All issues

This article explores a method of machine learning based on the theory of random functions. One of the main problems of this method is that decision rule of a model becomes more complicated as the number of training dataset examples increases. The decision rule of the model is the most probable realization of a random function and it's represented as a polynomial with the number of terms equal to the number of training examples. In this article we will show the quick way of the number of training dataset examples reduction and, accordingly, the complexity of the decision rule. Reducing the number of examples of training dataset is due to the search and removal of weak elements that have little effect on the final form of the decision function, and noise sampling elements. For each $(x_i,y_i)$-th element sample was introduced the concept of value, which is expressed by the deviation of the estimated value of the decision function of the model at the point $x_i$, built without the $i$-th element, from the true value $y_i$. Also we show the possibility of indirect using weak elements in the process of training model without increasing the number of terms in the decision function. At the experimental part of the article, we show how changed amount of data affects to the ability of the method of generalizing in the classification task.

The technique of simplification of mathematical model of a chemical reaction by reducing the number of steps of the reaction scheme, based on an analysis of sensitivity to changes in the objective function of the model parameters, is proposed. The reduced scheme of model reaction of formaldehyde oxidation is received. Functional characterizes the measure of proximity to the calculated values for the initial kinetic reaction scheme and the scheme resulting disturbance of its parameters. The advantage of this technique is the ability to analyze complex kinetic schemes and reduction of kinetic models to a size suitable for practical use. The results of computational experiments under different reaction conditions can be included in the functional and thus to receive the reduce scheme, which is consistent the detailed scheme for the desired range of conditions. Sensitivity analysis of the functional model allows to identify those parameters, which provide the largest (or smallest) the contribution to the result of the process simulation. The mathematical model can contain parameters, which change of values do not affect the qualitative and quantitative description of the process. The contribution of these parameters in the functional value won’t be of great importance. Thus it can be eliminated from consideration, which do not serve for modeling kinetic curves substances. The kinetic scheme of formaldehyde oxidation, the detailed mechanism which includes 25 stages and 15 substances, were investigated using this method. On the basis of the local and global sensitivity analysis, the most important stage of the process that affect the overall dynamics of the target concentrations of the reaction. The reduced scheme of model reaction of formaldehyde oxidation is received. This scheme also describes the behavior of the main substances, as detailed scheme, but has a much smaller number of reaction stages. The results of the comparative analysis of modeling of formaldehyde oxidation on detailed and reduced schemes are given. Computational aspects of the problems of chemical kinetics by Sobol’ global method an example of this reaction are specified. The comparison results are local, global and total sensitivity indices are given.

The different types of the reduced equations are known in the dynamics a heavy rigid body with a fixed point. Since the Euler−Poisson’s equations admit the three first integrals, then for the first approach the obtaining new forms of equations are usually based on these integrals. The system of six scalar equations can be transformed to a third-order system with them. However, in indicated approach the reduced system will have a feature as in the form of radical expressions a relatively the components of the angular velocity vector. This fact prevents the effective the effective application of numerical and asymptotic methods of solutions research. In the second approach the different types of variables in a problem are used: Euler’s angles, Hamilton’s variables and other variables. In this approach the Euler−Poisson’s equations are reduced to either the system of second-order differential equations, or the system for which the special methods are effective. In the article the method of finding the reduced system based on the introduction of an auxiliary variable is applied. This variable characterizes the mixed product of the angular momentum vector, the vector of vertical and the unit vector barycentric axis of the body. The system of four differential equations, two of which are linear differential equations was obtained. This system has no analog and does not contain the features that allows to apply to it the analytical and numerical methods. Received form of equations is applied for the analysis of a special class of solutions in the case when the center of mass of the body belongs to the barycentric axis. The variant in which the sum of the squares of the two components of the angular momentum vector with respect to not barycentric axes is constant. It is proved that this variant exists only in the Steklov’s solution. The obtained form of Euler−Poisson’s equations can be used to the investigation of the conditions of existence of other classes of solutions. Certain perspectives obtained equations consists a record of all solutions for which the center of mass is on barycentric axis in the variables of this article. It allows to carry out a classification solutions of Euler−Poisson’s equations depending on the order of invariant relations. Since the equations system specified in the article has no singularities, it can be considered in computer modeling using numerical methods.

