谱聚类算法在化学试剂库备优化中的应用

The application of spectral clustering algorithm on chemical inventory optimization

石油化工科学研究院建设和应用了化合物库存管理系统,实现了化学试剂使用情况的全生命周期记录,也为HSSE管理决策积累了丰富的数据。对化合物库存管理系统中某时间段内所有试剂的申领、申购数据进行了整理和清洗,利用谱聚类算法分析试剂使用频率,将试剂分为常用和非常用两个类别,并与试剂的实际库备情况进行交叉对比,从而指导库备试剂的种类优化,实现智能化管理。

The construction and application of Chemical Inventory Management System for Research Institute of Petroleum Processing has accomplished the lifecycle recording of chemical usage. It has also accumulated a large amount of data for HSSE management decisions. In this paper, the data of chemical requisition from Chemical Inventory Management System within a certain period of time was processed and cleaned. Spectral clustering was applied to analyze the usage of chemical, and to separate chemical into two categories: the frequently used chemical and the less frequently used chemical. The optimization of chemical inventory and intelligent management could be realized by comparing two categories to the actual chemical inventory situation.