当前位置: X-MOL 学术Arch. Math. Logic › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Knowledge, behavior, and rationality: rationalizability in epistemic games
Archive For Mathematical Logic ( IF 0.4 ) Pub Date : 2021-06-24 , DOI: 10.1007/s00153-021-00784-5
Todd Stambaugh , Rohit Parikh

In strategic situations, agents base actions on knowledge and beliefs. This includes knowledge about others’ strategies and preferences over strategy profiles, but also about other external factors. Bernheim and Pearce in 1984 independently defined the game theoretic solution concept of rationalizability, which is built on the premise that rational agents will only take actions that are the best response to some situation that they consider possible. This accounts for other agents’ rationality as well, limiting the strategies to which a particular agent must respond, enabling further elimination until the strategies stabilize. We seek to generalize rationalizability to account not only for actions, but knowledge of the world as well. This will enable us to examine the interplay between action based and knowledge based rationality. We give an account of what it means for an action to be rational relative to a particular state of affairs, and in turn relative to a state of knowledge. We present a class of games, Epistemic Messaging Games (EMG), with a communication stage that clarifies the epistemic state among the players prior to the players’ actions. We use a history based model, which frames individual knowledge in terms of local projections of a global history. With this framework, we give an account of rationalizability for subclasses of EMG.



中文翻译:

知识、行为和理性:认知博弈中的合理化

在战略情境中,代理人根据知识和信念采取行动。这包括了解其他人的战略和对战略概况的偏好,还包括其他外部因素。Bernheim 和 Pearce 于 1984 年独立定义了合理化的博弈论解决方案概念,其建立在理性主体只会对他们认为可能的某些情况做出最佳反应的前提下。这也解释了其他代理的合理性,限制了特定代理必须响应的策略,从而能够进一步消除,直到策略稳定。我们寻求概括合理性,不仅可以解释行为,还可以解释世界的知识。这将使我们能够检查基于行动的理性和基于知识的理性之间的相互作用。我们给出了一个行动相对于特定的事态,进而相对于知识的状态是合理的意味着什么。我们提出了一类游戏,即认知消息游戏(EMG),它有一个交流阶段,可以在玩家采取行动之前澄清玩家之间的认知状态。我们使用基于历史的模型,该模型根据全球历史的局部预测来构建个人知识。有了这个框架,我们给出了 EMG 子类的合理化。我们使用基于历史的模型,该模型根据全球历史的局部预测来构建个人知识。有了这个框架,我们给出了 EMG 子类的合理化。我们使用基于历史的模型,该模型根据全球历史的局部预测来构建个人知识。有了这个框架,我们给出了 EMG 子类的合理化。

更新日期:2021-06-24
down
wechat
bug