基于零件编码与聚类算法的零件族构造

Part Family Formation Based on Coding and Clustering Algorithm

大规模定制环境下,一个关键的问题是将众多客户个性化的产品依据相似性归类.针对客户个性化零件,依据其工艺特征的相似性划分零件族,可以减少产品内部加工的多样化,按零件族集中投入生产过程,从而达到增加生产批量、缩短交货期等效果.针对单纯应用GT编码分类法存在的不足,提出一种基于零件的加工属性编码采用聚类算法构造零件族的方法.首先从数学角度研究了基于零件加工工艺特征的编码方法,然后对K-means聚类算法作了一点改进,以期改善聚类结果,阐述了具体的算法步骤,并利用C语言进行了开发.最后通过一个实例,证明该方法切实可行.

One of the most important issues with respect to mass customization is to group numerous custom products into families according to similarity. Customer individual parts can be grouped into part families by the similarity of technical characteristics from the viewpoint of production. This is an effective way to decrease internal production variety. Accordingly, family based production will obtain mass production efficiency and faster delivery. To overcome the shortage of parts grouping only by GT coding, this paper presented a method by integrating coding and clustering algorithm. The quantitative part description by coding according to the technical characteristics was studied from the point of mathematics. K-means algorithm was then modified to improve the clustering results. The algorithm steps were elaborated and developed by C language. At last, a case study was also presented to verify the feasibility of this method.