基于GA-BP神经网络的智能制造系统评价模型

Evaluation model of intelligent manufacturing system based on GA-BP neural network

随着信息化与数字化的发展,制造业智能制造化转型已成为焦点。针对制造业智能制造发展水平的科学评价方法成为现实需要的情况,基于机器学习的方法建立了智能制造系统评价模型。通过专家调研方式获取制造业52个评价指标的样本数据,并使用SeqGAN生成对抗网络扩充真实样本。通过BP神经网络构建训练模型,结合遗传算法优化神经网络模型,将评价指标样本数据作为网络输入,工业1.0至工业4.0等7个标签作为网络输出,并进行神经网络的训练与验证。研究结果表明:笔者所提模型分类正确率达95%,较传统BP神经网络精度提升了2.1%。在案例验证中通过差距特征向量定位企业智能制造系统当前的优势与不足,该模型评价结果可为制造型企业的智能制造发展提供指导帮助。

With the development of informatization and digitization,the transformation of manufacturing intelligent manufacturing has become the focus.In view of the fact that the scientific evaluation method of the development level of intelligent manufacturing in the manufacturing industry has become an urgent need,an intelligent manufacturing system evaluation model was established based on the method of machine learning.The sample data of 52 evaluation indicators in the manufacturing industry were obtained through expert s survey,and the SeqGAN was used to expand real samples.The training model is constructed by BP neural network,and the neural network model is optimized combining genetic algorithm.The evaluation index sample data is used as the network input,and seven labels such as industry 1.0 to industry 4.0 are used as the network output,and the neural network is trained and verified.The experimental results show that the accuracy of this model classification proposed is 95%,which is 2.1%higher than the accuracy of the traditional BP neural network.In the case verification,the gap feature vector is used to illustrate the current advantages and disadvantages of enterprise intelligent manufacturing system.The evaluation result of this model can provide guidance and help for the development of intelligent manufacturing in manufacturing enterprises.