摘要
在短的time-scale中,在一个确定的系统拓扑结构下,用解析的方法确定复杂管理系统Agent的最优策略.构建了一个确定型多局域世界图博弈模型,对确定型多局域世界图,分两种情况——局域内Agent的合作博弈和局域之间的非合作博弈,讨论Agent行为均衡解;并将局域内Agent的合作随机微分博弈与局域Super-Agent之间的非合作随机微分博弈进行了耦合.以确定复杂管理系统Agent的最优策略.最后讨论了最优策略的稳定性.
In a short time-scale, the optimal strategy of Agent for complex management systems is determined by using the analytic method in a deterministic system topology structure. A game model in the deterministic multi-local-worlds graph is constructed and the equilibrium solutions of Agents behavior are discussed for deterministic Multi-Local-Worlds graphs in two cases, which are the cooperative game between agents in the same local-world and non-cooperative game between agents in different Local-Worlds; The two different solutions with corresponding game models are coupled together to get the optimal strategies of Agents and the stability of the optimal strategy is discussed at the end of paper.
出处
《管理科学学报》
CSSCI
北大核心
2016年第1期35-46,共12页
Journal of Management Sciences in China
基金
国家自然科学基金资助项目(71503188
71040001
70771083)
关键词
确定型
多局域世界图
Agent的最优策略
随机微分博弈
deterministic
multi-local-worlds graph
optimal strategies of agents
stochastic differential game
