SciELO - Scientific Electronic Library Online

 
vol.78 número165AGRUPAMIENTO HOMOG ÉNEO DE ELEMENTOS CON MÚLTIPLES ATRIBUTOS MEDIANTE ALGORITMOS GENÉTICOSMETODOLÓGIA PARA LA CONSTRUCCIÓN DE UN MIGRADOR UNIVERSAL DE BASES DE DATOS índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados

Revista

Articulo

Indicadores

Links relacionados

  • En proceso de indezaciónCitado por Google
  • No hay articulos similaresSimilares en SciELO
  • En proceso de indezaciónSimilares en Google

Compartir


DYNA

versión impresa ISSN 0012-7353versión On-line ISSN 2346-2183

Resumen

CORREA, RODRIGO; BEGAMBRE, OSCAR  y  CARRILLO E., JULIO C.. VALIDATION OF A PSO-SIMPLEX HYBRID ALGORITHM OF PARAMETRIC EVOLUTION TOPOLOGY. Dyna rev.fac.nac.minas [online]. 2011, vol.78, n.165, pp.255-265. ISSN 0012-7353.

This paper describes some of the most important aspects related to the numerical experimentation of a hybrid of the algorithm PSO (Particle Swarm Optimization) with the traditional modified simplex method of Nelder-Mead. The hybridization of these two techniques of optimization without restrictions was carried out with a topology that allows to optimize in each iteration the parameters of the algorithm PSO using the modified simplex method. Numerical experiments with this hybrid algorithm were carried out and applied to several of typical test functions to establish its effectiveness. The results obtained were compared with the simplex and the quadratic methods, which turned out to be very satisfactory since the point of view of their repeatability and reproducibility, although the time of computation was considerably longer. It stands out itself, nevertheless, that the precision of the hybrid algorithm was a hundred percent in all the trials for the test functions selected.

Palabras clave : Optimization without restriction; heuristic methods; stochastic methods.

        · resumen en Español     · texto en Español     · Español ( pdf )

 

Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons