Distributed model predictive control with a novel partition method

In this paper, we present a novel partition method of large-scale systems for distributed model predictive control. A distributed model predictive control strategy based on the novel partition method is proposed. The strategy requires decomposing the entire system into M subsystems based on the control input distribution. The proposed strategy has performance advantage over communication-based distributed model predictive control strategy. Global closed-loop system stability condition with applying the proposed distributed model predictive control strategy is showed. Performance is demonstrated via a numerical example.