A novel approach for pattern synthesis of arbitrary array

Abstract

In this paper, a novel approach, based on the genetic algorithm, for pattern synthesis of an arbitrary array is presented. Unlike conventional genetic algorithms dealing with binary coding and genetic operations, this algorithm avoids coding and uses decimal operation, directly dealing with real or complex coefficients. Furthermore, constraints on the phases of the complex coefficients can be imposed for practical implementation of digital phase shifters and digital attenuators. To demonstrate the effectiveness of the approach, two examples are presented, including a curved array with mutual coupling considered.