CAREER: Mechanism Design for Resource-Bounded Agents: Indirect Revelation and Strategic Approximations

职业：资源有限智能体的机制设计：间接启示和战略近似

• 批准号: 0238147
• 负责人: David Parkes
• 金额: $ 59.91万
• 依托单位: Harvard University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2003
• 资助国家: 美国
• 起止时间: 2003-02-15 至 2008-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0238147&HistoricalAwards=false
• 关键词: CAREER; Mechanism; Design; Resource; Bounded

Systems with distributed computation across open networks share many characteristics with economies, with individual computational devices that are largely autonomous and represent the self-interest of multiple users, including both individuals and businesses. A fundamental challenge is to design useful mechanisms to solve distributed problems in these open systems despite the self-interest of individual agents. This work proposes a foundational study into the issues that arise in computational mechanism design with computationally-bounded agents. One component proposes the study of indirect revelation mechanisms, in which agents are able to participate without computing, or revealing, complete information about their preferences or local constraints. Another component proposes the study of mechanisms that provide approximations to game-theoretic desiderata, such as strategy proofness. In both cases, the research agenda calls both for new theories and for new computational methods, that leverage the theory constructively within systems. The broader impact of the work promises the continued integration of game-theoretic and economic methods into computational systems, with applications ranging from automated mediation in e-marketplaces, to automated negotiation between devices in open systems. An important component involves the development of innovative graduate and undergraduate curricula, to help to educate students about problems at the interface between computer science and economics.

在开放网络上进行分布式计算的系统与经济体有许多共同的特征，其中个人计算设备在很大程度上是自主的，代表了多个用户（包括个人和企业）的自身利益。一个根本的挑战是设计有用的机制来解决这些开放系统中的分布式问题，尽管个体代理的自身利益。这项工作提出了一个基础性的研究，在计算机制设计中出现的问题与计算有界的代理。 其中一个组成部分提出了间接启示机制的研究，在这种机制中，代理人能够参与不计算，或透露，完整的信息，他们的喜好或当地的限制。另一个组成部分提出的机制，提供近似的博弈论的desiderata，如策略的证据的研究。在这两种情况下，研究议程都需要新的理论和新的计算方法，在系统内建设性地利用理论。 这项工作的更广泛的影响，承诺继续整合博弈论和经济方法到计算系统中，应用范围从电子市场的自动调解，在开放系统中的设备之间的自动谈判。一个重要的组成部分涉及创新的研究生和本科生课程的发展，以帮助教育学生在计算机科学和经济学之间的接口问题。