All issues

In the article a theory of the implicit iterative line-by-line recurrence method for solving the systems of finite-difference equations which arise as a result of approximation of the two-dimensional elliptic differential equations on a regular grid is stated. On the one hand, the high effectiveness of the method has confirmed in practice. Some complex test problems, as well as several problems of fluid flow and heat transfer of a viscous incompressible liquid, have solved with its use. On the other hand, the theoretical provisions that explain the high convergence rate of the method and its stability are not yet presented in the literature. This fact is the reason for the present investigation. In the paper, the procedure of equivalent and approximate transformations of the initial system of linear algebraic equations (SLAE) is described in detail. The transformations are presented in a matrix-vector form, as well as in the form of the computational formulas of the method. The key points of the transformations are illustrated by schemes of changing of the difference stencils that correspond to the transformed equations. The canonical form of the method is the goal of the transformation procedure. The correctness of the method follows from the canonical form in the case of the solution convergence. The estimation of norms of the matrix operators is carried out on the basis of analysis of structures and element sets of the corresponding matrices. As a result, the convergence of the method is proved for arbitrary initial vectors of the solution of the problem.

The norm of the transition matrix operator is estimated in the special case of weak restrictions on a desired solution. It is shown, that the value of this norm decreases proportionally to the second power (or third degree, it depends on the version of the method) of the grid step of the problem solution area in the case of transition matrix order increases. The necessary condition of the method stability is obtained by means of simple estimates of the vector of an approximate solution. Also, the estimate in order of magnitude of the optimum iterative compensation parameter is given. Theoretical conclusions are illustrated by using the solutions of the test problems. It is shown, that the number of the iterations required to achieve a given accuracy of the solution decreases if a grid size of the solution area increases. It is also demonstrated that if the weak restrictions on solution are violated in the choice of the initial approximation of the solution, then the rate of convergence of the method decreases essentially in full accordance with the deduced theoretical results.

The work belongs to the direction of experimental mathematics, which investigates the properties of mathematical objects by the computing facilities of a computer. The base is an exponential map, its topological properties (Cantor's bouquets) differ from properties of polynomial and rational complex-valued functions. The subject of the study are the character and features of the Fatou and Julia sets, as well as the equilibrium points and orbits of the zero of three iterated complex-valued mappings: $f:z \to (1+ \mu) \exp (iz)$, $g : z \to \big(1+ \mu |z - z^*|\big) \exp (iz)$, $h : z \to \big(1+ \mu (z - z^* )\big) \exp (iz)$, with $z,\mu \in \mathbb{C}$, $z^* : \exp (iz^*) = z^*$. For a quasilinear map g having no analyticity characteristic, two bifurcation transitions were discovered: the creation of a new equilibrium point (for which the critical value of the linear parameter was found and the bifurcation consists of “fork” type and “saddle”-node transition) and the transition to the radical transformation of the Fatou set. A nontrivial character of convergence to a fixed point is revealed, which is associated with the appearance of “valleys” on the graph of convergence rates. For two other maps, the monoperiodicity of regimes is significant, the phenomenon of “period doubling” is noted (in one case along the path $39\to 3$, in the other along the path $17\to 2$), and the coincidence of the period multiplicity and the number of sleeves of the Julia spiral in a neighborhood of a fixed point is found. A rich illustrative material, numerical results of experiments and summary tables reflecting the parametric dependence of maps are given. Some questions are formulated in the paper for further research using traditional mathematics methods.

In this paper, we consider conditional gradient methods for optimizing strongly convex functions. These are methods that use a linear minimization oracle, which, for a given vector $p \in \mathbb{R}^n$, computes the solution of the subproblem

\[ \text{Argmin}_{x\in X}{\langle p,\,x \rangle}. \]There are a variety of conditional gradient methods that have a linear convergence rate in a strongly convex case. However, in all these methods, the dimension of the problem is included in the rate of convergence, which in modern applications can be very large. In this paper, we prove that in the strongly convex case, the convergence rate of the conditional gradient methods in the best case depends on the dimension of the problem $ n $ as $ \widetilde {\Omega} \left(\!\sqrt {n}\right) $. Thus, the conditional gradient methods may turn out to be ineffective for solving strongly convex optimization problems of large dimensions.

