SPS Past Prize Winners

Celebrating excellence in Stochastic Programming—discover the past prize events that recognized outstanding contributions and achievements in the field.

2025: RogerJ-B Wets Junior Researcher Best Paper Prize in Stochastic Programming

First Prize: Bradley Sturt (University of Illinois at Chicago)

A nonparametric algorithm for optimal stopping based on robust optimization

Citation: This paper addresses the complex problem of optimal stopping within the framework of stochastic optimal control. Optimal stopping is a central topic in option pricing and mathematical finance, and it also plays a significant role in principal-agent problems. To tackle this challenge, Brad introduces a novel and highly innovative approach that enables approximation of the optimal stopping problem to arbitrary precision. His method involves a detailed analysis of the structure of Markovian stopping rules, revealing a totally unimodular structure within an underlying binary bilinear program. This key insight allows the problem to be solved efficiently using robust optimization techniques. As

the sole author, Brad brings fresh and impactful perspectives to this challenging class of problems. The jury believes this work has the potential to make a significant contribution to the field.

Second Prize: Rui Gao (University of Texas, Austin)

Finite-Sample Guarantees for Wasserstein Distributionally Robust Optimization: Breaking the Curse of Dimensionality

Citation: This paper addresses addresses a fundamental question in distributionally robust optimization as how to choose a small but reliable radius, such that the expected loss with respect to the worst-case distribution in the Wasserstein ball provides an upper bound for the expected loss with respect to the true distribution. The earlier guidelines on the radius choice were either asymptotic or over conservative, demanding the radius to scale in the order of n^(-1/d), where d denotes the dimension of the uncertain

parameters. Rui's paper shows that the radius is reliable when it scales in the order of 1/root(n), where n denotes the data size. This result waives the reliance of the exponent of n on d and provides the first finite-sample guarantee for generic Wasserstein DRO problems without suffering from the curse of dimensionality. Rui is a single author and a junior faculty member who independently worked on this challenging problem, and innovates both statistical analysis frameworks and optimization theories.

Prize Committee: Andrzej Ruszczynski (Rutgers University), Siqian Shen (University of Michigan at Ann Arbor), and Wim van Ackooij (EDF-Lab Paris-Saclay, chair).