Overlapping community detection on social networks has re-

ceived a lot of attention nowadays and it has been recently addressed

as Multi-objective Optimization Evolutionary Algorithms (MOEAs). In

this work we propose a multi-objective and evolutionary algorithm for

overlapping community detection in social networks, named MOGR-

PESA2, which builds an initial set of communities seeds.

Our algorithm employs the PESA-II framework and it proposes a new

probabilistic evolutionary operator, which uses the information contained

in the Pareto set in order to improve the heuristic search over the solution

space. Moreover, we also propose the use of the Spark GraphX API in

order to speeding up the building of the communities seeds.

The experimental evaluation over synthetic networks showed that our

proposal is promising and effective for overlapping community detection

in social networks. In addition, the inclusion of the Spark GraphX API

allows our proposal to significantly accelerate the identification of com-

munity seeds.