基于改进麻雀搜索算法的多维复杂函数优化问题的求解  

Solution of Multi-dimensional Complex Function Optimization Problems Based on Improved Sparrow Search Algorith

作　　者：张贺杰 赵茂先 ZHANG Hejie;ZHAO Maoxian(College of Mathematics and Systems Science,Shandong University of Science and Technology,Qingdao,Shandong 266590,China)

机构地区：[1]山东科技大学数学与系统科学学院,山东青岛266590

出　　处：《数学建模及其应用》2024年第1期20-29,51,共11页Mathematical Modeling and Its Applications

基　　金：国家自然科学基金(U22B2049)。

摘　　要：传统麻雀搜索算法在寻找最优解过程中,存在种群多样性不够高、易产生局部最优、收敛精度不稳定等问题.本文给出了一种改进的麻雀搜索算法,首先,用改进的Cubic和Bernoulli混合混沌映射初始化种群分布提高种群多样性;其次,在算法迭代过程中引入非线性自适应惯性权重和Levy飞行策略,调节算法的搜索范围和精度,改善算法的收敛速度和局部寻优能力;然后,引入鲸鱼优化算法的捕食策略进行扰动,避免陷入局部最优;最后,在12个基准测试函数上与传统麻雀搜索算法和其他算法进行评估,实验结果验证了改进的算法具有较好的收敛速度和求解精度,并提升了局部搜索能力.The Sparrow Search Algorithm is an intelligent optimization algorithm characterized by its simple structure and clear principles.The traditional sparrow search algorithm suffers from the problems of insufficient population diversity,the tendency to produce local optima and unstable convergence accuracy in the process of finding the optimal solution.In this paper,an improved sparrow search algorithm is given.Firstly,the population distribution is improved by utilizing an enhanced combination of Cubic and Bernoulli chaotic mappings to enhance population diversity.Secondly,nonlinear adaptive inertia weight and Lévy flight strategy are introduced during the algorithm iteration process to adjust the search range and precision,thus improving the convergence speed and local optimization capability of the algorithm.Furthermore,the predation strategy of the Whale Optimization Algorithm is incorporated to introduce perturbation and prevent getting trapped in local optimum.In conclusion,the improved algorithm was evaluated against traditional Sparrow Search Algorithm and other algorithms on twelve benchmark test functions.Experimental results confirmed that the enhanced algorithm exhibits superior convergence speed and solution accuracy,and it has enhanced local search capabilities.

关 键 词：麻雀搜索算法 混沌映射 非线性自适应惯性权重 Levy飞行 鲸鱼优化算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]