Publication View

Synchronization is optimal in non-diagonalizable networks (2006)

Abstract
We consider the problem of maximizing the synchronizability of oscillator networks by assigning weights and directions to the links of a given interaction topology. We first extend the well-known master stability formalism to the case of non-diagonalizable networks. We then show that, unless some oscillator is connected to all the others, networks of maximum synchronizability are necessarily non-diagonalizable and can always be obtained by imposing unidirectional information flow with normalized input strengths. The extension makes the formalism applicable to all possible network structures, while the maximization results provide insights into hierarchical structures observed in complex networks in which synchronization plays a significant role.. Comment: 4 pages, 1 figure; minor revision

Publication details
Download http://arxiv.org/abs/cond-mat/0605619
Repository arXiv (United States)
Keywords Condensed Matter - Disordered Systems and Neural Networks, Condensed Matter - Statistical Mechanics, Nonlinear Sciences - Adaptation and Self-Organizing Systems, Nonlinear Sciences - Chaotic Dynamics, Quantitative Biology - Neurons and Cognition
Type text