Evolutionary Algorithm and Surrogate-Assisted Optimisation of a Pattern Reconfigurable Antenna for Sub-6 GHz Applications

This paper presents the optimisation of the states of PIN diodes in a design of a pattern reconfigurable antenna operating at a centre frequency of 5 GHz. The proposed design comprises a patch antenna and a parasitic surface placed above it. The parasitic surface is a 4x4 metallic array connected by PIN diodes. Two optimisation techniques were used to find a set of PIN configurations that yielded satisfactory performance. The first technique involved an evolutionary algorithm, namely, Improving the Strength Pareto Evolutionary Algorithm (SPEA2). The second technique used a surrogate-assisted approach. Through simulation results, it was shown that SPEA2 outperforms the surrogate-assisted approach in terms of convergence speed. Furthermore, the proposed design achieved radiation efficiency up to 73% and broadside beam steering capabilities.