Also, the application of conditional gradient methods to minimization problems of a quadratic form is considered. The effectiveness of the Frank – Wolfe method for solving the quadratic optimization problem in the convex case on a simplex (PageRank) has already been proved. This work shows that the use of conditional gradient methods to solve the minimization problem of a quadratic form in a strongly convex case is ineffective due to the presence of dimension in the convergence rate of these methods. Therefore, the Shrinking Conditional Gradient method is considered. Its difference from the conditional gradient methods is that it uses a modified linear minimization oracle. It's an oracle, which, for a given vector $p \in \mathbb{R}^n$, computes the solution of the subproblem \[ \text{Argmin}\{\langle p, \,x \rangle\colon x\in X, \;\|x-x_0^{}\| \leqslant R \}. \] The convergence rate of such an algorithm does not depend on dimension. Using the Shrinking Conditional Gradient method the complexity (the total number of arithmetic operations) of solving the minimization problem of quadratic form on a $ \infty $-ball is obtained. The resulting evaluation of the method is comparable to the complexity of the gradient method.

One- and two-dimensional hydrodynamic models of turbulent transfer within the atmospheric surface layer under neutral thermal stratification are considered. Both models are based on the solution of system of the timeaveraged equations of Navier – Stokes and continuity using a 1.5-order closure scheme as well as equations for turbulent kinetic energy and the rate of its dissipation. The influence of the upper and lower boundary conditions on vertical profiles of wind speed and turbulence parameters within the atmospheric surface layer was derived using an one-dimensional model usually applied in case of an uniform ground surface. The boundary conditions in the model were prescribed in such way that the vertical wind and turbulence patterns were well agreed with widely used logarithmic vertical profile of wind speed, linear dependence of turbulent exchange coefficient on height above ground surface level and constancy of turbulent kinetic energy within the atmospheric surface layer under neutral atmospheric conditions. On the basis of the classical one-dimensional model it is possible to obtain a number of relationships which link the vertical wind speed gradient, turbulent kinetic energy and the rate of its dissipation. Each of these relationships can be used as a boundary condition in our hydrodynamic model. The boundary conditions for the wind speed and the rate of dissipation of turbulent kinetic energy were selected as parameters to provide the smallest deviations of model calculations from classical distributions of wind and turbulence parameters. The corresponding upper and lower boundary conditions were used to define the initial and boundary value problem in the two-dimensional hydrodynamic model allowing to consider complex topography and horizontal vegetation heterogeneity. The two-dimensional model with selected optimal boundary conditions was used to describe the spatial pattern of turbulent air flow when it interacted with the forest edge. The dynamics of the air flow establishment depending on the distance from the forest edge was analyzed. For all considered initial and boundary value problems the unconditionally stable implicit finite-difference schemes of their numerical solution were developed and implemented.

The study completes a series of the author’s works devoted to the spread of particles population in supercritical catalytic branching random walk (CBRW) on a multidimensional lattice. The CBRW model describes the evolution of a system of particles combining their random movement with branching (reproduction and death) which only occurs at fixed points of the lattice. The set of such catalytic points is assumed to be finite and arbitrary. In the supercritical regime the size of population, initiated by a parent particle, increases exponentially with positive probability. The rate of the spread depends essentially on the distribution tails of the random walk jump. If the jump distribution has “light tails”, the “population front”, formed by the particles most distant from the origin, moves linearly in time and the limiting shape of the front is a convex surface. When the random walk jump has independent coordinates with a semiexponential distribution, the population spreads with a power rate in time and the limiting shape of the front is a star-shape nonconvex surface. So far, for regularly varying tails (“heavy” tails), we have considered the problem of scaled front propagation assuming independence of components of the random walk jump. Now, without this hypothesis, we examine an “isotropic” case, when the rate of decay of the jumps distribution in different directions is given by the same regularly varying function. We specify the probability that, for time going to infinity, the limiting random set formed by appropriately scaled positions of population particles belongs to a set $B$ containing the origin with its neighborhood, in $\mathbb{R}^d$. In contrast to the previous results, the random cloud of particles with normalized positions in the time limit will not concentrate on coordinate axes with probability one.

