Towards a data-driven predictive-reactive production scheduling approach based on inventory availability

Abstract

To survive in a competitive business environment, manufacturing systems require the proper deployment of advanced technologies coming from Industry 4.0. These technologies allow access to quasi-real-time data that provide a continuously updated picture of the production system, including the state of available inventory. Data-driven predictive-reactive production scheduling has the potential to support the anticipation and prompt reaction to overcome different kinds of disruptions that occur in production execution nowadays. This research paper aims to propose a conceptual model for a data-driven predictive-reactive production scheduling approach combining machine learning and simulation-based optimization, considering current inventory of raw material, work in process and final products inventory to characterize a job-shop production execution state. The approach supports decision-making in dynamic situations related to inventory availability that can affect production schedules.