算子傅里叶变换的低秩加速
Low-Rank Acceleration of the Operator Fourier Transform
AI总结:
针对二维结构化网格上的亥姆霍兹方程,结合算子傅里叶变换与低秩交叉逼近(Cross-DEIM),将问题分解为薛定谔方程解的伪时间积分,显著降低计算成本。
摘要:
我们开发了一种数值算法,用于高效求解或近似求解二维结构化网格上的亥姆霍兹方程。利用算子傅里叶变换(OFT)和低秩交叉逼近方案(Cross-DEIM),将问题分解为薛定谔方程解在伪时间上的积分。OFT是求解算子方程(如分数阶拉普拉斯方程或亥姆霍兹方程,当后者写为两个近轴算子的乘积时)的框架。OFT的主要计算成本在于求解薛定谔方程,尤其是在维度或网格分辨率较高时。在本工作中,我们通过使用低秩方法降低了这一成本。当解中存在低秩结构时,这类方法旨在克服维数灾难。我们表明,对于某些类型的问题,这两种方法的结合可以大幅降低计算成本。
英文摘要:
We develop a numerical algorithm for the efficient solution or approximation of solutions to the Helmholtz equation on a structured grid in two dimensions. We make use of the Operator Fourier Transform (OFT) and a low-rank cross approximation scheme (Cross-DEIM) to decompose the problem into an integral over a pseudo-time of solutions to the Schrödinger equation. The OFT is a framework for solving operator equations like fractional Laplacian equations or the Helmholtz equation, when the latter is written as a product of two paraxial operators. The main computational cost in the OFT is the solution to the Schrödinger equation, especially when the dimension or mesh resolution is high. In this work, we alleviate this cost by utilizing a low-rank method. Such methods aim to beat the curse of dimensionality when low-rank structures are present in the solution. For a problem with a constant coefficient we show that the resulting solver converges at second order in the pseudo-time step and reaches the accuracy of the direct solution of the discretized linear system, which is limited by the second-order finite-difference discretization in space, while keeping the solution in low-rank form. The ranks of the accumulated solution change little as the grid is refined, which suggests a cost per time step that grows nearly linearly with the number of grid points in each direction.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。