Realistic network growth using only local information: From random to scale-free and beyond

Abstract

We introduce a simple one-parameter network growth algorithm which is able to reproduce a wide variety of realistic network structures but without having to invoke any global information about node degrees such as preferential-attachment probabilities. Scale-free networks arise at the transition point between quasi-random and quasi-ordered networks. We provide a detailed formalism which accurately describes the entire network range, including this critical point. Our formalism is built around a statistical description of the inter-node linkages, as opposed to the single-node degrees, and can be applied to any real-world network -- in particular, those where node-node degree correlations might be important.