All issues

We consider the approaches to the construction of methods for solving four-dimensional programming problems for calculating directions for multiple minimizations of smooth functions on a set of a given set of linear equalities. The approach consists of two stages.

At the first stage, the problem of quadratic programming is transformed by a numerically stable direct multiplicative algorithm into an equivalent problem of designing the origin of coordinates on a linear manifold, which defines a new mathematical formulation of the dual quadratic problem. For this, a numerically stable direct multiplicative method for solving systems of linear equations is proposed, taking into account the sparsity of matrices presented in packaged form. The advantage of this approach is to calculate the modified Cholesky factors to construct a substantially positive definite matrix of the system of equations and its solution in the framework of one procedure. And also in the possibility of minimizing the filling of the main rows of multipliers without losing the accuracy of the results, and no changes are made in the position of the next processed row of the matrix, which allows the use of static data storage formats.

At the second stage, the necessary and sufficient optimality conditions in the form of Kuhn–Tucker determine the calculation of the direction of descent — the solution of the dual quadratic problem is reduced to solving a system of linear equations with symmetric positive definite matrix for calculating of Lagrange's coefficients multipliers and to substituting the solution into the formula for calculating the direction of descent.

It is proved that the proposed approach to the calculation of the direction of descent by numerically stable direct multiplicative methods at one iteration requires a cubic law less computation than one iteration compared to the well-known dual method of Gill and Murray. Besides, the proposed method allows the organization of the computational process from any starting point that the user chooses as the initial approximation of the solution.

Variants of the problem of designing the origin of coordinates on a linear manifold, a convex polyhedron and a vertex of a convex polyhedron are presented. Also the relationship and implementation of methods for solving these problems are described.

Small practical value of many numerical methods for solving single-ended systems of linear equations with ill-conditioned matrices due to the fact that these methods in the practice behave quite differently than in the case of precise calculations. Historically, sustainability is not enough attention was given, unlike in numerical algebra ‘medium-sized’, and emphasis is given to solving the problems of maximal order in data capabilities of the computer, including the expense of some loss of accuracy. Therefore, the main objects of study is the most appropriate storage of information contained in the sparse matrix; maintaining the highest degree of rarefaction at all stages of the computational process. Thus, the development of efficient numerical methods for solving unstable systems refers to the actual problems of computational mathematics.

In this paper, the approach to the construction of numerically stable direct multiplier methods for solving systems of linear equations, taking into account sparseness of matrices, presented in packaged form. The advantage of the approach consists in minimization of filling the main lines of the 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. The storage format of sparse matrices has been studied and the advantage of this format consists in possibility of parallel execution any matrix operations without unboxing, which significantly reduces the execution time and memory footprint.

Direct multiplier methods for solving systems of linear equations are best suited for solving problems of large size on a computer — sparse matrix systems allow you to get multipliers, the main row of which is also sparse, and the operation of multiplication of a vector-row of the multiplier according to the complexity proportional to the number of nonzero elements of this multiplier.

As a direct continuation of this work is proposed in the basis for constructing a direct multiplier algorithm of linear programming to put a modification of the direct multiplier algorithm for solving systems of linear equations based on integration of technique of linear programming for methods to select the host item. Direct multiplicative methods of linear programming are best suited for the construction of a direct multiplicative algorithm set the direction of descent Newton methods in unconstrained optimization by integrating one of the existing design techniques significantly positive definite matrix of the second derivatives.

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.

We consider a numerically stable direct multiplicative algorithm of solving linear equations systems, which takes into account the sparseness of matrices presented in a packed form. The advantage of the algorithm is the ability to minimize the filling of the main rows of multipliers without losing the accuracy of the results. Moreover, changes in the position of the next processed row of the matrix are not made, what allows using static data storage formats. Linear system solving by a direct multiplicative algorithm is, like the solving with $LU$-decomposition, just another scheme of the Gaussian elimination method implementation.

In this paper, this algorithm is the basis for solving the following problems:

Problem 1. Setting the descent direction in Newtonian methods of unconditional optimization by integrating one of the known techniques of constructing an essentially positive definite matrix. This approach allows us to weaken or remove additional specific difficulties caused by the need to solve large equation systems with sparse matrices presented in a packed form.

Problem 2. Construction of a new mathematical formulation of the problem of quadratic programming and a new form of specifying necessary and sufficient optimality conditions. They are quite simple and can be used to construct mathematical programming methods, for example, to find the minimum of a quadratic function on a polyhedral set of constraints, based on solving linear equations systems, which dimension is not higher than the number of variables of the objective function.

