Social Network Optimization Based Procedure for Beam-Scanning Reflectarray Antenna Design

Alessandro Niccolai,Michele Beccaria,Riccardo E Zich,Paola Pirinoli,Andrea Massaccesi

doi:10.1109/ojap.2020.3022935

Abstract

Evolutionary algorithms can be successfully exploited for carrying on an effective design of beam-scanning passive reflectarrays, even if the problem is highly non-linear and multimodal. In this article, the Social Network Optimization (SNO) algorithm has been used for assessing an effective design procedure of a beam-scanning passive reflectarray (RA). For exploiting at most the optimization capabilities of SNO, the entire optimization environment has been deeply analyzed in all its parts. The performance of SNO and the beam-scanning capabilities of the optimized RA have been assessed through the comparison with other well established Evolutionary Algorithms.

Highlights

I N THE past years, Evolutionary Algorithms (EAs) have been successfully applied to antenna problems, thanks to their capability to find optimal solutions in nonlinear and multimodal problems [1], [2]
To improve the algorithms performance, especially related to their convergence capability, some hybrid approaches were proposed: in [7] a technique obtained by hybridizing Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) is presented, while in [8] and [9] the PSO alone or in conjunction with the GA is further hybridized with the Taguchi method
In [10], the genetic algorithm is combined with a local optimization approach, and in [11], [12] the convex programming is used to increase the performance of the single objective or the multi-objective PSO, respectively

Summary

Introduction

I N THE past years, Evolutionary Algorithms (EAs) have been successfully applied to antenna problems, thanks to their capability to find optimal solutions in nonlinear and multimodal problems [1], [2]. Among the various EAs, the most used, in particular for the array pattern synthesis, are the Genetic Algorithm (GA) [3], [4], the Differential Evolution (DE) [5] and Particle Swarm Optimization (PSO) [6]. The high efficiency of the EAs for array pattern problems has been exploited to design reflectarray configurations. Reflectarray antennas (RAs) have established themselves as powerful and efficient high-gain antennas, thanks to their numerous advantages, including the low profile, low cost, good radiation performance and ease of manufacturing [13], [14]. If more complex radiation patterns, as shaped or contoured beams, are required or if the focus is the steering of the pointing direction, alternative solutions must be exploited, mostly based on the use of an optimization algorithm [14]

Methods

Results

Conclusion