基于精英解学习的邻域搜索差分演化算法  

Neighborhood search differential evolution algorithm based on elitism learning

作　　者：吴克晴[1] 杜望 WU Keqing DU Wang(Faculty of Science, Jiangxi University of Science and Technology, Ganzhou 341000, China)

机构地区：[1]江西理工大学理学院,江西赣州341000

出　　处：《江西理工大学学报》2017年第5期99-106,共8页Journal of Jiangxi University of Science and Technology

基　　金：国家自然科学基金资助项目(61364015)

摘　　要：为了防止差分演化算法在求解复杂问题时容易陷入局部最优、收敛速度慢等问题,提出了一种基于精英解学习的邻域搜索差分演化算法(ELNDE).在该算法中利用多个精英解构建一个精英解池策略,并且对其进行反向学习,保证种群的多样性.在每一代种群演化计算过程中执行邻域搜索,通过精英解作为导向,加快算法的收敛速度的同时提高开采能力.使用13个基准测试函数对提出的算法进行了测试并且与几种知名的改进算法进行比较.实验表明,提出的算法在收敛速度和解的精度是具有较大的优势.In order to prevent the differential evolution（DE） from getting into the local optimization and slow convergence when solving complicated problems,the neighborhood search differential evolution algorithm based on elitism learning（ELNDE） has been proposed.The ELNDE utilizes elitism pool strategy for elitism opposition-based learning to enhance the population diversity.Neighborhood search operation is adopted in the evolutionary process,and the elitism is used as the guide of global optimization to accelerate the convergence speed and improve the exploitation ability.The proposed algorithm is tested by 13 benchmark functions and compared with other well-known DE algorithms.Experimental results show that the proposed algorithm have an advantage in both convergence rate and solution accuracy.

关 键 词：全局优化 差分演化 精英解学习 邻域搜索 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]