有人-无人作战智能体任务联盟形成策略方法  

Strategy of Manned-unmanned Combat Agents Task Coalition Formation

作　　者：万路军[1] 姚佩阳[1] 孙鹏[1] 邓长来 税冬东[1] 

机构地区：[1]空军工程大学信息与导航学院 [2]93307部队

出　　处：《空军工程大学学报（自然科学版）》2013年第3期10-14,共5页Journal of Air Force Engineering University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(70771157;61273048);空军工程大学信息与导航学院博士创新基金资助项目(KDY2011002)

摘　　要：为解决有人-无人作战智能体协同执行任务形成联盟问题,采取先任务分组后进行智能体到任务组匹配的联盟形成策略,即利用任务间相关程度具有传递性的特点,引入任务距离概念,设计了任务能力需求一致的任务分组策略,并给出基于可达矩阵变换的分组求解方法。针对联盟形成是分布式能力聚合过程的特点,在考虑智能体到任务组的匹配程度优先级和作战智能体能力利用效率的基础上,设计一种基于拍卖机制的作战智能体到任务组的匹配策略。通过案例计算,得到离岛突击想定下的任务联盟形成方案,分析所形成联盟的资源冗余性,验证了所提联盟形成策略和方法的有效性。To solve the manned-unmanned combat agents coalition formation problem in task enforcement, a method of the coalition formation strategy of task grouping first and the agents-group matching is adopted. The concept of task distance is introduced in order to make best use of the correlation transitivity between tasks. Then the tasks grouping strategy of coincident task requirement is designed and the grouping solu- tion-finding method based on reachable matrix transform is presented. Considering the match degree prior- ity and agent capability utilization efficiency in combat agents and task groups, the combat agent and task group match strategy based on auction mechanism is designed. A task coalition formation project under a given scenario is achieved by calculating, the resource redundancy of the formed task coalition is analyzed, which shows that the proposed coalition formation strategy and method are effective.

关 键 词：任务联盟 能力匹配 任务距离 拍卖机制 资源冗余 

分 类 号：V37[航空宇航科学与技术—航空宇航推进理论与工程] TP391.9[自动化与计算机技术—计算机应用技术]