A Data Scalable Augmented Lagrangian KKT Preconditioner for Large-Scale Inverse Problems

A Data Scalable Augmented Lagrangian KKT Preconditioner for Large-Scale Inverse Problems
复制标题

用于大规模反问题的数据可扩展增强拉格朗日 KKT 预处理器

DOI:
10.1137/16m1084365
复制
发表时间:
2017
影响因子:
3.1
通讯作者:
Ghattas, Omar
Ghattas, Omar
中科院分区:
数学2区
文献类型:
--
作者:
Alger, Nick;Villa, Umberto;Bui-Thanh, Tan;Ghattas, Omar

文献摘要

参考文献

被引文献

相似文献

目前最先进的预条件为减少海森和Karush-Kuhn-Tucker(KKT)运营商的大规模反问题通常是基于近似的减少海森与正则化算子。然而,这种近似的质量随着越来越多的信息观测或数据而降低。因此,从科学的角度来看,最好的情况(完全信息化的数据)是从计算的角度来看最坏的情况。在本文中,我们提出了一个增广拉格朗日型预条件子的基础上的块对角近似的增广左上块的KKT算子。预处理器需要求解两个线性子问题,出现在增广KKT算子,我们预计要容易得多的先决条件比减少海森。对预条件KKT算子的谱分析表明,当正则化选取适当时,预条件是有效的。特别是,它是有效的,当正则化不过度惩罚高度知情的参数模式,并不欠惩罚不知情的模式。最后,我们提出了一个大数据/低噪声泊松源反演问题的数值研究,证明了预条件的有效性。在这个例子中,使用我们的预处理器在KKT系统上进行三次MINRES迭代,得到的重建精度比使用正则化预处理的简化Hessian系统上的CG迭代50次更好。
Current state-of-the-art preconditioners for the reduced Hessian and the Karush--Kuhn--Tucker (KKT) operator for large-scale inverse problems are typically based on approximating the reduced Hessian with the regularization operator. However, the quality of this approximation degrades with increasingly informative observations or data. Thus the best case scenario from a scientific standpoint (fully informative data) is the worse case scenario from a computational perspective. In this paper we present an augmented Lagrangian-type preconditioner based on a block diagonal approximation of the augmented upper left block of the KKT operator. The preconditioner requires solvers for two linear subproblems that arise in the augmented KKT operator, which we expect to be much easier to precondition than the reduced Hessian. Analysis of the spectrum of the preconditioned KKT operator indicates that the preconditioner is effective when the regularization is chosen appropriately. In particular, it is effective when the regularization does not overpenalize highly informed parameter modes and does not underpenalize uninformed modes. Finally, we present a numerical study for a large data/low noise Poisson source inversion problem, demonstrating the effectiveness of the preconditioner. In this example, three MINRES iterations on the KKT system with our preconditioner results in a reconstruction with better accuracy than 50 iterations of CG on the reduced Hessian system with regularization preconditioning.
DOI: --
发表时间: 2013
期刊:
影响因子: --
作者:
T. Bui;O. Ghattas
通讯作者: O. Ghattas
基于 Hessian 的响应面近似,用于大规模统计反问题中的不确定性量化,并应用于地下水流
DOI: --
发表时间: 2013
期刊:
影响因子: --
作者:
H. P. Flath
通讯作者: H. P. Flath
DOI: 10.1137/120877532
发表时间: 2014
期刊: SIAM J. Optim.
影响因子: --
作者:
A. Schiela;S. Ulbrich
通讯作者: S. Ulbrich
DOI: 10.1137/120871547
发表时间: 2013-03
期刊: SIAM J. Sci. Comput.
影响因子: --
作者:
B. F. Nielsen;K. Mardal
通讯作者: B. F. Nielsen;K. Mardal
DOI: --
发表时间: 2009
影响因子: 4.1
作者:
E. Arian;A. Iollo
通讯作者: A. Iollo