Work in Progress Level Prediction with Long Short-Term Memory Recurrent Neural Network

Since the reliability of production plans drops largely within several days after plan creation, control faces huge challenges, when trying to foresee work in progress (WIP) level at bottleneck machines and react appropriately. Whereas researchers applied artificial intelligence predict lead times or transition improve planning reliability, only small efforts have been taken on time series prediction field control, especially topic WIP prediction. In this paper univarate approaches are used for predicting a machine one more step ahead. Long short-term memory recurrent neural networks, LSMT models show higher accuracy than classical methods. For ahead forecasting four different investigated. Systematical model tuning comparison various error measures presented real industrial use case from steal manufacturing industry.