Chung-Hsin Huang

Comparative Study of Some Population-Based Optimization Algorithms on Inverse Scattering of a Two-Dimensional Perfectly Conducting Cylinder in Dielectric Slab Medium

The application of four techniques for the shape reconstruction of a 2-D metallic cylinder buried in dielectric slab medium by measured the scattered fields outside is studied in the paper. The finite-difference time-domain (FDTD) technique is employed for electromagnetic analyses for both the forward and inverse scattering problems, while the shape reconstruction problem is transformed into optimization one during the course of inverse scattering. Then, four techniques including asynchronous particle swarm optimization (APSO), PSO, dynamic differential evolution (DDE) and self-adaptive DDE (SADDE) are applied to reconstruct the location and shape of the 2-D metallic cylinder for comparative purposes. The statistical performances of these algorithms are compared. The results show that SADDE outperforms PSO, APSO and DDE in terms of the ability of exploring the optima. However, these results are considered to be indicative and do not generally apply to all optimization problems in electromagnetics.

Time Domain Image Reconstruction for a Buried 2D Homogeneous Dielectric Cylinder Using NU-SSGA

This article presents an image reconstruction approach for a buried homogeneous cylinder with arbitrary cross-section in half space. The computational method combines the finite difference time domain (FDTD) method and non-uniform steady state genetic algorithm (NU-SSGA) to determine the shape and location of the subsurface scatterer with arbitrary shape. The FDTD-subgirdding technique is implemented for modeling the shape of the cylinder more closely. The inverse problem is reformulated into an optimization problem and the global searching scheme NU-SSGA with closed cubic-spline is then employed to search the parameter space. A set of representative numerical results is presented for demonstrating that the proposed approach is able to efficiently reconstruct the electromagnetic properties of homogeneous dielectric scatterer even when the initial guess is far from the exact one. In addition, the effects of Gaussian noises on imaging reconstruction are also investigated.

Time Domain Inverse Scattering for a Homogenous Dielectric Cylinder by Asynchronous Particle Swarm Optimization

In this paper, we propose a time domain inverse scattering technique for reconstructing the electromagnetic properties of a homogeneous dielectric cylinder based on the finite difference time domain method and the asynchronous particle swarm optimization (APSO). The homogeneous dielectric cylinder with unknown electromagnetic properties is illuminated by transverse magnetic pulse and the scattered field is recorded outside. By minimizing the discrepancy between the measured and estimated scattered field data, the location, shape, and permittivity of the dielectric cylinder are reconstructed. The inverse problem is resolved by an optimization approach and the global searching scheme APSO is then employed to search the parameter space. A set of representative numerical results is presented for demonstrating that the proposed approach is able to efficiently reconstruct the electromagnetic properties of the homogeneous dielectric scatterer even when the initial guess is far away from the exact one. In addition, the effects of Gaussian noises on imaging reconstruction are also investigated.

Time Domain Microwave Imaging of a Buried Homogeneous Dielectric Cylinder in a Slab Medium by Asynchronous Particle Swarm Optimization

A time-domain inverse scattering technique for estimating the location, shape, and permittivity of a dielectric cylinder in a slab medium is proposed. In this paper, the finite-difference time domain is employed for the analysis of the forward scattering part, and asynchronous particle swarm optimization (APSO) is applied for the reconstruction of the two-dimensional homogeneous dielectric cylinder. For the forward scattering, several electromagnetic pulses are launched to illuminate the unknown scatterers, and then the surrounding scattered electromagnetic fields are measured. In order to efficiently describe the details of the shape, a sub-gridding technique is implemented in the finite-difference time domain method. Then, the simulated electromagnetic fields are used for inverse scattering, in which APSO is employed to transform the inverse scattering problem into an optimization problem. APSO is a population-based optimization approach that aims to minimize a cost function between measurements and computer-simulated data. The numerical results presented for the two examples of scatterers under transverse-electric incidence demonstrate that the proposed method is capable of reconstructing a complicated shape with a rapid rate of convergence and robust immunity to noise.

Image reconstruction for buried inhomogeneous dielectric cylinders by evolutionary algorithms

This paper presents the studies of time domain inverse scattering for a two dimensional inhomogeneous dielectric cylinder buried in a half-space by the finite difference time domain (FDTD) method and evolutionary algorithms (EAs). For forward scattering, the FDTD method is employed to calculate the scattered E fields, while for the inverse scattering the evolutionary algorithms are utilized to determine the permittivity of the buried cylindrical scatterer with arbitrary cross section. The results obtained for different examples show that the dynamic differential evolution (DDE) algorithms outperform the non-uniform steady state genetic algorithm (NU-SSGA) variants in terms of finding best optima. The suitability and efficiency of applying these two methods for microwave imaging of 2-D inhomogeneous dielectric cylinders are examined.

Inverse scattering for a buried 2D homogeneous dielectric cylinder using by SSGA

This paper presents an image reconstruction approach for a buried homogeneous cylinder with arbitrary cross section in half space. The computational method combines the finite difference time domain (FDTD) method and non-uniform steady state genetic algorithm (NU-SSGA) to determine the shape and location of the subsurface scatterer with arbitrary shape. The subgirdding technique is implemented for modeling the shape of the cylinder more closely. The inverse problem is reformulated into an optimization problem and the global searching scheme NU-SSGA with closed cubic-spline is then employed to search the parameter space. A set of representative numerical results is presented for demonstrating that the proposed approach is able to efficiently reconstruct the electromagnetic properties of homogeneous dielectric scatterer even when the initial guess is far away from the exact one.