摘要
为了克服自适应和声算法求解多模函数时的缺陷,提出一种新颖的改进自适应和声算法。首先,新算法设计了一种新颖的自适应候选和声向量产生策略,提升了算法的搜索范围;其次,新算法提出了一种和声调整率PAR的设置方式,新方式随着进化代数增加逐渐增加PAR数值。针对五个标注测试函数的实验结果表明,与目前最有竞争力的自适应和声算法相比,新算法收敛速度更快寻优效果更好。
This paper presents a novel modified adaptive harmony search（MHS） algorithm for overcoming the shortcoming of self-adaptive harmony search（SHS） in solving multi-model function.First,in the new MHS algorithm a new adaptive candidate harmony vector generation strategy is designed so that a larger area can be searched.Secondly,the new algorithm presents a setting pattern for harmonic regulation rate PAR,in which PAR gradually increases along with the increase of evolutionary generations.It is illustrated by the results of experiment aiming at 5 benchmark test functions that compared with SHS,the most competitive algorithm at present,the MHS performs better in convergence speed and optimisation.
出处
《计算机应用与软件》
CSCD
北大核心
2012年第9期268-270,277,共4页
Computer Applications and Software
关键词
和声算法
自适应
优化
Harmony search, Self-adaptive, Optimisation
