A randomized algorithm for nonlinear model structure selection

The identification of polynomial Nonlinear Autoregressive Moving Average models with eXogenous variables (NARMAX) is typically carried out with incremental model building techniques that progressively select the terms to include in the model. The Model Structure Selection (MSS) turns out to be the hardest task of the identification process due to the difficulty of correctly evaluating the importance of a generic term. As a result, classical MSS methods sometimes yield unsatisfactory models, that are unreliable over long-range prediction horizons. The MSS problem is here recast into a probabilistic framework based on which a randomized algorithm for MSS is derived, denoted RaMSS. The method introduces a tentative probability distribution over models and progressively updates it by extracting useful information on the importance of each term from sampled model structures. The proposed method is validated over models with different characteristics by means of Monte Carlo simulations, which show its advantages over classical and competitor probabilistic MSS methods in terms of both reliability and computational efficiency.

[1]  Sean R. Anderson,et al.  Computational system identification for Bayesian NARMAX modelling , 2013, Autom..

[2]  Alberto Leva,et al.  NARX-based technique for the modelling of magneto-rheological damping devices , 2002 .

[3]  Peter J. Fleming,et al.  Time and frequency domain identification and analysis of a gas turbine engine , 2002 .

[4]  George W. Irwin,et al.  Prediction- and simulation-error based perceptron training: Solution space analysis and a novel combined training scheme , 2007, Neurocomputing.

[5]  Luigi Piroddi,et al.  A novel randomized approach to nonlinear system identification , 2014, 53rd IEEE Conference on Decision and Control.

[6]  Marcello Farina,et al.  Identification of polynomial input/output recursive models with simulation error minimisation methods , 2012, Int. J. Syst. Sci..

[7]  L. A. Aguirre,et al.  EFFECTS OF THE SAMPLING TIME ON THE DYNAMICS AND IDENTIFICATION OF NONLINEAR MODELS , 1995 .

[8]  Kang Li,et al.  Nonlinear modeling of NO/sub x/ emission in a coal-fired power generation plant , 2004, 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601).

[9]  S. Billings,et al.  VARIABLE SELECTION IN NON-LINEAR SYSTEMS MODELLING , 1999 .

[10]  Sean R. Anderson,et al.  Structure detection and parameter estimation for NARX models in a unified EM framework , 2012, Autom..

[11]  R. Tempo,et al.  Randomized Algorithms for Analysis and Control of Uncertain Systems , 2004 .

[12]  George W. Irwin,et al.  A fast nonlinear model identification method , 2005, IEEE Transactions on Automatic Control.

[13]  S. Billings,et al.  Algorithms for minimal model structure detection in nonlinear dynamic system identification , 1997 .

[14]  Lennart Ljung,et al.  Nonlinear black-box modeling in system identification: a unified overview , 1995, Autom..

[15]  Dale E. Seborg,et al.  Application of a general multi-model approach for identification of highly nonlinear processes-a case study , 1993 .

[16]  S. Geer.  Least Squares Estimation , 2005 .

[17]  Mario Sznaier,et al.  Randomized Algorithms for Analysis and Control of Uncertain Systems with Applications, Second Edition, Roberto Tempo, Giuseppe Calafiore, Fabrizio Dabbene (Eds.). Springer-Verlag, London (2013), 357, ISBN: 978-1-4471-4609-4 , 2014, Autom..

[18]  Sheng Chen,et al.  Model selection approaches for non-linear system identification: a review , 2008, Int. J. Syst. Sci..

[19]  Sheng Chen,et al.  Identification of MIMO non-linear systems using a forward-regression orthogonal estimator , 1989 .

[20]  C J Harris,et al.  Sparse Kernel Regression Modelling using combined locally regularised orthogonal least squares and D-Optimality , 2003 .

[21]  Luigi Piroddi,et al.  SEISMIC BEHAVIOUR OF BUTTRESS DAMS: NON-LINEAR MODELLING OF A DAMAGED BUTTRESS BASED ON ARX/NARX MODELS , 2001 .

[22]  Marcello Farina,et al.  Black box model identification of nonlinear input–output models: A Wiener–Hammerstein benchmark , 2012 .

[23]  Heinz Unbehauen,et al.  Structure identification of nonlinear dynamic systems - A survey on input/output approaches , 1990, Autom..

[24]  L. A. Aguirre,et al.  Prediction and simulation errors in parameter estimation for nonlinear systems , 2010 .

[25]  Johan A. K. Suykens,et al.  Wiener-Hammerstein Benchmark , 2009 .

[26]  David Rees,et al.  Nonlinear gas turbine modeling using NARMAX structures , 2001, IEEE Trans. Instrum. Meas..

[27]  L. Piroddi,et al.  A pruning method for the identification of polynomial NARMAX models , 2003 .

[28]  L. A. Aguirre,et al.  Dynamical effects of overparametrization in nonlinear models , 1995 .

[29]  Petre Stoica,et al.  Decentralized Control , 2018, The Control Systems Handbook.

[30]  S. Billings,et al.  Orthogonal parameter estimation algorithm for non-linear stochastic systems , 1988 .

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

[32]  Carlos M. Fonseca,et al.  'Identifying the structure of nonlinear dynamic systems using multiobjective genetic programming , 2004, IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans.

[33]  L. A. Aguirre,et al.  Use of a priori information in the identification of global nonlinear models-a case study using a buck converter , 2000 .

[34]  Stephen A. Billings,et al.  An iterative orthogonal forward regression algorithm , 2015, Int. J. Syst. Sci..

[35]  S. Billings.  Nonlinear System Identification: NARMAX Methods in the Time, Frequency, and Spatio-Temporal Domains , 2013 .

[36]  Stephen A. Billings,et al.  Model structure selection using an integrated forward orthogonal search algorithm assisted by squared correlation and mutual information , 2008, Int. J. Model. Identif. Control..

[37]  Chin-Hsiung Loh,et al.  ANALYSIS OF NONLINEAR SYSTEM USING NARMA MODELS , 1996 .

[38]  L. Piroddi,et al.  NARX model selection based on simulation error minimisation and LASSO , 2010 .

[39]  L. Piroddi,et al.  An identification algorithm for polynomial NARX models based on simulation error minimization , 2003 .