The Complementary Roles of Expert Systems and Optimization Tools in Dynamic System Identification

  • H. Nguyen-Phu
  • P. Weiler
  • R. T. Stefani
Part of the Microprocessor-Based and Intelligent Systems Engineering book series (ISCA, volume 9)


This paper describes a new approach of knowledge-based systems. There are three main phases: (1) an ‘intelligent’ diagnosis of the step response provides a starting set of parameters, (2) heuristic optimization improves those parameters and (3) the model is validated.

The starting set of parameters are provided by the IDGRA expert system in the usual plant form: G(s) = exp(-Tau.$) Nu(s)/De(s). The delay and lowest order of D(s) may be found from any of thirty models. The optimization phase uses a hierarchical method interfaced with the PC-Matlab identification toolbox. The optimum model is validated by an expert system solver. Examples are included.


Computer-Aided Dynamic System Design (CADSD) identification logic programming Prolog Knowledge-Based System (KBS) expert system solver intelligent front-end 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    P. Eykhoff and P.C. Parks (1990), “Identification and System Parameter Estimation; where do we stand now?”, Automatica, 26, pp. 3–5.CrossRefGoogle Scholar
  2. [2]
    M. Haest, G. Bastin, M. Gevers and V. Wertz (1990), “ESPION: an Expert System for System Identification”, Automatica, 26, pp. 85–95.MathSciNetzbMATHCrossRefGoogle Scholar
  3. [3]
    S. Gentil, A.Y. Barraud and K. Szafnicki (1990), “SEXI: An Expert Identification Package”, Automatica, 26, pp. 803–809.zbMATHCrossRefGoogle Scholar
  4. [4]
    J.E. Larsson and P. Persson (1988), “The knowledge database used in an expert system interface for IDPAC”, Proc. 1st IFAC Workshop on AI in Real-time Control, Swansea, UK, pp. 107–112.Google Scholar
  5. [5]
    M. Monsion, B. Bergeon, A. Khaddad and M. Bansard (1988), “An expert system for industrial process identification”, Proc. 1st IFAC Workshop on Artificial Intelligence in Real-time Control, Swansea, UK, pp. 95–99.Google Scholar
  6. [6]
    A. Betta and D.A. Linkens (1990), “Intelligent knowledge-based for dynamic system identification”, IEE Proc., 137 D, pp. 1–12.Google Scholar
  7. [7]
    H. Nguyen-Phu, T.H. Thanh, C. Rinck, M. Chaabane and C. Humbert (1989), “A knowledge-based system for use in process identification”, Proc. of the International Workshop ’E.S. and Electrical and Power Systems’ held in Avignon, France - May 29–30th.Google Scholar
  8. [8]
    H. Nguyen-Phu (1990), “Prolog Language - CADSD package interfacing procedures: Applications to heuristic identification and control of dynamic systems”, Proc. of the NATO Advanced Study Institute Conference on ’Expert Systems and Robotics’ held in Corfu, Greece - July 15–27, Eds. T. Jordanides and B. Torby, Springer-Verlag, vol. 71 ’Computer and Systems Sciences’, pp. 243–258.Google Scholar
  9. [9]
    L. Ljung (1986), “System Identification Toolbox User’s guide”, The Mathworks, Inc.Google Scholar
  10. [10]
    R.D. Strum and D.E. Kirk (1989), “Linear Systems: Be Discrete - Then Continuous”, IEEE Trans. on Education, 32, pp. 335–342.CrossRefGoogle Scholar
  11. [11]
    H. Nguyen-Phu (1984), Docteur d’Etat Thesis - Part III: “Computer-Aided Characterization Methods”, ESSTIN, University of Nancy I.Google Scholar
  12. [12]
    G.H. Hostetter, C.J. Savant, Jr. and R.T. Stefani (1989), Design feedback control systems, 2nd Ed., Saunders College Publishing (Holt, Rinehart and Winston, Inc.)Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 1991

Authors and Affiliations

  • H. Nguyen-Phu
    • 1
  • P. Weiler
    • 1
  • R. T. Stefani
    • 2
  1. 1.Laboratoire des Applications des Techniques de l’I.A. (L.A.T.I.A.)University of Nancy I, ESSTINVandoeuvreFrance
  2. 2.Department of Electrical EngineeringCalifornia State University, Long BeachUSA

Personalised recommendations