Closed-loop stability conditions of a non-cooperative distributed MPC configuration

Daniel D. Santana¹; Magnus M. F. Pereira¹; Márcio A. F. Martins¹

¹ Universidade Federal da Bahia

Baixar PDF doi:10.20906/CPS/CBA2018-0448

Resumo

This article evaluates the conditions of stability for a non-cooperative Distributed Model Predictive Controller configuration, which couple nominally stabilizing model predictive controllers. Two case studies using a four-tank system assess the characteristics of the distributed formulation. In the first case, for a set of system parameters, it is showed that if enough information is shared between local controllers, their cost function are non-increasing. In the second case, it is mapped the combination of system parameters, in which the conditions of stability of the distributed MPC formulation are held.

Palavras-chave: DMPC; IHMPC; feasibility; Lyapunov stability