基于管件编码和ART2神经网络聚类的船舶管件加工族构造

针对船舶管件加工族构造问题,提出一种基于编码和ART2神经网络的聚类方法。在管件编码中,综合考虑管件的托盘集配时间和结构工艺的特征,采用ART2神经网络实现对管件的自动聚类,形成合理的管件加工族。结果表明,ART2网络聚类方法对处理船舶管件加工族构造问题具有良好的适应性和稳定的分类能力。

Structuring Ship Pipe Piece Processing Family Based on Pipe Fittings Coding and ART2 Neural Network Clustering

To solve the problem of structuring ship pipe piece processing family, the paper pointed a clustering method based on coding and ART2 neural network. In the pipe fittings coding and comprehensive consideration of pipe pallet setting time and the structure technology characteristics, we achieved the automatic clustering of the pipe through ART2 neural network and reasonable pipe piece processing family. The result of experiment indicates that ART2 network has good adaptability and stable classification to deal with structuring ship pipe piece processing family.