MPC for max-plus-linear systems with an output cost criterion: steady-state behavior and guaranteed stability

In previous work we have extended the popular Model Predictive Control (MPC) design technique to a class of discrete event systems that can be described by a model that is ¿linear¿ in the max-plus algebra. In this paper we study the steady-state behavior of these so-called max-plus-linear systems in the case of an output cost criterion. We derive improved tuning rules for the controller parameters that guarantee us a feasible and stable operation of the controller in the unconstrained case. An example shows that violation of the tuning rules may destabilize the closed-loop system.