Structural Controllability

Structural controllability investigates the ability to control a dynamical system represented by a graph, regardless of the specific numerical values of its parameters. Current research focuses on efficiently identifying minimal sets of control inputs and crucial network connections (backbones) that guarantee controllability, often employing graph-theoretic methods and reinforcement learning with graph neural networks to solve this computationally challenging problem. These advancements are significant for understanding and controlling complex networked systems across diverse fields, from social networks to multi-agent systems, by providing efficient strategies for achieving desired system behaviors with limited control resources.