Rational approximation to trigonometric operators

We consider the approximation of trigonometric operator functions that arise in the numerical solution of wave equations by trigonometric integrators. It is well known that Krylov subspace methods for matrix functions without exponential decay show superlinear convergence behavior if the number of steps is larger than the norm of the operator. Thus, Krylov approximations may fail to converge for unbounded operators. In this paper, we propose and analyze a rational Krylov subspace method which converges not only for finite element or finite difference approximations to differential operators but even for abstract, unbounded operators. In contrast to standard Krylov methods, the convergence will be independent of the norm of the operator and thus of its spatial discretization. We will discuss efficient implementations for finite element discretizations and illustrate our analysis with numerical experiments. AMS subject classification (2000)  65F10, 65L60, 65M60, 65N22     

Matrix Krylov subspace methods for linear systems with multiple right-hand sides

In this paper, we first give a result which links any global Krylov method for solving linear systems with several right-hand sides to the corresponding classical Krylov method. Then, we propose a general framework for matrix Krylov subspace methods for linear systems with multiple right-hand sides. Our approach use global projection techniques, it is based on the Global Generalized Hessenberg Process (GGHP) – which use the Frobenius scalar product and construct a basis of a matrix Krylov subspace – and on the use of a Galerkin or a minimizing norm condition. To accelerate the convergence of global methods, we will introduce weighted global methods. In these methods, the GGHP uses a different scalar product at each restart. Experimental results are presented to show the good performances of the weighted global methods. AMS subject classification 65F10     

A new zero-finder for Tikhonov regularization

Tikhonov regularization with the regularization parameter determined by the discrepancy principle requires the computation of a zero of a rational function. We describe a cubically convergent zero-finder for this purpose. AMS subject classification (2000)  65F22, 65H05, 65R32     

On Lagrange multipliers of trust-region subproblems

Trust-region methods are globally convergent techniques widely used, for example, in connection with the Newton’s method for unconstrained optimization. One of the most commonly-used iterative approaches for solving trust-region subproblems is the Steihaug–Toint method which is based on conjugate gradient iterations and seeks a solution on Krylov subspaces. This paper contains new theoretical results concerning properties of Lagrange multipliers obtained on these subspaces. AMS subject classification (2000)  65F20     

Inexact GMRES for singular linear systems

Inexact Krylov subspace methods have been shown to be practical alternatives for the solution of certain linear systems of equations. In this paper, the solution of singular systems with inexact matrix-vector products is explored. Criteria are developed to prescribe how inexact the matrix-vector products can be, so that the computed residual remains close to the true residual, thus making the inexact method of practical applicability. Cases are identified for which the methods work well, and this is the case in particular for systems representing certain Markov chains. Numerical experiments illustrate the effectiveness of the inexact approach. AMS subject classification (2000)  65F10     

Analysis of the convergence of the minimal and the orthogonal residual methods

We consider two Krylov subspace methods for solving linear systems, which are the minimal residual method and the orthogonal residual method. These two methods are studied without referring to any particular implementations. By using the Petrov–Galerkin condition, we describe the residual norms of these two methods in terms of Krylov vectors, and the relationship between there two norms. We define the Ritz singular values, and prove that the convergence of these two methods is governed by the convergence of the Ritz singular values. AMS subject classification 65F10     

Stopping criteria for iterations in finite element methods

Summary. This work extends the results of Arioli [1], [2] on stopping criteria for iterative solution methods for linear finite element problems to the case of nonsymmetric positive-definite problems. We show that the residual measured in the norm induced by the symmetric part of the inverse of the system matrix is relevant to convergence in a finite element context. We then use Krylov solvers to provide alternative ways of calculating or estimating this quantity and present numerical experiments which validate our criteria.Mathematics Subject Classification (2000): 65N30, 65F10, 65F35     

Performance of ILU factorization preconditioners based on multisplittings

Summary. In this paper, we study the convergence of multisplitting methods associated with a multisplitting which is obtained from the ILU factorizations of a general H-matrix, and then we propose parallelizable ILU factorization preconditioners based on multisplittings for a block-tridiagonal H-matrix. We also describe a parallelization of preconditioned Krylov subspace methods with the ILU preconditioners based on multisplittings on distributed memory computers such as the Cray T3E. Lastly, parallel performance results of the preconditioned BiCGSTAB are provided to evaluate the efficiency of the ILU preconditioners based on multisplittings on the Cray T3E. Mathematics Subject Classification (2000):65F10, 65Y05, 65F50This work was supported by Korea Research Foundation Grant (KRF-2001-015-DP0051)     

