基于马氏距离及K最近邻算法的结构优化设计  

Structural Optimization Design Based on Mahalanobis Distance and K-nearest Neighbor Algorithm

作　　者：曹鸿猷[1] 孙文 CAO Hong-you;SUN Wen(School of Civil Engineering&Architecture,Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]武汉理工大学土木工程与建筑学院,武汉430070

出　　处：《武汉理工大学学报》2022年第10期60-71,79,共12页Journal of Wuhan University of Technology

摘　　要：针对结构优化效率问题,提出了基于解的特征的优化策略,通过KNN和马氏距离判别解的约束分类,用于取代结构分析,以提高优化效率。结合HS算法,该方法先计算新解与和声库中全部解的马氏距离,以最小马氏距离对应解的约束类别给定新解分类,然后对满足约束的解进行映射并更新和声库。最后通过算例验证了该方法的有效性和高效性。Aiming at the problem of structural optimization efficiency,the optimization strategy based on the characteristics of the solution was proposed.The constraint classification of the solution was determined by KNN and Mahalanobis distance,which was used to replace structural analysis to improve the optimization efficiency.Combined with HS algorithm,this method first calculated the Mahalanobis distance between the new solution and all solutions in the harmony memory,gave the new solution classification according to the constraint category of the solution corresponding to the minimum Mahalanobis distance,and then maps the solutions that meet the constrainted and updated the harmony memory.Finally,examples were given to verify the effectiveness and efficiency of the method.

关 键 词：结构优化设计 和声搜索算法 马氏距离 优化效率 K-最近邻算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TB21[自动化与计算机技术—控制科学与工程]