Exploring Evolutionary Multi-Objective Techniques in Self-Organizing Networks

Future networks are promising to solve current issues and provide new features. Self-organizing network (SON) paradigm is one of the anticipated solutions. It involves the use of cognition concept and optimization techniques to enhance the network performance. In this paper, we propose the use of two multi-objective optimization techniques, namely, the multi-objective particle swarm optimization (MOPSO) and the multi-objective central force optimization (MOCFO) in future SON to manage system resources efficiently. Therefore, the used system design and implementation are provided. In addition, the evaluation results of the proposed two methods are compared with those obtained using the non-dominated sorting genetic algorithm (NSGA-II). Extensive simulations carried out using MATLAB package showed that MOPSO is comparable to NSGA-II and outperforms MOCFO in the network throughput. In addition, considering the needed computation time for algorithm convergence, MOPSO is faster than NSGA-II and MOCFO by 5.8 times and 9.9 times on average, respectively. Moreover, this paper provides a study on algorithm convergence rate, solution diversity, and station load.