Complexity vs stability in small-world networks

Abstract

According to the May-Wigner stability theorem, increasing the complexity of a network inevitably leads to its destabilization, such that a small perturbation will be able to disrupt the entire system. One of the principal arguments against this observation is that it is valid only for random networks, and therefore does not apply to real-world networks, which presumably are structured. Here we examine how the introduction of small-world topological structure into networks affect their stability. Our results indicate that, in structured networks, the parameter values at which the stability-instability transition occurs with increasing complexity is identical to that predicted by the May-Wigner criteria. However, the nature of the transition, as measured by the finite-size scaling exponent, appears to change as the network topology transforms from regular to random, with the small-world regime as the cross-over region. This behavior is related to the localization of the largest eigenvalues along the real axis in the eigenvalue plain with increasing regularity in the network.

7 pages, 4 figures

References in corpus (4)

• Spectra of "Real-World" Graphs: Beyond the Semi-Circle Law
• Robust Patterns in Food Web Structure
• Diffusion, localization and dispersion relations on ``small-world'' lattices
• Stability of a neural network model with small-world connections

Cited by in corpus (10)

• Analyzing and Modeling Real-World Phenomena with Complex Networks: A Survey of Applications
• Evidence of universality for the May-Wigner stability theorem for random networks with local dynamics
• Modular networks emerge from multiconstraint optimization
• How much random a random network is : a random matrix analysis
• Modular networks with hierarchical organization: The dynamical implications of complex structure
• Stability and chaos in coupled two-dimensional maps on Gene Regulatory Network of bacterium E.Coli
• Network connectivity during mergers and growth: optimizing the addition of a module
• Balanced condition in networks leads to Weibull statistics
• Optimal interdependence enhances robustness of complex systems
• Stability as a natural selection mechanism on interacting networks