Input variable selection algorithms for HPC

Abstract

The paper refer to the problem of a holistic production control (HPC). The main idea of HPC is to derive an optimisation strategy which is based on a simple model of only a few production Key Performance Indicators (pKPI). The main challenge of the HPC approach is the derivation of an appropriate pKPI model, where the main steps are: data preprocessing, pKPI definition, input variables selection (IVS) and black-box modelling. In this paper we are focused on the problem of the selection of the most influential inputs, where the review of different IVS methods is given.