A GENETIC ALGORITHM FOR COMMUNITY FORMATION IN MULTI-AGENT SYSTEMS

Community formation has certainly gained more and more attention from both the researchers and practitioners in the fields of complex networks and multi-agent systems. Since the number of the possible communities is exponential in the number of agents, an efficient algorithm is needed. Genetic algorithms are very useful tools for obtaining high quality and optimal solutions for optimization problems, due to their self-organization, self-adaptation and high parallelism. The paper proposes a high performance genetic algorithm for community formation, the key concept in our algorithm being a new fitness index, which aims at being a trade-off between intelligence and cooperation, and allows not only community formation but also intelligence to be the driving principle in the community formation process.