A fuzzy dynamic programming solution to constrained nonlinear predictive control

A method for the on line solution of the constrained non linear predictive control is proposed. The method uses fuzzy dynamic programming which is based on the concepts of fuzzy constraints and fuzzy objective. The fuzzy objective used on this work is imposed on the error at the final instant of prediction and is expressed linguistically as: the error on the final instant of prediction should be zero. The fuzzy constraints are imposed on the control increment. The approach is tested in simulation for the control of concentration in continuously stirred tank reactor with encouraging results.