An algorithm for estimating upper bound horizon in model predictive control

The upper bound horizon in model predictive control (MPC) problem is computed by a new non-iterative algorithm. The global stability of the MPC problem is guaranteed by solving the infinite horizon constrained linear quadratic regulator (LQR) problem. While the infinite horizon constrained LQR problem can be transformed into the finite horizon constrained LQR problem based on the upper bound horizon equally. There are some algorithms for estimating the upper bound horizon, however, they need expensive computation or give a big value. Then an new algorithm to estimate the upper bound horizon is presented by the linear programming. It only need to solve a linear programming problem for online application. Finally, the comparison among some methods is given by an example. The proposed algorithm has less conservative than that of other algorithms in recent literatures.