基于集合覆盖理论Agent协作问题研究.pdf

24 6 Vo.l 24 No. 6 2009 12 JOURNAL OF SYSTEM S ENGINEERING Dec. 2009 do:i 10. 969/ j. issn. 1000- 5781. 2009. 06. 019 Agent 杨克巍, 李兴兵, 李孟军, 岑凯辉 (, 41007 ) : 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 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 logarith m ic ratio bound集合覆盖 论求解方法, 详细阐述了利用 SCP 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. : Agent; ; : TP18 : A : 1000- 5781( 2009) 06- 0759- 05 Resarch onAgent cooperation algorithm based on set covering problem YANG Kewei, LIXingbing, LIMengjun, EN Kaihui ( College of Information System and M anagement, NationalU niversity ofDefense Technology, Changsha 41007 , Ch ina) Abstract: Cooperation among intelligentAgents embod ies mu ltiAgent system s flex ibility and inte gration. A gents can accomp lish hard tasks effectively under group cooperation. Th is paper addresses an mi proved algorithm w hich applies the low logarithm ic ratio bound method to solve set covering problem ( SCP). This algorithm can be used inAgent task assigned problem d irectly to enhance the effectiveness of group cooperation in mu ltiAgent system ( MAS). The w hole modeling and solving process of combat Agents cooperation in the field is described by using the mi proved algorithm. Through a field fight case and the result, it is demostrated the mi proved algorithm can be used to solve the task assigned problem ofMAS. Key words: Agent cooperation; set covering problem ( SCP); task assigned problem 0 , MAS