On the choice of solution subspace for nonstationary iterated Tikhonov regularization

Tikhonov regularization is a popular method for the solution of linear discrete ill-posed problems with error-contaminated data. Nonstationary iterated Tikhonov regularization is known to be able to determine approximate solutions of higher quality than standard Tikhonov regularization. We investigate the choice of solution subspace in iterative methods for nonstationary iterated Tikhonov regularization of large-scale problems. Generalized Krylov subspaces are compared with Krylov subspaces that are generated by Golub–Kahan bidiagonalization and the Arnoldi process. Numerical examples illustrate the effectiveness of the methods.     

Matrix-free preconditioning using partial matrix estimation

We consider matrix-free solver environments where information about the underlying matrix is available only through matrix vector computations which do not have access to a fully assembled matrix. We introduce the notion of partial matrix estimation for constructing good algebraic preconditioners used in Krylov iterative methods in such matrix-free environments, and formulate three new graph coloring problems for partial matrix estimation. Numerical experiments utilizing one of these formulations demonstrate the viability of this approach. AMS subject classification (2000) 65F10, 65F50, 49M37, 90C06     

Hybrid enriched bidiagonalization for discrete ill‐posed problems

The regularizing properties of the Golub–Kahan bidiagonalization algorithm are powerful when the associated Krylov subspace captures the dominating components of the solution. In some applications the regularized solution can be further improved by enrichment, that is, by augmenting the Krylov subspace with a low‐dimensional subspace that represents specific prior information. Inspired by earlier work on GMRES, we demonstrate how to carry these ideas over to the bidiagonalization algorithm, and we describe how to incorporate Tikhonov regularization. This leads to a hybrid iterative method where the choice of regularization parameter in each iteration also provides a stopping rule.     

Relaxation strategies for nested Krylov methods

This paper studies computational aspects of Krylov methods for solving linear systems where the matrix–vector products dominate the cost of the solution process because they have to be computed via an expensive approximation procedure. In recent years, so-called relaxation strategies for tuning the precision of the matrix–vector multiplications in Krylov methods have proved to be effective for a range of problems. In this paper, we will argue that the gain obtained from such strategies is often limited. Another important strategy for reducing the work in the matrix–vector products is preconditioning the Krylov method by another iterative Krylov method. Flexible Krylov methods are Krylov methods designed for this situation. We combine these two approaches for reducing the work in the matrix–vector products. Specifically, we present strategies for choosing the precision of the matrix–vector products in several flexible Krylov methods as well as for choosing the accuracy of the variable preconditioner such that the overall method is as efficient as possible. We will illustrate this computational scheme with a Schur-complement system that arises in the modeling of global ocean circulation.     

Some results on the regularization of LSQR for large-scale discrete ill-posed problems

LSQR, a Lanczos bidiagonalization based Krylov subspace iterative method, and its mathematically equivalent conjugate gradient for least squares problems (CGLS) applied to normal equations system, are commonly used for large-scale discrete ill-posed problems. It is well known that LSQR and CGLS have regularizing effects, where the number of iterations plays the role of the regularization parameter. However, it has long been unknown whether the regularizing effects are good enough to find best possible regularized solutions. Here a best possible regularized solution means that it is at least as accurate as the best regularized solution obtained by the truncated singular value decomposition (TSVD) method. We establish bounds for the distance between the k-dimensional Krylov subspace and the k-dimensional dominant right singular space. They show that the Krylov subspace captures the dominant right singular space better for severely and moderately ill-posed problems than for mildly ill-posed problems. Our general conclusions are that LSQR has better regularizing effects for the first two kinds of problems than for the third kind, and a hybrid LSQR with additional regularization is generally needed for mildly ill-posed problems. Exploiting the established bounds, we derive an estimate for the accuracy of the rank k approximation generated by Lanczos bidiagonalization. Numerical experiments illustrate that the regularizing effects of LSQR are good enough to compute best possible regularized solutions for severely and moderately ill-posed problems, stronger than our theory predicts, but they are not for mildly ill-posed problems and additional regularization is needed.     

