Nonlinear dynamic matrix control based on inverse system method

BP neural network is used to approximate to the α th-order inverse system of a class of nonlinear discrete systems. Cascading the inverse model approximated by BP neural network with the original system to get the composite pseudo-linear system. According to the dynamic matrix control(DMC) method that was proposed based on linear system, the nonlinear dynamic matrix control based on inverse system method is proposed. Simulation not only show that the method has better performance, high accuracy and simple design but also validate the effectiveness of the method.