Dynamic Coalition Task Allocation of Heterogeneous Multiple Agents

异构多智能体联盟动态任务分配

¹ Naval Aviation University, Yantai 264001, China
² Unit 91213 of PLA, Yantai 264001, China

Received: 8 December 2019

Abstract

The dynamic coalition task allocation of heterogeneous multiple UAV agents is researched, which is divided into two parts. Firstly, the consensus based coalition algorithm(CBCA) is presented via consensus based bundle algorithm(CBBA), considering complex constraints of specific equipment requirements and coupling the relationships between the subtasks and the time windows. Secondly, three dynamic planning strategies are proposed in cope with appearance of new tasks during the allocation process. Finally, the feasibility and applicability of the present algorithm and dynamic planning strategies are validated in the scenario of a search and attack mission executed by multiple unmanned search aerial vehicles(USAVs) and unmanned combat aerial vehicles (UCAVs).

摘要

研究了突发新任务的动态场景下异构多无人机智能体分布式联盟任务分配问题，主要包括两方面内容：首先扩展了一致性包算法（consensus based bundle algorithm，CBBA），考虑任务载荷资源约束、子任务耦合关系约束及执行窗口约束等条件提出了一致性联盟算法（consensus based coalition algorithm，CBCA）；其次，针对新任务出现的动态应用需求，研究了3种动态任务分配策略，分别为无重规划动态分配策略（consensus based coalition algorithm with no resetting，NR-CBCA）、完全重规划动态分配策略（consensus based coalition algorithm with full resetting，FR-CBCA）及部分重规划动态分配策略（consensus based coalition algorithm with partial resetting，PR-CBCA）。最后，以侦察型无人机和攻击型无人机协同执行对地侦察攻击任务为例，验证了CBCA算法的可行性及3种分配策略对动态任务场景的适用性。