Communication, Beliefs, and Revenue Bounds

该奖项资助了博弈论的三个研究项目。第一种分析了参与者自由交流的战略情境，包括发表没有战略动机的声明。第二个项目寻求开发新的方法来确定哪些均衡结果与具体的一致性概念相一致。第三个项目分析了在需求信息不完全的情况下，利润最大化的多产品垄断者如何选择营销策略。这里的目标是制定最大可能销售额的界限，这可能会为这些公司为什么经常捆绑货物提供一些洞察。这些项目的结果可能会为分析讨价还价、谈判和重复的市场互动创造新的工具。第一个项目分析的模型没有对参与者所有可能的行动进行共同的显式列举；在实践中，这限制了沟通的可能性。取而代之的是，PI开发了一种模型，在这种模型中，玩家能够自由地宣布他们的私人信息。使用序列均衡的一种新的改进，结果通常在Kohlberg-Merten(1986)意义上是稳定的。第二个项目提出了一个论点，即范式完美，而不是向后归纳或顺序均衡，捕捉了游戏中动态一致性的决策理论后果。然后，研究人员开发了一种基于均衡的合理化模拟，以解决不变性问题。这意味着，不变性是作为结果寻求的，而不是作为公理强加的。第三个项目解决了多维优化机构设计中的一个复杂问题。虽然没有已知的方法来提供解决方案的有用特征，但在实践中，我们观察到公司使用简单的机制。该项目将研究良好的收入界限是否能够解释这些机制的广泛使用。

This award funds three research projects in game theory. The first analyzes strategic situations in which participants communicate freely, including making statements with no strategic motives. The second project seeks to develop new methods for determining which equilibrium results are consistent with specific concepts of consistency. The third project analyzes how a profit-maximizing multi-product monopolist will choose a marketing strategy when it has incomplete information about demand. The goal here is to develop bounds on the maximum possible sales, which could provide some insight into why such firms frequently bundle goods. The results of these projects may create new tools for analyzing bargaining, negotiation, and repeated market interactions.The first project analyzes a model without the common explicit enumeration of all possible actions for the players; in practice this limits communication possibilities. Instead the PI develops a model in which players are able to freely announce their private information. Using a new refinement of sequential equilibrium, the resulting outcomes are often stable in the Kohlberg-Mertens (1986) sense. The second project presents an argument that normal form perfection, rather than backward induction or sequential equilibrium, captures the decision-theoretic consequences of dynamic consistency in games. The researcher then develops an equilibrium-based analog of rationalizability to address the issues of invariance. This means that invariance is sought as a result rather than being imposed as an axiom. The third project tackles a complicated problem in multi-dimensional optimal mechanism design. While there are no known methods for providing useful characterizations of the solution, in practice we observe that firms use simple mechanisms. The project will examine whether good revenue bounds can explain the wide use of these mechanisms.