Detecting delivery delays early on: Analyzing process data and predicting lead times
Disruption to tightly scheduled supply chains often causes problems. At worst, delays in delivering supplier products can even bring production lines to a halt. It follows that shipments that might be delayed should be identified as early in the process as possible. But it is simply impossible to manually track the hundreds of thousands of shipments made each day.
This is where our ProDAB project comes in. Together with three different industry partners, we are examining their internal and intercorporate processes to develop software that can automatically predict both lead times and the filling of buffer stores in logistics processes. This makes it possible to identify early on any shipments that might be delayed. We use Bayesian networks to predict lead times, and neural networks to predict buffers. Based on this statistical modeling, we then have optimum control of the relevant process parameters, including the number of workers or the use of technical tools. In this way, the ProDAB project is enhancing the operational resilience of supply chains.