Symbiotic effect: A guideline for network modeling

¹  Institute of Computing Technology, Chinese Academy of Sciences - Beijing, 100190, PRC
²  Graduate University of Chinese Academy of Sciences - Beijing 100190, PRC
³  Department of Modern Physics, University of Science and Technology of China - Hefei 230026, PRC

Corresponding authors: gqzhang@ict.ac.cn guoqiang@ict.ac.cn

Received: 15 May 2009
Accepted: 7 September 2009

Abstract

Creating representative evolving model for a given real network is a challenging task because to capture the mechanism underlying the evolution is not easy. A widely accepted method is to make variants of the Barabási-Albert (BA) model. However, whether this approach can guarantee the establishment of accurate evolving models for real networks remains unknown, and even worse, it lacks the way to help make such a judgment. This letter reports the symbiotic effect which can be treated as a criterion telling whether it is reasonable to develop such variants. Different real networks can show either strong or weak symbiotic effects, while all the networks generated by well-known stochastic evolving models that belong to the BA variants display strong symbiotic effect. Thus, a real network with a strong symbiotic effect indicates the reasonability and possibility to develop a BA variant, while a weak symbiotic effect strongly signals the hardness or even impossibility to reproduce the evolving procedure by using a variant of the BA model.