一种改进的多agent分布式联盟形成算法

发布时间：2018-05-17 12:11

本文选题：多agent系统 + 联盟形成　；参考：《控制与决策》2017年04期

【摘要】：仅采用任务性价比作为多智能体任务分配过程中的任务选择标准,会产生时间消耗大、资源利用低等问题.为此,综合任务性价比和智能体资源的特点,提出了多任务准备度的概念.根据多智能体任务分配过程的收敛性和时效性,采用Learning Automata算法动态调整任务准备度各项的权重;进而利用该方法模拟解决了低、中、高3种任务需求下多智能体任务分配问题.仿真实验结果验证了所提出方法的有效性,资源冗余可至少减少20%.
[Abstract]:Only using the cost / performance ratio of the task as the task selection criterion in the task assignment process of multi-agent will lead to the problems of large time consumption and low utilization of resources. In this paper, the concept of multi-task readiness is put forward by synthesizing the characteristics of cost performance and agent resources. According to the convergence and timeliness of multi-agent task assignment process, Learning Automata algorithm is used to dynamically adjust the weight of task readiness, and then the multi-agent task assignment problem with low, medium and high task requirements is solved by using this method. The simulation results show that the proposed method is effective, and the resource redundancy can be reduced by at least 20%.
【作者单位】：北京理工大学自动化学院;
【基金】：国家基金委创新研究群体项目(61321002);国家基金委重大国际合作项目(61120106010) 国家自然科学基金项目(61573062)
【分类号】：TP18

【相似文献】