


Vol 64, No 8 (2024)
Articles
DEDICATED TO SERGEI KONSTANTINOVICH GODUNOV
Abstract
Sergei Konstantinovich Godunov was born on July 17, 1929, in Moscow. In 1946, after completing an Air Force special school, he enrolled in the Mechanics and Mathematics Faculty at Moscow State University, where he received a solid mathematical foundation and excellent training. His mentors included prominent Soviet scientists and educators B.N. Delone and I.G. Petrovsky. In 1951, after graduating from Moscow State University, he joined the Computational Bureau of the Steklov Mathematical Institute of the USSR Academy of Sciences. From his very first days working in the Computational Bureau, and later in the Department of Applied Mathematics at MIAS, S.K. Godunov became an active member of a large group of scientists dedicated to addressing critical practical tasks in mathematical modeling and calculations across various nuclear physics processes. This marked the beginning of his unwavering dedication to science, which continued for more than 70 years.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1337-1341



General numerical methods
APPROXIMATION AND SMOOTHING OF A FUNCTION BASED ON GODUNOV REGULARIZATION
Abstract
A new approach to function approximation is presented, based on S.K. Godunov’s ideas on the regularization of ill-conditioned systems. The proposed method allows for determining function values at nodes of a finer grid from data on a coarser grid while ensuring control over the smoothness of the resulting function. Convergence and smoothness estimates are substantiated, and results from computational experiments illustrate the effectiveness of the proposed method.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1342-1354



ON AN APPROACH TO IMPROVING THE ACCURACY AND EFFICIENCY OF THE ROMB METHOD FOR SOLVING THE UNSTEADY HEAT CONDUCTION EQUATION
Abstract
A modification of the ROMB method is presented for numerically solving unsteady nonlinear heat conduction equations. The modified difference schemes are conservative and, for smooth solutions, approximate the model differential equations with constant coefficients with second-order accuracy in both time and space. First differential approximations are provided for the modified difference schemes, enabling the assessment of approximation errors in the schemes.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1437-1444



CONVERGENCE ESTIMATES OF ITERATIVE METHODS FOR NUMERICAL MODELING OF THREE-DIMENSIONAL PROCESSES IN MAGNETOHYDRODYNAMICS
Abstract
This paper addresses the convergence of iterative processes applied to implicit, fully conservative difference schemes in three-dimensional magnetohydrodynamics, using both separate and combined solution methods for groups of difference equations split by physical processes. Convergence estimates for the iterative processes of the numerical methods considered in this study are obtained. The applicability of both combined and separate methods for solving three-dimensional difference equations in magnetohydrodynamics is examined. Given that the presented algorithm analysis is mainly qualitative, the validity of the obtained estimates was confirmed through numerical experiments on both model and real-world problems. Notably, the convergence estimates of the iterative processes allow for the selection of an optimal numerical method for solving difference equations in three-dimensional magnetohydrodynamic problems at any time step.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1424-1436



ESTIMATES FOR SOLUTIONS OF A BIOLOGICAL MODEL WITH INFINITE DISTRIBUTED DELAY
Abstract
A model of competition between several microorganism species described by a system of nonlinear differential equations with infinite distributed delay is considered. The case of asymptotic stability of the equilibrium point corresponding to the survival of only one species and the extinction of all others is studied. The conditions for the initial numbers of species and the initial concentration of the nutrient at which the system reaches an equilibrium state are specified, and estimates of the stabilization rate are established. The results are obtained using the Lyapunov–Krasovskii functional.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1409-1423



NUMERICAL SOLUTION OF THE IONIZATION EQUATIONS BY STEP FRACTIONS
Abstract
The fractional steps method is applied to solve electron-ion plasma ionization equations involving neutrals, with a focus on processes in stationary plasma thrusters (SPT). The approach is based on splitting by physical processes: plasma evolution is represented as a superposition of pure transport in an external field in phase space followed by ionization. The particle-in-cell method is used for calculating transport. In the case of planar symmetry, the ionization process is calculated using explicit analytical formulas. The derived formulas of the fractional steps method are compared with the numerical solution of the plasma evolution equation through a well-known physical ionization model.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1398-1408



ESTIMATES OF THE p-NORMS OF SOLUTIONS AND INVERSE MATRICES OF SYSTEMS OF LINEAR EQUATIONS WITH A CIRCULANT MATRIX
Abstract
The problem of estimating solutions and inverse matrices of systems of linear equations with a circulant matrix in the p-norm, 1 < p < w, is considered. An estimate is obtained for a circulant matrix with diagonal dominance. Based on this result and the idea of decomposing a matrix into a product of matrices related to the decomposition of the characteristic polynomial, an estimate is proposed for a general circulant matrix.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1388-1397



