华中科技大学学报（自然科学版）2017，Vol.45Issue(5)：23-27,44,6.DOI:10.13245/j.hust.170505
基于正交设计的反向学习差分进化算法
Differential evolution algorithm using orthogonaldesign opposition-based learning
摘要
Abstract
When opposition-based learning technique is used to generate an opposite individual, it is usually taking the inverse value of all dimensions and makes some dimensions have the potential to degenerate, however, all the partial opposite solutions will greatly increase the function evaluation times, thus a differential evolution algorithm was proposed based on orthogonal design.The orthogonal design was used to generate a number of representative partial opposite solutions, which can achieve a good balance between the ability to enhance the algorithm and decrease the number of function evaluation.Experimental results show that the proposed algorithm has better convergence accuracy and speed than other opposition-based learning differential evolution algorithms, and it is not sensitive to the change of function dimension with a good robustness.关键词
差分进化算法/反向学习/正交设计/部分反向解/维度退化Key words
differential evolution algorithm/opposition-based learning/orthogonal design/partial opposite solutions/dimension degeneration分类
信息技术与安全科学引用本文复制引用
閤大海,李元香,祝婕..基于正交设计的反向学习差分进化算法[J].华中科技大学学报（自然科学版）,2017,45(5):23-27,44,6.基金项目
国家自然科学基金资助项目(61672391) （61672391）
国家重大仪器专项资助项目 (2011YQ170065.4). （2011YQ170065.4）
