Evolutionary optimization as applied to inverse scattering problems

This review is aimed at presenting an overview of evolutionary algorithms (EAs) as applied to the solution of inverse scattering problems. The focus of this work is on the use of different population-based optimization algorithms for the reconstruction of unknown objects embedded in an inaccessible region when illuminated by a set of microwaves. Starting from a general description of the structure of EAs, the classical stochastic operators responsible for the evolution process are described. The extension to hybrid implementations when integrated with local search techniques and the exploitation of the 'domain knowledge', either a priori obtained or collected during the optimization process, are also presented. Some theoretical discussions concerned with the convergence issues and a sensitivity analysis on the parameters influencing the stochastic process are reported as well. Successively, a review on how various researchers have applied or customized different evolutionary approaches to inverse scattering problems is carried out ranging from the shape reconstruction of perfectly conducting objects to the detection of the dielectric properties of unknown scatterers up to applications to sub-surface or biomedical imaging. Finally, open problems and envisaged developments are discussed.