A Novel Technique for ARMA Modelling with Order and Parameter Estimation Using Genetic Algorithms

Abstract

AbstractA new method for determining simultaneously the order and parameters of Auto Regressive Moving Average (ARMA) models is presented in this paper. ARMA models, which can be present in different fields such as communication systems, control systems, internet software and hardware models are determined using genetic algorithms (GAs). Given ARMA (p, q) model input/output data with the absence of any information about the order, the correct model (p, q) is determined (order and parameters). The proposed method works on the principle of minimizing the overall deviation between the actual plant output, with or without noise, and the estimated plant output. The algorithm does not use complex mathematical procedures in its detection. Simulation results covered in this paper show in detail the efficiency and the generality of the proposed approach. In addition to that, the new method is compared with other well known methods for ARMA model order and parameter estimation.KeywordsGenetic AlgorithmsAuto Regressive Moving AverageOrder and parameters estimation