Bottleneck prediction and data-driven discrete-event simulation for a balanced manufacturing line

Bottleneck identification is a relevant tool for continuous optimization of production lines. In this work, we implement data-driven discrete-event simulator (DDS) based on experimental distributions, obtained from real historical data. The DDS allows to analyse the behavior balanced manufacturing line at Bosch Thermotechnology, under different hypotheses. It shows that some scenarios perceived as likely increase output may actually decrease metrics, reveals importance injection rates, and leads need adequate time bottleneck forecasting tools, which allow shift managers intervention in useful frame. Eleven prediction models are tested, where random forest multi-layer perceptron attain best performances (above 95% all metrics). This data flow operationalized through micro-services pipeline briefly discussed.