Strategic Learning in Games of Incomplete Information: Theory and Experiments

不完全信息博弈中的策略学习：理论与实验

• 批准号: 9011828
• 负责人: Thomas Palfrey
• 金额: $ 19.11万
• 依托单位: California Institute of Technology
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1990
• 资助国家: 美国
• 起止时间: 1990-07-01 至 1993-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9011828&HistoricalAwards=false
• 关键词: Strategic; Learning; Games; Incomplete; Information

This project involves theoretical and experimental work on a problem of great importance for game theory generally and for the evaluation of game-theoretic experiments. The problem to be addressed deals with learning and models of learning. Of particular interest is the learning which occurs when individuals play the same game against different opponents. The theoretical part of the project will focus on studying those conditions under which individuals do or do not converge in such games to the Nash equilibrium. The experimental portion of the project will pursue a combination of two classes of designs in an attempt to evaluate both the theory of strategic learning and some of the myopic learning models developed by the investigators. The first design involves repeated play of a game with incomplete information between fixed opponents, while the second involves repeated play between changing opponents. These designs will enable the investigators to study learning explicitly within a dynamic framework.

这个项目涉及一个对博弈论和博弈论实验评估都非常重要的问题的理论和实验工作。要解决的问题涉及学习和学习模式。特别令人感兴趣的是当个人与不同的对手玩同一个游戏时发生的学习。该项目的理论部分将专注于研究在哪些条件下，个体在这种博弈中可以或不会收敛到纳什均衡。该项目的实验部分将采用两类设计的组合，试图评估策略学习理论和研究人员开发的一些短视学习模型。第一种设计涉及在固定对手之间信息不完全的情况下重复博弈，而第二种设计涉及变化对手之间的重复博弈。这些设计将使研究人员能够在动态框架内明确地研究学习。