Detection of community structures in networks via global optimization

We present an analysis of communality structure in networks based on the application of simulated annealing techniques. In this case we use as “cost function” the already introduced modularity Q (1), which is based on the relative number of links within a commune against the number of links that would correspond in case the links were distributed randomly. We compare the results of our approach against other methodologies based on betweenness analysis and show that in all cases a better community structure can be attained.