基于属性约简的核电企业供应商选择方法研究

作为我国核电行业的改革目标之一,核电供应链体系的构建具有重要意义,其中供应商作为供应链体系的基础,其评价与选择尤为关键。由于核电行业的一些特殊性质,核电企业的供应商评价十分复杂,需要综合考虑供应商交货能力等许多因素。因此使用了粗糙集与神经网络相结合的方法对核电企业的供应商评价指标进行分析,利用粗糙集理论的属性约简过程将冗余属性以及冲突对象删除,然后基于神经网络进行供应商分类,有效降低数据维数以及分类过程的计算复杂度,提高供应商选择的准确性。

Study on Nuclear Power Stations Vendor Selection Method Based on Attribute Reduction

As an object of nuclear power stations, the optimization of supply chain is imperative. So the selection of vendor becomes important. Because of some special properties, many factors are critical which make selection procedure complicated. Therefore this paper presents a method based on rough sets and neural network. First, evaluation criteria are analyzed. Then redundant attributes will be deleted from decision table according to an attribute reduction algorithm. At last neural network will be used to classification. This method will reduce the complexity of data dimension and the classification process, which also improve prediction precision.