Stable multivariable model based predictive control

The stability of model based predictive control (MBPC) is studied in a multiple-input, multiple-output framework in the presence of active state and input constraints, and implemented in both an input-output and a state space format. The constrained generalized predictive control method of Rossiter and Konvaritakis is extended from the input-output to the state space case. In this approach, stability is established by demonstrating the existence of a monotonically decreasing cost function.<
>