Abstract:In this paper, we present a numerical algorithm for the accurate and efficient computation of the convolution of the frequency domain layered media Green's function with a given density function. Instead of compressing the convolution matrix directly as in the classical fast multipole method, fast direct solvers, and fast H-matrix algorithms, the new algorithm considers a translated form of the original matrix so that many existing building blocks from the highly optimized free-space fast multipole method can be easily adapted to the Sommerfeld integral representations of the layered media Green's function. An asymptotic analysis is performed on the Sommerfeld integrals for large orders to provide an estimate of the decay rate in the new "multipole" and "local" expansions. In order to avoid the highly oscillatory integrand in the original Sommerfeld integral representations when the source and target are close to each other, or when they are both close to the interface in the scattered field, mathematically equivalent alternative direction integral representations are introduced. The convergence of the multipole and local expansions and formulas and quadrature rules for the original and alternative direction integral representations are numerically validated.
What problem does this paper attempt to address?
This paper attempts to address two major challenges in efficient numerical simulation of wave propagation in layered media:
1. **Fast application of the Green's function in layered media**: Develop an optimal fast algorithm to apply the Green's function in layered media to a given density function. Traditional fast multipole method (FMM), fast direct solvers, and fast H - matrix algorithms usually achieve accelerated computation by directly compressing the coefficient matrix. However, the new algorithm proposed in this paper transforms the original matrix, enabling the use of existing well - optimized free - space FMM tools, thereby achieving more efficient computation.
2. **Handling the case where source and target points are close**: An effective numerical scheme for calculating the Green's function in layered media when the source and target points are close to each other or close to the media interface. In this case, the original Sommerfeld integral representation will lead to a highly oscillatory integrand, making it difficult to solve accurately. To this end, the article introduces different integration paths to ensure that the interaction can also be effectively evaluated in such special cases.
### Specific problem description
- **Challenge 1**: Traditional methods such as classical FMM, fast direct solvers, and H - matrix algorithms directly compress the coefficient matrix, but are not very efficient when dealing with the Green's function in layered media. The new algorithm achieves more efficient computation by transforming the original matrix and using existing free - space FMM tools.
- **Challenge 2**: When the source and target points are close to the same interface, for example, both are close to the upper interface \(y = 0\) and \(|x_1 - x_0|\gg|y_1 - y_0|\), the original Sommerfeld integral representation will lead to computational difficulties. To solve this problem, the article introduces a mathematically equivalent alternative - direction integral representation, thus avoiding the highly oscillatory integrand.
### Solution overview
- **Algorithm framework**: The new algorithm transforms the original matrix so that existing free - space FMM tools can be used without significant modification. This includes operations such as multipole expansion, multipole - to - multipole transformation (M2M), multipole - to - local transformation (M2L), and local - to - local transformation (L2L).
- **Alternative - direction integral representation**: To handle the case where the source and target points are close, the article introduces different integration paths so that the interaction can be accurately evaluated in these special cases. Specifically, by changing the integration path, the highly oscillatory integrand is avoided, thereby improving the stability and accuracy of the computation.
### Summary
The main contribution of this paper is to provide a new algorithm framework that can not only efficiently calculate the Green's function in layered media, but also maintain the stability and accuracy of the computation when the source and target points are close. This provides important technical support for the numerical simulation of wave propagation in layered media.