Particle-swarm optimization and its application to antenna far-field-pattern prediction from planar scanning

Particle-swarm optimization (PSO) is a high-performance optimizer, recently introduced to the electromagnetics community, which seems to be an attractive alternative to other stochastic optimization methods such as genetic algorithms. An overview of the main characteristics and schemes related to the PSO technique available in the literature are presented and discussed. Furthermore, in this paper, equivalent magnetic-surface currents and PSO are used together to model the radiation of an antenna under test (A UT)from planar near-field data. The results of a comparison of the performance of both global and local PSOs with synchronous and asynchronous updates, as well as the results of reconstructed far-field patterns from near-field samples, art, included in order to demonstrate the usefulness of the method. (C) 2005 Wiley Periodicals, Inc.