Besov regularization,thresholding and wavelets for smoothing data

Smoothing of data is a problem very important for many applications ranging from 1-D signals (e.g., speech) to 2-D and 3-D signals (e.g., images). Many methods exist in the literature for facing the problem; in the present paper we point our attention on regularization. We shall treat regularization methods in a general framework which is well suited for wavelet analysis; in particular we shall investigate on the relation existing between thresholding methods and regularization. We shall also introduce a new regularization method (Besov regularization), which includes some known regularization and thresholding methods as particular cases. Numerical experiments based on some test problems will be performed in order to compare the performance of some methods of smoothing data.

AMS (MOS) Subject Classifications: 65R30, 41A60.     

Deflated and Augmented Krylov Subspace Techniques

We present a general framework for a number of techniques based on projection methods on ‘augmented Krylov subspaces’. These methods include the deflated GMRES algorithm, an inner–outer FGMRES iteration algorithm, and the class of block Krylov methods. Augmented Krylov subspace methods often show a significant improvement in convergence rate when compared with their standard counterparts using the subspaces of the same dimension. The methods can all be implemented with a variant of the FGMRES algorithm. © 1997 by John Wiley & Sons, Ltd.     

Alternatives for parallel Krylov subspace basis computation

Numerical methods related to Krylov subspaces are widely used in large sparse numerical linear algebra. Vectors in these subspaces are manipulated via their representation onto orthonormal bases. Nowadays, on serial computers, the method of Arnoldi is considered as a reliable technique for constructing such bases. However, although easily parallelizable, this technique is not as scalable as expected for communications. In this work we examine alternative methods aimed at overcoming this drawback. Since they retrieve upon completion the same information as Arnoldi's algorithm does, they enable us to design a wide family of stable and scalable Krylov approximation methods for various parallel environments. We present timing results obtained from their implementation on two distributed-memory multiprocessor supercomputers: the Intel Paragon and the IBM Scalable POWERparallel SP2. © 1997 John Wiley & Sons, Ltd.     

Large-scale Tikhonov regularization via reduction by orthogonal projection

This paper presents a new approach to computing an approximate solution of Tikhonov-regularized large-scale ill-posed least-squares problems with a general regularization matrix. The iterative method applies a sequence of projections onto generalized Krylov subspaces. A suitable value of the regularization parameter is determined by the discrepancy principle.     

Multi-parameter regularization techniques for ill-conditioned linear systems

Summary.  When a system of linear equations is ill-conditioned, regularization techniques provide a quite useful tool for trying to overcome the numerical inherent difficulties: the ill-conditioned system is replaced by another one whose solution depends on a regularization term formed by a scalar and a matrix which are to be chosen. In this paper, we consider the case of several regularizations terms added simultaneously, thus overcoming the problem of the best choice of the regularization matrix. The error of this procedure is analyzed and numerical results prove its efficiency. Received January 15, 2002 / Revised version received July 31, 2002 / Published online October 29, 2002 Mathematics Subject Classification (1991): 65F05 – 65F22     

Martingale solutions and Markov selections for stochastic partial differential equations

We present a general framework for solving stochastic porous medium equations and stochastic Navier–Stokes equations in the sense of martingale solutions. Following Krylov [N.V. Krylov, The selection of a Markov process from a Markov system of processes, and the construction of quasidiffusion processes, Izv. Akad. Nauk SSSR Ser. Mat. 37 (1973) 691–708] and Flandoli–Romito [F. Flandoli, N. Romito, Markov selections for the 3D stochastic Navier–Stokes equations, Probab. Theory Related Fields 140 (2008) 407–458], we also study the existence of Markov selections for stochastic evolution equations in the absence of uniqueness.     

Vibro-acoustic Investigations Using Finite and Infinite Elements

The solution of large systems of equations often involves the use of iterative solution procedures, mostly by means of Krylov subspace methods. The performance of these methods, however, strongly depends on the spectrum of the resulting system matrices, which is affected by the polynomial approximation function space. The current work shows that finite elements based on Bernstein polynomials yield particularly good performance in combination with commonly employed Krylov solvers. Adapting the basis of Astley-Leis infinite elements, the Bernstein polynomials may also be used for exterior acoustic simulations. The efficiency of such elements for interior acoustic problems as well as for sound radiation analyses is assessed. (© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)