基于鸽群智能行为的大规模无人机集群聚类优化算法

Clustering optimization algorithm for large-scale unmanned aerial vehicle based on intelligent behavior of pigeons

未来空中战场,大规模无人机集群系统将成为主导力量.而对大规模无人机集群系统进行分组聚类是完成作战任务规划的必要步骤.在实际战场中无人机受到有限通信约束,无法得到全面而有效的全局作战信息.因此本文提出一种基于鸽群智能行为的大规模无人机集群聚类优化算法.根据聚类模型设计鸽群优化算法,研究分析导航能力优异的鸽群智能行为,将鸽群飞行过程中的层级网络机制映射到鸽群优化算法中,解决有限交互环境下的信息不完整问题.一方面,依据鸽群在飞行过程中来自临近个体的引导更为有效直接,因而在有限交互环境下,基本鸽群优化算法中的全局最优信息由交互范围内的最优个体信息替代;另一方面,鸽群的中心位置更新包括三部分:增量惯性部分、模仿部分、环境影响部分.为验证改进后鸽群优化算法在有限交互范围下的有效性,本文采用三种算法针对三个数据集进行聚类分组,仿真结果表明改进后的鸽群优化算法在最优解与平均最优解上均有改善,为实际作战环境下的无人机集群系统聚类分组提供了有效的解决方法.

In future air battlefield,clusters of large-scale unmanned aerial vehicle(UAV)will become the dominant force.The effective grouping and clustering of large-scale UAV cluster systems are necessary steps to complete combat tasks.Due to the limited communication constraints in the battlefield,UAVs cannot obtain comprehensive and effective global combat information.Thus,this paper proposed a large-scale UAV clustering optimization algorithm based on the intelligent behavior of pigeons.Also,this paper studied and analyzed the intelligent behavior of the flock of pigeons with excellent navigation ability and mapped the hierarchical network mechanism in the flight process of the pigeon flock into the pigeon-inspired optimization(PIO)algorithm.Hence,it solved the problem of incomplete information in a limited interactive environment.On one hand,it is more effective and direct to guide the pigeon flock from adjacent individuals during flight.Therefore,under the limited interaction condition,the global optimal information of the basic PIO algorithm is replaced by the optimal individual information within the interaction range.On the other hand,the central position renewal of the pigeon flock consists of three parts:inertia part,imitation part,and environmental impact part.In order to verify the effectiveness of the improved PIO algorithm in a limited interactive range,this paper adopted three algorithms to cluster three datasets.Simulation results reveal that the improved PIO algorithm achieved significant improvement in the optimal solution and the average optimal solution;notably,the computation time does not increase significantly,thus providing an effective solution for the clustering of UAV cluster systems in the actual combat environment.