Approximate explicit robust model predictive control of a CSTR with fast reactions

This paper presents a novel approach to the construction of explicit representation of robust model predictive feedback laws for constrained linear systems with parametric uncertainties. The approach is based on employing approximate dynamic programming where a large optimal control problem is split into a sequence of problems of smaller size. To be able to solve each such problem explicitly, we employ a bounded approximation of optimal value functions. Such a procedure is demonstrated to be superior, both in terms of construction time as well as in the complexity of the resulting controllers, to traditional approaches based on the one-shot approach. The proposed method is demonstrated on a simulation case study involving a continuous stirred tank reactor (CSTR) where a fast multi-component chemical reaction takes place.