基于多层次划分的服装产品族构造方法

Clothing Product Family Construction Method Based on Multi-Level Division

服装生产是典型的多品种小批量模式,为减少服装生产中产品内部多样性的影响,实现用工业化方式进行个性化生产,提出了基于多层次划分的产品成组分类方法。通过服装构成分解和成组技术,基于特征编码构建适用于服装生产的产品信息模型;采用改进的K-means聚类算法对同一品类下不同款式的服装进行产品族划分,并引入有效性评价指标CSI确定最佳聚类数,降低人为因素的干扰;最后采用BP神经网络实现新产品的族匹配。通过实例验证,提出的方法能有效构建相似件的产品族并进行新产品归类,有助于组织精益化生产,实现服装生产的快速反应与柔性化。

Clothing production is a typical mode of multi-variety and small-batch.In order to reduce the influence of the product diversity in clothing production and realize personalized production in an industrialized way,an clothing grouping classification method based on multi-level division is proposed.Through clothing composition decomposition and grouping techniques and based on feature coding,a product information model applicable to clothing production is constructed.The improved K-means clustering algorithm is used to divide the product families of different styles of clothing under the same category,and the validity evaluation index CSI is introduced to determine the optimal number of clusters.Finally,BP neural network is used to realize the family matching of new products.Through example testing,the proposed method can effectively construct product families of similar pieces and categorize new products,which helps to organize lean production and achieve rapid response and flexibility in clothing production.