https://sites.google.com/view/row-series/home
ROW by Dr. Amine Bennouna
Time: Oct 16, 2026 at 15:00 CET
Title: What Data Enables Optimal Decisions? A Study of Data Informativeness in Optimization Under Uncertainty
Abstract:
We study the fundamental question of how informative a dataset is for solving a given optimization problem under uncertainty. In our setting, a dataset provides partial information about unknown parameters of the optimization problem. Focusing on linear programs with uncertainty in the cost vector, we characterize exactly when a dataset is sufficient to recover an optimal decision. Our main contribution is a sharp geometric characterization that reveals how the task structure and the uncertainty set jointly determine data requirements. Building on this characterization, we develop a practical algorithm that identifies, for a given task, a minimal sufficient dataset whose acquisition guarantees finding the optimal decision.
Our results demonstrate that small, carefully selected data can often fully determine optimal decisions, providing a principled foundation for decision-aware experimentation, data selection and field experiments.
Bio:
Amine Bennouna is an Assistant Professor of Operations at the Kellogg School of Management, Northwestern University. He earned his PhD in Operations Research from MIT and his BSc and MSc in Applied Mathematics from École Polytechnique. His research is in learning and optimization under uncertainty. His recent work studies how data informs decisions in operations problems, with the goal of developing decision-focused experimental design and data collection methods.
--------------------------------------------
Register for access to zoom link.