摘要
为了解决零件分组中各特征的重要性和分组数目难以确定的难题,提出了一种用可调节的权重系数向量和改进的ISODATA模糊聚类算法来实施零件分组的方法.根据零件分组特征的重要性合理调节权重系数向量,用该向量修正ISODATA模糊聚类算法中的零件特征矩阵,在算法运行结束后计算聚类中心的模糊距离,将距离小于设定阈值的类别合并为一组.用VB开发了一个零件分组系统,采用该分组方法在系统上进行反复实验和比较,结果与人们的分组偏好和习惯一致,验证了分组方法的有效性和可靠性.
To solve the difficulty in determining the importance of various part attributes and the number of part families in part family identification, a new part family identification method using an adjustable weight coefficient vector and an improved ISODATA fuzzy clustering algorithm is proposed in this paper. The proposed method begins with adjusting the weight coefficient vector according to the importance of part attributes ; then, the feature matrix in ISODATA fuzzy clustering algorithm is amended using the vector; furthermore, the fuzzy distance between the fuzzy clustering centers of different part families is calculated at the end of computing ; finally, part families with distance between each other less than the threshold are merged into one group. The method is implemented in VB, and the empirical results are identical with those identified according to people＇ s preference and habit, which verifies the availability and feasibility of the proposed method.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2009年第3期113-116,共4页
Journal of Harbin Institute of Technology
基金
国家高技术研究发展计划资助项目(863-511-910-403)
关键词
ISODATA
模糊聚类
零件分组
特征码
ISODATA
fuzzy clustering
part family identification
characteristic .code
