基于混合遗传算法的多无人机巡逻路径优化  

The optimization of multi-UAVs patrol path with hybrid genetic algorithm

作　　者：李国军[1] 郑滋椀 范英盛 卢甜甜[1] 徐志江 LI Guojun;ZHENG Ziwan;FAN Yingsheng;LU Tiantian;XU Zhijiang(Basic Courses Department,Zhejiang Police College,Hangzhou 310053,China;School of Big-Data and Network Security,Zhejiang Police College,Hangzhou 310053,China;School of Automation,Zhejiang Institute of Mechanical&Electrical Engineering,Hangzhou 310053,China)

机构地区：[1]浙江警察学院公共基础部,浙江杭州310053 [2]浙江警察学院大数据与网络安全研究院,浙江杭州310053 [3]浙江机电职业学院自动化学院,浙江杭州310053

出　　处：《浙江大学学报（理学版）》2024年第1期21-28,共8页Journal of Zhejiang University（Science Edition）

基　　金：2023JC32;浙江省“尖兵”“领雁”攻关计划项目(2023C01030);国家自然科学基金青年项目(41901160).

摘　　要：假设无人机巡逻的起、终点均为派出所,提出了一种融合传统遗传算法和爬山算法的警用无人机巡逻路径优化模型——混合遗传算法。按照轮盘赌法则,进行种群个体的选择,以增大优秀种群个体被选中的概率,达到较好的优化效果。同时定义了与路径优化相适应的基因交叉和变异规则。仿真结果表明,提出的混合遗传算法在寻优效果上明显优于传统遗传算法。Aiming at the optimization of multi-UAVs patrol path,a patrol model of multi-UAVs based on hybrid genetic algorithm is proposed.When constructing the patrol model,each UAV must start from the police station and return to the police station at the end of the patrol.The algorithm is designed by combining traditional genetic algorithm and hill-climbing algorithm.In order to achieve a better optimization effect,the roulette wheel method is employed to select the excellent individuals with higher probability when selecting individuals of the population.In the application of genetic algorithm,the rules of gene crossover and mutation adapted to path optimization are defined.The simulation results show that the proposed hybrid genetic algorithm is significantly better than the traditional genetic algorithm on the optimization effect.

关 键 词：遗传算法 爬山算法 巡逻 路径优化 

分 类 号：O224[理学—运筹学与控制论] U492.2[理学—数学]