聚类分析结合BP神经网络的钢坯库堆垛问题算法

在考虑了钢坯库实际堆垛的基础上,设计了一种将订单K-means聚类分析后通过BP神经网络生成入库计划的算法。该算法主要分为2个阶段,首先通过K-means聚类将轧制计划中的合同按照轧制出厂日期等条件形成类别;然后通过BP神经网络算法生成入库计划。利用钢厂实际生产数据对本算法进行验证。结果表明,本算法能够有效地减少倒垛次数,提高垛位空间利用率。

An algorithm for storage of steel billet based on BP neural network

Considering the actual stacking of the steel billet,a algorithm was designed for the storage of the steel billet,which was based on the K-means clustering analysis and BP neural network to make the storage plan of raw steel. The algorithm was mainly divided into two stages. First,the order of the rolling steel was re-combining by K-means clustering based on the stacking date and other factors;second,the BP neural network was supposed to generate the storage plan of the steel billet. Finally the algorithm was validated by the actual production data of the steel plant. The results show that the algorithm can effectively reduce the number of bil-let stacking and improve the stack space utilization.