Diseño optimizado de antenas mediante algoritmos genéticos

Resumo

This dissertation presents a study of the application of the state of the art of genetic algorithm based optimizers in the design of electromagnetic devices, in particular antenna arrays and ultra broad band wire antennas. First, the canonical genetic algorithm is reviewed revisiting some of the most common genetic operators and optimization strategies. Later on the Hybrid Taguchi Genetic Algorithm is explained and systematically analyzed. In this chapter a thorough analysis of the performance of this algorithm is carried out under variations of the parameters which define its dynamic behaviour. Some modifications to improve its performance are proposed and tested. Following to the presentation and analysis of the optimization algorithms used in this work, a new version of the multiobjective non dominated sorting genetic algorithm is presented, discussed and analyzed as well. As an application of the proposed algorithms, the mono and multiobjective optimization of antenna arrays and thin-wire antennas are presented. In particular, a study of the applicability of the Hybrid Taguchi Genetic Algorithm on the array synthesis using a extension of the Schelkunoff method is carried out. Later on a multiobjective antenna array design, based on a low-cost radiating element, is presented and discussed. Finally, thin-wire antenna designs are optimized based on bow-tie and log-periodic geometries.