Supervised learning to support the process planning of contract logistics projects

Abstract

Due to the outsourcing trend, contract logistics is a constantly growing industry. Especially for the essential and time-consuming planning of logistics processes in a contract logistics project, experienced planners are required. However, the growing shortage of skilled workers makes recruiting these planners increasingly difficult. Hence, a supervised learning approach will be investigated to support especially inexperienced planners in process planning. This article explores how supervised learning can extract the process knowledge contained in legacy contract logistics project documentation to suggest process steps during a new project process planning. The investigation results in boosted decision trees predicting the next process step correctly in 81% of the cases. In addition, the article guides what data should be collected today for even better results in future applications.