A Modified Particle Swarm Optimizer Applied to Linear Array Synthesis

In this work, classical PSO schemes are modified by introducing a selection operator typically used in evolutionary algorithms, the tournament selection strategy, and the novel hybrid algorithms are applied to the complex synthesis of linear arrays with directive element patterns, in order to weigh up the improvements achieved. Several classical PSO schemes are addressed by the PSO suite tested in this work, including PSO with synchronous and asynchronous updates of the swarm along with global and local topologies for both, real-valued and binary encoding algorithms. Moreover, results comparing both, classical PSO schemes and the new hybrid approaches are reported and discussed, demonstrating that the selection operator clearly increases search pressure over the swarm, speeding up convergence for certain specific classical PSO schemes, those in which the selection pressure do not lead them to a premature convergence, making diversity vanish.