机会网络中基于信誉度惩罚的重复博弈模型  

REPEATED GAME MODEL BASED ON CREDIBILITY PENALTY FOR OPPORTUNISTIC NETWORK

作　　者：赵宇红[1] 包凤莲 王静宇[1] Zhao Yuhong;Bao Fenglian;Wang Jingyu(School of Information Engineering,Inner Mongolia University of Science and Technology,Baotou 014010,Inner Mongolia,China;Third Staff and Workers Hospital of Baotou Steel Group,Information Management Division,Baotou 014010,Inner Mongolia,China)

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010 [2]包钢集团第三职工医院,内蒙古包头014010

出　　处：《计算机应用与软件》2021年第12期121-127,共7页Computer Applications and Software

基　　金：国家自然科学基金项目(61662056)。

摘　　要：节点的自私行为将严重影响机会网络的传输性能。为激励节点协作,提出一种基于信誉度惩罚策略的重复博弈模型。惩罚策略以信誉度度量节点的历史行为,并设计不同程度的惩罚,重复博弈中节点考虑未来的长久收益以及对自私表现下惩罚的恐惧而选择协作转发。利用演化博弈理论分析并证明了节点由自私向协作行为转变的动态过程中的演化稳定性。仿真结果表明,该模型可有效激励节点参与协作,在自私节点较多时,也能保证较高节点传输成功率和较低的网络延迟。The selfish behavior of nodes will seriously affect the transmission performance of opportunity network.In order to stimulate node cooperation,a repeated game model based on credit penalty strategy is proposed.Credibility was used to measure the historical behavior of nodes in punishment strategy,and various degrees of punishment were designed accordingly.In the repeated game,nodes chose cooperative forwarding in consideration of long-term future benefits and fear of punishment in the face of selfishness.Evolutionary game theory was used to analyze and prove the evolutionary stability of nodes in the process of changing from selfish behavior to cooperative behavior.Simulation results show that the model can effectively stimulate nodes to participate in cooperation,and can also guarantee a higher transmission success rate and a lower network delay when there are more selfish nodes.

关 键 词：机会网络 自私节点 重复博弈 信誉度 演化博弈 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]