Suboptimal model predictive control of hybrid systems based on mode-switching constraints

Model predictive control (MPC) is recognized as a very versatile and effective way of controlling constrained hybrid dynamical systems in closed-loop. The main drawback of hybrid MPC is the heavy computation burden of the associated on-line mixed-integer optimization. Explicit MPC solutions overcome such a problem by rewriting the control law in piecewise affine form, but are limited to relatively simple hybrid control problem setups. This paper presents an alternative approach for reducing the complexity of computations by suitably constraining the mode sequence over the prediction horizon, so that on-line optimization is solved more quickly. While tracking performance of the feedback loop may be affected because of the suboptimality of the approach, closed- loop stability is guaranteed. The effectiveness of the method is demonstrated by an example.