Numerical methods of approximate solution of the Cauchy problem for coupled systems of evolution equations are considered. Separating simpler subproblems for individual components of the solution achieves simplification of the problem at a new level in time. The decoupling method, a significant approach to simplifying the problem, is based on the decomposition of the problem's operator matrix. The approximate solution is constructed based on the linear composition of solutions to auxiliary problems. The paper investigates decoupling variants based on extracting the diagonal part of the operator matrix and the lower and upper triangular submatrices. The study introduces a new decomposition approach, which involves splitting the operator matrix into rows and columns. The composition stage utilizes various variants of splitting schemes, showcasing the versatility of the approach. In additive operator-difference schemes, we can distinguish explicit-implicit schemes, factorized schemes for two-component splitting, and regularized schemes for general multicomponent splitting. The study of stability of two- and three-level decoupling composition schemes is carried out using the theory of stability (correctness) of operator-difference schemes for finite-dimensional Hilbert spaces. The theoretical results of the decoupling technique for systems of evolution equations are illustrated on a test two-dimensional problem for a coupled system of two diffusion equations with inhomogeneous self- and cross-diffusion coefficients.

Formulating boundary value problems for multidimensional partial derivative equations in terms of invariant operators of vector (tensor) analysis is convenient. Computational algorithms for approximate solutions are based on constructing grid analogs of vector analysis operators. This is most easily done by dividing the computational domain into rectangular cells when the grid nodes coincide with the cell vertices or are the cell centers. Grid operators of vector analysis for irregular regions are constructed using Delaunay triangulations or Voronoi partitions. This paper uses two-dimensional merged Voronoi-Delaunay grids to represent the grid cells as orthodiagonal quadrilaterals. Consistent approximations of the gradient, divergence, and rotor operators are proposed. On their basis, operator-difference approximations for typical stationary scalar and vector problems are constructed.

Additive (splitting) schemes are used to contruct effective computational algorithms for approximately solving initial-boundary value problems for nonstationary partial differential equations. Splitting schemes are normally used when the main operator of a problem has an additive representation. Problems where the operator at the time derivative of a solution is split are also of interest. For first-order evolution equations, we propose splitting schemes based on transforming the original equation to an equivalent system of equations.

Difference methods are widely used for the approximate solution of boundary value problems for partial differential equations. Grid approximations are most simply constructed when the computational domain is divided into rectangular cells. Typically, the grid nodes coincide with the vertices of the cells. In addition to such node-center approximations, grids with nodes at the centers of cells are also used. It is convenient to formulate boundary value problems in terms of invariant operators of vector (tensor) analysis, which are associated with corresponding grid analogs. In this work, analogs of the gradient and divergence operators are constructed on non-standard rectangular grids the nodes of which consist of both the vertices of the computational cells and their centers. The proposed approach is illustrated using approximations of a boundary value problem for a stationary two-dimensional convection–diffusion equation. The key features of constructing approximations for vector problems are discussed with a focus on applied problems of the mechanics of solids.

Many applied problems lead to the necessity of solving boundary value problems for evolutionary integrodifferential equations of the first order. Such models with memory belong to the class of nonlocal when we have to work with a solution for all previous moments of time. A transformation to local problems when approximating the difference kernel of an integral operator by a sum of exponents was proposed earlier. In this case, the Cauchy problem for the system of evolution equations with additional unknowns is solved. Numerical-analytical schemes for the approximate solution of such problems have been constructed in the paper at the analytical representation of the additional unknowns. Corresponding stability estimates have been obtained using two-level time approximations.

Abstract Many non-local processes are modeled using mathematical models that include fractional powers of elliptic operators. The approximate solution of stationary problems with fractional powers of operators is most often based on rational approximations introduced in various versions for a fractional power of the self-adjoint positive operator. The purpose of this work is to use such approximations for the approximate solution of nonstationary problems. We consider Cauchy problems for the first and second order differential-operator equations in finite-dimensional Hilbert spaces. Estimates for the proximity of an approximate solution to an exact one are obtained when specifying the absolute and relative errors of the approximation of the fractional power of the operator. We construct splitting schemes based on the additive representation with a rational approximation of the operator’s fractional power. The stability and accuracy of factorized two-level additive operator-difference schemes for the first order evolution equation and three-level schemes for a second order equation are established.

