A1291
Title: Dynamic assortment selection with position effects
Authors: Yiyun Luo - School of Statistics and Management, Shanghai University of Finance and Economics (China) [presenting]
Abstract: In online retailing and advertising, the seller aims to offer the customers an assortment of items that incur maximal expected revenue. A new online decision-making problem, Dynamic Assortment Selection with Positioning (DAP)d, is proposed. There are two key characteristics of the DAP problem. Firstly, customers' preferences for the items are unknown at the beginning and thus need to be learned from interactions with customers. Secondly, the selected assortment and the positioning of items in the assortment would influence the customers' purchasing behaviours and, thus, the revenues. The goal is maximising overall revenue in a finite horizon by making sequential assortment and positioning decisions. Specifically, for the DAP problem, a UCB-based policy with sublinear regrets is developed. The proposed policy delivers superior performances in various simulation settings by handling the position effects well.