Problem 3. Construction of a continuous analogue of the problem of minimizing a real quadratic polynomial in Boolean variables and a new form of defining necessary and sufficient conditions of optimality for the development of methods for solving them in polynomial time. As a result, the original problem is reduced to the problem of finding the minimum distance between the origin and the angular point of a convex polyhedron, which is a perturbation of the $n$-dimensional cube and is described by a system of double linear inequalities with an upper triangular matrix of coefficients with units on the main diagonal. Only two faces are subject to investigation, one of which or both contains the vertices closest to the origin. To calculate them, it is sufficient to solve $4n – 4$ linear equations systems and choose among them all the nearest equidistant vertices in polynomial time. The problem of minimizing a quadratic polynomial is $NP$-hard, since an $NP$-hard problem about a vertex covering for an arbitrary graph comes down to it. It follows therefrom that $P = NP$, which is based on the development beyond the limits of integer optimization methods.

A set of implicit difference schemes on the five-pointwise stensil is under construction. The analysis of properties of difference schemes is carried out in a space of undetermined coefficients. The spaces were introduced for the first time by A. S. Kholodov. Usually for properties of difference schemes investigation the problem of the linear programming was constructed. The coefficient at the main term of a discrepancy was considered as the target function. The optimization task with inequalities type restrictions was considered for construction of the monotonic difference schemes. The limitation of such an approach becomes clear taking into account that approximation of the difference scheme is defined only on the classical (smooth) solutions of partial differential equations.

The functional which minimum will be found put in compliance to the difference scheme. The functional must be the linear on the difference schemes coefficients. It is possible that the functional depends on net function – the solution of a difference task or a grid projection of the differential problem solution. If the initial terms of the functional expansion in a Taylor series on grid parameters are equal to conditions of classical approximation, we will call that the functional will be the generalized condition of approximation. It is shown that such functionals exist. For the simple linear partial differential equation with constant coefficients construction of the functional is possible also for the generalized (non-smooth) solution of a differential problem.

Families of functionals both for smooth solutions of an initial differential problem and for the generalized solution are constructed. The new difference schemes based on the analysis of the functionals by linear programming methods are constructed. At the same time the research of couple of self-dual problems of the linear programming is used. The optimum monotonic difference scheme possessing the first order of approximation on the smooth solution of differential problem is found. The possibility of application of the new schemes for creation of hybrid difference methods of the raised approximation order on smooth solutions is discussed.

The example of numerical implementation of the simplest difference scheme with the generalized approximation is given.

A numerically stable direct multiplicative method for solving systems of linear equations that takes into account the sparseness of matrices presented in a packed form is considered. The advantage of the method is the calculation of the Cholesky factors for a positive definite matrix of the system of equations and its solution within the framework of one procedure. And also in the possibility of minimizing the filling of the main rows of multipliers without losing the accuracy of the results, and no changes are made to the position of the next processed row of the matrix, which allows using static data storage formats. The solution of the system of linear equations by a direct multiplicative algorithm is, like the solution with LU-decomposition, just another scheme for implementing the Gaussian elimination method.

The calculation of the Cholesky factors for a positive definite matrix of the system and its solution underlies the construction of a new mathematical formulation of the unconditional problem of quadratic programming and a new form of specifying necessary and sufficient conditions for optimality that are quite simple and are used in this paper to construct a new mathematical formulation for the problem of quadratic programming on a polyhedral set of constraints, which is the problem of finding the minimum distance between the origin ordinate and polyhedral boundary by means of a set of constraints and linear algebra dimensional geometry.

To determine the distance, it is proposed to apply the known exact method based on solving systems of linear equations whose dimension is not higher than the number of variables of the objective function. The distances are determined by the construction of perpendiculars to the faces of a polyhedron of different dimensions. To reduce the number of faces examined, the proposed method involves a special order of sorting the faces. Only the faces containing the vertex closest to the point of the unconditional extremum and visible from this point are subject to investigation. In the case of the presence of several nearest equidistant vertices, we investigate a face containing all these vertices and faces of smaller dimension that have at least two common nearest vertices with the first face.

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.

In this paper, we consider the problem of restoring the correspondence matrix based on the observations of real correspondences in Moscow. Following the conventional approach [Gasnikov et al., 2013], the transport network is considered as a directed graph whose edges correspond to road sections and the graph vertices correspond to areas that the traffic participants leave or enter. The number of city residents is considered constant. The problem of restoring the correspondence matrix is to calculate all the correspondence from the $i$ area to the $j$ area.

