A Complex Image Reduction Technique Using Genetic Algorithm for the MoM Solution of Half-Space MPIE

The discrete complex image method (DCIM) has been widely used to analytically approximate the Sommerfeld integrals (SIs) of Green's functions in layered media. This method has been greatly improved in accuracy and generality in the past years, but has resulted in a large amount of complex images. It is found that using so many complex images for SI approximation in method of moment (MoM) applications is redundant and inefficient. To reduce the number of complex images, a technique by means of genetic algorithm is proposed and applied to the MoM solution of the half-space mixed-potential integral equation (MPIE).