Optimizing social group homogeneity in online communities

The optimization of homogeneity in social interest groups in Online Social Networks involves their formation and evolution and it is based on both users’ preferences and Choices of the administrators’ groups. To consider the satisfaction of group members, also taking into account their similarity, the homogeneity of a social group it is often regarded as a basilar condition. In this paper we introduce a group homogeneity measure based on users’ behavioral information. Such a measure is optimized by a novel algorithm that by operating in a fully-distributed multi-agent framework performs users and groups profiles matching. Such an approach produces clear and significant advantages in groups formation in an online social network context as it is shown by some experi-mental campaigns performed on simulated social networks