Title
Assortment Optimization under Non-Parametric Choice Models
Abstract
Assortment optimization is a fundamental challenge for firms seeking to maximize revenue by selecting the best subset of products to offer customers with heterogeneous preferences. A key aspect of this problem is the choice model, which captures customer behavior and decision-making. In this talk, we explore assortment optimization under two non-parametric choice models: the decision forest model, a tree-based ensemble approach, and the consideration set model, a set-based ensemble framework. We analyze the approximability of the corresponding assortment optimization problems and propose exact solution methods based on integer programming and Benders decomposition. Using a real-world dataset, we demonstrate the effectiveness of our framework, showing how it can seamlessly incorporate consumer behavioral anomalies into a firm’s assortment planning strategies.
Bio
Yi-Chun Akchen is an Assistant Professor at the School of Management, University College London. His research leverages optimization and machine learning to analyze consumer choice from a data-driven perspective and offer operational insights. He earned his Ph.D. from the University of California, Los Angeles (UCLA).