基于DBSCAN聚类的异构多智能体分层任务分配方法

A Hierarchical Task Assignment Method for Heterogeneous Multi-agents Based on DBSCAN Clustering

针对多约束条件下大规模探测/通信智能体集群协同探测任务分配问题,从全局与局部相结合的角度,提出了一种分层任务分配求解方法。首先,根据通信距离约束对所有任务节点进行聚类预分组,将集群任务分配问题划分为上层全局任务分配和底层局部任务分配。然后,根据聚类结果采用启发式算法求解探测/通信智能体组间全局任务分配结果。随后,根据探测智能体的全局任务分配结果,采用遗传算法对探测智能体组内任务进行分配。最后,通信智能体根据探测智能体的组内任务分配结果,采用基于虚拟节点的方法进行组内任务分配。实验结果表明,相较于直接求解方法,分层任务分配方法不仅解决了大规模集群协同任务分配问题,还可以在保证优化目标值相近的情况下,缩短70%以上的求解时间,较快得到相对最优的任务分配结果。

Large-scale area reconnaissance detection is a common operational task for drones,unmanned vehicles,and other intelligent agents.It has significant application potential in various fields such as military,fishery,and energy.This paper aims to address the problem of collaborative detection task allocation for large-scale detection/communication agent clusters with multiple constraints.To accomplish this,a hierarchical task allocation solution method is proposed,considering both global and local aspects.First,all task nodes are clustered and pre-grouped based on communication distance constraints.This divides the cluster task allocation problem into upper-level global task allocation and bottom-level local task allocation.Next,a heuristic algorithm is employed to solve the global task allocation results among detection/communication agent groups,taking into account the clustering outcomes.Following this,a genetic algorithm is used to allocate tasks within the detection agent group,leveraging the global task allocation results.Finally,the communication agent employs a method based on virtual nodes to allocate tasks within the group,utilizing the intra-group task allocation results of the detection agent.Experimental results demonstrate that the hierarchical task allocation method significantly reduces the solution time by over 70%without compromising the optimization target values.