A computational technique based on a real-coded genetic algorithm for microwave imaging purposes

TLDR

The study proposes a genetic‑algorithm–based computational method to solve a nonlinear inverse scattering problem in short‑range microwave imaging, aiming to recover the locations, shapes, and dielectric distributions of cylindrical scatterers and introduces a hybrid GA‑CG variant for preliminary testing. The method discretizes the integral‑equation formulation into a nonlinear optimization problem and applies a real‑coded genetic algorithm to minimize a chosen functional. The algorithm reconstructs strong scatterers with sub‑Rayleigh resolution, yields internal electric field distributions, and outperforms approximated formulations and binary‑coded GA, with a hybrid GA‑CG variant showing promising preliminary results.

Abstract

A computational approach based on a genetic algorithm is proposed for the solution of a nonlinear inverse scattering problem for short-range microwave imaging applications. Starting from an integral-equation formulation, the aim is to derive locations, shapes, and distributions of the dielectric parameters of cylindrical scatterers. Simultaneously, the approach also provides the distributions of the internal total electric field. After discretization, the problem is recast as a nonlinear optimization problem. The paper exploits the application of a real-coded genetic algorithm in order to minimize a suitable functional. The reconstruction of strong scatterers with a resolution beyond the Rayleigh criterion is shown, and computational aspects are discussed. Comparisons with results obtained by using approximated formulations and a binary-coded genetic algorithm are also provided. Finally, a hybrid version of the approach (based on the combined strategy of the genetic algorithm and a conjugate gradient method) is presented and preliminarily tested.