Sparse Systems Of Linear Equations Research Articles

Computing the positive solutions of the discrete third-order three-point right focal boundary-value problems

The existence of positive solutions to the discrete third-order three-point boundary-value problems (BVPs) was recently established in Ji and Yang [Positive solutions of discrete third-order three-point right focal boundary value problems, J. Differ. Equat. Appl. 15 (2009), pp. 185–195]. In this paper, we propose an algorithm for the computation of such positive solutions. The method is based on the power method for the dominant eigenvalue and the Crout-like factorization algorithm for the sparse system of linear equations. At each iteration of the method, it calls for a linear solver with linear computational complexity. The proposed method is extremely effective for large-scale problems. A numerical example is also included to demonstrate the effectiveness of the algorithm when applied to the third-order three-point BVPs of differential equation.

FIRE4, LiteRed and accompanying tools to solve integration by parts relations

We present the new version of the Mathematica code FIRE and ideas behind it. It can be applied together with the recently developed code LiteRed by Lee in order to provide an integration by parts reduction to master integrals for quite complicated families of Feynman integrals. As an example, we consider four-loop massless propagator integrals for which LiteRed provides reduction rules and FIRE assists to apply these rules. So, as a by-product one obtains a four-loop variant of the well-known algorithm MINCER. The existence of these explicit reduction rules shows that any four-loop massless propagator integral can be reduced to a linear combination of master integrals in the sense of a mathematical theorem. We also describe various algebraic ways to find additional relations between master integrals associated with several families of Feynman integrals. Program summaryProgram title: FIRE4Catalogue identifier: AEPW_v1_0Program summary URL:http://cpc.cs.qub.ac.uk/summaries/AEPW_v1_0.htmlProgram obtainable from: CPC Program Library, Queen’s University, Belfast, N. IrelandLicensing provisions: GNU General Public License, version 2No. of lines in distributed program, including test data, etc.: 102107No. of bytes in distributed program, including test data, etc.: 3539091Distribution format: tar.gzProgramming language: Wolfram Mathematica 6.0 or higher.Computer: Workstation or PC.Operating system: Unix, Linux, Windows, Mac OS X.RAM: Depends on the complexity of the problemClassification: 4.4, 4.8, 5, 20.External routines: QLink [1], FLink [2], LiteRed [3]Nature of problem:Reducing Feynman integrals to master integrals can be treated as a task to solve a huge system of sparse linear equations with polynomial coefficients.Solution method:Since the matrix of equations is very specific, none of the standard methods of solving linear equations can be applied efficiently. The program approaches solving those equations with a special version of Gauss elimination. It stores data on disk with the use of the QLink [1] routine and can speed up calculations (evaluation of GCD) with FLink [2]. The external package LiteRed [3] can be used to produce additional rules for reduction.Restrictions:The complexity of the problem is mostly restricted by the CPU time required to perform the reduction of integrals.Running time:Depends on the complexity of the problem

Adaptive nonparametric empirical Bayes estimation via wavelet series: The minimax study

In the present paper, we derive lower bounds for the risk of the nonparametric empirical Bayes estimators. In order to attain the optimal convergence rate, we propose generalization of the linear empirical Bayes estimation method which takes advantage of the flexibility of the wavelet techniques. We present an empirical Bayes estimator as a wavelet series expansion and estimate coefficients by minimizing the prior risk of the estimator. As a result, estimation of wavelet coefficients requires solution of a well-posed low-dimensional sparse system of linear equations. The dimension of the system depends on the size of wavelet support and smoothness of the Bayes estimator. An adaptive choice of the resolution level is carried out using Lepski et al. (1997) method. The method is computationally efficient and provides asymptotically optimal adaptive EB estimators. The theory is supplemented by numerous examples.

Solving sparse linear systems of equations over finite fields using bit-flipping algorithm

Let Fq be the finite field with q elements. We give an algorithm for solving sparse linear systems of equations over Fq when the coefficient matrix of the system has a specific structure, here called relatively connected. This algorithm is based on a well-known decoding algorithm for low-density parity-check codes called bit-flipping algorithm. We modify and extend this hard decision decoding algorithm. The complexity of this algorithm is linear in terms of the number of columns n and the number of nonzero coefficients ω of the matrix per iteration. The maximum number of iterations is bounded above by m, the number of equations.

