A bottom-up algorithm to create structurally balanced social networks by modifying the sources of tension

The study of social structure and the effect of it on social members is an attractive area in social networks. Structural balance theory focuses on patterns of signed links and frequency/popularity of them. In recent years several works try to define some approximations to calculate the distance of one unbalanced graph from nearest balanced one. But these works don't have any idea about the links with unstable signs that changing their sign makes the network more balanced. Also, some works introduce a centralized algorithm to detect these links. In this paper, we have introduced a localized algorithm for detecting and changing the sign of these links as a source of tension. The results of simulation for several scale-free networks with different features show that proposed algorithm has the ability to move the network to a balanced one. AS the proposed algorithm focuses on components of the social network to calculate localized measures, it is appropriate for agent-based models to study other social phenomena.