Cooperative distributed constrained model predictive control for uncertain nonlinear large scale systems

In this paper two linear constrained cooperative distributed extended dynamic matrix control (CDEDMC) and adaptive generalized predictive control (CDGPC) are proposed to control the uncertain nonlinear large-scale systems. In these approaches, a proposed cooperative optimization is employed which improves the global cost function. The cost values and convergence time are reduced using the proposed cooperative optimization strategy. The proposed approaches are designed based on the compensation of the mismatch between linearized and nominal nonlinear models. In CDEDMC the mismatch is considered as a disturbance and compensated; Also in CDGPC it is compensated using online identification of the linearized model. The typical distributed linear algorithms like DMC leads to an unstable response if the reference trajectory is a little far from the equilibrium point. This problem will be partially solved using the CDEDMC and will be completely solved using the CDGPC even if the reference trajectory is too far from the equilibrium point. The performance and effectiveness of proposed approaches are demonstrated through simulation of a typical uncertain nonlinear large-scale system.

Keywords:

Cooperative optimization approach , Cooperative distributed extended dynamic matrix control , Cooperative distributed adaptive generalized predictive control