基于混沌映射和混合变异的蜻蜓算法  

Dragonfly algorithm based on chaotic mapping and mixed mutation

作　　者：姜封国 韩雪松 岳攀 陈雨 Jiang Fengguo;Han Xuesong;Yue Pan;Chen Yu(School of Architecture&Civil Engineering,Heilongjiang University of Science&Technology,Harbin 150022,China)

机构地区：[1]黑龙江科技大学建筑工程学院,哈尔滨150022

出　　处：《黑龙江科技大学学报》2025年第1期103-109,共7页Journal of Heilongjiang University of Science And Technology

基　　金：黑龙江省自然科学基金项目(LH2022E108)。

摘　　要：针对标准蜻蜓算法存在的全局探索能力不足,收敛精度不高等自身缺陷问题,提出一种采用混沌映射和高斯分布及非均匀分布混合优化的蜻蜓算法。通过引入混沌映射使种群初始化时产生更好的混沌种群,采用高斯分布和非均匀分布来优化蜻蜓分布位置,避免其陷入局部最优,提高算法寻优能力和收敛精度。将改进后的算法应用于8种不同的测试函数和一个13杆桁架结构的优化设计,并与原始蜻蜓算法和其他群智能算法计算得到的结果比较,在算例中与原始蜻蜓算法相比结构质量减少了5.86%,改进后的蜻蜓算法收敛速度更快,收敛精度更高。This paper is aimed at addressing the defects of insufficient global exploration ability and low convergence accuracy in standard dragonfly algorithm,and proposes a hybrid optimization dragonfly algorithm with chaotic mapping,Gaussian distribution and non-uniform distribution.The study involves introducing chaotic mapping to generate a better chaotic population as initialized;using Gaussian distribution and non-uniform distribution to optimize the distribution position of dragonflies for avoiding falling into local optimality and improving the optimization ability and convergence accuracy;and applying the improved algorithm to 8 different test functions and an optimization design of a 13-bar truss structure to compare with the results calculated by the original dragonfly algorithm and other swarm intelligence algorithms.The final result shows that the structural mass is reduced by 5.86%compared with the original dragonfly algorithm,indicating that the improved dragonfly algorithm has faster convergence speed and higher convergence accuracy.

关 键 词：蜻蜓算法 tent混沌映射 高斯分布 非均匀分布 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]