基于遗传算法的覆盖近似空间最优近似求解算法  

Optimal Approximation Algorithm of Covering Approximation Space Based on Genetic Algorithm

作　　者：武慧栋 冯琴荣[1] WU Huidong;FENG Qinrong(College of Mathematics and Computer Science,Shanxi Normal University,Taiyuan 030031,Shanxi,China)

机构地区：[1]山西师范大学数学与计算机科学学院,山西太原030031

出　　处：《山西师范大学学报（自然科学版）》2022年第1期10-15,共6页Journal of Shanxi Normal University(Natural Science Edition)

摘　　要：经典粗糙集用一对精确集来近似粗糙集.在其扩展模型中,有学者提出了从相似度出发用一个精确集来刻画粗糙集的模型.这种模型可以更贴近目标集合,在规则提取方面也有很好的应用.在划分情形下最优近似的算法已经相对完善,但是对于覆盖情形下的最优近似其算法还很匮乏.针对这一问题,提出了一种基于改进的遗传算法的最优近似求解算法.通过采用子群更替和强制变异等方法,增加了种群多样性,提高了准确率,加快了收敛速度.仿真结果表明该算法可以简单有效的找到覆盖近似空间中的最优近似.Classical rough set uses a pair of exact sets to approximate rough set.In its extended model,some scholars proposed a rough set model based on similarity.This model can be closer to the target set and has a good application in rule extraction.In the case of partition,the optimal approximation algorithm has been rela⁃tively perfect,However,there is still a lack of optimal approximation algorithm in the case of coverage.To solve this problem,this paper proposes an optimal approximation algorithm based on improved genetic algorithm,Sim⁃ulation results show that the algorithm can find the optimal approximation in the covering approximation space simply and effectively.

关 键 词：覆盖近似空间 最优近似 遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]