复杂约束条件下异构多智能体联盟任务分配

Coalition task allocation of heterogeneous multiple agents with complex constraints

中文关键词  多智能体系统  分布式决策  一致性包算法  联盟生成  无人机

英文关键词  multi agent systems  distributed decision making  consensus-based bundle algorithm(CBBA)  coalition formation  unmanned aerial vehicles

中文摘要

      分布式任务决策是提高多智能体系统自主性的关键. 以异构多智能体协同执行复杂任务为背景, 首先建立 了一种考虑任务载荷资源约束、任务耦合关系约束及执行窗口约束等条件的异构多智能体分布式联盟任务分配模 型; 其次, 对一致性包算法(CBBA)进行了扩展, 提出了基于改进冲突消解原则的一致性联盟算法(CBCA), 以实现异 构多智能体协同无冲突任务分配, 并进一步证明了在一定条件下CBCA算法收敛于改进顺序贪婪算法(ISGA). 最后 通过数值仿真, 验证了CBCA算法求解复杂约束条件下异构多智能体联盟任务分配问题的可行性和快速性.

英文摘要

      Distributed decision making is the key of enhancing the autonomy of multi-agent systems. Firstly, based on the problem of multi agents cooperating to perform complicated tasks, a decentralized coalition task assignment model of heterogeneous multiple agents with consideration of specific equipment requirements of tasks, coupled constraints and the time window constraint is proposed. Secondly, the consensus-based bundle algorithm (CBBA) is extended to the consensusbased coalition algorithm (CBCA) with the improved principles of conflicts resolution to achieve the conflict-free solution of the task assignment problem. Further CBCA is proved to be convergent to the improved sequential greedy algorithm (ISGA). Finally, simulation results of multiple unmanned search and combat aerial vehicles performing search ad attack missions demonstrate the feasibility and rapidity of the proposed algorithm.