Simultaneous Optimization of Weights and Structure of an RBF Neural Network

We propose here a new evolutionary algorithm, the RBF-Gene algorithm, to optimize Radial Basis Function Neural Networks. Unlike other works on this subject, our algorithm can evolve both the structure and the numerical parameters of the network: it is able to evolve the number of neurons and their w...

Full description

Saved in:
Bibliographic Details
Published inArtificial Evolution pp. 49 - 60
Main Authors Lefort, Virginie, Knibbe, Carole, Beslon, Guillaume, Favrel, Joël
Format Book Chapter
LanguageEnglish
Published Berlin, Heidelberg Springer Berlin Heidelberg 2006
SeriesLecture Notes in Computer Science
Subjects
Genetic Algorithm
Genetic Code
Good Individual
Hide Neuron
Radial Basis Function Neural Network
Online AccessGet full text
ISBN3540335897
9783540335894
ISSN0302-9743
1611-3349
DOI10.1007/11740698_5

Cover

More Information
Summary:We propose here a new evolutionary algorithm, the RBF-Gene algorithm, to optimize Radial Basis Function Neural Networks. Unlike other works on this subject, our algorithm can evolve both the structure and the numerical parameters of the network: it is able to evolve the number of neurons and their weights. The RBF-Gene algorithm’s behavior is shown on a simple toy problem, the 2D sine wave. Results on a classical benchmark are then presented. They show that our algorithm is able to fit the data very well while keeping the structure simple – the solution can be applied generally.
ISBN:3540335897
9783540335894
ISSN:0302-9743
1611-3349
DOI:10.1007/11740698_5