New updates of incomplete LU factorizations and applications to large nonlinear systems

In this paper, we address the problem of preconditioning sequences of large sparse indefinite systems of linear equations and present two new strategies to construct approximate updates of factorized preconditioners. Both updates are based on the availability of an incomplete factorization for one matrix of the sequence and differ in the approximation of the so-called ideal update. For a general treatment, an incomplete LU (ILU) factorization is considered, but the proposed approaches apply to incomplete factorizations of symmetric matrices as well. The first strategy is an approximate diagonal update of the ILU factorization; the second strategy relies on banded approximations of the factors in the ideal update. The efficiency and reliability of the proposed preconditioners are shown in the solution of nonlinear systems of equations by preconditioned Newton–Krylov methods. Nearly matrix-free implementations of the updating strategy are provided, and numerical experiments are carried out on application problems.

A parallel finite-element method for three-dimensional controlled-source electromagnetic forward modelling

SUMMARY We present a nodal finite-element method that can be used to compute in parallel highly accurate solutions for 3-D controlled-source electromagnetic forward-modelling problems in anisotropic media. Secondary coupled-potential formulation of Maxwell’s equations allows to avoid the singularities introduced by the sources, while completely unstructured tetrahedral meshes and mesh refinement support an accurate representation of geological and bathymetric complexity and improve the solution accuracy. Different complex iterative solvers and an efficient pre-conditioner based on the sparse approximate inverse are used for solving the resulting large sparse linear system of equations. Results are compared with the ones of other researchers to check the accuracy of the method. We demonstrate the performance of the code in large problems with tens and even hundreds of millions of degrees of freedom. Scalability tests on massively parallel computers show that our code is highly scalable.

Convergence analysis of modified Newton-HSS method for solving systems of nonlinear equations

Hermitian and skew-Hermitian splitting(HSS) method has been proved quite successfully in solving large sparse non-Hermitian positive definite systems of linear equations. Recently, by making use of HSS method as inner iteration, Newton-HSS method for solving the systems of nonlinear equations with non-Hermitian positive definite Jacobian matrices has been proposed by Bai and Guo. It has shown that the Newton-HSS method outperforms the Newton-USOR and the Newton-GMRES iteration methods. In this paper, a class of modified Newton-HSS methods for solving large systems of nonlinear equations is discussed. In our method, the modified Newton method with R-order of convergence three at least is used to solve the nonlinear equations, and the HSS method is applied to approximately solve the Newton equations. For this class of inexact Newton methods, local and semilocal convergence theorems are proved under suitable conditions. Moreover, a globally convergent modified Newton-HSS method is introduced and a basic global convergence theorem is proved. Numerical results are given to confirm the effectiveness of our method.

Improved Magnetotelluric Zohdy-Oldenburg Direct Inversion

Based on researches of Baiyao (1995) and Zhang and Xu (2001), this paper proposes an improved 2D MT Zohdy-Oldenburg direct inversion method, in the least-square sense, embodying the features of Zohdy’s ratio method and Oldenburg’s difference method, in the condition of rugged topography, with phase information. It bypasses large calculations of the Jacobian matrix and large sparse linear systems of equations and enables direct modifications and comparisons of the model parameters. According to the calculation and analysis of examples, it shows faster convergence and higher precision. In contrast with the conventional linear inversion, the calculation speed of this new method can be increased by more than 10 times.

