基于博弈理论的移动自组网激励机制研究  

Study on game-based incentive mechanism of mobile ad hoc network

作　　者：闻英友[1,2] 赵博[1] 赵宏 

机构地区：[1]东北大学医学影像计算教育部重点实验室,辽宁沈阳110819 [2]东软集团研究院,辽宁沈阳110719

出　　处：《通信学报》2014年第4期44-52,共9页Journal on Communications

基　　金：国家自然科学基金资助项目(60903159;61173153);沈阳市科技计划基金资助项目(1091176-1-00);中央高校基本科研业务费基金资助项目(N110818001;N100218001)~~

摘　　要：针对移动自组网中节点在报文转发过程中的表现出的自私行为,利用博弈理论,从静态和动态2个方面对其进行了完整的建模与分析。首先,提出了一种严厉针锋相对策略,并建立了一个无限重复报文转发博弈模型,求得了激励一致性条件。然后,利用演化博弈理论对节点由自私向协作转变的动态过程进行了分析,并证明了严厉针锋相对策略的演化稳定性。仿真结果表明,即使在自私节点比率为1的条件下,只要合理选择惩罚参数,均可有效激励自私节点的协作转发行为,整体网络性能最多可提升80%。According to the selfishness of rational mobile ad hoc network nodes showed during the packet forwarding, the selfish behavior statically and dynamically based on game theory were analyzed and modeled. A stern tit for tat strat-egy（STFT） was proposed to motivate node cooperation, and an infinite repeated game model was established to analyze the node behavior. Then, an incentive-compatible condition was obtained analytically. The dynamic process of selfish node turning to cooperate using the evolutionary game theory was studied, and the evolutionary stability of STFT was proved. Simulation results show that, even if the ratio of selfish nodes is one, by setting punishment parameters reasona-bly, the overall network performance can be improved 80%at most.

关 键 词：移动自组网 重复博弈 演化博弈 激励机制 自私性 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]