We have considered the Cauchy problem for a first-order evolutionary equation with fractional powers of an operator. Such nonlocal mathematical models are used, for example, to describe anomalous diffusion processes. We want the transition to a new level in time to be solved usual problems. Computational algorithms are constructed based on some approximations of operator functions. Currently, when solving stationary problems with fractional powers of an operator, the most attention is paid to rational approximations. In the approximate solution of nonstationary problems, we come to equations with an additive representation of the prob-lem operator. Additive-operator schemes are constructed by using different variants of splitting schemes. In the present work, the time approximations are based on approximations of the tran-sition operator by the product of exponents. We use exponent splitting schemes of the first and second-order accuracy. The results of numerical experiments for a two-dimensional model problem with fractional powers of the elliptic operator are presented.

The approximate solution of the Cauchy problem for second-order evolution equations is performed, first of all, using three-level time approximations. Such approximations are easily constructed and relatively uncomplicated to investigate when using uniform time grids. When solving applied problems numerically, we should focus on approximations with variable time steps. When using multilevel schemes on non-uniform grids, we should maintain accuracy by choosing appropriate approximations and ensuring stability of the approximate solution. In this paper, we construct unconditionally stable schemes of the first- and second-order accuracy on a non-uniform time grid for the approximate solution of the Cauchy problem for a second-order evolutionary equation. The novelty of the paper consists in the fact that these stability estimates are obtained without any restrictions on the magnitude of the step change and on the number of step changes. We use a special transformation of the original second-order differential-operator equation to a system of first-order equations. For the system of first-order equations, we apply standard two-level time approximations. We obtained stability estimates for the initial data and the right-hand side in finite-dimensional Hilbert space. Eliminating auxiliary variables leads to three-level schemes for the initial second-order evolution equation. Numerical experiments were performed for the test problem for a one-dimensional in space bi-parabolic equation. The accuracy and stability properties of the constructed schemes are demonstrated on non-uniform grids with randomly varying grid steps.

In this paper, we consider a class of convection-diffusion equations with memory effects. These equations arise as a result of homogenization or upscaling of linear transport equations in heterogeneous media and play an important role in many applications. First, we present a dememorization technique for these equations. We show that the convection-diffusion equations with memory effects can be written as a system of standard convection-diffusion-reaction equations. This allows removing the memory term and simplifying the computations. We consider a relation between dememorized equations and micro-scale equations, which do not contain memory terms. We note that dememorized equations differ from micro-scale equations and constitute a macroscopic model. Next, we consider both implicit and partially explicit methods. The latter is introduced for problems in multiscale media with high-contrast properties. Because of high contrast, explicit methods are restrictive and require time steps that are very small (scales as the inverse of the contrast). We show that, by appropriately decomposing the space, we can treat only a few degrees of freedom implicitly and the remaining degrees of freedom explicitly. We present a stability analysis. Numerical results are presented that confirm our theoretical findings of partially explicit schemes applied to dememorized systems of equations.

Nonstationary problems are solved numerically by applying multilevel (with more than two levels) time approximations. They are easy to construct and relatively easy to study in the case of uniform grids. However, the numerical study of application-oriented problems often involves approximations with a variable time step. The construction of multilevel schemes on nonuniform grids is associated with maintaining the prescribed accuracy and ensuring the stability of the approximate solution. In this paper, three-level schemes for the approximate solution of the Cauchy problem for a second-order evolution equation are constructed in the special case of a doubled or halved step size. The focus is on the approximation features in the transition between different step sizes. The study is based on general results of the stability (well-posedness) theory of operator-difference schemes in a finite-dimensional Hilbert space. Estimates for stability with respect to initial data and the right-hand side are obtained in the case of a doubled or halved time step.

