Identification Methods Based on Associative Search Procedure

Received: 20 April 2011 Abstract Accepted: 19 August 2011 In modern control systems, identification is an integral part of adaptive control where process models are adjusted using real-time operation data and control actions optimal with respect to some performance criterion are developed. A variety of identification methods based on mathematical statistics techniques have been developed. Algorithms optimal for certain classes of objects and external disturbances were categorized dependent on the available a priori information about the control object. The limits of approximating models development and application were outlined. Against this background, the paper presents novel associative search techniques enabling the development of a new dynamic object’s model on each time step rather than plant approximation pertaining to time. The model is build using the data samples from process history (associations) developed at the learning phase. The new techniques employs the models of human individual’s (process operator’s, stock analyst’s or trader’s) behavior based on professional knowledge formalization. Application examples from oil refining and chemical industries, power engineering, and banking are adduced.

[1]  Michio Sugeno,et al.  Fuzzy identification of systems and its applications to modeling and control , 1985, IEEE Transactions on Systems, Man, and Cybernetics.

[2]  Anil K. Jain,et al.  Data clustering: a review , 1999, CSUR.

[3]  Vimla L. Patel,et al.  Cognitive models of directional inference in expert medical reasoning , 1997 .

[4]  Vladimir A. Lototsky,et al.  ASSOCIATIVE SEARCH MODELS IN INDUSTRIAL SYSTEMS , 2007 .

[5]  Andrey V. Gavrilov,et al.  The model of associative memory of intelligent system , 2002, Proceedings 6th Russian-Korean International Symposium on Science and Technology. KORUS-2002 (Cat. No.02EX565).

[6]  Vladimir A. Lototsky,et al.  ASSOSIATIVE SEARCH MODELS IN TRADING , 2008 .

[7]  José María Carazo,et al.  Smoothly distributed fuzzy c-means: a new self-organizing map , 2001, Pattern Recognit..

[8]  H. Simon,et al.  The sciences of the artificial (3rd ed.) , 1996 .

[9]  P. Ramadge,et al.  Discrete time stochastic adaptive control , 1979, 1979 18th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.

[10]  Han-Fu Chen.  Recursive System Identification and Adaptive Control by Use of the Modified Least Squares Algorithm , 1984 .

[11]  Ann Macintosh,et al.  Applications and Innovations in Intelligent Systems IX , 2002, Springer London.

[12]  E. Hunt.  COGNITIVE SCIENCE: Definition, Status, and Questions , 1989 .

[13]  Natalia Bakhtadze,et al.  FUZZY SOFT SENSORS FOR CHEMICAL AND OIL REFINING PROCESSES , 2008 .

[14]  A. Juditsky,et al.  On minimax approach to non‐parametric adaptive control , 2001 .

[15]  V. M. Chadeev.  Digital Identification of Nonlinear Dynamic Systems , 2004 .

[16]  Robert E. Kalaba,et al.  Dynamic Programming and Modern Control Theory , 1966 .

[17]  R. Tinos,et al.  Reinforcement learning schemes for control design , 1997, Proceedings of the 1997 American Control Conference (Cat. No.97CH36041).

[18]  Kaddour Najim,et al.  Learning automata and stochastic optimization , 1997 .

[19]  Rex V. Brown,et al.  Numerical and verbal decision analysis: comparison on practical cases , 2000 .

[20]  Lennart Ljung,et al.  Perspectives on system identification , 2010, Annu. Rev. Control..

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