Optimal inventory planning under uncertainty

Optimizing inventory planning in wholesale with artificial intelligence

Al-based inventory planning in the warehouse: Use of the latest forecasting models from research to determine the best possible ordering strategy
© bannafarsai - Adobe Stock
Al-based inventory planning in the warehouse: Use of the latest forecasting models from research to determine the best possible ordering strategy

“Out of stock” has been a frequently heard refrain in recent times. Wood, bathroom fittings and toilet paper are not in stock when customers need them, while goods that are not in demand are taking up valuable space. This makes managing inventory levels a key challenge in the wholesale sector. In the OBER project, researchers at our Center for Applied Research on Supply Chain Services are combining forecasting models with mathematical optimization to create an artificial intelli-gence (AI) that also takes into account restrictions such as purchasing conditions, storage capacity and capital commitment.

While forecasts are already in use in some places, they often provide only a mean of previous sales and thus deliver point forecasts. Often, however, it is precisely the uncertainty factor that is critical in allowing statements to be made about how likely it is for a predicted sales volume to be exceeded or undercut. The Fraunhofer Al uses the latest forecasting models from research, quantifies this uncertainty factor and, on this basis, determines the best possible ordering strategy for dispatchers in terms of order quantity and timing. The project is funded by the Bavarian Ministry of Economic Affairs, Regional Development and Energy. Project partners are Trevisto, Eisen-Fischer and FIS Informationssysteme und Consulting.