摘要
多价值链协同发展背景下,制造企业没有充分考虑服务链、营销链等其他价值链对产品需求的影响。为提高制造企业产品需求预测的精度,提出了产品数据空间和一维卷积神经网络(one-dimensional convolutional neural networks,1D-CNN)-长短期记忆(long short-term memory,LSTM)神经网络的深度学习算法。首先,整合不同价值链对产品需求影响的相关数据构建产品数据空间。其次,从数据空间中获取多链数据集用于1D-CNN-LSTM模型的预测。其中,1D-CNN通过两次卷积池化操作获取数据的深层次特征,LSTM则通过进一步学习数据特征中的重要信息来进行时间序列预测。最后,通过某电气设备制造企业生产销售的环网柜产品的相关数据进行算例分析,并与其他几种模型进行预测结果比较。结果表明:1D-CNN-LSTM模型的预测效果优于神经网络模型和单一的LSTM模型。可见提出的1D-CNN-LSTM深度学习模型更具优越性,预测效果好。
Under the background of multi-value chains collaborative development,the impact of other value chains such as service chain and marketing chain on product demand has not been fully considered by manufacturing enterprises.In order to improve the accuracy of product demand forecasting in manufacturing enterprises,a product data space and a deep learning algorithm of one-dimensional convolutional neural networks(1D-CNN)-long short-term memory(LSTM)were proposed.Firstly,the relevant data of the impact of different value chains on product demand were integrated to build a product data space.Secondly,the multi-value chain data set was obtained from the data space for the prediction of 1D-CNN-LSTM model.Among them,the deep-seated features of the data were obtained through two convolution and pooling operations of 1D-CNN.And the next step of time series prediction was realized by furthering learning the important information in the data deep-seated features through LSTM.Finally,an example was analyzed through the relevant data of ring main units by an electrical equipment manufacturing enterprise,and the prediction results were compared with other models.The results show that the prediction effect of 1D-CNN-LSTM model is better than that of neural network models and a single LSTM model.It is concluded that the 1D-CNN-LSTM deep learning model proposed in this paper has more advantages and good prediction effect.
作者
吴庚奇
牛东晓
耿世平
张焕粉
WU Geng-qi;NIU Dong-xiao;GENG Shi-ping;ZHANG Huan-fen(School of Economics and Management, North China Electric Power University, Beijing 102206, China;Beijing Key Laboratory of New Energy Power and Low Carbon Development, Beijing 102206, China;Beijing Qingchang Power Technology Co., Ltd., Beijing 100089, China)
出处
《科学技术与工程》
北大核心
2021年第31期13413-13420,共8页
Science Technology and Engineering
基金
国家重点研发计划(2020YFB1707800)。
