Deep multistage multi-task learning for quality prediction of multistage manufacturing systems
Deep multistage multi-task learning for quality prediction of multistage manufacturing systems
复制标题
DOI：
10.1080/00224065.2021.1903822
复制
发表时间：
2021-04
影响因子：
2.5
通讯作者：
Hao Yan;Nurrettin Dorukhan Sergin;William A. Brenneman;Steve J. Lange;Shan Ba
中科院分区：
文献类型：
--
作者：
Hao Yan;Nurrettin Dorukhan Sergin;William A. Brenneman;Steve J. Lange;Shan Ba
摘要在多阶段制造系统中，基于过程传感变量的多质量指标建模是非常重要的。然而，经典的建模技术预测每个质量变量一次一个，这没有考虑阶段内或阶段之间的相关性。我们提出了一个深度多级多任务学习框架，根据MMS中的顺序系统架构，在统一的端到端学习框架中联合预测所有输出感知变量。我们的数值研究和真实的案例研究表明，新的模型具有上级性能相比，许多基准方法，以及通过开发的变量选择技术的巨大的可解释性。
Abstract In multistage manufacturing systems, modeling multiple quality indices based on the process sensing variables is important. However, the classic modeling technique predicts each quality variable one at a time, which fails to consider the correlation within or between stages. We propose a deep multistage multi-task learning framework to jointly predict all output sensing variables in a unified end-to-end learning framework according to the sequential system architecture in the MMS. Our numerical studies and real case study have shown that the new model has a superior performance compared to many benchmark methods as well as great interpretability through developed variable selection techniques.