Symbolic regression: an alternative method to model the optical response of photonic biological and bio-inspired structures

Abstract

In this contribution, we assess the performance of symbolic regression (SR) when used to model the optical response of biological and bio-inspired structures. To this end, we search for analytical closed-form expressions that model the reflectance spectra related to the Tersina viridis bird’s plumage and the porous structure inspired on the Cyphochilus insulanus beetle. Our numerical results demonstratethe high prediction accuracy of the employed symbolic regression scheme. The retrieved models not only capture the dependency of the optical response on various relevant geometrical and illumination parameters, but they are also dimensionally homogeneous.