The rapid develop of industry requires the high control quality, which promotes the production of many kinds of advanced control method, predictive control is one of successful method of them. The advantages of predictive control are the low requirement to model, convenience of calculation and good control effect, but it also has the disadvantages of the insufficient in some aspects such as the theory research of control stability and robustness. Because of the great success of predictive control in process industry, more and more researchers show interest in this field. There are many kinds of predictive control algorithm now, three main are Dynamic Matrix Control(DMC), Model Algorithm Control(MAC) and Generalized Predictive Control(GPC). Although there are so many kinds predictive control algorithm, there is trouble to choose specialized algorithm, and which one is better? This dissertation is carried out under a unified frame called the unified predictive control. The main content of this dissertation is based on the unified predictive control. The dissertation is composed of three parts. The first part is the review of the predictive control history, the second is the unified predictive control research based of input-output model, The third is the unified predictive control based of state-space model. The main contribution of this dissertation are as follows: 1, After reviewing the developing process of predictive control, the character of the predictive control and the developing level are introduced, and some latest research achievements in this field are also presented. Some problems are pointed. 2, Two main predictive control-DMC and GPC including SISO and MIMO system with constraints are introduced, and the compare of these two methods is given. 3, Through analyzing of the SISO and MIMO unified predictive control model, configuration method of cost function is given and the control law is obtained. On the basis of the unified predictive control algorithm, the internal model controller and filter of unified predictive control are obtained. The internal model structure is used to analyze the influence of SISO unified predictive control parameters to closed-loop performance. Some basic conclusions are also achieved in MIMO unified predictive control. All these results are obtained based on the input-output model. 4, Because of the shortcomings of input-output model especially the complicated expression in the multivariable unified predictive control, the state-space model is used to give research in predictive control. The unified state-space model, modified cost function and state-space unified control law are both obtained for SISO and MIMO system. Simulation shows the reasonability of the statespace unified predictive control
修改评论