Model-Based and Data-Driven Hierarchical Control and Topology Co-Design for Robust Networked Systems

Researchers propose a hierarchical control strategy for networked systems using both model-based and data-driven approaches to ensure robust performance while optimizing network topology. The method leverages dissipativity theory and linear matrix inequality problems to design distributed controllers without requiring centralized computation, with applications demonstrated in DC microgrid voltage regulation.

This research addresses a fundamental challenge in managing complex interconnected systems by proposing control strategies that balance robustness with computational efficiency. The dual approach—offering both model-based and data-driven pathways—acknowledges the reality that real-world networked systems often lack complete dynamic knowledge, making the data-driven variant particularly valuable for practical deployment. The use of dissipativity theory provides a mathematically rigorous framework for guaranteeing system stability and performance across distributed subsystems without requiring global system information.

The hierarchical design methodology represents an evolution in control theory by decomposing complex network problems into manageable local optimization tasks through LMI formulations. This approach sidesteps the computational intractability of centralized design while maintaining formal guarantees, a significant advantage for large-scale infrastructure systems. The relaxation of conventional disturbance bounds using quadratic matrix inequalities demonstrates practical sophistication in handling real-world uncertainty.

The application to DC microgrids reveals the timing and relevance of this work to modern energy systems transitioning toward distributed renewable sources. Robust voltage regulation and current sharing are critical for grid stability, and autonomous, decentralized control mechanisms reduce single points of failure and improve resilience. The research bridges theoretical control advancement with infrastructure-critical applications.

Future implementations may extend this framework to power systems with faster dynamics, higher renewable penetration, or more heterogeneous component characteristics. The compositionality properties suggest scalability to larger networks, making this foundational work relevant for engineers designing next-generation smart grids and distributed energy management systems.

• →Hierarchical control design uses dissipativity theory to guarantee robust performance while maintaining computational efficiency through decentralized LMI optimization.
• →Data-driven approach enables control design from input-output trajectories alone, eliminating the requirement for complete subsystem dynamic models.
• →Topology co-design optimizes interconnection costs alongside controller synthesis, addressing both performance and infrastructure efficiency simultaneously.
• →Application to DC microgrids demonstrates practical utility for distributed renewable energy systems requiring autonomous voltage regulation and current sharing.
• →Compositionality and decentralizability properties enable scalable deployment in large-scale networked infrastructure without centralized computation.