Computational Analysis of Decomposition Strategies for MPC of Resource-Constrained Dynamic Systems

Abstract

This paper presents two decomposition approaches, Bilevel Optimization, and Benders Decomposition, to a model predictive control of resource-constrained dynamic systems. The proposed methods yield a distributed solution that converges to the same optimum that would be obtained by a centralized controller. It is also showed that the decompositions enable the use of multi-core or distributed architectures. Computational analyses from experiments of synthetics computing problems are reported and discussed, showing the capacity of the decomposition approaches in combination with parallel computation. The results showed that Bilevel optimization is more appropriate than Benders decomposition for the considered formulation. Bilevel optimization was also shown to benefit to a greater degree from parallelism.