Novel optimizing algorithm of superdirective multi-layered cylindric antenna

Abstract

We propose the design of a superdirective antenna based on a multi-layered dielectric cylinder. A novel optimization goal is proposed based on the physical characteristics of the Dirac delta function. The side-lobe properties of superdirective antenna is discussed and treated as the optimization goal for antenna design for the first time. A multi-stage genetic algorithm and trust region reflective algorithm are utilized to shorten the time duration for optimization. The optimizing efficiency is higher and the cost of time is significantly reduced compared with proposed designs. The effect of weight of different algorithms is discussed. Both the dielectric loss and metallic loss are also discussed for practical cases. The optimization is of low computing cost and can be accomplished by personal computers.