Commande prédictive approximante.

Abstract

Predictive control is by far the best known method in industry for its tolerance towards different types of systems and the constraints imposed. Nevertheless, it is these limitations upwind speed systems. We propose in this work an approach that is based on the approximate predictive control by fuzzy learning. The optimal predictive control law obtained off line using simulation model is used as data base for training the fuzzy system. Simulations were performed to illustrate the interest of our approach and its validation.