Analysis of GP-designed metamaterial using equivalent circuit model and Prony’s method

This is the authors' abstract. We don't add key points for this paper.

Prior Artificial Magnetic Conductor (AMC) designs using Genetic Programming (GP) showed that GP-based designs are competitive and often exceed those based on human expertise. However, the resultant typologies are of wild shapes and beyond intuitive explanation. To gain better physical insight into GP-designed AMCs, we develop equivalent circuit models for them and use Prony’s method to reveal their deep physical characteristics. Ansys Q3D Extractor is employed to help determine the RLC component values of the equivalent circuits. The natural resonant frequencies extracted by Prony’s method provide further explanation of the bandwidth of the structures. The equivalent circuit models and related natural frequencies of the GP designs can also provide guidance for human experts to reduce the optimization time and improve the performance of the designs.