View source for Structural Controllability

'''Structural controllability''' is the property of a network that determines whether its dynamics can be driven to any desired state by controlling only a subset of its nodes — the driver nodes — regardless of the specific parameters assigned to the edges. The concept, formalized by Lin (1974), separates the topological question of whether control is possible from the parametric question of how to achieve it. A network is structurally controllable if and only if there exists a choice of parameters that makes it controllable; the topology determines the possibility, the parameters determine the realization. This distinction is the foundation of [[Control Graph Theory|control graph theory]]: the graph structure constrains the dynamical possibilities, and structural controllability is the constraint that matters most.

The minimum number of driver nodes required to structurally control a network is determined by the graph's maximum matching — a purely combinatorial property that can be computed efficiently. For directed networks, the result is striking: scale-free networks require a vanishingly small fraction of driver nodes, but the identity of those nodes is highly sensitive to topology. A hub node may be a critical driver or entirely irrelevant, depending on whether it is part of a dilation — a set of nodes with fewer incoming edges than nodes. This sensitivity means that structural controllability is not a robust property: small topological changes can transfer controllability from one node to another, making the network's control points unpredictable without exact knowledge of its structure. The design implication is that control architecture must be designed with structural controllability in mind, not merely with the assumption that enough actuators will suffice.

[[Category:Systems]]
[[Category:Mathematics]]
[[Category:Cybernetics]]