基于证据理论的多Agent重复囚徒困境博弈研究

在多Agent系统中,利用证据理论与重复囚徒困境博弈对其中的合作行为进行了研究.为加强Agent之间的合作行为,构建了基于信度函数的综合信任模型.该综合信任度的建立来于两个方面:一是由以前交互历史中观察的直接证据而得到的直接信任度,二是由来自于第三方的推荐证据而得到的推荐信任度.基于综合信任模型,给出了多Agent重复囚徒困境博弈算法,并对所研究的问题进行了仿真.仿真结果表明,基于综合信任度的博弈策略能够有效地加强系统中的合作行为,提高系统的收益.

Iterated prisoner's dilemma game with multi-Agent based on evidence theory

Cooperation behavior in multi-Agent systems is studied by using evidence theory and iterated prisoner's dilemma games.To strengthen the cooperation between Agents,this paper constructed a combined trust model using the belief structure.The combined trust was composed of the direct trust based on the direct evidence and the recommended trust based on other Agents'recommended evidence.Based on the constructed model,a combined trust strategy and a game algorithm in iterated prisoner's dilemma game were presented.At last.an application caLse was simulated in a multiAgent system.The simulation results demonstrate that the proposed strategy can effectively strengthen the cooperation behavior and increase the revenues of the system.