EFFICIENT HYDRAULIC AND THERMAL ANALYSIS OF HEAT SINKS USING VOLUME AVERAGING THEORY AND GALERKIN METHODS

Air- and water-cooled heat sinks are still the most common heat rejection devices in electronics, making their geometric optimization a key issue in thermal management. Because of the complex geometry, the use of finite-difference, finite-volume, or finite-element methods for the solution of the governing equations becomes computationally expensive. In this work, volume averaging theory is applied to a general heat sink with periodic geometry to obtain a physically accurate, but geometrically simplified, system model. The governing energy and momentum equations are averaged over a representative elementary volume, and the result is a set of integro-partial differential equations. Closure coefficients are introduced, and their values are obtained from data available in the literature. The result of this process is a system of closed partial differential equations, defined on a simple geometry, which can be solved to obtain average velocities and temperatures in the system. The intrinsic smoothness of the solution and the simplified geometry allow the use of a modified Fourier‐Galerkin Method for efficient solutions to the set of differential equations. Modified Fourier series are chosen as the basis functions because they satisfy the boundary conditions a priori and lead to a sparse system of linear equations for the coefficients. The validity of the method is tested by applying it to model the hydraulic and thermal behavior of an air-cooled pin-fin and a water-cooled micro-channel heat sink. The convergence was found to be O(N i3:443 ), while the runtime was »0.25 s for N = 56. The numerical results were validated against the experimental results, and the agreement was excellent with an average error of »4% and a maximum error of »5%.

Preconditioning for Sparse Linear Systems at the Dawn of the 21st Century: History, Current Developments, and Future Perspectives

Iterative methods are currently the solvers of choice for large sparse linear systems of equations. However, it is well known that the key factor for accelerating, or even allowing for, convergence is the preconditioner. The research on preconditioning techniques has characterized the last two decades. Nowadays, there are a number of different options to be considered when choosing the most appropriate preconditioner for the specific problem at hand. The present work provides an overview of the most popular algorithms available today, emphasizing the respective merits and limitations. The overview is restricted to algebraic preconditioners, that is, general-purpose algorithms requiring the knowledge of the system matrix only, independently of the specific problem it arises from. Along with the traditional distinction between incomplete factorizations and approximate inverses, the most recent developments are considered, including the scalable multigrid and parallel approaches which represent the current frontier of research. A separate section devoted to saddle-point problems, which arise in many different applications, closes the paper.

Energy efficiency vs. performance of the numerical solution of PDEs: An application study on a low-power ARM-based cluster

Power consumption and energy efficiency are becoming critical aspects in the design and operation of large scale HPC facilities, and it is unanimously recognised that future exascale supercomputers will be strongly constrained by their power requirements. At current electricity costs, operating an HPC system over its lifetime can already be on par with the initial deployment cost. These power consumption constraints, and the benefits a more energy-efficient HPC platform may have on other societal areas, have motivated the HPC research community to investigate the use of energy-efficient technologies originally developed for the embedded and especially mobile markets. However, lower power does not always mean lower energy consumption, since execution time often also increases. In order to achieve competitive performance, applications then need to efficiently exploit a larger number of processors. In this article, we discuss how applications can efficiently exploit this new class of low-power architectures to achieve competitive performance. We evaluate if they can benefit from the increased energy efficiency that the architecture is supposed to achieve. The applications that we consider cover three different classes of numerical solution methods for partial differential equations, namely a low-order finite element multigrid solver for huge sparse linear systems of equations, a Lattice-Boltzmann code for fluid simulation, and a high-order spectral element method for acoustic or seismic wave propagation modelling. We evaluate weak and strong scalability on a cluster of 96 ARM Cortex-A9 dual-core processors and demonstrate that the ARM-based cluster can be more efficient in terms of energy to solution when executing the three applications compared to an x86-based reference machine.

