CONTROL OF KEY PERFORMANCE INDICATORS THROUGH AN APPROXIMATED PREDICTIVE MODEL

Abstract

Key performance indicators (KPI) play an important role in the manufacturing system. In practices, it is often needed to control certain KPI value to be at a specified level. However, the relationship between system inputs and KPIs and among KPIs are very complex. In this paper, we focus on the KPI control for optimizing the system inputs to reach desired outputs values based on an approximated predictive model for KPIs. The model can achieve high flexibility for analysis, thus providing accurate KPI control. We demonstrate the effectiveness of the KPI control through simulation studies.