Metasurface parameter optimization of Fano resonance based on a BP-PSO algorithm

An all-dielectric metasurface is proposed, and the transmission spectrum is analyzed by numerical simulation. The Fano resonance line appears in the transmission spectrum. The mechanism of Fano resonance is analyzed based on multipole coupling theory. The mathematical model between structural parameters and spectral performance is established by the back propagation (BP) neural network. Then, the genetic algorithm, sparrow search algorithm, and particle swarm optimization (PSO) algorithms are used to find the structural parameters corresponding to the optimal performance. The result shows that the quality factor is increased by three times, reaching 3805, and the modulation depth is close to 100% after PSO optimization. Our study provides a new method for the design of metasurfaces and parameter optimization of optical micro-nano structures.