We consider a finite-dimensional optimization problem, the formulation of which in addition to the required variables contains parameters. The solution to this problem is a dependence of optimal values of variables on parameters. In general, these dependencies are not functions because they can have ambiguous meanings and in the functional case be nondifferentiable. In addition, their domain of definition may be narrower than the domains of definition of functions in the condition of the original problem. All these properties make it difficult to solve both the original parametric problem and other tasks, the statement of which includes these dependencies. To overcome these difficulties, usually methods such as non-differentiable optimization are used.

This article proposes an alternative approach that makes it possible to obtain solutions to parametric problems in a form devoid of the specified properties. It is shown that such representations can be explored using standard algorithms, based on the Taylor formula. This form is a function smoothly approximating the solution of the original problem for any parameter values, specified in its statement. In this case, the value of the approximation error is controlled by a special parameter. Construction of proposed approximations is performed using special functions that establish feedback (within optimality conditions for the original problem) between variables and Lagrange multipliers. This method is described for linear problems with subsequent generalization to the nonlinear case.

From a computational point of view the construction of the approximation consists in finding the saddle point of the modified Lagrange function of the original problem. Moreover, this modification is performed in a special way using feedback functions. It is shown that the necessary conditions for the existence of such a saddle point are similar to the conditions of the Karush – Kuhn – Tucker theorem, but do not contain constraints such as inequalities and conditions of complementary slackness. Necessary conditions for the existence of a saddle point determine this approximation implicitly. Therefore, to calculate its differential characteristics, the implicit function theorem is used. The same theorem is used to reduce the approximation error to an acceptable level.

Features of the practical implementation feedback function method, including estimates of the rate of convergence to the exact solution are demonstrated for several specific classes of parametric optimization problems. Specifically, tasks searching for the global extremum of functions of many variables and the problem of multiple extremum (maximin-minimax) are considered. Optimization problems that arise when using multicriteria mathematical models are also considered. For each of these classes, there are demo examples.

Verification of the physical-statistical approach within reliability theory for the simplest cases was carried out, which showed its validity. An analytical solution of the one-dimensional basic equation of the physicalstatistical approach is presented under the assumption of a stationary degradation rate. From a mathematical point of view this equation is the well-known continuity equation, where the role of density is played by the density distribution function of goods in its characteristics phase space, and the role of fluid velocity is played by intensity (rate) degradation processes. The latter connects the general formalism with the specifics of degradation mechanisms. The cases of coordinate constant, linear and quadratic degradation rates are analyzed using the characteristics method. In the first two cases, the results correspond to physical intuition. At a constant rate of degradation, the shape of the initial distribution is preserved, and the distribution itself moves equably from the zero. At a linear rate of degradation, the distribution either narrows down to a narrow peak (in the singular limit), or expands, with the maximum shifting to the periphery at an exponentially increasing rate. The distribution form is also saved up to the parameters. For the initial normal distribution, the coordinates of the largest value of the distribution maximum for its return motion are obtained analytically.

In the quadratic case, the formal solution demonstrates counterintuitive behavior. It consists in the fact that the solution is uniquely defined only on a part of an infinite half-plane, vanishes along with all derivatives on the boundary, and is ambiguous when crossing the boundary. If you continue it to another area in accordance with the analytical solution, it has a two-humped appearance, retains the amount of substance and, which is devoid of physical meaning, periodically over time. If you continue it with zero, then the conservativeness property is violated. The anomaly of the quadratic case is explained, though not strictly, by the analogy of the motion of a material point with an acceleration proportional to the square of velocity. Here we are dealing with a mathematical curiosity. Numerical calculations are given for all cases. Additionally, the entropy of the probability distribution and the reliability function are calculated, and their correlation is traced.

