Footnote:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments February 16, 2022 erstellt
Description:
Demand estimation and assortment optimization are two critical problems in revenue management. When dealing with these two problems, a discrete choice model (DCM) is usually used to capture customers' demands. A critical assumption for DCM is that each customer can purchase at most one product. However, purchasing multiple products is ubiquitous in real life, especially in e-commerce. To depict the multi-choice behavior, this paper proposes the multi-choice random utility model and the multi-choice rank list model, and establishes the relationship between these two models. In addition, we incorporate the behavior-reveal-preference (BRP) framework into our multi-choice rank list model by using customers' behavior data (including viewing, clicking, adding to shopping carts, and purchasing) to improve the modeling power. Based on numerical experiments on real-world data from Cainiao Network, we demonstrate that considering customers' multi-choice behavior and combining the behavior-reveal-preference framework can drastically improve the model's predictive power. Moreover, we propose a mixed-integer linear programming (MILP) formulation for the assortment optimization under the multi-choice rank list model. Besides, the personalized assortment can be updated for each customer if the corresponding behavior information is available