基于任务需求匹配的网络连通质量控制方法  

Network connectivity quality control method based on matching task requirement

作　　者：何杏宇 王怡 杨桂松[1] 贾明权[3] 周玥 徐心玥 He Xingyu;Wang Yi;Yang Guisong;Jia Mingquan;Zhou Yue;Xu Xinyue(School of Optical-Electrical&Computer Engineering,University of Shanghai for Science&Technology,Shanghai 200093,China;College of Communication&Art Design,University of Shanghai for Science&Technology,Shanghai 200093,China;Southwest China Institute of Electronic Technology,Chengdu 610036,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海理工大学出版印刷与艺术设计学院,上海200093 [3]中国西南电子技术研究所,成都610036

出　　处：《计算机应用研究》2023年第10期3125-3131,共7页Application Research of Computers

基　　金：国家自然科学基金资助项目(61602305,61802257);上海市自然科学基金资助项目(18ZR1426000,19ZR1477600);南通市科技局社会民生计划资助项目(MS12021060);浦东新区科技发展基金产学研专项资助项目(PKX2021-D10);敏捷智能计算四川省重点实验室开放式基金资助项目。

摘　　要：现有多智能体系统网络研究方法缺乏对局部网络质量与区域性任务需求之间的匹配分析,这将影响任务的完成效率和质量,为此提出一种基于任务需求匹配的网络连通质量控制方法。首先,利用智能体移动状态分析系统网络的间歇连通性;然后,结合任务需求以及系统的间歇网络连通性构建基于主从模式的多智能体子网集合,进而从三个方面评估子网集合的连通质量;最后,提出基于网络连通质量控制的多智能体移动优化模型,用最大的子网集合连通质量表示多智能体网络连通质量,在智能体移动距离和网络连通质量的约束下求解任务完成率最大化的多智能体移动策略,通过该策略形成多智能体网络以执行任务。实验结果表明,该方法可以有效控制网络连通质量,维持任务完成率并提高智能体移动效用。The existing multi-agent network researches ignore the matching between qualities of local networks and the requirements of regional tasks,which influences task completion efficiency and quality.In view of this,this paper proposed a network connectivity quality control method based on matching task requirement.Firstly,this method analyzed the intermittent connectivity of multi-agent network by using states of agents.Secondly,it constructed a subnet set based on leader-follower model to execute tasks,according to the requirements of tasks and the intermittent connectivity of network,and it further evaluated the connectivity quality of the subnet set.Finally,this method designed a multi-agent mobility optimization model based on network connectivity quality,in which the maximum connectivity quality of subnet set represented to the network connectivity quality.Under the constraints of agent’s mobile distance and network connectivity quality,this model obtained multi-agent mobility strategy with optimal task completion rate.Experimental results verify that the proposed method can effectively control the network connectivity quality,maintain the task completion rate and improve the mobile utility of agents.

关 键 词：多智能体系统 任务需求 网络连通性分析 网络连通质量 

分 类 号：TP393.03[自动化与计算机技术—计算机应用技术]