Vertex degree distribution and arc endpoints degree distribution of graphs with a linear rule of preferential attachment and Pennock graphs

TL;DR

This paper derives exact degree distributions for two classes of preferential attachment graphs, demonstrating their isomorphism and providing formulas for graph generation with power-law degree distributions, validated by simulations.

Contribution

It provides exact formulas for degree distributions of L-graphs and Pennock graphs, establishing their relationship and practical methods for graph calibration.

Findings

Exact degree distributions for L-graphs and Pennock graphs.

Proved isomorphism between Pennock and certain L-graphs.

Validated results through numerical calculations and simulations.

Abstract

The article deals with two classes of growing random graphs following the preferential attachment rule with a linear weight function, L-graphs, and hybrid Pennock graphs. We determine the exact final vertex degree distribution and the exact final two-dimensional arcs degree distributions of graphs under consideration. The study proves that each hybrid Pennock graph is isomorphic to a certain L graph and that the converse does not hold since there are no Pennock graphs isomorphic to L graphs with negative displacements in the expression for the linear weight function. A formula is derived that makes it possible to determine the weight functions, which are used to generate graphs with the required asymptotic power-law vertex degree distribution. The reliability of the obtained results is confirmed by calculations using accurate numerical methods and simulation modeling, i.e. by direct generating of the graphs. The practical value of the results is demonstrated by an example of their effective application for accurate calibration of a growing graph that simulates a network of Internet at the level of autonomous systems.