摘要
基于混合蛙跳算法(SFLA)具有寻优精度不高、算法收敛速度较慢等不足,提出了一种基于自适应同步因子的混合蛙跳算法(AS_SFLA)。当种群个体根据更新公式进行位置更新时,通过引入一个同步因子,改变组内迭代时的蛙跳规则,对青蛙个体更新位置进行扰动,从而增加种群位置的多样性并调整算法的搜索规模,从而改进算法的局部搜索能力。经使用两种不同参数设置对9个基准测试函数进行仿真实验后,比较SFLA、AS_SFLA和ISFLA1结果表明,自适应同步因子可以提高算法的局部搜索精度,避免算法陷入局部最优,实验结果证明了AS_SFLA具有更好的求解质量和局部搜索能力,适合高维复杂函数的优化。
Basic shuffled frog leaping algorithm(SFLA)has a slow convergence speed and a low precision. To overcome theseshortcomings,this paper proposes an improved algorithm-adaptive synchronized factor shuffled frog leaping algorithm(AS_SFLA).In this algorithm,the adaptive synchronized factor is introduced to change frog update rule in local iterations to improve the abilityin local search. Each species update according to the corresponding position updating formula. The factor disturbs the individualwhen the position updates,which increases the diversity of population location and adjusts the search scope. Each individual adjuststhe factor dynamically in the local iterations. The rule of updating positions is more reasonable. Compared simulation results of exper-iments on nine benchmark functions with two different groups of factors among SFLA,AS_SFLA and ISFLA1,the results show thatthe adaptive synchronized factor strategy balances the searching ability of AS_SFLA in the local and global iteration processes,which makes the algorithm avoid to fall into local optimum. Finally,AS_SFLA is proved to act better in solution quality,searching ability and can be more suitable for high-dimensional optimization of complex functions.
作者
李敏楠
刘升
LIMinnan;LIU Sheng(School of Management,Shanghai University of Engineering Science,Shanghai 201620)
出处
《计算机与数字工程》
2018年第6期1083-1088,共6页
Computer & Digital Engineering
基金
国家自然科学基金项目(编号:61075115)
上海市教委科研创新基金重点项目(编号:12ZZ185)
上海工程技术大学研究生科研创新项目(编号:E3-0903-16-01304)资助
关键词
混合蛙跳算法
自适应同步因子
更新规则
局部搜索策略
组内迭代
shuffled frog leaping algorithm
adaptive synchronized factor
update rule
local searching strategy
intra groupiteration