Production Bottleneck Prediction of Multi-Variety and Small Batch Production Workshop

Aiming at the bottleneck dynamic drift problem of production line, a dynamic bottleneck prediction method based on Markov chain was proposed to predict the bottleneck before the bottleneck appeared and predict the bottleneck change state. First, Markov state probability transition matrix is determined by analyzing the past data. Second, the current operation state of the process is brought into the Markov chain to calculate the probability of each process bottleneck at the next moment. Finally, the bottleneck index model is used to accurately locate the bottleneck location, and then get the exact bottleneck process.