All issues

Multiplicative methods for sparse matrices are best suited to reduce the complexity of operations solving systems of linear equations performed on each iteration of the simplex method. The matrix of constraints in these problems of sparsely populated nonzero elements, which allows to obtain the multipliers, the main columns which are also sparse, and the operation of multiplication of a vector by a multiplier according to the complexity proportional to the number of nonzero elements of this multiplier. In addition, the transition to the adjacent basis multiplier representation quite easily corrected. To improve the efficiency of such methods requires a decrease in occupancy multiplicative representation of the nonzero elements. However, at each iteration of the algorithm to the sequence of multipliers added another. As the complexity of multiplication grows and linearly depends on the length of the sequence. So you want to run from time to time the recalculation of inverse matrix, getting it from the unit. Overall, however, the problem is not solved. In addition, the set of multipliers is a sequence of structures, and the size of this sequence is inconvenient is large and not precisely known. Multiplicative methods do not take into account the factors of the high degree of sparseness of the original matrices and constraints of equality, require the determination of initial basic feasible solution of the problem and, consequently, do not allow to reduce the dimensionality of a linear programming problem and the regular procedure of compression — dimensionality reduction of multipliers and exceptions of the nonzero elements from all the main columns of multipliers obtained in previous iterations. Thus, the development of numerical methods for the solution of linear programming problems, which allows to overcome or substantially reduce the shortcomings of the schemes implementation of the simplex method, refers to the current problems of computational mathematics.

In this paper, the approach to the construction of numerically stable direct multiplier methods for solving problems in linear programming, taking into account sparseness of matrices, presented in packaged form. The advantage of the approach is to reduce dimensionality and minimize filling of the main rows of multipliers without compromising accuracy of the results and changes in the position of the next processed row of the matrix are made that allows you to use static data storage formats.

As a direct continuation of this work is the basis for constructing a direct multiplicative algorithm set the direction of descent in the Newton methods for unconstrained optimization is proposed to put a modification of the direct multiplier method, linear programming by integrating one of the existing design techniques significantly positive definite matrix of the second derivatives.

In this paper we describe an algorithm of numerical solving of an inverse problem on a hyperbolic heat equation with additional second time derivative with a small parameter. The problem in this case is finding an initial distribution with given final distribution. This algorithm allows finding a solution to the problem for any admissible given precision. Algorithm allows evading difficulties analogous to the case of heat equation with inverted time. Furthermore, it allows finding an optimal grid size by learning on a relatively big grid size and small amount of iterations of a gradient method and later extrapolates to the required grid size using Richardson’s method. This algorithm allows finding an adequate estimate of Lipschitz constant for the gradient of the target functional. Finally, this algorithm may easily be applied to the problems with similar structure, for example in solving equations for plasma, social processes and various biological problems. The theoretical novelty of the paper consists in the developing of an optimal procedure of finding of the required grid size using Richardson extrapolations for optimization problems with inexact gradient in ill-posed problems.

In this paper the problem of searching for the design of the magnetic system for creation a magnetic field with the required characteristics in the given area is solved. On the basis of analysis of the mathematical model of the magnetic system rather a general approach is proposed to the solving of the inverse problem, which is written by the Fredgolm equation H(z) = ∫_SIJ(s)G(z, s)ds, z ∈ S _H, s ∈ S _I . It was necessary to define the current density distribution function J(s) and the existing winding geometry for creation of a required magnetic field H(z). In the paper a method of solving those by means of regularized iterative processes is proposed. On the base of the concrete magnetic system we perform the numerical study of influence of different factors on the character of the magnetic field being designed.

