学术动态 | 讲座预告

报告摘要

In this study, we investigate the performance of parsimonious discrete choice models in the context of demand estimation and assortment optimization. Specifically, we examine three well-known models—the Basic Attraction Model (BAM), the Exponomial Model (EM), and the Random Consideration Set (RCS) model—and introduce their generalized versions that incorporate shadow attractions: the Generalized Attraction Model (GAM), the Generalized Exponomial Model (GEM), and the Generalized Random Consideration Set Model (G-RCS). Additionally, we develop kappa-versions (kappa-BAM and kappa-RCS) to account for non-purchasing visitors or ``browsers", adding a single parameter to the base models.

Our empirical analysis encompasses both synthetic and real-world datasets. Synthetic data experiments utilize random utility models and decision forests to simulate various consumer behaviors, including irrational choices. Real-world data sources include hotel bookings, sushi preferences, and retail sales from the IRI Academic Dataset. The findings reveal that generalized models often outperform their basic counterparts, particularly as dataset size increases. Model performance is context-dependent; for example, EM excels with deep preference lists, while RCS performs better with irrational consumer behavior. Importantly, accounting for browsers significantly improves model performance in scenarios with low market share. Enhanced predictive accuracy leads to better assortment decisions, underscoring the practical value of selecting appropriate parsimonious models for retail applications.

报告人简介

Guillermo Gallego is X.Q. Deng Presidential Chair Professor at CUHK-SZ, the Liu Family Professor Emeritus at Columbia University, and the Crown Worldwide Professor Emeritus at the Hong Kong University of Science and Technology. He is an INFORMS Fellow (2012), MSOM Distinguish Fellow (2013), HKIE Fellow (2016), and an international scholar that is widely recognized as one of the pioneers of modern Dynamic Pricing. He is the recipient of the 2011 Historical Award of the INFORMS Revenue Management & Pricing Section, the 2012 INFORMS Practice Award, the 2016 INFORMS Impact Prize, the INFORMS Revenue Management & Pricing Section Prize (both in 2005 and 2021), and the INFORMS Lanchester Prize Honorable Mention (2024). He is the only scholar in Operations Management to win best paper awards from both Management Science and Operations Research, the two top journals in his field. Professor Gallego was the Chairman of the Industrial Engineering and Operations Research (IEOR) department at Columbia University (2002-2008) and the Head of the department of Industrial Engineering and Decision Analytics at the Hong Kong University of Science and Technology (2016-2022). Prof Gallego’s research interests are Dynamic Pricing and Revenue Optimization, Supply Chain Management, Electronic Commerce, and Inventory Theory. He has published influential papers in the leading journals of his field where he has also occupied a variety of editorial positions.