Opinion Dynamics with Rational Agents

理性代理人的意见动态

• 批准号: 438508461
• 负责人: Professor Dr. Martin Hoefer
• 金额: --
• 依托单位: Arbeitsgruppe Algorithmen und Komplexität
• 依托单位国家: 德国
• 项目类别: Research Units
• 资助国家: 德国
• 项目状态: 未结题
• 来源: https://gepris.dfg.de/gepris/projekt/438508461?language=en
• 关键词: Opinion; Dynamics; Rational; Agents

Dynamic opinion formation is an important aspect of modern societies. In the age of internet platforms and social media, opinion dynamics have had significant impact on societal and political issues. In this context, the amount of information available to the agents crucially influences the dynamics of opinion formation. Opinion dynamics are a current area of interest in algorithmic game theory. In this project, we are interested in a number of important algorithmic problems in the context of standard opinion formation models. In the second phase of the project, we study questions arising from the restriction of available information and erratic behavior of agents, the influence of lacking coordination and asynchronous dynamics, as well as opinion formation in dynamic environments. In addition, we are interested in the impact of information design, i.e., strategic signaling by information providers. Our main goals are to characterize existence and structure of (meta-)stable states, as well as to derive provable bounds on convergence times for dynamics. Moreover, we will analyze approximation algorithms for information design, and prove corresponding hardness of approximation results.

动态的意见形成是现代社会的一个重要方面。在互联网平台和社交媒体时代，舆论动态对社会和政治问题产生了重大影响。在这种情况下，代理人可获得的信息量对意见形成的动态有着至关重要的影响。意见动态是算法博弈论的一个当前感兴趣的领域。在这个项目中，我们感兴趣的是一些重要的算法问题的背景下，标准的意见形成模型。在该项目的第二阶段，我们研究的问题所产生的限制，可用的信息和不稳定的行为的代理，缺乏协调和异步动态的影响，以及在动态环境中的意见形成。此外，我们对信息设计的影响感兴趣，即，信息提供商的战略信号。我们的主要目标是描述（Meta）稳定状态的存在性和结构，以及推导出可证明的动态收敛时间的界限。此外，我们还将分析信息设计的逼近算法，并证明逼近结果的相应硬度。