A Strategy to Solve MaOPs with Multiple Swarms Based on Broadcast Communication

The application of multi-objective evolutionary algorithms to solve many-objective optimization problems face several problems, most derived from the fact that there is no single best solution, but a set of solutions. To obtain this set of solutions, the Pareto Optimality Theory is often used. It is difficult for an algorithm to converge to the Pareto Front and at the same time to guarantee the diversity of the obtained solutions. A practical approach to deal with these issues is to employ multi-swarm strategies. Multi-swarm strategies have already been proven to be good approaches to solve mono-objectives problems and in this work we propose a multi-swarm strategy to solve many-objective problems. We designed a PSO strategy based on independent swarms that interact by means of particle migration policies implemented with asynchronous broadcast communication. We empirically evaluated the performance of the proposed strategy, in particular the convergence and diversity of the obtained solutions, as well as the scalability with respect to the number of objectives. To the best of our knowledge this is the first work that evaluates the application of PSO based on multiple swarms to solve problems with a large number of objectives. Performance metrics were employed as quality indicators and results show that the multi-swarm execution does bring advantages to the convergence and diversity.