一种可容错的覆盖网节点合作激励策略  

A Fault-Tolerant Cooperation Incentive Strategy for Overlay Network Nodes

作　　者：王锐[1] 朱青林[1] 钱德沛[1,2] 刘涛[2] 

机构地区：[1]北京航空航天大学计算机学院,北京100191 [2]西安交通大学电子与信息工程学院,陕西西安710049

出　　处：《电子学报》2010年第2期327-332,共6页Acta Electronica Sinica

基　　金：国家自然科学基金(No.60673180);国家863高技术研究发展计划(No.2006AA01A106;2006AA01A118;No.2009AA01Z144;No.2009AA01A131)

摘　　要：为了促进覆盖网节点之间的合作,以任意相邻的两个节点为博弈参与者,用纯策略博弈建立了问题模型,证明了静止状态下的节点相互转发博弈是典型的囚徒困境,以及在无限重复囚徒困境博弈情况下,基于针锋相对策略的合作均衡是脆弱的,偶然的网络故障会引发惩罚行为,从而导致节点间的不合作.提出一种可容错的针锋相对策略TTFT(Tolerant Tit-for-tat),使节点能够在一定时间内容忍网络故障的发生.证明了使用该策略可以在发生网络故障的情况下,在有限时间内使节点达到稳定的合作状态,并证明了作弊节点的收益增量相对较低,能够有效降低节点作弊的可能性.模拟试验表明,在多个节点交互的环境下,该策略可以容忍一定比率的故障,促进节点合作,显著提高全体节点的收益总额,并可以降低作弊节点的收益,有效防止作弊.Aiming at improving the cooperation of nodes in oveday networks,a novel incentive strategy which is used in the packet forwarding game between two neighboring nodes is proposed. An infinitely iterated game model is established to promote the cooperation.We proof that the cooperative equilibrium based on Tit-for-tat strategy is vulnerable in the realistic network environment.A fault Tolerant Tit-for-tat （TTFT） strategy is proposed to cope with the network failure. We proof that this strategy can maintain cooperation in case of network failure. Cheating towards this strategy can get few extra payoffs so that there is little motivation to cheat. The simulation shows that 1 ） the TIFF strategy can maintain the cooperation and significantly improve the overall payoff when unexpected network failure occurs;2） the TIFF can keep the payoff of cheating nodes in a relatively low level so that it can effectively avoid cheating.

关 键 词：覆盖网 节点合作 无限重复博弈 针锋相对策略 容错 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]