arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1309.1815cs.GT
收藏

战略代理人网络中的信息共享

Information Sharing in Networks of Strategic Agents

Jie Xu, Yangbo Song, Mihaela van der Schaar

摘要:

确保社交网络(如意见一致、合作估计、分布式学习和适应等)随着智能体数量的激增和有效运作,参与智能体需要通过重复共享信息来相互协作。然而,分享信息往往是昂贵的代理人,而导致他们没有直接的直接利益。因此,缺乏合作的激励,旨在最大化自己的个体效用的战略代理人将保留而不是共享信息,导致网络的低效运行甚至崩溃。在本文中,我们开发了一个系统的框架,设计分布式评级协议,旨在激励战略代理相互合作,共享信息。建议的激励协议利用代理人的相互作用的持续性质来分配评级,并通过它们来确定未来的奖励和惩罚:代理人的行为是直接享受高评级-因此更多的未来访问其他人的信息;代理人没有直接享受低评级-因此更少的未来访问其他人的信息。与现有的评级协议不同,该协议以分布式方式在线运行,并考虑到代理的底层互连性以及它们的异质性。我们证明,在许多部署方案的无政府代价(PoA)通过采用建议的评级协议是一个。在PoA大于1的设置中,我们表明,建议的评级协议仍然显着优于现有的激励机制,如针锋相对。重要的是,所提出的评级协议也可以有效地运行在部署方案中,战略代理人的互动随着时间的推移,新的代理加入网络的网络拓扑结构。

英文摘要:

To ensure that social networks (e.g. opinion consensus, cooperative estimation, distributed learning and adaptation etc.) proliferate and efficiently operate, the participating agents need to collaborate with each other by repeatedly sharing information. However, sharing information is often costly for the agents while resulting in no direct immediate benefit for them. Hence, lacking incentives to collaborate, strategic agents who aim to maximize their own individual utilities will withhold rather than share information, leading to inefficient operation or even collapse of networks. In this paper, we develop a systematic framework for designing distributed rating protocols aimed at incentivizing the strategic agents to collaborate with each other by sharing information. The proposed incentive protocols exploit the ongoing nature of the agents' interactions to assign ratings and through them, determine future rewards and punishments: agents that have behaved as directed enjoy high ratings -- and hence greater future access to the information of others; agents that have not behaved as directed enjoy low ratings -- and hence less future access to the information of others. Unlike existing rating protocols, the proposed protocol operates in a distributed manner, online, and takes into consideration the underlying interconnectivity of agents as well as their heterogeneity. We prove that in many deployment scenarios the price of anarchy (PoA) obtained by adopting the proposed rating protocols is one. In settings in which the PoA is larger than one, we show that the proposed rating protocol still significantly outperforms existing incentive mechanisms such as Tit-for-Tat. Importantly, the proposed rating protocols can also operate efficiently in deployment scenarios where the strategic agents interact over time-varying network topologies where new agents join the network over time.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