摘要
针对噪声环境下多模函数的优化,本文理论上分析了噪声对多模函数优化的全局收敛性和收敛精度的影响,并通过全局区域搜索率和全局区域收敛精度分析噪声对算法的影响程度.实验结果和分析表明,增加多模函数寻优难度和噪声强度,遗传算法的全局区域搜索率都在下降,全局区域收敛精度总体变差;重采样的方法能够有效提高算法的全局区域搜索率,总体改善算法的全局区域收敛精度;确定性排挤遗传算法(Deterministic Crowding Genetic A-lgorithm,DCGA)和多种群遗传算法(Mult-i Population Genetic Algorithm,MPGA)的全局区域搜索率和全局区域收敛精度要优于杰出保留遗传算法(Elist Genetic Algorithm,EGA).
An in-depth study was carded out on the genetic algorithm for MFO（Multi-modal function optimization） in noise environment. The effect of noise on MFO was theoretically analyzed. The probability of searching global area and the precision of global convergence were proposed to analyze the global convergence of genetic algorithm for MFO. It was fotmd that the complexity of Multi-modal function and the strength of noise have influence on the performance of genetic algorithm for MFO. The result shows resampling method could iower the effect of noise,and the performance of MPGA and DCGA was better than that of EGA.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2012年第2期327-330,共4页
Acta Electronica Sinica
基金
国家自然科学基金(No.60963002)
江西省自然科学基金(No.2009GZS0090
No.2010GZS0169)
关键词
遗传算法
多模函数优化
噪声环境
genetic algorithm
Multi-modal function optim zafion
noise environment
