Recent advances in adaptive MPC

It is well known that an origin of MPC is the controller in adaptive control scheme. In other words, MPC is popularly employed as the control law in adaptive control. In primitive adaptive MPC, constraints are not taken into account. However, after this adaptive MPC, tremendous progress has been made in MPC theory independently in a way that various constraints on the state and input variables are considered and closed loop stability is guaranteed. Ironically, it turns out that such a well developed MPC suffers from being employed as the control law in adaptive control scheme because it is difficult to guarantee closed loop stability when the MPC is combined with adaptation algorithm and feasibility at all times owing to estimated parameters. This paper presents several important issues in the problem of adaptive MPC and introduces recent results and future work.