Multi-Objective Optimization of NARX Model for System Identification Using Genetic Algorithm

The problem of constructing an adequate and parsimonious Nonlinear Autoregressive model process with eXogenous input (NARX) structure for modeling nonlinear dynamic system is studied. NARX has been shown to perform function approximation and represent dynamic systems. The structures are usually guessed or selected in accordance with the designer prior knowledge, however the multiplicity of the model parameters make it troublesome to get an optimum structure. The trial and error approach is not efficient and may not arrive to an optimum structure. An alternative algorithm based on multi-objective optimization algorithm is proposed. The developed model should fulfill two criteria or objectives namely good predictive accuracy and optimum model structure. The result shows that the proposed algorithm is able to correctly identify the simulated examples and adequately model real data structure and based on a set of solutions called the Pareto optimal set, from which the best network is selected.

[1]  Derrick K. Rollins,et al.  Continuous-time Hammerstein and Wiener modeling under second-order static nonlinearity for periodic process signals , 2006, Comput. Chem. Eng..

[2]  Rich Caruana,et al.  Overfitting in Neural Nets: Backpropagation, Conjugate Gradient, and Early Stopping , 2000, NIPS.

[3]  G-C Luh,et al.  Inversion control of non-linear systems with an inverse NARX model identified using genetic algorithms , 2000 .

[4]  I. J. Leontaritis,et al.  Input-output parametric models for non-linear systems Part II: stochastic non-linear systems , 1985 .

[5]  Sheng Chen,et al.  Orthogonal least squares methods and their application to non-linear system identification , 1989 .

[6]  R. K. Ursem.  Multi-objective Optimization using Evolutionary Algorithms , 2009 .

[7]  P. Young,et al.  Time series analysis, forecasting and control , 1972, IEEE Transactions on Automatic Control.

[8]  Robiah Ahmad.  Identification of discrete-time dynamic systems using modified genetic algorithm , 2004 .

[9]  Kalyanmoy Deb,et al.  Muiltiobjective Optimization Using Nondominated Sorting in Genetic Algorithms , 1994, Evolutionary Computation.

[10]  Lennart Ljung,et al.  System Identification: Theory for the User , 1987 .

[11]  Hishamuddin Jamaluddin,et al.  Optimum grouping in a modified genetic algorithm for discrete-time, non-linear system identification , 2007 .

[12]  Pakkiriswamy Aadaleesan,et al.  Nonlinear system identification using Wiener type Laguerre–Wavelet network model , 2008 .

[13]  Stephen A. Billings,et al.  Non-linear system identification using neural networks , 1990 .

[14]  David Roze,et al.  Sound synthesis of a nonlinear string using Volterra series , 2007 .

[15]  J. D. Schaffer,et al.  Some experiments in machine learning using vector evaluated genetic algorithms (artificial intelligence, optimization, adaptation, pattern recognition) , 1984 .