Concise Iterative Algorithms On the State Feedback Form for Model Predictive Control and Stability Analysis of Regime Switching Systems

摘要

In this paper we present iterative algorithms to compute the state feedback form for model predictive control (MPC) of regime switching systems and perform stability analysis. We consider cases involving classic quadratic cost functions, costs with a penalty on sudden changes of control, time-invariant and time-variant systems, and Markovian jump systems (MJS). The algorithms are efficient and easy to implement, and the stability analysis of the closed loop system becomes straight forward using them, as illustrated with numerical examples. We then show, by varying the depth of the prediction horizon, that a very long horizon in MPC does not significantly improve stability. This result has instructive value on system design using MPC.