The Committee on Stochastic Programming is pleased to announce the winers of the triennial Dupačová-Prékopa Best Student Paper Prize in Stochastic Programming. The prize was awarded at the 16th International Conference on Stochastic Programming (ICSP) held in Davis, California, USA, July 24-July 28, 2023.
The award committee consists of Johannes O. Royset (Naval Postgraduate School, chair), Werner Römisch (Humboldt-Universität Berlin), and Claudia Sagastizábal (University of Campinas).
The prize is kindly supported by the European Working Group on Stochastic Optimization (EWGSO).
Finalists:
- Kayla Cummings, MIT: Activated Benders Decomposition for Day-ahead Paratransit Itinerary Planning (with Alexandre Jacquillat, Vikrant Vaze)
- Fenpeng Li, Columbia University: General Feasibility Bounds for Sample Average Approximation via Vapnik–Chervonenkis Dimension (with Henry Lam)
- Hanyang Li, University of Minnesota: A Decomposition Method for Two-Stage Stochastic Programs with Nonconvex Recourse (with Ying Cui)
- Niels van der Laan, University of Groningen: A Converging Benders’ Decomposition Algorithm for Two-Stage Mixed-Integer Recourse Models (with Ward Romeijnders)
Winner: Niels van der Laan
Citation: The paper describes a new solution methodology for the two-stage mixed-integer recourse problem with general mixed-integer variables in both stages. While recognized as a fundamental problem in stochastic programming since the early days of the field, the problem has eluded the development of satisfactory solution methodologies. The paper constructs a Benders’ decomposition algorithm using a new family of optimality cuts that, in turn, rely on parametrically solving extended formulations of the second-stage problems. The authors establish theoretical convergence and demonstrate computational efficiency on investment planning and capacity expansion problems. The paper appears in Operations Research. The committee judges the paper a seminal contribution to the field of stochastic programming and anticipates several efficient solvers being constructed based on its ideas.