A Novel Implicit Generalized Predictive Control Algorithm

The conventional generalized predictive control (GPC) derives the future control-increment vector by recursively solving the Diophantine equations. In order to reduce the on-line computation time, a short predictive horizon or a short control horizon is used. However, this is against the basic principle of the GPC to some extent, which sometimes leads to poor control performance. This paper proposes a novel method to both compute the future control-increment vector and choose the control horizon for a class of generalized predictive control algorithms. The proposed method does not involve the Diophantine equations and in the meantime can determine the control horizon, which not only reduces the computational complexity but also avoids selecting the control horizon by the rule of thumb. The efficiency of the algorithm is demonstrated with a comparative simulation study.