Advertisement

Genetic improvements of feedforward nets for approximating functions

  • Klaus-U. Höffgen
  • H. Peter Siemon
  • Alfred Ultsch
Neural Networks Neural Networks And Genetic Algorithms
Part of the Lecture Notes in Computer Science book series (LNCS, volume 496)

Abstract

Theoretical approaches to construct feedforward nets for the approximation of functions often do not take into account the number of units needed. For practical reasons this number may be very limited, however. In this paper we propose a systematic way to construct neural networks and then use a genetic algorithm to minimize the number of units needed. This leads to "optimal" networks.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

4. References

  1. [Goldberg 89]
    Goldberg, D.E.; Genetic Algorithms in Search, Optzimization & Machine Learning Addison-Wesley Publishing Company, 1989Google Scholar
  2. [Hecht-Nielsen 89]
    Hecht-Nielsen,R.: Theory of the Backpropagation Neural Network, Proc. Int. Joint Conf. on Neural Networks, Vol I, Washington DC 1987, pp. 593–601Google Scholar
  3. [Hornik et.al. 89]
    Hornik, K., Stinchcombe, M., White, H.: Multilayer Feedforward Networks are Universal Approximators, Neural Networks, Vol.2, 1989, pp. 359–366Google Scholar
  4. [Höffgen/Siemon 90]
    Höffgen, K.-U., Siemon, H.P.: Approximation of Functions with Feedforward Nets, Forschungsberichte Informatik Nr. 346, Universität Dortmund 1990Google Scholar
  5. [Schiffmann et.al. 90]
    Schiffmann, W., Mecklenburg, K.: Genetic Generation of Backpropagation Trained Neural Networks, Int. Conf. on Parallel Processing in Neural Systems and Computers, 1990, pp. 205–208Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Klaus-U. Höffgen
    • 1
  • H. Peter Siemon
    • 1
  • Alfred Ultsch
    • 1
  1. 1.Department of Computer ScienceUniversity of DortmundDortmundFederal Republic of Germany

Personalised recommendations