Domain decomposition methods are used for the approximate solution of boundary value problems for partial differential equations on parallel computing systems. The specifics of nonstationary problems is most completely taken into account when using noniterative domain decomposition schemes. Regionally additive schemes are constructed on the basis of various classes of splitting schemes. A new class of domain decomposition schemes with an additive representation of the solution on a new time level is distinguished that is based on splitting the domain into subdomains based on a partition of unity. An example of the Cauchy problem for first-order evolution equations with a positive self-adjoint operator in a finite-dimensional Hilbert space is considered. Unconditionally stable two- and three-level splitting schemes are constructed for the corresponding system of equations.

In computational practice, most attention is paid to rational approximations of functions and approximations by the sum of exponents. We consider a wide enough class of nonlinear approximations characterized by a set of two required parameters. The approximating function is linear in the first parameter; these parameters are assumed to be positive. The individual terms of the approximating function represent a fixed function that depends nonlinearly on the second parameter. A numerical approximation minimizes the residual functional by approximating function values at individual points. The second parameter's value is set on a more extensive set of points of the interval of permissible values. The proposed approach's key feature consists in determining the first parameter on each separate iteration of the classical nonnegative least squares method. The computational algorithm is used to rational approximate the function x-& alpha;,0<& alpha;<1,x & GE;1$$ {x}<^>{-\alpha },\kern0.3em 0<\alpha <1,\kern0.3em x\ge 1 $$. The second example concerns the approximation of the stretching exponential function exp(-x & alpha;),0<& alpha;<1$$ \exp \left(-{x}<^>{\alpha}\right),0 at x & GE;0$$ x\ge 0 $$ by the sum of exponents.

The Cauchy problem for a first-order evolutionary equation with memory with the time derivative of the Volterra integral term and difference kernel in the finite-dimensional Banach space is considered. The fundamental difficulties of the approximate solution of such problems are caused by nonlocality with respect to time when the solution at the current time depends on the entire history. Transformation of the first-order integrodifferential equation to a system of evolutionary local equations with the approximation of the difference kernel by a sum of exponential functions is used. For the weakly coupled system of local equations with additional ordinary differential equations, estimates of stability of solution with respect to initial data and right-hand side are obtained using the concept of logarithmic norm. Similar estimates are obtained for the approximate solution using two-level time approximations.

In approximating solutions of nonstationary problems, various approaches are used to compute the solution at a new time level from a number of simpler (sub-)problems. Among these approaches are splitting methods. Standard splitting schemes are based on one or another additive splitting of the operator into "simpler" operators that are more convenient/easier for the computer implementation and use inhomogeneous (explicitly-implicit) time approximations. In this paper, a new class of splitting schemes is proposed that is characterized by an additive representation of the solution instead of the operator corresponding to the problem (called problem operator). A specific feature of the proposed splitting is that the resulting coupled equations for individual solution components consist of the time derivatives of the solution components. The proposed approaches are motivated by various applications, including multiscale methods, domain decomposition, and so on, where spatially local problems are solved and used to compute the solution. Unconditionally stable splitting schemes are constructed for a first-order evolution equation, which is considered in a finite-dimensional Hilbert space. In our splitting algorithms, we consider the decomposition of both the main operator of the system and the operator at the time derivative. Our goal is to provide a general framework that combines temporal splitting algorithms and spatial decomposition and its analysis. Applications of the framework will be studied separately.

We consider the Cauchy problem for a first-order evolution equation with memory in a finite-dimensional Hilbert space when the integral term is related to the time derivative of the solution. The main problems of the approximate solution of such nonlocal problems are due to the necessity to work with the approximate solution for all previous time moments. We propose a transformation of the first-order integrodifferential equation to a system of local evolutionary equations. We use the approach known in the theory of Voltaire integral equations with an approximation of the difference kernel by the sum of exponents. We formulate a local problem for a weakly coupled system of equations with additional ordinary differential equations. We have given estimates of the stability of the solution by initial data and the right-hand side for the solution of the corresponding Cauchy problem. The primary attention is paid to constructing and investigating the stability of two-level difference schemes, which are convenient for computational implementation. The numerical solution of a two-dimensional model problem for the evolution equation of the first order, when the Laplace operator conditions the dependence on spatial variables, is presented.

