摘要
针对多无人机协同路径规划问题,提出了一种决策学习型蜣螂优化算法(DLDBO)。传统蜣螂优化算法(DBO)种群之间缺乏信息互换,容易陷入局部最优解。因此,利用Pearson相关系数计算个体之间的相似性,通过相似性指标判断并作出决策:若不相似,利用折射反向学习计算得到候选解,在一定程度上提高个体之间影响的同时增强算法跳出局部最优的能力;若相似,利用所提出的链式邻近学习引导蜣螂个体,增加影响个体更新的因素,充分促进个体之间的信息交流。在CEC2017测试套件的29个测试函数上进行了充分的对比实验,结果表明,DLDBO性能明显优于其他六种先进的变体算法。利用DLDBO规划无人机群的飞行路径,最终能够得到较为理想的协同路径并且有效避开威胁,优于其余三种优秀的协同路径规划算法,满足了无人机协同飞行的需求。
A decision learning dung beetle optimization algorithm(DLDBO)solves the problem of multi-UAV collaborative path planning.The traditional dung beetle optimization algorithm(DBO)lacks information exchange among populations and easily falls into local optimal solutions.Therefore,this paper used the Pearson correlation coefficient to calculate the similarity between individuals and used the similarity index to make decisions.If individuals were not similar,it applied refraction reverse learning to calculate the candidate solution,which improved interaction among individuals and enhanced the algorithm’s ability to escape local optima.If individuals were similar,it guided dung beetle individuals using the proposed chain proximity learning,increasing factors affecting individual renewal and promoting information exchange.Comparative experiments on 29 test functions of the CEC2017 test suite show that the DLDBO algorithm significantly outperforms six other advanced variants.Using DLDBO for UAV flight path planning obtains an ideal collaborative path,effectively avoiding threats and surpassing three other excellent collaborative path planning algorithms,meeting the needs of UAV collaborative flight.
作者
张乐
胡毅文
杨红
杨超
马宏远
Zhang Le;Hu Yiwen;Yang Hong;Yang Chao;Ma Hongyuan(School of Intelligent Science&Engineering,Shenyang University,Shenyang 110044,China;School of Information Engineering,Shenyang University,Shenyang 110044,China)
出处
《计算机应用研究》
北大核心
2025年第1期196-204,共9页
Application Research of Computers
基金
辽宁省兴辽英才计划资助项目(XLYC1807138)
辽宁省教育厅面上项目(LJKZ1172)。
关键词
蜣螂优化算法
折射反向学习
链式邻近学习
无人机协同路径规划
dung beetle optimization algorithm
refraction reverse learning
chain proximity learning
UAV collaborative path planning
作者简介
张乐(1975-),男(回族),辽宁沈阳人,教授,硕导,博士,主要研究方向为复杂非线性系统智能控制等;通信作者:胡毅文(2000-),男,湖北潜江人,硕士研究生,主要研究方向为智能优化算法、路径规划等(hyw0012102023@163.com);杨红(1978-),女(回族),辽宁沈阳人,教授,硕导,博士,主要研究方向为非线性系统、模糊切换系统、鲁棒控制等;杨超(1999-),男,江西宜春人,硕士研究生,主要研究方向为智能优化算法、路径规划等;马宏远(1999-),男,河北石家庄人,硕士研究生,主要研究方向为切换正系统、模糊控制等.
