Modeling a production system by parametric identification approach

Abstract

This paper refers to possibility of utilization a new method for modeling an industrial production system using the identification approach.The production system studied is considered as a black box, which does not require any knowledge of the internal parameters. Two structures of linear parametric models are outlined and compared i.e. Autoregressive with Exogenous (ARX) model and Autoregressive Moving Average with Exogenous (ARMAX) model. The best linear model is ARMAX, due to its higher percentage of best fit.

System Identification of Inventory System Using ARX and ARMAX Models

“…[5] The identification system is a fairly mature field that has seen an interesting development and various contributions over the years, that we could not all include in this work, but just a few references that dealt with a parametric models which are ARX, ARMAX, OE and BJ. [9][10][11][12] Many works has been accomplished at this level, however, there's a number of issues in industry, which have not been satisfactorily dealt [4]. In previous study, we proposed "black box" modeling of a production line with three machines from the observed inputs and outputs, using ARX and ARMAX models.…”

“…In previous study, we proposed "black box" modeling of a production line with three machines from the observed inputs and outputs, using ARX and ARMAX models. [12] In this paper, we study the case of an inventory system from a distribution company. The inventory system is considered as a MISO system with two input variables and one output variable.…”

“…The inventory system is considered as a MISO system with two input variables and one output variable. The same modeling method in [12], is used to identify the internal properties of the studied system.…”