摘要
多Agent联盟生成是多Agent系统的关键问题之一,主要研究如何在多Agent系统中动态生成面向任务的最优联盟.为使Agent能稳定的组织起来完成单Agent不能完成的任务并在成本、资源、利益等方面达到一个良好的平衡性能并达到全局最优,提出了联盟多目标综合评价模型,并将量子进化多目标算法应用于多目标多任务Agent联盟问题,运用编码的映射,将资源组合和任务分配合并为一个过程,降低了问题的复杂性.对比实验结果表明该算法求得的解的质量高,平衡性好,能有效避免了联盟死锁和资源浪费.
Multi-agent coalition formation is one of the key problems in multi-agent systems. The main research is how to dynamically generate task-oriented coalition and the optimal structure in multi-agent systems. For the establishment of a mechanism to organize agent stability completes the task that can not be completed by a single agent and reach the global optimum. This paper presents a multi-objective comprehensive evaluation model, and a multi-objective quantum evolutionary algorithm is proposed to solve multi-agent coalition formation problem. The algorithm uses coding mapping, the combination of resources and task allocation are combined to a process, which reducing the complexity of the problem. Experimental results show that the proposed algorithm can solve the multi-task agent coalition formation problem effectively and efficiently.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2012年第10期2253-2261,共9页
Systems Engineering-Theory & Practice
基金
广东省自然科学基金(10252500002000001)
广东省教育部产学研结合项目(2010B090400235)
湖南省教育厅优秀青年项目(10B062)
国家自然科学基金(60903168)
关键词
多AGENT
联盟生成
多目标优化
量子多目标进化算法
multi-agent
coalition formation
multi-objective optimization
quantum multi-objective evo-lutionary algorithm
