English

A Joint Bidiagonalization Based Algorithm for Large Scale Linear Discrete Ill-posed Problems in General-Form Regularization

Numerical Analysis 2020-07-21 v2

Abstract

Based on the joint bidiagonalization process of a large matrix pair {A,L}\{A,L\}, we propose and develop an iterative regularization algorithm for the large scale linear discrete ill-posed problems in general-form regularization: minLx \mboxsubjectto xS={x Axbτe}\min\|Lx\| \ \mbox{{\rm subject to}} \ x\in\mathcal{S} = \{x|\ \|Ax-b\|\leq \tau\|e\|\} with a Gaussian white noise ee and τ>1\tau>1 slightly, where LL is a regularization matrix. Our algorithm is different from the hybrid one proposed by Kilmer {\em et al.}, which is based on the same process but solves the general-form Tikhonov regularization problem: minx{Axb2+λ2Lx2}\min_x\left\{\|Ax-b\|^2+\lambda^2\|Lx\|^2\right\}. We prove that the iterates take the form of attractive filtered generalized singular value decomposition (GSVD) expansions, where the filters are given explicitly. This result and the analysis on it show that the method must have the desired semi-convergence property and get insight into the regularizing effects of the method. We use the L-curve criterion or the discrepancy principle to determine kk^*. The algorithm is simple and effective, and numerical experiments illustrate that it often computes more accurate regularized solutions than the hybrid one.

Keywords

Cite

@article{arxiv.1807.08419,
  title  = {A Joint Bidiagonalization Based Algorithm for Large Scale Linear Discrete Ill-posed Problems in General-Form Regularization},
  author = {Zhongxiao Jia and Yanfei Yang},
  journal= {arXiv preprint arXiv:1807.08419},
  year   = {2020}
}

Comments

25 pages, 4 figures

R2 v1 2026-06-23T03:10:18.089Z