【作者】 曾燕； 周佩玲； 周涛； 马婧； 杨春霞；

【Author】 ZENG Yan ZHOU Pei-ling ZHOU Tao MA Jing YANG Chun-xia (Department of Electronic Science and Technology,USTC,Hefei 230026,China)

【机构】 中国科学技术大学电子科学与技术系；

【摘要】 建立了一个经纪人既利用全局信息,又通过局域信息进行决策的少数者博弈模型,研究了系统的动力学行为。数值模拟结果显示,系统总收益与网络拓扑结构及记忆长度 m 有关,且在引入局域信息之后,系统总收益高于基本少数者博弈。结果还表明系统中个体的成功率与网络中节点的度存在一定相关性,节点的度越大,能获取的局域信息越多,经纪人成功的几率也就越大,反之越小．更多还原

【Abstract】 A model of minority game where the agents make decisions according to both global information and local information is proposed.In the model,the agents make decisions to be in minority of their neighboring agents with probability p,and act by the basic minority game with probability 1-p at each time step.The simulation results indicate that the system profit is related to the network topology and memory length m,and after the inducting of local information,the system performs better than the basic minority game with p in a considerable range.In addition,it is found that the winning rate of the agents in the system is related to the degree of the nodes in the network,the greater the degree is,the more information the agent obtains,the more he wins.更多还原