Customer-individual bundling and auction design in revenue management

这个项目的重点是动态定价，更具体地说，多产品市场的拍卖设计。组合拍卖是最常见的拍卖机制类型，因为它们允许不同类型的效用函数，包括补充和替代。在本研究计画中，我们希望探讨组合拍卖中两个密切相关的研究问题。首先，我们要分析边约束（如限制赢家的数量）对拍卖格式的效率的影响，并分析考虑边约束的新的价格规则。这种边约束是该领域大多数应用的前提条件，但在理论上很少受到关注。其次，我们想研究在线环境，在这种环境中，商品定期出售，但每个时期的投标人的确切数量是未知的，有时甚至没有关于其他竞争对手估值的分布信息。这种环境不同于独立的私人价值模型，这是大部分理论的基础，但类似的环境，因为他们在收入管理文献中分析。然而，在线设置可能为收入管理中的许多新应用提供更好的描述，其中组合拍卖被定期用于购买或出售易腐物品。例如电视和广播广告的销售、许多采购拍卖或飞往特定目的地的航班的销售。除了理论开发，我们计划在实验室中分析投标人在这两种情况下的行为，并开发相应的软件工具。

This project focuses on dynamic pricing, more specifically auction designs for multi-product markets. Combinatorial auctions are the most general types of auction mechanisms, as they allow for different types of utility functions including complements and substitutes. In this research project, we want to explore two closely related research questions in combinatorial auctions. First, we want to analyze the impact of side constraints (such as limits on the number of winners) on the efficiency of auction formats and analyze new pricing rules for ascending auctions, which take side constraints into account. Such side constraints are a precondition for most applications in the field, but they have received little attention in the theory. Second, we want to study online environments, where goods are sold periodically, but the exact number of bidders is unknown in each period and sometimes not even distributional information about the valuations of other competitors is available. This environment is different to the independent private values model, which has been the basis for much of the theory, but similar to environments as they are analyzed in the revenue management literature. Online settings might, however, provide a much better description for many new applications in revenue management, where combinatorial auctions are used periodically to buy or sell perishable items. Examples are the sale of TV and Radio ads, many procurement auctions, or the sale of flights to particular destinations. In addition to theory development, we plan to analyze bidder behavior in both settings in the lab and develop respective software tools.