Calculations of radiation transport in the shockwave layer of a descent space vehicle cause essential difficulties due to complex multi-resonance dependence of the absorption macroscopic cross sections from the photon energy. The convergence of two approximate spectrum averaging methods to the results of exact pointwise spectrum calculations is investigated. The first one is the well known multigroup method, the second one is the Lebesgue averaging method belonging to methods of the reduction of calculation points by means of aggregation of spectral points which are characterized by equal absorption strength. It is shown that convergence of the Lebesgue averaging method is significantly faster than the multigroup approach as the number of groups is increased. The only 100–150 Lebesgue groups are required to achieve the accuracy of pointwise calculations even in the shock layer at upper atmosphere with sharp absorption lines. At the same time the number of calculations is reduced by more than four order. Series of calculations of the radiation distribution function in 2D shock layer around a sphere and a blunt cone were performed using the local flat layer approximation and the Lebesgue averaging method. It is shown that the shock wave radiation becomes more significant both in value of the energy flux incident on the body surface and in the rate of energy exchange with the gas-dynamic flow in the case of increasing of the vehicle’s size.

Optical systems with two-dimensional feedback demonstrate wide possibilities for studying the nucleation and development processes of dissipative structures. Feedback allows to influence the dynamics of the optical system by controlling the transformation of spatial variables performed by prisms, lenses, dynamic holograms and other devices. A nonlinear interferometer with a mirror image of a field in two-dimensional feedback is one of the simplest optical systems in which is realized the nonlocal nature of light fields.

A mathematical model of optical systems with two-dimensional feedback is a nonlinear parabolic equation with rotation transformation of a spatial variable and periodicity conditions on a circle. Such problems are investigated: bifurcation of the traveling wave type stationary structures, how the form of the solution changes as the diffusion coefficient decreases, dynamics of the solution’s stability when the bifurcation parameter leaves the critical value. For the first time as a parameter bifurcation was taken of diffusion coefficient.

The method of central manifolds and the Galerkin’s method are used in this paper. The method of central manifolds and the Galerkin’s method are used in this paper. The method of central manifolds allows to prove a theorem on the existence and form of the traveling wave type solution neighborhood of the bifurcation value. The first traveling wave born as a result of the Andronov –Hopf bifurcation in the transition of the bifurcation parameter through the сritical value. According to the central manifold theorem, the first traveling wave is born orbitally stable.

Since the above theorem gives the opportunity to explore solutions are born only in the vicinity of the critical values of the bifurcation parameter, the decision to study the dynamics of traveling waves of change during the withdrawal of the bifurcation parameter in the supercritical region, the formalism of the Galerkin method was used. In accordance with the method of the central manifold is made Galerkin’s approximation of the problem solution. As the bifurcation parameter decreases and its transition through the critical value, the zero solution of the problem loses stability in an oscillatory manner. As a result, a periodic solution of the traveling wave type branches off from the zero solution. This wave is born orbitally stable. With further reduction of the parameter and its passage through the next critical value from the zero solution, the second solution of the traveling wave type is produced as a result of the Andronov –Hopf bifurcation. This wave is born unstable with an instability index of two.

Numerical calculations have shown that the application of the Galerkin’s method leads to correct results. The results obtained are in good agreement with the results obtained by other authors and can be used to establish experiments on the study of phenomena in optical systems with feedback.

The article deals with the development of the noise-reduction algorithm based on anisotropic nonlinear data filtering of computed tomography (CT). Analysis of domestic and foreign literature has shown that the most effective algorithms for noise reduction of CT data use complex methods for analyzing and processing data, such as bilateral, adaptive, three-dimensional and other types of filtrations. However, a combination of such techniques is rarely used in practice due to long processing time per slice. In this regard, it was decided to develop an efficient and fast algorithm for noise-reduction based on simplified bilateral filtration method with three-dimensional data accumulation. The algorithm was developed on C ++11 programming language in Microsoft Visual Studio 2015. The main difference of the developed noise reduction algorithm is the use an improved mathematical model of CT noise, based on the distribution of Poisson and Gauss from the logarithmic value, developed earlier by our team. This allows a more accurate determination of the noise level and, thus, the threshold of data processing. As the result of the noise reduction algorithm, processed CT data with lower noise level were obtained. Visual evaluation of the data showed the increased information content of the processed data, compared to original data, the clarity of the mapping of homogeneous regions, and a significant reduction in noise in processing areas. Assessing the numerical results of the algorithm showed a decrease in the standard deviation (SD) level by more than 6 times in the processed areas, and high rates of the determination coefficient showed that the data were not distorted and changed only due to the removal of noise. Usage of newly developed context dynamic threshold made it possible to decrease SD level on every area of data. The main difference of the developed threshold is its simplicity and speed, achieved by preliminary estimation of the data array and derivation of the threshold values that are put in correspondence with each pixel of the CT. The principle of its work is based on threshold criteria, which fits well both into the developed noise reduction algorithm based on anisotropic nonlinear filtration, and another algorithm of noise-reduction. The algorithm successfully functions as part of the MultiVox workstation and is being prepared for implementation in a single radiological network of the city of Moscow.