The article is devoted to modeling the shallow water hydrodynamic processes using the example of the Azov Sea. The article presents a mathematical model of the hydrodynamics of a shallow water body, which allows one to calculate three-dimensional fields of the velocity vector of movement of the aquatic environment. Application of regularizers according to B.N.Chetverushkin in the continuity equation led to a change in the method of calculating the pressure field, based on solving the wave equation. A discrete finite-difference scheme has been constructed for calculating pressure in an area whose linear vertical dimensions are significantly smaller than those in horizontal coordinate directions, which is typical for the geometry of shallow water bodies. The method and algorithm for solving grid equations with a tridiagonal preconditioner are described. The proposed method is used to solve grid equations that arise when calculating pressure for the three-dimensional problem of hydrodynamics of the Azov Sea. It is shown that the proposed method converges faster than the modified alternating triangular method. A parallel implementation of the proposed method for solving grid equations is presented and theoretical and practical estimates of the acceleration of the algorithm are carried out taking into account the latency time of the computing system. The results of computational experiments for solving problems of hydrodynamics of the Sea of Azov using the hybrid MPI + OpenMP technology are presented. The developed models and algorithms were used to reconstruct the environmental disaster that occurred in the Sea of Azov in 2001 and to solve the problem of the movement of the aquatic environment in estuary areas. Numerical experiments were carried out on the K-60 hybrid computing cluster of the Keldysh Institute of Applied Mathematics of Russian Academy of Sciences.

In many social groups, for example, in technical committees for standardization, at the international, regional and national levels, in European communities, managers of ecovillages, social movements (occupy), international organizations, decision-making is based on the consensus of the group members. Instead of voting, where the majority wins over the minority, consensus allows for a solution that each member of the group supports, or at least considers acceptable. This approach ensures that all group members’ opinions, ideas and needs are taken into account. At the same time, it is noted that reaching consensus takes a long time, since it is necessary to ensure agreement within the group, regardless of its size. It was shown that in some situations the number of iterations (agreements, negotiations) is very significant. Moreover, in the decision-making process, there is always a risk of blocking the decision by the minority in the group, which not only delays the decisionmaking time, but makes it impossible. Typically, such a minority is one or two odious people in the group. At the same time, such a member of the group tries to dominate in the discussion, always remaining in his opinion, ignoring the position of other colleagues. This leads to a delay in the decision-making process, on the one hand, and a deterioration in the quality of consensus, on the other, since only the opinion of the dominant member of the group has to be taken into account. To overcome the crisis in this situation, it was proposed to make a decision on the principle of «consensus minus one» or «consensus minus two», that is, do not take into account the opinion of one or two odious members of the group.

The article, based on modeling consensus using the model of regular Markov chains, examines the question of how much the decision-making time according to the «consensus minus one» rule is reduced, when the position of the dominant member of the group is not taken into account.

The general conclusion that follows from the simulation results is that the rule of thumb for making decisions on the principle of «consensus minus one» has a corresponding mathematical justification. The simulation results showed that the application of the «consensus minus one» rule can reduce the time to reach consensus in the group by 76–95%, which is important for practice.

The average number of agreements hyperbolically depends on the average authoritarianism of the group members (excluding the authoritarian one), which means the possibility of delaying the agreement process at high values of the authoritarianism of the group members.

The probabilistic topic model of a text document collection finds two matrices: a matrix of conditional probabilities of topics in documents and a matrix of conditional probabilities of words in topics. Each document is represented by a multiset of words also called the “bag of words”, thus assuming that the order of words is not important for revealing the latent topics of the document. Under this assumption, the problem is reduced to a low-rank non-negative matrix factorization governed by likelihood maximization. In general, this problem is ill-posed having an infinite set of solutions. In order to regularize the solution, a weighted sum of optimization criteria is added to the log-likelihood. When modeling large text collections, storing the first matrix seems to be impractical, since its size is proportional to the number of documents in the collection. At the same time, the topical vector representation (embedding) of documents is necessary for solving many text analysis tasks, such as information retrieval, clustering, classification, and summarization of texts. In practice, the topical embedding is calculated for a document “on-the-fly”, which may require dozens of iterations over all the words of the document. In this paper, we propose a way to calculate a topical embedding quickly, by one pass over document words. For this, an additional constraint is introduced into the model in the form of an equation, which calculates the first matrix from the second one in linear time. Although formally this constraint is not an optimization criterion, in fact it plays the role of a regularizer and can be used in combination with other regularizers within the additive regularization framework ARTM. Experiments on three text collections have shown that the proposed method improves the model in terms of sparseness, difference, logLift and coherence measures of topic quality. The open source libraries BigARTM and TopicNet were used for the experiments.