Distributed/Hierarchical Control Architecture Design

Distributed and/or hierarchical control strategies occupy a central role in controlling complex networks effectively. A systematic decomposition method for designing the corresponding architectures is currently lacking. In this paper, we propose a network theoretic method to design distributed/hierarchical control architectures. The system digraph is first simplified through unimportant edge removal and node agglomeration. The connected components of the simplified network form a subsystem in a distributed control architecture. For each subsystem, hierarchical input-output subnetworks are detected using the Pichai-Sezer-Siljak algorithm, thus generating a hierarchical architecture. The proposed method is illustrated through a reaction-separation process.