CAJ | 학술논문

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.
위료촉진복개망절점지간적합작,이임의상린적량개절점위박혁삼여자,용순책략박혁건립료문제모형,증명료정지상태하적절점상호전발박혁시전형적수도곤경,이급재무한중복수도곤경박혁정황하,기우침봉상대책략적합작균형시취약적,우연적망락고장회인발징벌행위,종이도치절점간적불합작.제출일충가용착적침봉상대책략TTFT(Tolerant Tit-for-tat),사절점능구재일정시간내용인망락고장적발생.증명료사용해책략가이재발생망락고장적정황하,재유한시간내사절점체도은정적합작상태,병증명료작폐절점적수익증량상대교저,능구유효강저절점작폐적가능성.모의시험표명,재다개절점교호적배경하,해책략가이용인일정비솔적고장,촉진절점합작,현저제고전체절점적수익총액,병가이강저작폐절점적수익,유효방지작폐.
Aiming at improving the cooperation of nodes in overlay 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 environ-mert.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 moti-vation to cheat. The simulation shows that 1) the TTFT strategy can maintain the cooperation and significantly improve the overall payoff when unexpected network failure occurs;2) the TTFT can keep the payoff of cheating nodes in a relatively low level so that it can effectively avoid cheating.