Using Different Many-Objective Techniques in Particle Swarm Optimization for Many Objective Problems: An Empirical Study

Pareto based Multi-Objective Evolutionary Algorithms face several problems when dealing with a large number of objectives. In this situation, almost all solutions become nondominated and there is no pressure towards the Pareto Front. The use of Particle Swarm Optimization algorithm (PSO) in multi-objective problems grew in recent years. The PSO has been found very efficient in solve Multi-Objective Problems (MOPs) and several Multi-Objective Particle Swarm Optimization algorithms (MOPSO) have been proposed. This work has the goal to study how PSO is affected when dealing with ManyObjective Problems. Recently, some many-objective techniques have been proposed to avoid the deterioration of the search ability of multi-objective algorithms. Here, two many-objective techniques are applied in PSO: Controlling the Dominance Area of Solutions and Average Ranking. An empirical analysis is performed to identify the influence of these techniques on convergence and diversity of the MOPSO search in different manyobjective scenarios. The experimental results are analyzed applying some quality indicators and some statistical tests.