Stable Distributed Model Predictive Control strategy based on agent coordination

Many systems composed by several interacting subsystems are usually controlled by a distributed control framework. Distributed Model Predictive Control (DMPC) strategy, in which each subsystem is controlled by a local MPC controller, has advantages of accommodating constraints, less computational cost and high flexibility. In order to improve the global performance and guarantee the system stability, a stabilized DMPC strategy is proposed in this paper, in which subsystems interact through inputs. At first, local initial feasible solutions are achieved based on a Minkowski functional to guarantee the local closed-loop system stabilization. And then the global optimal solutions are obtained through coordination strategy for the sake of reducing iteration time and accelerating the convergence speed efficiently. Finally, the accuracy and efficiency of the proposed scheme is put to test through simulation.