Abstract
Neural network model predictive controllers have demonstrated high potential in the non-conventional branch of nonlinear control. However, the major issue in process control of nonlinear systems is the sensitivity to parameters variations and uncertainties. Indeed, when the process is controlled by neural network model predictive control (NNMPC) and subject to parameters variations or uncertainties, unsatisfactory tracking performances are obtained. To overcome this problem, we propose in this paper an adaptive neural network model predictive control (ANNMPC) where a neural model identification block is incorporated in the scheme and online update of the weights is provided when the process is subject to parameters variations and uncertainties. Simulations have been carried out to show the robustness of this control algorithm.