Efficient and Accurate Electromagnetic Angular Sweeping of Rough Surfaces by MPI Parallel Randomized Low-Rank Decomposition

Abstract

A message passing interface (MPI) parallel scheme on distributed memory platforms is developed for the low-rank decomposition to accommodate the memory requirement during angular sweeping of a rough surface in terms of tapered wave incidence. Numerical examples, including that conducted on a 160 × 160 square wavelength rough surface, are carried out to demonstrate the performance of the proposed MPI angular sweeping with respect to accuracy, efficiency, scalability, and the peak memory requirement. Index Terms-Distributed memory parallel platform, message passing interface (MPI) parallelization, randomized low-rank decomposition, rank deficiency. I. INTRODUCTION T HE electromagnetic simulation of scattering from a random rough surface plays a fundamental role in many areas, such as remote sensing, target recognition, and radar surveillance [1]-[15]. The so-called analytical methods, such as the Kirchhoff approximation and the small perturbation method [16], are effective for simulating scattering from targets with rough surfaces. However, their accuracy is sometimes uncontrollable, and full-wave simulation may be required to calibrate them. In full-wave simulations, the randomness can be treated by the Monte Carlo (MC) method [13], [17], the polynomial chaos expansion method [18]-[20], or other approaches [21]. Among them, the MC method performs well when the surface variation is large. The electromagnetic wave scattering by a large rough surface may be highly dependent on the direction of the incident wave.