自适应调整差分进化算法在优化问题中的应用  

Self-adaptive Differential Evolution Algorithm and Its Application to Complex Optimization Problems

作　　者：肖文显[1] 王俊阁[1] 马孝琴[1] 

机构地区：[1]河南科技学院网络信息中心,河南新乡453003

出　　处：《哈尔滨理工大学学报》2015年第1期71-74,共4页Journal of Harbin University of Science and Technology

基　　金：国家自然科学基金(71171151);河南省教育厅自然科学研究计划(13B520011)

摘　　要：差分进化算法在求解优化问题时,进化后期由于种群多样性急剧下降,算法全局搜索能力被削弱,极易陷入局部最优解而"早熟"收敛.针对该问题定义了算法停滞系数和个体相似系数.根据算法停滞系数自适应调整算法的缩放系数.同时,根据个体相似系数判定种群普通个体与最优个体的相似性,并以此为基础对相似个体实施基因重构操作,从而避免种群个体严重趋同造成的种群多样性下降问题.将改进算法应用于标准测试函数和车辆路径问题的优化.模拟计算结果表明:改进算法的优化结果优于标准差分进化算法,改进的差分进化算法具有更强的全局寻优能力,适于求解复杂优化问题.When differential evolutionary algorithm is used in solving complex optimization problems, capability of global search is decreased in the later evolution period, and the ability of global search is weakened and algorithm can easily fall into local optimum. In response to these problems, stagnation factor and similarity factor are defined. Scaling factor F is adaptive adjusted according to the stagnation factor. Similarity between individuals is determined by the similarity factor, the gene reconstruction is taken. So that the problem of decrease of population diversity can be alleviated. The improved algorithm is applied to standard test function optimization and vehicle routing problem, the results show that improved differential evolution suitable for solving complex optimization problems. algorithm has better global search ability. And the algorithm is

关 键 词：优化问题 差分进化 自适应调整 基因重构 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]