Asynchronous hybrid architecture for parametric system identification using a fuzzy real coded evolutionary algorithm

  • Oscar Montiel
  • Oscar Castillo
  • Patricia Melin
  • Roberto Sepulveda

Abstract

The aim of this paper is to present an asynchronous hybrid architecture for parametric system identification. It consists in combining the advantage that the evolutionary algorithms have in exploring a huge multidimensional search space with traditional gradient methods who have proved to be excellent exploiting search space regions. The evolutionary algorithm is real coded and uses fuzzy recombination. It is shown a mathematical analysis of errors and stability.
Published
2004-02-01