A mathematical programming approach to overlapping community detection

We propose a new optimization model to detect overlapping communities in networks. The elaborates suggestions contained Zhang et al. (2007), which were identified through the use of fuzzy membership function, calculated as outcome mathematical programming problem. In our approach, we retain idea using both and communities, but replace objective function proposed there with another one, based on Newman Girvan’s definition modularity. Next, formulate mixed-integer linear calculate optimal communities. After some computational tests, provide evidence that proposal can fix biases previous model, is, its tendency calculating composed almost all nodes. Conversely, reveal other structural properties, such nodes or acting bridges between Finally, be used only for moderate size networks due computation time, two heuristic algorithms solve largest instances, compare favourably methodologies.