Machine learning technologies for order flowtime estimation in manufacturing systems

Abstract

The problem of order due date assignment is an important issue for many companies and especially SMEs, which typically rely on production managers’ best estimates to assign customer order due-dates. This paper investigates the use of machine learning (ML) technologies for order flowtime estimation in dynamic job shops, utilising a discrete event simulation framework for modelling manufacturing operations. The data generated via simulation is used by a series of ML technologies for predicting when orders could be completed. A series of experiments are conducted, and the performance of the proposed approach is compared with conventional due date assignment methods.