有限次重复囚徒博弈中的合作机制研究  

Cooperation mechanism in finite iterated prisoner's dilemma:revisit based on real-life decision making simulation

作　　者：杨城[1] 吕峻闽[1,2] 缪春池[1] 

机构地区：[1]西南财经大学经济信息工程学院,成都611130 [2]西南财经大学天府学院,成都610051

出　　处：《计算机应用研究》2012年第4期1322-1325,共4页Application Research of Computers

摘　　要：模仿现实中人们的决策方式,提出类"触发策略"的策略思想,将原问题由双策略的多阶段博弈转换为多策略的一次性博弈,并建立起扩展的支付矩阵;然后运用进化博弈理论,将随机扰动引入复制子动态,从理论上说明有限次重复囚徒博弈之所以能够涌现合作是复制效应和变异效应共同作用的结果;最后通过建立多主体系统的仿真模型,进一步分析和验证了合作涌现的门限条件和稳定状态。This article proposed a new decision-making process based quasi trigger strategy approach to convert FIPD study from the dual-strategy multistage game into multi-strategy one-short game,and then established an extension payoff matrix and adopted the replicator dynamics equation with stochastic disturbance to explain theoretically that the emergence of cooperation in FIPD was a common result of imitation effects and mutation effects.This paper developed a multi-agent system simulation model in the end to further discuss the threshold of cooperation emergence in FIPD and its stabilization.

关 键 词：有限次重复囚徒博弈 进化博弈 演化稳定策略 类“触发策略” 随机扰动 

分 类 号：F224.32[经济管理—国民经济]