ESTRATEGIA HÍBRIDA AGA-SIMPLEX PARA LA ESTIMACIÓN DE PARÁMETROS DE MODELOS DINÁMICOS A PARTIR DE LA RESPUESTA AL ESCALÓN

Abstract

It is presented a hybrid strategy based on an Adaptive Genetic Algorithm and the Simplex method for the open-loop dynamic models parameter estimation. It is analyzed the behavior of the operators implicated in the genetic optimizer, as well as the performance of the Simplex in a second estimation stage. The proposal is compared with the conventional techniques which have better performance in the task. The comparison is based on the sum of square error committed in simulation from a test batch composed by typical dynamical process models. It is also stated the solution for systems with special dynamics