求解逆运动学的多策略蜻蜓算法
- 分享到:
摘要:
蜻蜓算法(Dragonfly Algorithm,DA)是一种新型群智能算法,存在求解精度不高、收敛速度慢、优化不稳定等不足.基于此,提出了一种多策略改进的蜻蜓算法(Multi-strategy Improved Dragonfly Algorithm,MIDA).首先通过Logistic混沌映射改善算法的初始种群,使算法能更快锁定最优解区域;其次融合共生生物搜索算法(Symbiotic Organisms Search,SOS)来增加个体间的信息交流;最后采用了余弦扰动避免蜻蜓算法过早陷入局部最优解,并且采用正交实验验证该算法的可行性.通过12个基准函数和求解逆运动学这一实际应用来验证MIDA的有效性.结果表明,MIDA在函数优化方面遥遥领先,在求解逆运动学问题中,优化效果比其他对比算法高10%~35%.
Dragonfly algorithm is a new type of swarm intelligence algorithm,which has shortcomings such as low solution accuracy,slow convergence speed,and unstable optimization.In this regard,a multi-strategy improved Dragonfly algorithm is proposed in this paper.Firstly,the initial population of the algorithm is improved through logistic chaotic mapping,so that the algorithm can lock the optimal solution area faster.secondly,the symbiotic search algorithm is integrated to increase the information exchange between individuals.Finally,the cosine disturbance is used to avoid the premature trapping of the dragonfly algorithm local optimal solution.And the feasibility of the algorithm is verified by orthogonal experiments.This paper verifies the effectiveness of MIDA with 12 benchmark functions and the practical application of solving inverse kinematics.The results show that MIDA is far ahead in function optimization,and its optimization effect is 10%-35% higher than other comparison algorithms in solving inverse kinematics problems.
作者:
黄华娟 闵峰
Huang Huajuan;Min Feng(College of Artificial Intelligence,Guangxi Minzu University,Nanning 530006,China;College of Electronic Information,Guangxi Minzu University,Nanning 530006,China)
机构地区:
广西民族大学人工智能学院 广西民族大学电子信息学院
出处:
《betway官方app 学报:自然科学版》 CAS 北大核心 2023年第5期46-56,I0004,I0005,共13页
Journal of Henan Normal University(Natural Science Edition)
基金:
国家自然科学基金(62266007,61662005) 广西自然科学基金(2021GXNSFAA220068,2018GXNSFAA294068)。
关键词:
蜻蜓算法 混沌映射 共生生物搜索算法 余弦扰动 函数优化 逆运动学
dragonfly algorithm chaotic mapping symbiotic organisms search cosine perturbation function optimization inverse kinematics
分类号:
TP18 [自动化与计算机技术—控制理论与控制工程]
