A two-level model predictive control formulation for stabilization and optimization

In this paper, we present a novel MPC algorithm, which has a two-level hierarchical structure. For the lower level control objective of stabilization, no optimization is involved, making it computationally efficient. For the higher-level control objective of achieving an economic target, on-line optimization is performed with any desired objective function and control horizon without affecting the stability of the closed-loop system. This higher-level optimization problem does not have to be solved within one sampling period, making the overall algorithm computationally attractive. The proposed two-level algorithm is illustrated with a benchmark problem.