Networks are presented as controls in controlled dynamic systems. Viability is the property for a state x that there exists a trajectory starting from x and satisfying the constraints until trhe time horizon. To obtain this, connection matrices must be selected at each time and each visited satte among a specific set, the regulation map, which is carefully defined and built.
Two examples, Sampson's monks and Padgett and Ansell's Florentines, illustrate the viability approach of dynamic networks. Notably, the relationship with centrality is studied. Historical processes involving networks are discussed.