基于无限重复博弈的P2P网络信任模型研究  

Research on trust model based on infinitely repeated games theory in P2P networks

作　　者：王保玉[1] 高承实[1] 戴青[1] 陈景林 刘洋[3] 

机构地区：[1]解放军信息工程大学,郑州450004 [2]舟山警备区,浙江舟山316000 [3]济南军区司令部,济南250000

出　　处：《计算机应用研究》2013年第9期2802-2804,共3页Application Research of Computers

基　　金：国家部委基金资助项目(9140c130103120c13062)

摘　　要：为解决P2P电子商务环境中存在的安全问题,针对现有一些信任模型的局限性,提出一种基于无限重复博弈理论的信任模型。通过建立奖惩策略,对节点的不合作行为,依据博弈策略对节点进行惩罚,同时对选择合作策略的节点进行奖励,并根据欺诈行为次数设定不同的惩罚周期。理论分析和仿真实验表明,合作策略会成为节点博弈的帕累托最优策略,该策略模型能有效遏制和惩罚恶意行为,提高节点诚实交易的积极性,从而有效提高网络环境的安全性和稳定性。To resolve the security problems in P2P networks, this paper proposed a trust model based on infinitely repeated game theory. Reward and punishment strategy in this paper not only punished the malicious behavior, but also incentivized the cooperation nodes, and the duration of penalty was determined in the times of fraudulent conduct. Mathematic analysis and simulation experiments show that the strategy of cooperation will be Pareto optimal strategy, which can punish and stifle availa- bly malicious behavior, and improve incentives of the nodes trading honestly, then the security and stability of the P2P network environment will be enhanced efficiently.

关 键 词：P2P网络 信任模型 重复博弈 奖惩策略 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]