融合自适应t分布和随机游走策略的松鼠优化算法的研究  

Research on Squirrel Optimization Algorithm Integrating Adaptive t Distribution and Random Walk Strategy

作　　者：张莲 贾浩 张尚德 赵梦琪 赵娜 黄伟 ZHANG Lian;JIA Hao;ZHANG Shangde;ZHAO Mengqi;ZHAO Na;HUANG Wei(Chongqing Energy Internet Engineering Technology Research Center,Chongqing 400054;School of Electrical and Electronic Engineering,Chongqing University of Technology,Chongqing 400054)

机构地区：[1]重庆市能源互联网工程技术研究中心,重庆400054 [2]重庆理工大学电气与电子工程学院,重庆400054

出　　处：《计算机与数字工程》2024年第8期2343-2347,2410,共6页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:61402063)资助。

摘　　要：针对松鼠优化算法在后期寻优能力不足、容易陷入局部最优以及种群多样性损失较大的问题,提出了一种融合自适应t分布和随机游走策略的松鼠优化算法(TRWSSA)。该算法利用折射反向学习策略进行种群初始化,增强了种群的整体多样性;引入非线性搜索因子并且在每一次松鼠位置更新中加入自适应t分布扰动位置,减少算法陷入局部最优的概率,增强全局寻优能力;在最后的位置更新中加入随机游走策略对最优松鼠位置进行扰动更新,提高算法后期的收敛精度和速度。通过在8个基准函数上的仿真实验,对比其他智能算法以及改进算法,实验结果和分析表明TRWSSA在收敛速度、收敛精度上有明显提升,且能较好地解决寻优不足问题。To solve the problem of limited searching ability,easy falling into local optimum,and huge loss of population variety,a novel squirrel optimization algorithm(TRWSSA)is developed,which combines adaptive t-distribution with random walk strategy.For population initialization,the method employs a refraction reverse learning technique,which increases the population's total variety.The chance of the method falling into a local optimum is lowered and the global optimization ability is boosted by incorporating a non-linear search factor and adding an adaptive t-distribution perturbation site to each squirrel location update.A random walk approach is introduced to the algorithm's last position update to update the ideal squirrel location,which improves the algorithm's convergence accuracy and speed.The experimental findings and analysis reveal that TRWSSA has a considerable improvement in convergence speed and accuracy,and it can better tackle the problem of insufficient optimization,when compared to other intelligent algorithms and improved algorithms.

关 键 词：智能优化算法 松鼠算法 算法改进 融合策略 折射反向学习 自适应t分布 随机游走 基准函数 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]