Evolutionary Optimization of the Structure of Neural Networks by a Recursive Mapping as Encoding

  • B. Sendhoff
  • M. Kreutz
Conference paper


The determination of the appropriate structure of artificial neural networks for a specific problem or problem domain remains an open question. One attempt to solve this optimization problem is the application of evolutionary algorithms and the choice of an appropriate coding, the genotype → phenotype mapping. We employ a coding procedure from the class of recursive coding methods and apply the optimization process to the problem of prediction and modeling of chaotic time series. The network structure and the inital weight setting are determined by an evolutionary process and the ‘fine tuning’ of weights is achieved by a standard back-propagation algorithm. We focus on the properties of the coding procedure and the understanding of the network structures in this context.


Evolutionary Algorithm Small Chromosome Large Chromosome Learning Cycle Chaotic Time Series 
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]
    P.J. Angeline, G.M Saunders, and J.B. Pollack. An evolutionary algorithm that constructs recurrent neural networks. IEEE Trans. Neural Networks, 5(1):54–65, 1994.CrossRefGoogle Scholar
  2. [2]
    G. Deco and B. Schürmann. Neural learning of chaotic dynamics. Neural Processing Letters, 2(2):23–26, 1995.CrossRefGoogle Scholar
  3. [3]
    F. Gruau. Genetic synthesis of modular neural networks. In S. Forrest, editor, Proc. 5th Int. Conf. on Genetic Algorithms, pages 318–325, San Mateo, CA, 1993. Morgan Kaufmann Publishers.Google Scholar
  4. [4]
    H. Kitano. Designing neural networks using genetic algorithms with graph generation system. Complex Systems, 4:461–476, 1990.MATHGoogle Scholar
  5. [5]
    H. Kitano. Neurogenetic learning:an integrated method of designing and training neural networks using genetic algorithms. Physica D, 75:225–238, 1994.MATHCrossRefGoogle Scholar
  6. [6]
    E.N. Lorenz. Irregularity: A fundamental property of the atmosphere. Tellus, A(36):98, 1984.Google Scholar
  7. [7]
    J. Principe, A. Rathie, and J. Kuo. Prediction of chaotic time series with neural networks and the issue of dynamic modeling. International Journal of Bifurcation and Chaos, 2(4):989, 1992.MATHCrossRefGoogle Scholar
  8. [8]
    B. Sendhoff, M. Kreutz, and W. von Seelen. Causality and the analysis of evolutionary algorithms. 1997. Submitted to IEEE Transactions on Evolutionary Computation.Google Scholar

Copyright information

© Springer-Verlag Wien 1998

Authors and Affiliations

  • B. Sendhoff
    • 1
  • M. Kreutz
    • 1
  1. 1.Institut für NeuroinformatikRuhr-Universität BochumBochumGermany

Personalised recommendations