EFFICIENT AND STABLE TIME INTEGRATION OF THE CAHN-HILLARD EQUATIONS: EXPLICIT, IMPLICIT, AND EXPLICIT-ITERATIVE SCHEMES
Abstract
The article proposes a new algorithm for numerical integration over time of the Cahn-Hilliard equation, based on the combined application of the Eyre splitting method and the local iteration modified (LIM) scheme for solving a finite-dimensional problem at each time step. The proposed method is gradient-stable and allows calculations with large time steps and has an explicit nature of calculations. The results of numerical calculations are presented, demonstrating the capabilities of the proposed method and its comparison with common methods of time integration of the Cahn– Hilliard equation.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1366-1387



A METHOD FOR SEPARATING THE MATRIX SPECTRUM BY A STRAIGHT LINE AND AN INFINITE STRIP FLUTTER PROBLEM
Abstract
A new method for separating the matrix spectrum relative to the straight line is proposed, based on a fractional-linear transformation. It is noted that it has a number of advantages over approaches based on an exponential transformation, namely, its applicability area is wider, and the number of iterations required for its convergence is significantly smaller. The proposed method is used to study the problems of infinite strip flutter with various edge fixing conditions, which after suitable discretization of differential operators are reduced to spectral problems for matrices. The study of stability regions by the method of spectrum dichotomy relative to the imaginary axis allows one to construct neutral curves in the plane of parameters of the flutter problem.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1355-1365



Partial Differential Equations
ENERGY ESTIMATES FOR A CLASS OF PSEUDO-HYPERBOLIC OPERATORS WITH VARIABLE COEFFICIENTS
Abstract
A class of strictly pseudo-hyperbolic fourth-order operators with variable coefficients is considered. Energy estimates are established under certain conditions on the coefficients. These estimates imply the uniqueness of the solution to the Cauchy problem, as well as a priori estimates.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1466-1475



STRUCTURED PSEUDOSPECTRA IN PROBLEMS OF SPATIAL STABILITY OF BOUNDARY LAYERS
Abstract
The paper is devoted to the numerical analysis of the sensitivity of the characteristics of spatial stability of boundary layers to the errors with which the main flow is specified. It is proposed to use structured pseudospectra for this purpose. It is shown that the obtained estimates are significantly more accurate than the estimates based on the unstructured pseudospectrum. The presentation is carried out using the example of a viscous incompressible fluid flow over a concave surface of small curvature with flow parameters favorable for the development of Goertler vortices and Tollmien–Schlichting waves.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1476-1485



ESTIMATES OF SOLUTIONS FOR A CLASS OF NONAUTONOMOUS SYSTEMS OF NEUTRAL TYPE WITH CONCENTRATED AND DISTRIBUTED DELAYS
Abstract
A class of systems of nonautonomous differential equations of neutral type with concentrated and distributed delays is considered. By using a Lyapunov–Krasovskii functional, estimates are established imply whether the solutions are stable. In the case of exponential stability, estimates for the stabilization rate of the solutions at infinity are given.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1486-1499



SPHERICAL SPLINE SOLUTIONS OF THE INHOMOGENEOUS BIHARMONIC EQUATION
Abstract
An inhomogeneous biharmonic equation is considered on the unit sphere in three-dimensional space. The solution of this equation, belonging to the Sobolev space on the sphere, is approximated by a sequence of solutions of the same equation but with specific right-hand sides, represented as linear combinations of shifts of the Dirac delta function. It is proven that, given specified nodes on the sphere determining the shifts, special solutions of the equation — spherical biharmonic splines — exist, and the weights corresponding to each are solutions of an associated non-degenerate system of linear algebraic equations. The connection between the approximation quality of the differential problem solution by spherical biharmonic splines and the problem of the convergence rate of optimal weighted spherical cubature formulas is established.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1456-1465



CONDITIONS FOR SOLVABILITY OF THE CAUCHY PROBLEM FOR ONE PSEUDOHYPERBOLIC SYSTEM
Abstract
The Cauchy problem for one system, not resolved with respect to the highest derivative with respect to time, is considered. The system under study belongs to the class of pseudohyperbolic systems. The system describes transverse bending-torsional vibrations of an elastic rod. Conditions for the solvability of the Cauchy problem in Sobolev spaces and estimates of the solution are obtained.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1445-1455



