基于分布式协商进化算法的多Agent目标冲突消解  

Goal conflict resolution of multi-agent systems based on distributed negotiation evolution algorithm

作　　者：杨萍[1] 刘卫东[1] 毕义明[1] 

机构地区：[1]第二炮兵工程学院,陕西西安710025

出　　处：《系统工程与电子技术》2009年第8期1918-1922,共5页Systems Engineering and Electronics

基　　金：第二炮兵工程学院创新人才资助项目(xy200703)资助课题

摘　　要：针对多Agent系统研究中的目标冲突消解问题,建立了在多个Agent的局部目标和系统全局目标间进行协调优化的多目标优化模型。在多Agent分布式规划的框架下,提出了一种基于遗传算法(genetic algorithm,GA)的分布式协商进化算法,用于求解多目标规划模型。针对GA搜索中保持解的多样性、提高收敛速度等问题,对选择算子进行了设计。通过仿真实验,证明新的选择算子能有效提高解的质量。最后将该算法应用于部队机动协同路线规划的目标冲突消解问题,验证了其有效性。In order to deal with goal conflict resolution in the research of multi agent systems, a multiobjective optimization model is presented to search for some compromise among coordinate local goals of multiagent and collective goals of systems. Under the distributed planning frame of multi-agent, a distributed negotiation evolution algorithm based on the genetic algorithm （DNEAGA）which is used for solving multi-objective programming model is proposed. In order to keep the solutions diversity and increase the convergence speed, a selection operator in genetic algorithm is designed. Simulation experiments show that the new selection operator can improve the solutions quality effectively. Finally, the DNEAGA is applied to solving the goal conflict in maneuver route planning of operations, and the effectiveness of the proposed algorithm is verified.

关 键 词：多AGENT系统 目标冲突消解 多目标优化 分布式协商进化算法 遗传算法 PARETO最优解 

分 类 号：N945.25[自然科学总论—系统科学]