— A receding horizon control algorithm, originally proposed for tracking best-possible steady-states in the presence of overly stringent state and/or input constraints, is analyzed for the case of nonlinear plant models and possibly nonconvex cost functionals. Unlike the linear case (with convex cost functionals), convergence to equilibrium is not always possible and only average performance bounds are guaranteed in general.
David Angeli, Rishi Amrit, James B. Rawlings