基于蚁群劳动分工的无人机群搜寻打击策略

Search and Strike Strategy for Unmanned Aircraft Swarm Based on Ant Colony Labor Division

针对无人机集群在信息不对称条件下对地固定目标察打任务分配问题,受蚁群劳动分工群智能进化思想启发,设计了一种融合差分进化算法和动态蚁群劳动分工模型的动态搜寻打击策略。首先,将无人机搜寻打击目标的分配过程映射为蚁群觅食的劳动分工过程,综合分析无人机与目标相对距离、发现时间、暴露状态等因素对目标选择的影响,提出目标“诱惑度”。然后,充分考虑任务分配的马尔可夫性质,引入带有先验知识的精英保留策略的差分进化算法,在每次目标选择前,实时更新“诱惑度”,形成动态环境刺激更新机制。最后,建立信息不对称条件下对地固定目标察打任务分配仿真环境,采取对比实验方法,在确定的12种实验条件下对无人机搜寻打击方案进行验证。仿真结果表明,所提出的策略与两种传统策略相比,蓝方平均损失数提升3.78%、3.90%,红方任务平均消耗时间下降6.26%、6.39%,能够有效解决无人机集群在信息不对称条件下的任务分配问题,为提升无人机集群动态决策能力提供了算法支撑。

Aiming at the problem of unmanned aerial vehicle(UAV)cluster's task allocation for detecting and striking fixed targets on the ground under the condition of information asymmetry,inspired by the idea of intelligent evolution of ant colony labor division group,a dynamic search and strike strategy integrating differential evolution algorithm and dynamic ant colony labor division model is designed.Firstly,the assignment process of UAV searching and striking targets is mapped as the labor division process of ant colony foraging,and the influence of the relative distance between the UAV and the target,the discovery time,the exposure state and other factors on the target selection are comprehensively analyzed,so that the target's"temptation degree"is proposed.Then,the Markov nature of task allocation is fully considered,and a differential evolutionary algorithm with a priori knowledge of the elite retention strategy is introduced to update the"seductiveness"in real time before each target selection,forming a dynamic environmental stimulus update mechanism.Finally,a simulation environment for the task assignment of detecting and striking fixed targets on the ground under the condition of information asymmetry is established,and a comparative experimental method is adopted to validate the UAV search and strike program under the 12 experimental conditions identified.The simulation results show that,this paper's strategy improves by 3.78%and 3.90%and decreases by 6.26%and 6.39%over the two traditional strategies in the evaluation metrics of the average number of losses on the blue side and the average time consumed on the red side of the task,respectively,it can effectively solve the task allocation problem of UAV cluster under the condition of information asymmetry,and provides algorithmic support for improving the dynamic decision-making ability of UAV cluster.