An Efficient Surrogate Assisted Particle Swarm Optimization for Antenna Synthesis

By virtue of the prediction abilities of machine learning (ML) methods, the ML-assisted evolutionary algorithm has been treated as an efficient solution for antenna design automation. This article presents an efficient ML-based surrogate-assisted particle swarm optimization (SAPSO). The proposed algorithm closely combines the particle swarm optimization (PSO) with two ML-based approximation models. Then, a novel mixed prescreening (mixP) strategy is proposed to pick out promising individuals for full-wave electromagnetic (EM) simulations. As the optimization procedure progresses, the ML models are dynamically updated once new training data are obtained. Finally, the proposed algorithm is verified by three real-world antenna examples. The results show that the proposed SAPSO-mixP can find favorable results with a much smaller number of EM simulations than other methods.

I. Introduction

Designing antennas by experience-based rules of thumb and parameter sweeping methods are very time-consuming. What is more, it is possible that failing to find a satisfactory solution even after many trials and errors [1]. Antenna synthesis, with the aid of several methods, such as finite-difference time-domain (FDTD), space mapping, and evolutionary algorithms (EAs), has attracted intense interest in the past decades [2]–[4]. Herein, the EAs, which include genetic algorithm (GA), particle swarm optimization (PSO), differential evolution (DE), and others, have been regarded as a promising optimization framework for finding high-quality solutions in electromagnetics (EMs) problems [4]. However, caused by the inherent mechanism of a population-based iterative algorithm, the computational budget of EA-based antenna synthesis is remarkably high. Because this method requires dozens of EM simulations at each iteration, and hundreds or even thousands of iterations are necessary for obtaining a near-optimal solution. Therefore, it is challenging to design an EA-based algorithm for antenna synthesis when considering the computational budgets.