供应链需求预测中的神经网络预测技术应用分析

The Application on Neural Network Technology in the Supply Chain Demand Prediction

本文认为,解决供应链上游企业准确理解下游企业所提供的订货信息,进行科学的市场需求预测是供应链管理中具有战略性和规划性决策的问题。基于市场准确预测的协作计划预测和补货技术(CPFR技术)能及时准确地预测由各项促销措施或异常变化带来的销售高峰波动,从而使销售商和供应商做好准备,赢得主动。而构筑基于CPFR技术的供应链需要良好的预测方法和技术以及信息系统的支撑。文章指出,基于误差反传算法的神经网络(BP神经网络)作为一种对复杂经济现象进行分析和预测的有效工具,在有大量数据和长期学习过程的情况下,能发挥出更大的优势和性能,加之其具备自学习和自适应能力,能适时调整流程企业产品销售的信息误差,对供应链的决策系统起到精确的辅助作用。文章建立的基于协作计划预测和补货(CPFR)数据库的供应链需求预测支持系统的流程框架,以及在此基础上提出的基于BP神经网络的供应链需求预测模型,可以提高预测精度,有效避免评价过程中的人为失误,取得令人满意的预测结果。

The author believes that scientific demand prediction is a key issue in supply chain management. CPFRC is a kind of technology that can help correctly predict market demand. Demand prediction also needs good method, technology and information system to construct CPFR supply chains. In addition, Neural Network is one of the most popular prediction methods. After the construction of supply chain demand prediction supporting system which is based CPFR data-house technology, the author proposes a BP neural network prediction model, which can avoid human mistakes in the process of evaluation. The result shows that the method is satisfactory.