摘要
针对传统方法难以有效协调无人机编队间的任务规划和频谱资源分配导致任务收益低下及频谱资源浪费的问题,构建了一个由无人机编队规模、任务执行顺序和带宽分配构成的混合整数非线性规划问题系统模型,提出了一种结合遗传算法与梯度投影法的多起点多无人机编队联合优化算法。利用遗传算法生成离散变量的候选解,通过梯度投影法对连续变量进行高效优化。所提算法能有效应对多起点多无人机编队规划场景,从多个初始点同时搜索解空间,避免算法陷入局部最优解,提升了解的多样性和鲁棒性。仿真结果表明,所提算法在任务分配合理性、无人机资源利用率和任务收益等方面显著优于传统算法,并且在复杂场景下展现了更高的任务收益和收敛性。
In the field of task planning and spectrum resources allocation for multiple unmanned aerial vehicle(UAV)formations,traditional approaches often face challenges in effectively coordinating task planning and spectrum resources allocation across UAV formations,resulting in inefficient task execution and underutilization of spectrum resources.To address this issue,a mixed-integer nonlinear programming model is formulated,which consists of UAV formation sizes,task execution sequences,and bandwidth allocation strategies.In addition,a multi-start joint optimization algorithm integrating a genetic algorithm with the gradient projection method is proposed.The genetic algorithm is employed to generate candidate solutions for discrete variables,while the gradient projection method efficiently optimizes continuous variables.The proposed algorithm can effectively handle multi-start multi-UAV formation planning scenarios,which explores the solution space from multiple initial points to avoid falling into local optimum,thereby enhancing solution diversity and robustness.The simulation results validate the superiority of the proposed algorithm over the traditional methods in terms of the rationality of task allocation,utilization efficiency of UAV resources and overall task yield.Furthermore,the algorithm demonstrates enhanced convergence and significantly higher task rewards in complex scenarios.
作者
钱鹏智
王轶宇
张佳栋
张余
QIAN Pengzhi;WANG Yiyu;ZHANG Jiadong;ZHANG Yu(The 63rd Research Institute,National University of Defense Technology,Nanjing 210007,China)
出处
《陆军工程大学学报》
2025年第6期72-80,共9页
Journal of Army Engineering University of PLA
关键词
无人机编队
任务规划
频谱资源分配
梯度投影法
遗传算法
UAV formation
task planning
spectrum resources allocation
gradient projection method
genetic algorithm
