Supervised Robustly Feasible Soft Switching Model Predictive Control with Bounded Disturbances

It is common that an efficient constrained plant operation under full range of disturbance inputs requires meeting different sets of control objectives. This calls for application of multiple model predictive controllers each of them being best fit into specific operating conditions. It is inevitable then to switch between the controllers during the plant operation. A simple hard switching may introduce unwanted transients and more importantly it may not achieve robustly feasible controller operation. In this paper, a softly switched model predictive control scheme is presented based on the invariant set theory. The algorithm is derived for the switching mechanism parameter design that minimizes the soft switching duration time. Numerical examples with simulation results show that the proposed methods can be useful in practical applications