Are randomly grown graphs really random?

TL;DR

This paper investigates a minimal model of a growing network, revealing that such graphs exhibit a rich phase transition behavior and are fundamentally different from static random graphs due to degree correlations.

Contribution

It introduces a simple growing network model and demonstrates that its structural properties differ significantly from static random graphs, especially in phase transition behavior.

Findings

Giant component emerges at delta=1/8 with an infinite-order phase transition.

Average component size jumps discontinuously at the transition point.

Grown graphs differ fundamentally from static random graphs due to degree correlations.

Abstract

We analyze a minimal model of a growing network. At each time step, a new vertex is added; then, with probability delta, two vertices are chosen uniformly at random and joined by an undirected edge. This process is repeated for t time steps. In the limit of large t, the resulting graph displays surprisingly rich characteristics. In particular, a giant component emerges in an infinite-order phase transition at delta = 1/8. At the transition, the average component size jumps discontinuously but remains finite. In contrast, a static random graph with the same degree distribution exhibits a second-order phase transition at delta = 1/4, and the average component size diverges there. These dramatic differences between grown and static random graphs stem from a positive correlation between the degrees of connected vertices in the grown graph--older vertices tend to have higher degree, and to link with other high-degree vertices, merely by virtue of their age. We conclude that grown graphs, however randomly they are constructed, are fundamentally different from their static random graph counterparts.