Parallel adaptive cross approximation for accelerating time-domain method of moments

In this paper, the parallel Adaptive Cross Approximation (ACA) algorithm is introduced to reduce the memory and CPU time requirements for solving the Time-Domain Integral Equation (TDIE). Marching-On-in-Degree (MOD) is selected to be the equation solver so as to avoid the late-time oscillation in Marching-On-in-Time (MOT); ACA is a purely algebraic algorithm whose formulation and implementation are independent of the integral kernel (Green's function) so that it can be easily applied to TDIE to reduce the memory requirement. Meanwhile, parallel technique based on partitioning the octree's group is employed either to accelerate the computation. Through some numerical examples, the proposed method is proven feasible and efficient.