A Fast GPC Algorithm with Output Penalty

Aiming at decreasing the input and output fluctuations of controlled plant, an improved GPC algorithm is proposed. Firstly, an output increment penalty term is added into objective function for inhibiting output fluctuations and the overshoot. Secondly, the predictive control law is softened to restrain input fluctuations and improve control performance. Finally, a Toeplitz prediction equation is introduced to avoid the iterative calculation resulting from solving Diophantine equation in the traditional GPC algorithm. The Toeplitz prediction equation gives the linear relationship among past input/output, future input and the predictive output, and then the future control variables can be solved directly. The simulation results show that the algorithm can decrease the output and input fluctuations, increase the anti-interference ability and reduce online calculation time.