基于集合覆盖理论的Agent协作问题研究  

Resarch on Agent cooperation algorithm based on set covering problem

作　　者：杨克巍[1] 李兴兵[1] 李孟军[1] 岑凯辉[1] 

机构地区：[1]国防科学技术大学信息系统与管理学院,湖南长沙410073

出　　处：《系统工程学报》2009年第6期759-763,共5页Journal of Systems Engineering

摘　　要：多智能Agent系统中的协作体现多Agent系统(MAS)的灵活性、整体性,通过协作提高Agent群体完成任务的效率.将集合覆盖理论(SCP)引入MAS系统协作行为中的任务分配问题求解,使用改进的低logarithmic ratio bound集合覆盖理论求解方法,详细阐述了利用SCP理论求解Agent任务分配问题的算法,并根据一个战场作战Agent任务分配实例进行了计算,有效地解决战场作战Agent的任务分配问题.Cooperation among intelligent Agents embodies multi-Agent system＇ s flexibility and integration. Agents can accomplish hard tasks effectively under group cooperation. This paper addresses an improved algorithm which applies the low logarithmic ratio bound method to solve set covering problem （SCP）. This algorithm can be used in Agent task assigned problem directly to enhance the effectiveness of group cooperation in multi-Agent system （MAS）. The whole modeling and solving process of combat Agents cooperation in the field is described by using the improved algorithm. Through a field fight case and the result, it is demostrated the improved algorithm can be used to solve the task assigned problem of MAS.

关 键 词：AGENT协作 集合覆盖理论 任务分配问题 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]