Multiple effects of gradient coupling on network synchronization

Abstract

Recent studies have shown that the synchronizability of complex networks can be significantly improved by gradient or asymmetric couplings, and increase of the gradient strength could enhance the network synchronizability monotonically. Here we argue and demonstrate that, for a typical complex network, there could be an optimal gradient where the maximum network synchronizability is achieved. That is, large gradient may deteriorate synchronization. We attribute the suppressing effect of gradient coupling to the phenomenon of network breaking and show that, comparing with sparse homogeneous networks, dense heterogeneous networks suffer less from network breaking and, consequently, benefit more from large gradient in improving synchronization. The findings are supported by indirect simulations of eigenvalue analysis and direct simulations of coupled nonidentical oscillators. © 2008 The American Physical Society.