中国制造业数字化供应链资源属性特征分析及需求预测研究

Analysis of Resource Characteristics and Demand Forecasting of China’s Manufacturing Digital Supply Chain

摘要: 现阶段制造业数字化经验不足导致数字化供应链资源浪费现象严重，对数字化供应链资源属性特征进行分析并预测资源需求是实现数字化供应链降本增效的重要途径。本文基于文本分析对制造业数字化供应链资源属性特征进行分析，剖析了时序性、智能化以及可预测性是数字化供应链资源的主要特征，资源需求预测对企业绩效影响不断增大。随后基于LSTM构建CNN-BiLSTM-Attention的资源需求预测模型进行数字化供应链资源需求预测。在数值实验中，本文所构建的预测模型具有99.62%的预测精度，高于文中所列其他机器学习预测模型，具有较高精确性及较强适应性。

Abstract: At this stage, the lack of digital experience in the manufacturing industry has led to a serious waste of digital supply chain resources, so it has become an urgent problem to analyze the characteristics of digital supply chain resources and predict resource demand to reduce costs and increase efficiency of digital supply chain. Therefore, this paper analyzes the attribute characteristics of digital supply chain resources in the manufacturing industry based on text analysis, and analyzes that timing, intelligence and predictability are the main characteristics of digital supply chain resources, and the impact of resource demand prediction on enterprise performance is increasing. Then, based on LSTM, a CNN-BiLSTM-Attention resource demand forecasting model was constructed to forecast the resource demand of the digital supply chain. In numerical experiments, the prediction model constructed in this paper has a prediction accuracy of 99.62%, which is higher than that of other machine learning prediction models listed in this paper, and has higher accuracy and strong adaptability.