Mathematical physics
DENSITY GRADIENT MODEL IN A SPHERICALLY SYMMETRIC FORMULATION AND ITS EXPLICIT-IMPLICIT DISSIPATIVE DISCRETIZATION FOR STUDYING INTERFACE DYNAMICS
Abstract
This work is dedicated to the development of an unconditionally gradient-stable (dissipative) numerical method for solving a conservative density gradient model in a spherically symmetric formulation. The algorithm is constructed using the Eyre method based on convex splitting of the system’s free energy. The gradient stability of the algorithm is proven in both semi-discrete and fully discrete cases. Theoretical results are validated through several test calculations. The proposed numerical method is applied to analyze the impact of the specified diffusion mobility on the nature of interface evolution.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1500-1516



SCHEME FOR CALCULATING UNSTEADY FLOWS OF HEAT-CONDUCTING GAS IN A THREE-TEMPERATURE APPROXIMATION
Abstract
A numerical modeling technique for unsteady flows of heat-conducting gas in a three-temperature approximation is presented. The methodology is based on the foundational principles of S.K. Godunov. For time integration, each time step is computed by splitting the governing equations into hyperbolic and parabolic subsystems. The first subsystem is solved using a generalized Godunov scheme, while the second uses an explicit-iterative Chebyshev scheme. Adaptive, curvilinear moving grids are used for discretization, and the discrete scheme is formulated in curvilinear coordinates, preserving the symmetries of the differential problem. The methodology is implemented as a parallel code for multiprocessor computers. Its primary purpose is to support computational studies in controlled thermonuclear fusion, though it can also be applied to other areas of computational aerogas dynamics.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1517-1528



COMPUTATION OF UNSTEADY SWIRLING FLOWS IN NOZZLES AND PIPES BY APPLYING A NEW LOCALLY IMPLICIT GODUNOV-TYPE SCHEME
Abstract
A new class of numerical schemes for calculating unsteady swirling flows in nozzles and pipes based on compressible inviscid gas equations is presented. The main advantage of these schemes is their ability to efficiently handle unsteady problems with multiple scales. The construction of such schemes is based on the well-known Godunov approach, which involves calculating fluxes at the faces of computational cells (volumes) by solving auxiliary one-dimensional problems in the vicinity of each face and approximating conservation laws. The scheme switches from an explicit method to an implicit one for flux calculation based on an analysis of the current solution near the cell face. The scheme is absolutely stable and does not generate spurious oscillations. Its effectiveness is demonstrated in the calculation of unsteady swirling flows in nozzles and pipes. Specifics of setting up problems of this type are investigated, and options for proper problem formulation are proposed. Properties of the solution for swirling flows with a central body covering only part of the symmetry axis in the computational domain are also studied.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1529-1545



FREE BOUNDARY METHOD FOR COUPLED PROBLEMS OF GAS AND SOLID DYNAMICS
Abstract
This paper presents a new approach to numerical modeling of gas flow around stationary and moving rigid bodies, allowing for the use of Eulerian grids that are not tied to the geometry of the body. The bodies are assumed to be absolutely rigid and undeformable, with their elastic properties disregarded. The gas is inviscid and non-heat-conducting, described by compressible fluid equations. The proposed approach is based on averaging the equations of the original model over a small spatial filter. This results in a system of averaged equations that includes an additional quantity — the solid volume fraction parameter — whose spatial distribution digitally represents the geometry of the body (analogous to an order function). This system of equations operates across the entire space. Under this approach, the standard boundaryvalue problem within the gas region is effectively reduced to a Cauchy problem over the entire space. For a one-dimensional model, the numerical solution of the averaged equations is considered using Godunov’s method. In intersected cells, a discontinuous solution is introduced, leading to a compound Riemann problem that describes the decay of the initial discontinuity in the presence of a confining wall. It is shown that the approximation of the numerical flux for the compound Riemann problem solution ensures transport of the order function without numerical dissipation.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1546-1560



Computer science
ON SPECTRAL PORTRAITS OF NEAREST NEIGHBOR GRAPH INCIDENCE MATRICES
Abstract
This paper explores the possibility of applying S.K. Godunov’s method of constructing spectral portraits of matrices to estimate the rank of special types of matrices that appear in various applications, such as nearest neighbor graph structure analysis, finite automata theory, and sparse matrix spectrum estimation. A computational algorithm for generating an ensemble of random distance matrices and the associated nearest neighbor graphs is described. Based on computational experiments, parameters of the vertex degree distribution of random nearest neighbor graphs are evaluated. These estimates are feasible because the distribution is independent of the random distance function and follows a multivariate normal distribution. It is proven that the rank of the incidence matrix of a nearest neighbor graph equals the total number of vertices with in-degree 0 and 1, and the rank distribution of such a matrix is derived. It is also shown that, in this context, a method based on analyzing the vertex degree distribution is highly effective for determining the matrix rank.
Žurnal vyčislitelʹnoj matematiki i matematičeskoj fiziki. 2024;64(8):1561-1570


