Probabilistic prediction in scale-free networks: Diameter changes

TL;DR

This paper investigates how the diameter of scale-free networks changes probabilistically when nodes are removed, revealing a robust algebraic decay pattern with specific exponents across different networks.

Contribution

It identifies a universal algebraic decay in diameter changes with a consistent exponent, and highlights the existence of a different decay pattern in certain real-world networks.

Findings

Diameter changes follow an algebraic decay with exponent ~2.2 in most networks.

The exponent is robust across degree exponents 2<γ≤3.

Some networks, like the Internet, exhibit a different decay exponent ~1.7.

Abstract

In complex systems, responses to small perturbations are too diverse to predict how much they would be definitely, and then such diverse responses can be predicted in a probabilistic way. Here we study such a problem in scale-free networks, for example, the diameter changes by the deletion of each node for various in silico and real world scale-free networks. We find that the diameter changes are indeed diverse and exhibit an algebraic decay with an exponent \zeta asymptotically. Interestingly, the exponent \zeta is robust as \zeta \simeq 2.2(1) for most scale-free networks, insensitive to the degree exponents \gamma as long as 2 < \gamma \le 3. However, there is another type with \zeta \simeq 1.7(1) and its examples include the Internet and its related in silico model.