To restore the matrix, we propose to use one of the most popular methods of calculating the correspondence matrix in urban studies — the entropy model. In our work, which is based on the work [Wilson, 1978], we describe the evolutionary justification of the entropy model and the main idea of the transition to solving the problem of entropy-linear programming (ELP) in calculating the correspondence matrix. To solve the ELP problem, it is proposed to pass to the dual problem. In this paper, we describe several numerical optimization methods for solving this problem: the Sinkhorn method and the Accelerated Sinkhorn method. We provide numerical experiments for the following variants of cost functions: a linear cost function and a superposition of the power and logarithmic cost functions. In these functions, the cost is a combination of average time and distance between areas, which depends on the parameters. The correspondence matrix is calculated for multiple sets of parameters and then we calculate the quality of the restored matrix relative to the known correspondence matrix.

We assume that the noise in the restored correspondence matrix is Gaussian, as a result, we use the standard deviation as a quality metric. The article provides an overview of gradient-free optimization methods for solving non-convex problems. Since the number of parameters of the cost function is small, we use the grid search method to find the optimal parameters of the cost function. Thus, the correspondence matrix calculated for each set of parameters and then the quality of the restored matrix is evaluated relative to the known correspondence matrix. Further, according to the minimum residual value for each cost function, we determine for which cost function and at what parameter values the restored matrix best describes real correspondence.

We consider an application of the Message Passing Interface (MPI) technology for parallelization of the program code which solves equation of the linear elasticity theory. The solution of this equation describes the propagation of elastic waves in demormable rigid bodies. The solution of such direct problem of seismic wave propagation is of interest in seismics and geophysics. Our implementation of solver uses grid-characteristic method to make simulations. We consider technique to reduce time of communication between MPI processes during the simulation. This is important when it is necessary to conduct modeling in complex problem formulations, and still maintain the high level of parallelism effectiveness, even when thousands of processes are used. A solution of the problem of effective communication is extremely important when several computational grids with arbirtrary geometry of contacts between them are used in the calculation. The complexity of this task increases if an independent distribution of the grid nodes between processes is allowed. In this paper, a generalized approach is developed for processing contact conditions in terms of nodes reinterpolation from a given section of one grid to a certain area of the second grid. An efficient way of parallelization and establishing effective interprocess communications is proposed. For provided example problems we provide wave fileds and seismograms for both 2D and 3D formulations. It is shown that the algorithm can be realized both on Cartesian and on structured (curvilinear) computational grids. The considered statements demonstrate the possibility of carrying out calculations taking into account the surface topographies and curvilinear geometry of curvilinear contacts between the geological layers. Application of curvilinear grids allows to obtain more accurate results than when calculating only using Cartesian grids. The resulting parallelization efficiency is almost 100% up to 4096 processes (we used 128 processes as a basis to find efficiency). With number of processes larger than 4096, an expected gradual decrease in efficiency is observed. The rate of decline is not great, so at 16384 processes the parallelization efficiency remains at 80%.

This article reviews both historical achievements and modern results in the field of Markov Decision Process (MDP) and convex optimization. This review is the first attempt to cover the field of reinforcement learning in Russian in the context of convex optimization. The fundamental Bellman equation and the criteria of optimality of policy — strategies based on it, which make decisions based on the known state of the environment at the moment, are considered. The main iterative algorithms of policy optimization based on the solution of the Bellman equations are also considered. An important section of this article was the consideration of an alternative to the $Q$-learning approach — the method of direct maximization of the agent’s average reward for the chosen strategy from interaction with the environment. Thus, the solution of this convex optimization problem can be represented as a linear programming problem. The paper demonstrates how the convex optimization apparatus is used to solve the problem of Reinforcement Learning (RL). In particular, it is shown how the concept of strong duality allows us to naturally modify the formulation of the RL problem, showing the equivalence between maximizing the agent’s reward and finding his optimal strategy. The paper also discusses the complexity of MDP optimization with respect to the number of state–action–reward triples obtained as a result of interaction with the environment. The optimal limits of the MDP solution complexity are presented in the case of an ergodic process with an infinite horizon, as well as in the case of a non-stationary process with a finite horizon, which can be restarted several times in a row or immediately run in parallel in several threads. The review also reviews the latest results on reducing the gap between the lower and upper estimates of the complexity of MDP optimization with average remuneration (Averaged MDP, AMDP). In conclusion, the real-valued parametrization of agent policy and a class of gradient optimization methods through maximizing the $Q$-function of value are considered. In particular, a special class of MDPs with restrictions on the value of policy (Constrained Markov Decision Process, CMDP) is presented, for which a general direct-dual approach to optimization with strong duality is proposed.