# $O(N)$ Iterative and $O(NlogN)$ Fast Direct Volume Integral Equation Solvers with a Minimal-Rank ${\cal H}^2$-Representation for Large-Scale $3$-D Electrodynamic Analysis

## TL;DR

This paper introduces linear and log-linear complexity volume integral equation solvers for large-scale 3D electrodynamics, utilizing a minimal-rank ${\
m H}^2$-matrix representation to achieve fast computation and storage.

## Contribution

It develops a novel cluster-based multilevel low-rank representation and minimal-rank ${\cal H}^2$-matrix algorithms for efficient large-scale electrodynamic analysis.

## Key findings

- Achieves $O(N)$ iterative solver complexity.
- Develops $O(N\log N)$ direct solver for large-scale problems.
- Demonstrates superior performance on large 3D structures.

## Abstract

Linear complexity iterative and log-linear complexity direct solvers are developed for the volume integral equation (VIE) based general large-scale electrodynamic analysis. The dense VIE system matrix is first represented by a new cluster-based multilevel low-rank representation. In this representation, all the admissible blocks associated with a single cluster are grouped together and represented by a single low-rank block, whose rank is minimized based on prescribed accuracy. From such an initial representation, an efficient algorithm is developed to generate a minimal-rank ${\cal H}^2$-matrix representation. This representation facilitates faster computation, and ensures the same minimal rank's growth rate with electrical size as evaluated from singular value decomposition. Taking into account the rank's growth with electrical size, we develop linear-complexity ${\cal H}^2$-matrix-based storage and matrix-vector multiplication, and thereby an $O(N)$ iterative VIE solver regardless of electrical size. Moreover, we develop an $O(NlogN)$ matrix inversion, and hence a fast $O(NlogN)$ \emph{direct} VIE solver for large-scale electrodynamic analysis. Both theoretical analysis and numerical simulations of large-scale $1$-, $2$- and $3$-D structures on a single-core CPU, resulting in millions of unknowns, have demonstrated the low complexity and superior performance of the proposed VIE electrodynamic solvers. %The algorithms developed in this work are kernel-independent, and hence applicable to other IE operators as well.