摘要
为了评价遗传算法在噪声环境下的优化性能,提出"平均最优解"和"最优解分布标准差"两个指标,实验结果表明新指标可以有效地评价噪声环境下遗传算法的优化性能.研究了实数编码遗传算法在噪声强度递增环境下的性能.结果表明小生境策略和多种群策略可以改善遗传算法在噪声环境下的性能,单点交叉在噪声环境下的性能要优于混合交叉.
＂The average of best＂ and ＂the standard deviation of best＂ were proposed to evaluate the performance of genetic algorithms in noisy environment.Results show the evaluation indexes could effectively evaluate the performance of genetic algorithm in the environment.Based on this,the performances of real-coded genetic algorithms with different crossover operator in noisy environment have been investigated.Results show the performance of deterministic crowding genetic algorithm and multi population genetic algorithm in the environment are better than that of elist genetic algorithm;the behaviors of alone point crossover outperform that of blended crossover.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2010年第9期2090-2094,共5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)
航空基金(No.2009ZD56003)
江西省教育厅科技项目(No.GJJ09483)
关键词
遗传算法
噪声环境
性能评价
genetic algorithm
noisy environment
performance evaluation
