摘要
多尺度量子谐振子优化算法(MQHOA)是一种利用量子谐振子的概率解释构造的新智能优化算法,其运行框架包含2个互相嵌套的过程:量子谐振子收敛和多尺度收敛。采样运算是MQHOA算法的基本运算单元,其高斯随机数的生成效率直接影响算法的执行效率,采用Box-Muller方法实现高斯随机数的高效生成,大幅提升MQHOA算法的执行效率。通过对MQHOA算法运行框架与方法的分析,给出MQHOA算法的详细实现方法;对10个优化测试函数进行实验分析,其结果与10个相关算法的结果进行对比,表明MQHOA算法可以准确地求解一维和多维函数优化问题。
MQHOA is a novel intelligent optimization algorithm which is constructed by quantum harmonic oscillator＇s wave function. MQHOA＇s operation architecture includes two embedded convergence processes： QHO convergence and M convergence. As sampling is basic computing unit of MQHOA,Gaussian number generation efficiency has tremendous influence on MQHOA＇s running efficiency. Box-Muller method is used to generate Gaussian number,and MQHOA can run more efficiently. MQHOA implementation is specified in detail by analyzing MQHOA＇s operation architecture. Experimental results show that one-dimensional and high-dimensional function＇s global optimization can be tackled precisely by MQHOA.
出处
《成都信息工程学院学报》
2015年第5期433-438,共6页
Journal of Chengdu University of Information Technology
基金
国家自然科学基金资助项目(60702075)
广东省科技厅高新技术产业化科技攻关资助项目(2011B010200007)
