Solving Inverse Scattering Problems through BA-based Optimization

An efficient reconstruction scheme for dielectric targets is proposed based on a kind of stochastic optimization method - Bat Algorithm (BA). Traditional inversion methods, e.g., Born Iterative Method and Contrast Source Inversion Method, are deterministic. They are completely dependent on gradient information and could get stuck in local minimum in complex imaging problems where a priori information is insufficient. BA is based on swarm intelligence. It was inspired by the echolocation behavior of microbats. The process of searching for optimized result is similar to bats' flying with changing velocity and location. Some random factors are designed to ensure its global searching ability. The numerical example for reconstructing a two-dimensional (2-D) dielectric object is presented to illustrate the proposed BA-based inversion approach.