n人随机重复博弈下的策略选择研究

扩展了重复博弈下的策略选择研究.针对由于技术进步导致的新形式博弈,提出了n人随机重复博弈概念(n-person stochastic iterated prisoner dilemma,NSIPD).针对NSIPD,设计了一种博弈策略的计算机竞赛,并研究了策略选择问题.采用随机匹配选择两个局中人进行囚徒困境博弈,且每人的策略历史可被所有人知道.详细介绍了该竞赛的设计思想和具体实验,实验结果表明在此博弈环境下TIT-FOR-TAT(TFT)和背叛策略都不是最优策略,最优策略与博弈代数以及局中人策略分布密切相关.

Research on strategy selection in n-person stochastic iterated game

The strategy selection in 2-person iterated prisoner dilemma was extended and the concept of n-person stochastic iterated prisoner dilemma(NSIPD) was proposed to accommodate to new development arising from technology advancement.In NSIPD everyone's history information of choice could be seen by all players.At each game round,two players were selected randomly to play PD game.A computer tournament was used to search for the best strategy in NSIPD.The experimental results show that TIT-FOR-TAT and ALWAYS_DEFECT are not always the best strategy,and that the best strategy is closely correlated with repeated times and strategy distribution.