Optimizations of Dielectric Slab Surfaces with Genetic Algorithms and Machine Learning

This paper presents an efficient combination of genetic algorithms and machine-learning techniques to optimize scattering properties of 3D dielectric (PLA) slabs with surface deformations. The top and bottom surfaces of slabs are optimized to obtain the desired electromagnetic properties while they are represented by Bezier functions so that final designs can be fabricated via 3D printers. The numerical results demonstrate the superiority of the developed optimization environment to reach the desired scattering characteristics.