediction accuracy compared to existing ML benchmarks. This is a comprehensive paper with novel contributions to computing and important practical implications. The committee members commend the authors for this excellent work.
Honorable Mention: Eyyüb Y. Kıbıs ̧ ̇I. Esra Büyüktahtakın, Robert G. Haight, Najmaddin Akhundov, Kathleen Knight, and Charles E. Flower, for their paper "A Multistage Stochastic Programming Approach to the Optimal Surveillance and Control of the Emerald Ash Borer in Cities," INFORMS Journal on Computing, 33(2), 2021.
Committee: Archis Ghate, Chair (Washington), Ariela Sofer (George Mason), David Woodruff (UC Davis)
2021 Harvey J Greenberg Research Award
Winners: Hamidreza Validi, Austin Buchanan, and Eugene Lykhovyd for their paper "Imposing Contiguity Constraints in Political Districting Models," to appear, Operations Research, 2021.
A classical problem in operations research that concerns the generation of political districting maps; early approaches using integer programming date back to the 1960s. In this well-rounded and timely paper, the authors survey modern attempts to add contiguity constraints to these classical models and propose two new formulations that are shown to impose contiguity in an easier way. The authors investigate the theoretical relationships between these models, discover new sets of cutting planes, and develop Lagrangian techniques to fix variables. With these innovations, they solve, for the first time, optimally compact districting maps for 21 US states at the census tract level. Their source code, models, and data are publicly available. The selection committee believes that this paper represents the spirit of Harvey Greenberg's work: tackling problems of societal impact with real-life data using state-of-the-art operations research techniques, providing an intriguing example of OR in practice.
Committee: Pascal van Hentenryck (Chair), Sven Leyffer, Alice Smith
2020 Harvey J Greenberg Research Award
Winners: Danial Davarnia and Willem-Jan van Hoeve for their paper "Outer Approximation for Integer Nonlinear Programs via Decision Diagrams," forthcoming in Mathematical Programming Series A.
Committee: Dorit Hochbaum (chair), Pascal van Hentenryck, Karla Hoffman