Modeling for Dynamic Systems with Fuzzy Sequential Knowledge

  • I. Takeuchi
  • T. Furuhashi
Part of the Studies in Fuzziness and Soft Computing book series (STUDFUZZ, volume 59)


This paper presents a fuzzy modeling method for dynamic systems, focusing on the knowledge representation framework. The basic idea lies in the use of “fuzzy” sequential knowledge for the description of dynamic characteristics of a system. Symbolic Dynamic System(SDS), a model for symbolic sequences, is extended to deal with “fuzzy” symbolic sequences. This approach introduces topological nature into the symbolic sequences, which allows an interpretation of the knowledge in numerical forms. The model consists of a mixture of sub-models and represents dynamic characteristics by fuzzy transitions among those sub-models. The parameter estimation algorithm, which performs steepest gradient descent in cost functions defined with fuzzy constraints, is presented. Numerical experiments demonstrate the feasibility of the proposed method.


Membership Function Dynamic Characteristic Knowledge Representation Target System Convex Polyhedron 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Metropolis, N. and Stein, M. and Stein, P, “On finite limit sets for transformations on the unit interval”, Journal of Combinatorial Theory, vol. 15, pp. 2544, 1973MathSciNetGoogle Scholar
  2. 2.
    T.Takagi and M.Sugeno, “Fuzzy Identification of Systems and Its Applications to Modeling and Control”, IEEE Trans. on SMC,vol.15, num.1, pp.116132,1985Google Scholar
  3. 3.
    R. Kruse and R. Buck-Emden and R. Cordes, “Processor Power Considerations -An Application of Fuzzy Markov Chains”, Fuzzy Sets and Systems, vol.21, num. 3, pp. 289–299, 1987MathSciNetCrossRefGoogle Scholar
  4. 4.
    T.W.Cacciatore and S.J.Nowlan, “Mixtures of controllers for jump linear and nonlinear plants”, Advances in Neural Information Processing Systems, vol. 6, pp. 719–726, 1994Google Scholar
  5. 5.
    Y. Bengio and P. Frasconi, “Input-Output HMM’s for Sequence Processing”, IEEE Trans. on Neural Networks, vol.7, num. 5, pp. 1231–1249, 1996CrossRefGoogle Scholar
  6. 6.
    A. P. Dempster and N. M. Laird and D. B. Rubin, “Maximum likelihood from incomplete data via the EM algorithm”, Journal of the Royal Statistical Society series B vol. 39, pp. 1–38, 1977MathSciNetMATHGoogle Scholar
  7. 7.
    G. E. Hinton and T. J. Sejnowski, “Learning and relearning in Boltzmann machines”, in D. E. Rumelhart and J. L. Mclelland, editors, Parallel Distributed Processing, MIT Press, Cambridge, MA, 1986Google Scholar

Copyright information

© Physica-Verlag Heidelberg 2001

Authors and Affiliations

  • I. Takeuchi
    • 1
  • T. Furuhashi
    • 2
  1. 1.The Institute of Physical and Chemical ResearchNagoya UnivesityJapan
  2. 2.Nagoya UniversityJapan

Personalised recommendations