Evolving the Topology and the Weights of Neural Networks Using a Dual Representation

  • Authors:
  • João Carlos Figueira Pujol;Riccardo Poli

  • Affiliations:
  • School of Computer Science, The University of Birmingham, Birmingham B15 2TT, UK.;School of Computer Science, The University of Birmingham, Birmingham B15 2TT, UK.

  • Venue:
  • Applied Intelligence
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

Evolutionary computation is a class of global search techniquesbased on the learning process of a population of potential solutions to agiven problem, that has been successfully applied to a variety of problems.In this paper a new approach to the construction of neural networks based onevolutionary computation is presented. A linear chromosome combined to agraph representation of the network are used by genetic operators, whichallow the evolution of the architecture and the weights simultaneouslywithout the need of local weight optimization. This paper describes theapproach, the operators and reports results of the application of thistechnique to several binary classification problems.