The paper presents results of experimental, computational and analytical study of the effect of nonequilibrium chemical activation of air-fuel mixture on effectiveness of Diesel process. The generation of a high-voltage multi-streamer discharge in combustion chamber at the compression phase is considered as the method of the activation. The description of electrical discharge system, results of measurement and visualization are presented. The plasma-chemical kinetics of nonequilibrium ignition is analyzed to establish a passway for a proper reduction of chemical kinetics scheme. The results of numerical simulation of gas dynamic processes at presence of plasma-assisted combustion in a geometrical configuration close to the experimental one are described.

Stability of finite difference schemes of lattice Boltzmann method for modelling of 1D diffusion for cases of D1Q2 and D1Q3 lattices is investigated. Finite difference schemes are constructed for the system of linear Bhatnagar–Gross–Krook (BGK) kinetic equations on single particle distribution functions. Brief review of articles of other authors is realized. With application of multiscale expansion by Chapman–Enskog method it is demonstrated that system of BGK kinetic equations at small Knudsen number is transformated to scalar linear diffusion equation. The solution of linear diffusion equation is obtained as a sum of single particle distribution functions. The method of linear travelling wave propagation is used to show the unconditional asymptotic stability of the solution of Cauchy problem for the system of BGK equations at all values of relaxation time. Stability of the scheme for D1Q2 lattice is demonstrated by the method of differential approximation. Stability condition is written in form of the inequality on values of relaxation time. The possibility of the reduction of stability analysis of the schemes for BGK equations to the analysis of special schemes for diffusion equation for the case of D1Q3 lattice is investigated. Numerical stability investigation is realized by von Neumann method. Absolute values of the eigenvalues of the transition matrix are investigated in parameter space of the schemes. It is demonstrated that in wide range of the parameters changing the values of modulas of eigenvalues are lower than unity, so the scheme is stable with respect to initial conditions.

While modeling seismic wave propagation, it is important to take into account nontrivial topography, as this topography causes multiple complex phenomena, such as diffraction at rough surfaces, complex propagation of Rayleigh waves, and side effects caused by wave interference. The primary goal of this research is to construct a method that implements the free surface on topography, utilizing an overset curved grid for characterization, while keeping the main grid structured rectangular. For a combination of the regular and curve-linear grid, the workability of the grid characteristics method using overset grids (also known as the Chimera grid approach) is analyzed. One of the benefits of this approach is computational complexity reduction, caused by the fact that simulation in a regular, homogeneous physical area using a sparse regular rectangle grid is simpler. The simplification of the mesh building mechanism (one grid is regular, and the other can be automatically built using surface data) is a side effect. Despite its simplicity, the method we propose allows us to increase the digitalization of fractured regions and minimize the Courant number. This paper contains various comparisons of modeling results produced by the proposed method-based solver, and results produced by the well-known solver specfem2d, as well as previous modeling results for the same problems. The drawback of the method is that an interpolation error can worsen an overall model accuracy and reduce the computational schema order. Some countermeasures against it are described. For this paper, only two-dimensional models are analyzed. However, the method we propose can be applied to the three-dimensional problems with minimal adaptation required.

The paper is devoted to convex-concave saddle point problems where the objective is a sum of a large number of functions. Such problems attract considerable attention of the mathematical community due to the variety of applications in machine learning, including adversarial learning, adversarial attacks and robust reinforcement learning, to name a few. The individual functions in the sum usually represent losses related to examples from a data set. Additionally, the formulation admits a possibly nonsmooth composite term. Such terms often reflect regularization in machine learning problems. We assume that the dimension of one of the variable groups is relatively small (about a hundred or less), and the other one is large. This case arises, for example, when one considers the dual formulation for a minimization problem with a moderate number of constraints. The proposed approach is based on using Vaidya’s cutting plane method to minimize with respect to the outer block of variables. This optimization algorithm is especially effective when the dimension of the problem is not very large. An inexact oracle for Vaidya’s method is calculated via an approximate solution of the inner maximization problem, which is solved by the accelerated variance reduced algorithm Katyusha. Thus, we leverage the structure of the problem to achieve fast convergence. Separate complexity bounds for gradients of different components with respect to different variables are obtained in the study. The proposed approach is imposing very mild assumptions about the objective. In particular, neither strong convexity nor smoothness is required with respect to the low-dimensional variable group. The number of steps of the proposed algorithm as well as the arithmetic complexity of each step explicitly depend on the dimensionality of the outer variable, hence the assumption that it is relatively small.