Dynamic matrix control based on Elman neural network

Abstract

Dynamic matrix control (for short DMC) algorithm is presented based on Elman neural network (for short ENN), because the ENN can character the dynamic nonlinear system. The basic idea is to predict the system further output by ENN, and to estimate the system dynamic matrix, and then to carry out the dynamic matrix control algorithm. The method can control the nonlinear dynamic system and break through the limitation of mathematical model. The simulation shows that the ENN has good prediction ability and the proposed method is effective.