A finite difference Hartree–Fock program for atoms and diatomic molecules

The newest version of the two-dimensional finite difference Hartree–Fock program for atoms and diatomic molecules is presented. This is an updated and extended version of the program published in this journal in 1996. It can be used to obtain reference, Hartree–Fock limit values of total energies and multipole moments for a wide range of diatomic molecules and their ions in order to calibrate existing and develop new basis sets, calculate (hyper)polarizabilities (αzz, βzzz, γzzzz, Az,zz, Bzz,zz) of atoms, homonuclear and heteronuclear diatomic molecules and their ions via the finite field method, perform DFT-type calculations using LDA or B88 exchange functionals and LYP or VWN correlations ones or the self-consistent multiplicative constant method, perform one-particle calculations with (smooth) Coulomb and Krammers–Henneberger potentials and take account of finite nucleus models. The program is easy to install and compile (tarball+configure+make) and can be used to perform calculations within double- or quadruple-precision arithmetic. Program summaryProgram title: 2dhfCatalogue identifier: ADEB_v2_0Program summary URL:http://cpc.cs.qub.ac.uk/summaries/ADEB_v2_0.htmlProgram obtainable from: CPC Program Library, Queen’s University, Belfast, N. IrelandLicensing provisions: GNU General Public License version 2No. of lines in distributed program, including test data, etc.: 171196No. of bytes in distributed program, including test data, etc.: 9481802Distribution format: tar.gzProgramming language: Fortran 77, C.Computer: any 32- or 64-bit platform.Operating system: Unix/Linux.RAM: Case dependent, from few MB to many GBClassification: 16.1.Catalogue identifier of previous version: ADEB_v1_0Journal reference of previous version: Comput. Phys. Comm. 98(1996)346Does the new version supersede the previous version?: YesNature of problem: The program finds virtually exact solutions of the Hartree–Fock and density functional theory type equations for atoms, diatomic molecules and their ions. The lowest energy eigenstates of a given irreducible representation and spin can be obtained. The program can be used to perform one-particle calculations with (smooth) Coulomb and Krammers–Henneberger potentials and also DFT-type calculations using LDA or B88 exchange functionals and LYP or VWN correlations ones or the self-consistent multiplicative constant method.Solution method: Single-particle two-dimensional numerical functions (orbitals) are used to construct an antisymmetric many-electron wave function of the restricted open-shell Hartree–Fock model. The orbitals are obtained by solving the Hartree–Fock equations as coupled two-dimensional second-order (elliptic) partial differential equations (PDEs). The Coulomb and exchange potentials are obtained as solutions of the corresponding Poisson equations. The PDEs are discretized by the eighth-order central difference stencil on a two-dimensional single grid, and the resulting large and sparse system of linear equations is solved by the (multicolour) successive overrelaxation ((MC)SOR) method. The self-consistent-field iterations are interwoven with the (MC)SOR ones and orbital energies and normalization factors are used to monitor the convergence. The accuracy of solutions depends mainly on the grid and the system under consideration, which means that within double precision arithmetic one can obtain orbitals and energies having up to 12 significant figures. If more accurate results are needed, quadruple-precision floating-point arithmetic can be used.Reasons for new version: Additional features, many modifications and corrections, improved convergence rate, overhauled code and documentation.Summary of revisions: see ChangeLog found in tar.gz archiveRestrictions: The present version of the program is restricted to 60 orbitals. The maximum grid size is determined at compilation time.Unusual features: The program uses two C routines for allocating and deallocating memory. Several BLAS (Basic Linear Algebra System) routines are emulated by the program. When possible they should be replaced by their library equivalents.Additional comments: automake and autoconf tools are required to build and compile the program; checked with f77, gfortran and ifort compilersRunning time: Very case dependent — from a few CPU seconds for the H2 defined on a small grid up to several weeks for the Hartree–Fock-limit calculations for 40–50 electron molecules.

Boundary element formulations for the numerical solution of two-dimensional diffusion problems with variable coefficients

This paper presents new formulations of the radial integration boundary integral equation (RIBIE) and the radial integration boundary integro-differential equation (RIBIDE) methods for the numerical solution of two-dimensional diffusion problems with variable coefficients. The methods use either a specially constructed parametrix (Levi function) or the standard fundamental solution for the Laplace equation to reduce the boundary-value problem (BVP) to a boundary–domain integral equation (BDIE) or boundary–domain integro-differential equation (BDIDE). The radial integration method (RIM) is then employed to convert the domain integrals arising in both BDIE and BDIDE methods into equivalent boundary integrals. The resulting formulations lead to pure boundary integral and integro-differential equations with no domain integrals. Furthermore, a subdomain decomposition technique (SDBDIE) is proposed, which leads to a sparse system of linear equations, thus avoiding the need to calculate a large number of domain integrals. Numerical examples are presented for several simple problems, for which exact solutions are available, to demonstrate the efficiency of the proposed approaches.

Ordered coloring of grids and related graphs

We investigate a coloring problem, called ordered coloring, in grids and some other families of grid-like graphs. Ordered coloring (also known as vertex ranking) has applications, among other areas, in efficient solving of sparse linear systems of equations and scheduling parallel assembly of products. Our main technical results improve upper and lower bounds for the ordered chromatic number of grids and related graphs.

