PrABCast

Companies with an early customer decoupling point in production are faced with the conflict of objectives of ensuring a high degree of flexibility while at the same time fulfilling individualised customer orders. Reliable forecasts of specific customer demands make it possible to initiate production orders ahead of time.

For decades, various forecasts based on statistical methods have been used to predict future orders, whereby their suitability depends on the type of order initiation.

In this research project, the extent to which a high degree of flexibility can be guaranteed while at the same time maximising the fulfilment of individual customer orders will be investigated in various use cases. The focus is on whether machine learning (ML) forecasting can deliver better results than traditional methods.

The goal of the PrABCast project is to develop a user-friendly IT tool for the simple use of ML methods for forecasting sales and demand in SMEs. This is intended to lower the threshold for SMEs to apply ML in their own companies by adapting successful pilot projects.
This will be achieved by exploiting the potential of predictive sales and demand planning for customer order-related or hybrid production. The aim is to adapt demand quantities and reduce order lead times.

At the beginning of the project, the requirements of contract manufacturing are recorded and weighted. Then the possible use cases of the project partners are classified and pilot projects are generated. The next step will be to identify unused data sources and evaluate their impact on the quality of the forecasts.
To compare the different forecasts, known analysis methods will be applied and compared with the results of the ML forecasts in combination with the newly found data sources.
The developed forecasting methods will then be incorporated into a general model. This should reduce the time needed to develop such procedures, but also evaluate the dependency on the classes of use cases.
Finally, the model will be transformed into a user-friendly IT tool. In order to check the practical applicability, the application in further use cases will be accompanied.

The project PrABCast (Nr. 22180 N) of the German Logistics Association e.V. (BVL) - Schlachte 31, 28195 Bremen - is funded by the AiF within the scope of the program for the promotion of joint industrial research and development (IGF).