题　目：Component Pricing with a Bundle Size Discount

嘉　宾：王纯，副教授，清华大学经济管理学院

主持人：孙海龙 助理教授 上海交通大学安泰经济与管理学院

时　间：2023年3月24日（周五）14:00-15:30pm

地　点：安泰楼B201

内容简介:

Firms selling multiple products usually adopt bundle pricing in their marketing strategy for the purpose of extracting large consumer surplus. In this work, we propose and analyze a bundling mechanism, referred to as component pricing with a bundle size discount (CPBSD), which sells bundles for the summed prices of the included products (component pricing) minus a discount based on the number of products purchased (bundle size discount). CPBSD is conceptually simple and has been widely used in real-world business settings. Theoretically, we show that CPBSD subsumes several most well-studied bundling mechanisms including component pricing, pure bundling, and bundle size pricing as special cases. We further prove that, under a general condition, CPBSD attains the optimal profit asymptotically for a large number of products among all bundle pricing mechanisms. From a practical perspective, we formulate a mixed-integer linear program for the optimal pricing scheme of CPBSD, and we also develop an approximation algorithm for efficiently solving CPBSD in large-scale problems. Through comprehensive numerical experiments, we show that CPBSD demonstrates superior performance in contrast to other bundling mechanisms. In particular, compared to bundle size pricing whose outstanding empirical performance has been extensively tested in the literature, CPBSD performs especially well when products are heterogeneous and the production costs are high. Furthermore, the performance of CPBSD increases significantly when the potential surplus provided by products are negatively correlated with the product valuations. We also show that our approximation algorithm may achieve a performance that is very close to the optimal CPBSD. Given the theoretical guarantee and the computational efficiency, CPBSD presents an appealing selling mechanism for retailers to improve their profitability.

演讲人简介：

王纯，清华大学经济管理学院管理科学与工程系副教授, 哥伦比亚大学运筹学博士。主要研究兴趣包括：金融工程、库存管理、收益管理中的随机控制与不确定性决策；强化学习算法；机器学习与商务数据分析。

欢迎广大师生参加！