Simulation error minimization identification based on multi‐stage prediction

Classical prediction error minimization (PEM) methods are widely used for model identification, but they are also known to provide satisfactory results only in specific identification conditions, e.g. disturbance model matching. If these conditions are not met, the obtained model may have quite different dynamical behavior compared with the original system, resulting in poor long range prediction or simulation performance, which is a critical factor for model analysis, simulation, model-based control design. In the mentioned non-ideal conditions a robust and reliable alternative is based on the minimization of the simulation error. Unfortunately, direct optimization of a simulation error minimization (SEM) criterion is an intrinsically complex and computationally intensive task. In this paper a low-complexity approximate SEM approach is discussed, based on the iteration of multi-step PEM methods. The soundness of the proposed approach is demonstrated by showing that, for sufficiently high prediction horizons, the k-steps ahead (single- or multi-step) PEM criteria converge to the SEM one. Identifiability issues and convergence properties of the algorithm are also discussed. Some examples are provided to illustrate the mentioned properties of the algorithm. Copyright © 2010 John Wiley & Sons, Ltd.

[1]  Rong Chen,et al.  Nonparametric multistep‐ahead prediction in time series analysis , 2004 .

[2]  Guy Albert Dumont,et al.  System identification and control using genetic algorithms , 1992, IEEE Trans. Syst. Man Cybern..

[3]  Feng Ding,et al.  Gradient based and least-squares based iterative identification methods for OE and OEMA systems , 2010, Digit. Signal Process..

[4]  S. Shah,et al.  A control-relevant identification strategy for GPC , 1992 .

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

[6]  Kumpati S. Narendra,et al.  Identification and control of dynamical systems using neural networks , 1990, IEEE Trans. Neural Networks.

[7]  Marcello Farina,et al.  Simulation Error Minimization–Based Identification of Polynomial Input–Output Recursive Models , 2009 .

[8]  Rong-Song He,et al.  Improving real-parameter genetic algorithm with simulated annealing for engineering problems , 2006, Adv. Eng. Softw..

[9]  Chao-Ming Huang,et al.  A particle swarm optimization to identifying the ARMAX model for short-term load forecasting , 2005 .

[10]  M. Small,et al.  Towards long-term prediction , 2000 .

[11]  L. Ljung,et al.  Design variables for bias distribution in transfer function estimation , 1986, The 23rd IEEE Conference on Decision and Control.

[12]  Ioan Doré Landau,et al.  Recursive output error identification algorithms theory and evaluation , 1980, Autom..

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

[14]  Sirish L. Shah,et al.  MPC relevant identification––tuning the noise model , 2004 .

[15]  S. Shah,et al.  Identification for long-range predictive control , 1991 .

[16]  Sten Bay Jørgensen,et al.  MPC-Relevant Prediction-Error Identification , 2007, 2007 American Control Conference.

[17]  J. A. Rossiter,et al.  Modelling and implicit modelling for predictive control , 2001 .

[18]  Petre Stoica,et al.  Some properties of the output error method , 1982, Autom..

[19]  Marcello Farina,et al.  Some convergence properties of multi-step prediction error identification criteria , 2008, 2008 47th IEEE Conference on Decision and Control.

[20]  Thierry Poinot,et al.  Initialization of output-error identification methods – Comparison between ARX and RPM models , 2009 .

[21]  John B. Moore.  Global convergence of output error recursions in colored noise , 1981, CDC 1981.