Network growth under the constraint of synchronization stability.

While it is well recognized that realistic networks are typically growing with time, the dynamical features of their growing processes remain to be explored. In the present paper, incorporating the requirement of synchronization stability into the conventional models of network growth, we will investigate how the growing process of a complex network is influenced by, and also will influence, the network collective dynamics. Our study shows that, constrained by the synchronization stability, the network will be growing in a selective and dynamical fashion. In particular, we find that the chance for a new node to be accepted by the growing network could have a large variation, i.e., it follows roughly a power-law distribution. Furthermore, we find that, with the dynamical growth, the network is always developed into structures of clear scale-free features, despite the form of the link attachment (preferential or random). The dynamical properties of network growth are studied using the method of eigenvalue analysis, and they are verified by direct simulations of coupled chaotic oscillators. Our study implies that, driven by the network collective dynamics, network growth could also be highly dynamical.