An essential class of problems describes physical processes occurring in non-convex domains containing a corner greater than 180 degrees on the boundary. The solution in a neighborhood of a corner is singular and its finding using classical approaches entails a loss of accuracy. In the paper, we consider stationary, linearized by Picard’s iterations, Navier – Stokes equations governing the flow of a incompressible viscous fluid in the convection form in $L$-shaped domain. An $R_\nu$-generalized solution of the problem in special sets of weighted spaces is defined. A special finite element method to find an approximate $R_\nu$-generalized solution is constructed. Firstly, functions of the finite element spaces satisfy the law of conservation of mass in the strong sense, i.e. at the grid nodes. For this purpose, Scott – Vogelius element pair is used. The fulfillment of the condition of mass conservation leads to the finding more accurate, from a physical point of view, solution. Secondly, basis functions of the finite element spaces are supplemented by weight functions. The degree of the weight function, as well as the parameter $\nu$ in the definition of an $R_\nu$-generalized solution, and a radius of a neighborhood of the singularity point are free parameters of the method. A specially selected combination of them leads to an increase almost twice in the order of convergence rate of an approximate solution to the exact one in relation to the classical approaches. The convergence rate reaches the first order by the grid step in the norms of Sobolev weight spaces. Thus, numerically shown that the convergence rate does not depend on the corner value.

Finding the global minimum of a nonconvex function is one of the key and most difficult problems of the modern optimization. In this paper we consider special classes of nonconvex problems which have a clear and distinct global minimum.

In the first part of the paper we consider two classes of «good» nonconvex functions, which can be bounded below and above by a parabolic function. This class of problems has not been widely studied in the literature, although it is rather interesting from an applied point of view. Moreover, for such problems first-order and higher-order methods may be completely ineffective in finding a global minimum. This is due to the fact that the function may oscillate heavily or may be very noisy. Therefore, our new methods use only zero-order information and are based on grid search. The size and fineness of this grid, and hence the guarantee of convergence speed and oracle complexity, depend on the «goodness» of the problem. In particular, we show that if the function is bounded by fairly close parabolic functions, then the complexity is independent of the dimension of the problem. We show that our new methods converge with a linear convergence rate $\log(1/\varepsilon)$ to a global minimum on the cube.

In the second part of the paper, we consider the nonconvex optimization problem from a different angle. We assume that the target minimizing function is the sum of the convex quadratic problem and a nonconvex «noise» function proportional to the distance to the global solution. Considering functions with such noise assumptions for zero-order methods is new in the literature. For such a problem, we use the classical gradient-free approach with gradient approximation through finite differences. We show how the convergence analysis for our problems can be reduced to the standard analysis for convex optimization problems. In particular, we achieve a linear convergence rate for such problems as well.

Experimental results confirm the efficiency and practical applicability of all the obtained methods.

The paper provides the mathematical and numerical models of the interrelated thermo- and hydrodynamic processes in the operational mode of development the unified oil-producing complex during the hydrogel flooding of the non-uniform oil reservoir exploited with a system of arbitrarily located injecting wells and producing wells equipped with submersible multistage electrical centrifugal pumps. A special feature of our approach is the modeling of the special ground-based equipment operation (control stations of submersible pumps, drossel devices on the head of producing wells), designed to regulate the operation modes of both the whole complex and its individual elements.

The complete differential model includes equations governing non-stationary two-phase five-component filtration in the reservoir, quasi-stationary heat and mass transfer in the wells and working channels of pumps. Special non-linear boundary conditions and dependencies simulate, respectively, the influence of the drossel diameter on the flow rate and pressure at the wellhead of each producing well and the frequency electric current on the performance characteristics of the submersible pump unit. Oil field development is also regulated by the change in bottom-hole pressure of each injection well, concentration of the gel-forming components pumping into the reservoir, their total volume and duration of injection. The problem is solved numerically using conservative difference schemes constructed on the base of the finite difference method, and developed iterative algorithms oriented on the parallel computing technologies. Numerical model is implemented in a software package which can be considered as the «Intellectual System of Wells» for the virtual control the oil field development.