基于监察博弈的对等网激励机制的研究  

Game Theoretic Framework Based Supervise for Incentives in P2P Systems

作　　者：霍英[1,2] 陈志刚[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083 [2]韶关学院计算机科学学院,广东韶关512005

出　　处：《小型微型计算机系统》2010年第8期1509-1513,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60573127;60873082)资助;湖南省自然科学基金项目(06JJ30032)资助

摘　　要：在对等网络中,存在着大量的Free-Ride节点,会严重影响到系统的性能,如何激励这些节点提供服务是提高对等网性能的重要因素.本文在节点理性假设的基础上,借助博弈论为工具,提出一种采用监察策略的对等网激励机制,并计算出了采用监察博弈的混合策略纳什均衡策略,重点分析了监察成本与参与人收益之间的博弈关系.仿真实验表明,通过采用适当的监察概率,能保证在较小的网络开销下,达到激励节点主动参与贡献资源的目的.The large numbers of free-ride have decreased severely on the performance of P2P networks.How to incentive nodes to provide service is an important factor to improve the performance of P2P networks.Based on the hypothesis of the nodes are rational and in virtue of game theory,an incentive mechanism adopting supervise politic in P2P networks was bring forward.The mixture Nash equilibrium adopting supervise politic was figured out.And the game relationship between supervise costing and participate income was mainly analyzed.Simulation results show that,the nodes would initiative contribute their resource and the cost of P2P network is small at the same time,by carefully configuration of the probability of supervise in our incentive mechanism.

关 键 词：对等网 博弈论 激励 监察 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]