摘要
针对基本混合蛙跳算法(shuffled frog leaping algorithm,SFL)在求解高维复杂问题时的不足,本文提出一种自适应参数调整的改进策略。首先,利用变公比数列分析了SFL更新轨迹的收敛性;在此基础上,利用系统稳定性分析方法,提出在SFL更新公式中基于比例系数和适应度标准差来自适应调整更新的方法。最后,基于3组共8个标准测试函数将本文改进SFL与基本SFL和4个改进型粒子群优化算法(particle swarm optimization,PSO)作对比,验证了本文改进策略对各类复杂函数的高效性;同时,对比了改进SFL与基本SFL和wPSO在求解高维问题时的性能,验证了改进SFL对高维问题求解的有效性。
An improvement strategy of adaptive parameter adjustment is proposed to improve the efficiency of the shuffled frog leaping algorithm（SFL）in solving high dimensional complex problems.First of all,the convergence feature of the SFL is analyzed based on the theory of geometrical sequence.Then,an improvement strategy of adaptive parameter adjustment based on proportional coefficient and fitness standard deviation is proposed to the update the formula.Finally,based on three groups of eight criteria functions,the performance of the modified SFL with basic SFL and four modified particle swarm optimization（PSO）is compared,and the results verify the high-efficiency of the improvement strategy for various complex functions.Meanwhile,the performance of the modified SFL with basic SFL and wPSO on solving high dimension problems is compared,and the results verify the validity of the modified SFL.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2016年第8期1939-1950,共12页
Systems Engineering and Electronics
基金
国家高技术研究发展计划(863计划)(2015AA042101)资助课题
关键词
混合蛙跳算法
收敛性
自适应参数调整
智能计算
shuffled frog leaping（SFL）algorithm
convergence feature
adaptive parameter adjustment
intelligent computing
