Home – Home – Jornals – Advanced Search

The demands of accuracy in measurements and engineering models today render the condition number of problems larger. While a corresponding increase in the precision of floating point numbers ensured a stable computing, the uncertainty in convergence when using residue as a stopping criterion has increased. We present an analysis of the uncertainty in convergence when using relative residue as a stopping criterion for iterative solution of linear systems, and the resulting over/under computation for a given tolerance in error. This shows that error estimation is significant for an efficient or accurate solution even when the condition number of the matrix is not large. An Ο(1) error estimator for iterations of the CG algorithm was proposed more than two decades ago. Recently, an Ο(κ²) error estimator was described for the GMRES algorithm which allows for non-symmetric linear systems as well, where κ is the iteration number. We suggest a minor modification in this GMRES error estimation for increased stability. In this work, we also propose an Ο(n) error estimator for A-norm and l₂-norm of the error vector in Bi-CG algorithm. The robust performance of these estimates as a stopping criterion results in increased savings and accuracy in computation, as condition number and size of problems increase.

The relation between complex matrices H and A, given by the equality H A = ĀH is called the pseudo-commutation. The set S_H of all A that pseudo-commute with a nonsingular n × n matrix H is called the pseudo-commutation class defined by H . Every class S_H is a subspace of the space M_n(C) interpreted as a real vector space of dimension 2n². Under the assumption dim_R S_H = n², we find a necessary and sufficient condition for the possibility to decomplexify all the matrices in S_H by one and the same similarity transformation.

In the article, formulas for exact calculation of the approximation error of multiple Itô stochastic integrals based on their orthogonal expansion are obtained. As an example, stochastic Itô integrals with multiplicities 2-4 are considered, which are used in the numerical methods for solving stochastic differential equations with orders of strong convergence 1-2.

The paper discusses the question of why intervals, which are the main object of Interval Analysis, have exactly the form that we know well and habitually use, and not some other. In particular, we investigate why traditional intervals are closed, i.e. contain their endpoints, and also what is wrong with an empty interval. A second question considered in the work is how expedient it is to expand the set of traditional intervals by some other objects. We show that improper («reversed») intervals and the arithmetic of such intervals (the Kaucher complete interval arithmetic) are very useful from many different points of view.

The lattice Boltzmann method (LBM) is a numerical scheme for solving fluid dynamics problems. One of the important and actively developing areas of LBM is correct construction of the scheme on non-uniform spatial grids. With non-uniform grids the total number of calculations can be significantly reduced. However, at the moment the construction of an LBM scheme near a boundary of grids with different spatial steps inevitably requires data interpolation, which can reduce the LBM approximation order and lead to violation of conservation laws. In this work, for the first time, we have developed and tested a method for constructing an athermal node-based LBM on non-uniform grids without interpolation, with the same time step for grids of different scales. The method is based on a two-stage transformation of populations corresponding to different on-grid stencils.

We have developed stochastic algorithms to simulate signals detected by a receiver of a laser navigation system designed for safe aircraft landing. Radiant flux and radiance at the receiver, as well as the contribution of radiation of different orders of scattering are estimated by a Monte Carlo method. Computation results show that the proposed algorithms allow one to study the efficiency of the laser navigation system in various conditions.

In this paper, an efficient numerical method which is a collocation method is considered in order to obtain numerical solutions of the KdV-Kawahara equation. The numerical method is based on a finite element formulation and a spline interpolation by trigonometric quintic B-spline basis. Firstly, the KdV-Kawahara equation is split into a coupled equation via an auxiliary variable as υ=u_xxx. Subsequently, a collocation method is applied to the coupled equation together with the forward difference and the Cranck-Nicolson formula. This application leads us to obtain an algebraic equation system in terms of time variables and trigonometric quintic B-spline basis. In order to measure the error between numerical solutions and exact ones, the error norms L₂ and L_∞. are calculated successfully. The results are illustrated by means of two numerical examples with their graphical representations and comparisons with other methods.

A comparative analysis of two algorithms for estimation of weighted mean particle flux - «by particles» and «by collisions» - is made on the basis of test problem solving for a single-speed particle propagation process with scattering and multiplication in a random medium. It is shown that the first algorithm is preferable for a simple estimation of the mean flux and the second one, for estimation of the parameters of a possible superexponential flux growth. Two models of the random medium are considered: a chaotic «Voronoi mosaic» and «a spherically layered mosaic». For a fixed mean correlation radius, the superexponential growth has been stronger for the layered mosaic.

The article considers the problem of calculating the phase probability density function of a phase-shift keying signal received under conditions of distortion and additive noise. This problem is reduced to an inverse problem, namely, to solving an integral equation of the convolution type. The functions included in the integral equation are analyzed. The case of equiprobable symbols, which is important from a practical point of view, is considered separately. Numerical simulation results are presented.

Based on the use of the geometric series theorem, we transform a linear Fredholm integral equation of the second kind defined on a large interval into an equivalent linear system of Fredholm integral equations of the second kind; then, we inflict a refinement in the way the investigated generalised iterative scheme approximates the sought-after solution. By avoiding to inverse a bounded linear operator, and computing a truncated geometric sum of the former's associated sequence of bounded linear operators instead, we notice that our approach furnishes a better performance in terms of computational time and error efficiency.