基于物理DSM(design structure matrix)模型给出了产品结构的建模方法,并提出了对物理DSM产品结构模型进行四向量降维和四点刻度数值化的方法。以行列变换算法为基础,给出了对物理DSM产品结构模型进行聚类划分的方法。把物理DSM产品结构模型的联系信息流量作为标准,开发了对聚类划分方案进行评估与择优的算法。通过一个工程实例验证了算法的可行性。
Clustering algorithms of product architecture play the important role during product development lifecyele. Based on physical DSM(design structure matrix,DSM), the product architecture is modelled. The method is developed to reduce the dimension of the physical DSM product architecture model. Four-point scale is improved to numeralize the DSM model. The clustering algorithms of the product structure model based on rowcolumn transform are developed. The evaluation and selection algorithms of clustering results are developed based on the quantity of relation information.These algorithms are applied to an engineering example. The resuit validates them.