Decentralized switched model-based predictive control for distributed large-scale systems with topology switching

Ahandani M. A., Kharrati H., Hashemzadeh F., Baradarannia M.

Nonlinear Analysis: Hybrid Systems, vol.38, 2020 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 38
  • Publication Date: 2020
  • Doi Number: 10.1016/j.nahs.2020.100912
  • Journal Name: Nonlinear Analysis: Hybrid Systems
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Communication Abstracts, Compendex, INSPEC, MathSciNet, Metadex, zbMATH, Civil Engineering Abstracts
  • Keywords: Decentralized switched model-based predictive control, Distributed switched large-scale system, Network topology, Switching signal
  • Istanbul Technical University Affiliated: No


This paper proposes a decentralized switched model-based predictive control (DeSwMPC) for handling coupling among subsystems in a distributed switched large-scale system composed of physically interconnected subsystems. In the distributed switched large-scale systems, interactions among subsystems vary over time according to an exogenous input signal named switching signal. The proposed controller aims at stabilizing the origin of the whole closed-loop system while guaranteeing the satisfaction of constraints in the presence of a switching signal. In the DeSwMPC, to consider switching signal effect in variation of network topology, a robust tube-based switched model-based predictive control (SwMPC) is employed as local controller. The SwMPC controllers with switch-robust control invariant (switch-RCI) set as its target set are robust to unknown mode switching. In the employed decentralized model-based predictive control (DeMPC), by assuming interconnections as the additive disturbances, the effect of switch is only reflected on local constraint sets of the nominal subsystems. Simulations are performed on two typical examples. In the first case, the switching times are unknown a priori but the neighborhood sets after switch are known a priori. In the second case, both of them are assumed to be unknown a priori. The obtained results demonstrate that the proposed DeSwMPC satisfies the input and state constraints at all times. They also validate that the closed-loop system converges to the origin.