Model predictive control: Theory and practice—A survey

In the past decade, Model Predictive Control (MPC) has become a very attractive feedback strategy, particularly for difficult multivariable control problems with constraints. This paper is a survey of what has been accomplished—both theoretical developments and practical applications. The MPC family of methods (DMC, QDMC, IDCOM, MAC, etc.) is unified in a common framework, stability and robustness properties are discussed, and the relationships to optimal control, LQR, and H-infinity control are elaborated.