当前位置: X-MOL 学术SIAM J. Sci. Comput. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Sparse Hierarchical Preconditioners Using Piecewise Smooth Approximations of Eigenvectors
SIAM Journal on Scientific Computing ( IF 3.0 ) Pub Date : 2020-12-14 , DOI: 10.1137/20m1315683
Bazyli Klockiewicz , Eric Darve

SIAM Journal on Scientific Computing, Volume 42, Issue 6, Page A3907-A3931, January 2020.
When solving linear systems arising from PDE discretizations, iterative methods (such as conjugate gradient (CG), GMRES, or MINRES) are often the only practical choice. To converge in a small number of iterations, however, they have to be coupled with an efficient preconditioner. One approach to preconditioning is provided by the hierarchical approximate factorization methods. However, to guarantee sufficient accuracy on the eigenvectors corresponding to the smallest eigenvalues, these methods typically have to be performed at very stringent accuracies, making the preconditioner expensive to apply. On the other hand, for a large class of problems, including many elliptic equations, the eigenvectors corresponding to small eigenvalues are smooth functions of the PDE grid. In this paper, we describe a hierarchical approximate factorization approach which focuses on improving accuracy on the smooth eigenvectors. The improved accuracy is achieved by preserving the action of the factorized matrix on piecewise polynomial functions of the grid. Based on the factorization, we propose a family of sparse preconditioners with $\mathcal{O}({n})$ or $\mathcal{O}({n \log{n}})$ construction complexities. Our methods exhibit rapid convergence of CG in benchmarks run on large elliptic problems, arising for example in flow or mechanical simulations. In the case of the linear elasticity equation the preconditioners are exact on the near-kernel rigid body modes.


中文翻译:

使用特征向量的分段光滑逼近的稀疏分层预处理器

SIAM科学计算杂志,第42卷,第6期,第A3907-A3931页,2020年1月。
在求解由PDE离散化产生的线性系统时,迭代方法(例如共轭梯度(CG),GMRES或MINRES)通常是唯一的实际选择。但是,要收敛于少量的迭代,必须将它们与有效的预处理器结合在一起。分层近似因子分解方法提供了一种预处理的方法。但是,为了保证对应于最小特征值的特征向量具有足够的精度,通常必须非常严格地执行这些方法,这使得预处理器的使用成本很高。另一方面,对于一大类问题,包括许多椭圆方程,对应于小特征值的特征向量是PDE网格的光滑函数。在本文中,我们描述了一种分层的近似因式分解方法,该方法着重于提高平滑特征向量的准确性。通过保留分解矩阵对网格的分段多项式函数的作用,可以提高精度。基于分解,我们提出了一个构造复杂度为$ \ mathcal {O}({n})$或$ \ mathcal {O}({n \ log {n}})$的稀疏预处理器族。我们的方法显示出在大型椭圆问题上运行的基准中CG的快速收敛性,例如在流动或机械仿真中。在线性弹性方程的情况下,预处理器在近核刚体模式下是精确的。通过保留分解矩阵对网格的分段多项式函数的作用,可以提高精度。基于分解,我们提出了一个构造复杂度为$ \ mathcal {O}({n})$或$ \ mathcal {O}({n \ log {n}})$的稀疏预处理器族。我们的方法显示出在大型椭圆问题上运行的基准中CG的快速收敛,例如在流动或机械模拟中出现的情况。在线性弹性方程的情况下,预处理器在近核刚体模式下是精确的。通过保留分解矩阵对网格的分段多项式函数的作用,可以提高精度。基于分解,我们提出了一个构造复杂度为$ \ mathcal {O}({n})$或$ \ mathcal {O}({n \ log {n}})$的稀疏预处理器族。我们的方法显示出在大型椭圆问题上运行的基准中CG的快速收敛,例如在流动或机械模拟中出现的情况。在线性弹性方程的情况下,预处理器在近核刚体模式下是精确的。
更新日期:2020-12-15
down
wechat
bug