Community structure identification has been an important research area for biology, physics, information systems, and social sciences for studying properties of networks representing complex relationships. Lately, Genetic Algorithms (GAs) are being utilized for community detection. GAs are machine-learning methods that mimic natural selection. However, previous approaches suffer from some deficiencies: redundant representation and linearity assumption, that we will try to address. in. The algorithm presented here is a novel framework that addresses both of these above issues. This algorithm is also flexible as it is easily adapted to any given mathematical objective. Additionally, our approach doesn’t require prior information about the number of true communities in the network. Overall, our efficient approach holds potential for sifting out communities representing complex relationships in networks of interest across different domains.