Domain decomposition methods are used for the approximate solution of boundary value problems for partial differential equations on parallel computing systems. The specifics of nonstationary problems is most completely taken into account when using noniterative domain decomposition schemes. Regionally additive schemes are constructed on the basis of various classes of splitting schemes. A new class of domain decomposition schemes with an additive representation of the solution on a new time level is distinguished that is based on splitting the domain into subdomains based on a partition of unity. An example of the Cauchy problem for first-order evolution equations with a positive self-adjoint operator in a finite-dimensional Hilbert space is considered. Unconditionally stable two- and three-level splitting schemes are constructed for the corresponding system of equations.

In a finite-dimensional Hilbert space, we consider the Cauchy problem for a second-order integro-differential evolution equation with memory where the integrand is the product of a difference kernel by a linear operator of the time derivative of the solution. The main difficulties in finding the approximate value of the solution of such nonlocal problems at a given point in time are due to the need to work with approximate values of the solution for all previous points in time. A transformation of the integro-differential equation in question to a system of weakly coupled local evolution equations is proposed. It is based on the approximation of the difference kernel by a sum of exponentials. We state a local problem for a weakly coupled system of equations with additional ordinary differential equations. To solve the corresponding Cauchy problem, stability estimates of the solution with respect to the initial data and the right-hand side are given. The main attention is paid to the construction and stability analysis of three-level difference schemes and their computational implementation.

This work continues a line of work on developing partially explicit methods for multiscale problems. In our previous works, we considered linear multiscale problems where the spatial heterogeneities are at the subgrid level and are not resolved. In these works, we have introduced contrast-independent, partially explicit time discretizations for linear equations. The contrast-independent, partially explicit time discretization divides the spatial space into two components: contrast dependent (fast) and contrast independent (slow) spaces defined via multiscale space decomposition. Following this decomposition, temporal splitting was proposed, which treats fast components implicitly and slow components explicitly. The space decomposition and temporal splitting are chosen such that they guarantees stability, and we formulated a condition for the time stepping. This condition was formulated as a condition on slow spaces. In this paper, we extend this approach to nonlinear problems. We propose a splitting approach and derive a condition that guarantees stability. This condition requires some type of contrast-independent spaces for slow components of the solution. We present numerical results and show that the proposed methods provide results similar to implicit methods with a time step that is independent of the contrast.

At present, splitting schemes of various types are available for evolution equations of the first and second order in the case when the basic elliptic operator of the problem has an additive representation. Numerous applications lead to boundary value problems for nonstationary Sobolev-type equations with an elliptic operator at the time derivative. When splitting schemes are used to find an approximate solution of such problems, it is necessary to use an additive representation for both the basic elliptic operator and the operator at the time derivative. This paper deals with the Cauchy problem for a first-order evolution equation in the special case when the operator at the derivative can be represented in terms of the basic operator. The equation is written as a differential-algebraic system of two equations. Unconditionally stable multicomponent splitting schemes are constructed.

Several applied problems are characterized by the need to numerically solve equations with an operator function (matrix function). In particular, in the last decade, mathematical models with a fractional power of an elliptic operator and numerical methods for their study have been actively discussed. Computational algorithms for such non-standard problems are based on approximations by the operator function. The most widespread are the approaches using various options for rational approximation. Also, we note the methods that relate to approximation by exponential sums. In this paper, the possibility of using approximation by exponential products is noted. The solution of an equation with an operator function is based on the transition to standard stationary or evolutionary problems. General approaches are illustrated by a problem with a fractional power of the operator. The first class of methods is based on the integral representation of the operator function under rational approximation, approximation by exponential sums, and approximation by exponential products. The second class of methods is associated with solving an auxiliary Cauchy problem for some evolutionary equation.