Understanding the Structure of Power Law Networks

摘要

The degree distribution of scale-free networks follow power laws. There continues to be disagreement, however, as to what additional properties these networks share. A wide range of techniques useful as aids in understanding common structure as well as in differentiating between elements of this class of networks are explored. First, the utility of a polar coordinate plot for power law distributions is explained. Second, computational experience with two procedures to uncover shortest paths in scale-free networks based solely on locally available data is provided. Next, a tabu search is developed to find high quality solutions for two bi-objective models. For specific objective weights, networks whose degree distributions follow a power law are shown to arise. Lastly, links between the clustering coefficient distribution and modularity are described. Computational experiments supporting the connection between the first nontrivial eigenvalue of a Laplacian matrix and network synchrony are conducted.