Analysis of the Effects of Irregular Terrain on Radio Wave Propagation Based on a Three-Dimensional Parabolic Equation

The effects from lateral variations of irregular terrain on the propagation of radio waves are considered by the representation of the vector fields in terms of two scalar Hertz potentials in spherical coordinates. The combination of three-dimensional parabolic equations for these potentials with an impedance boundary condition for the ground, followed by a transformation of variables, will define a boundary-condition problem characterized by equations displaying coefficients that depend on the terrain height function and its partial derivatives. The problem solution through the Crank-Nicolson scheme will lead to a sparse system of linear equations, which will be solved by a direct method. The resulting numerical model will be applied to test cases to show similar features to those from examples described in the literature, taken as references for validation purposes.

Software rejuvenation and resource reservation policies for optimizing server resource availability using cyclic nonhomogeneous Markov chains

Software rejuvenation is modeled in a client–server system, which provides resources to priority classes of users. To assure availability, resource reservation policies are adopted for the higher priority classes. In addition software rejuvenation is proposed to optimize resource availability. The system is modeled by a cyclic nonhomogeneous Markov chain to capture the variation of the arrival and service rates during a day period. An optimization problem is solved based on a similar previous work and given the optimal resource reservation policy obtained by its solution, rejuvenation is performed and the optimal rejuvenation policy is determined. As a measure of resource availability the blocking probability of each priority class is used. Performability indicators expressing the total cost are also derived, with respect to the optimal resource reservation and optimal rejuvenation policies, to examine whether rejuvenation benefits the system in terms of cost. To derive the blocking probabilities, the limiting probability distribution is computed using explicit generalized approximate inverse preconditioning for solving efficiently sparse linear systems of algebraic equations. Copyright © 2012 John Wiley & Sons, Ltd.

FINITE ELEMENT MATRIX GENERATION ON A GPU

This paper presents an e-cient technique for fast gener- ation of sparse systems of linear equations arising in computational electromagnetics in a flnite element method using higher order ele- ments. The proposed approach employs a graphics processing unit (GPU) for both numerical integration and matrix assembly. The per- formance results obtained on a test platform consisting of a Fermi GPU (1x Tesla C2075) and a CPU (2x twelve-core Opterons), indicate that the GPU implementation of the matrix generation allows one to achieve speedups by a factor of 81 and 19 over the optimized single- and multi-threaded CPU-only implementations, respectively.

Newton-MHSS methods for solving systems of nonlinear equations with complex symmetric Jacobian matrices

Modified Hermitian and skew-Hermitian splitting (MHSS) method is an unconditionally convergent iterative method for solving large sparse complex symmetric systems of linear equations. By making use of the MHSS iteration as the inner solver for the inexact Newton method, we establish a class of inexact Newton-MHSS methods for solving large sparse systems of nonlinear equations with complex symmetric Jacobian matrices at the solution points. The local and semi-local convergence properties are analyzed under some proper assumptions. Moreover, by introducing a backtracking linear search technique, a kind of global convergence inexact Newton-MHSS methods are also presented and analyzed. Numerical results are given to examine the feasibility and effectiveness of the inexact Newton-MHSS methods.

Iterative Krylov solution methods for geophysical electromagnetic simulations on throughput-oriented processing units

Many geo-scientific applications involve boundary value problems arising in simulating electrostatic and electromagnetic fields for geophysical prospecting and subsurface imaging of electrical resistivity. Modeling complex geological media with three-dimensional finite-difference grids gives rise to large sparse linear systems of equations. For such systems, we have implemented three common iterative Krylov solution methods on graphics processing units and compared their performance with parallel host-based versions. The benchmarks show that the device efficiency improves with increasing grid sizes. Limitations are currently given by the device memory resources.

Convergence of a generalized MSSOR method for augmented systems

Recently, Wu et al. [S.-L. Wu, T.-Z. Huang, X.-L. Zhao, A modified SSOR iterative method for augmented systems, J. Comput. Appl. Math. 228 (1) (2009) 424–433] introduced a modified SSOR (MSSOR) method for augmented systems. In this paper, we establish a generalized MSSOR (GMSSOR) method for solving the large sparse augmented systems of linear equations, which is the extension of the MSSOR method. Furthermore, the convergence of the GMSSOR method for augmented systems is analyzed and numerical experiments are carried out, which show that the GMSSOR method with appropriate parameters has a faster convergence rate than the MSSOR method with optimal parameters.