Fast algorithms for indices of nested split graphs approximating real complex networks

We present a method based on simulated annealing to obtain a nested split graph that approximates a real complex graph. This is used to compute a number of graph indices using very efficient algorithms that we develop, leveraging the geometrical properties of nested split graphs. Practical results are given for six graphs from such diverse areas as social networks, communication networks, word associations, and molecular chemistry. We present a critical analysis of the appropriate perturbation schemes that search the whole space of nested split graphs and the distance functions that gauge the dissimilarity between two graphs.