All issues

The study of complex processes in various spheres of human activity is traditionally based on the use of mathematical models. In modern conditions, the development and application of such models is greatly simplified by the presence of high-speed computer equipment and specialized tools that allow, in fact, designing models from pre-prepared modules. Despite this, the known problems associated with ensuring the adequacy of the model, the reliability of the original data, the implementation in practice of the simulation results, the excessively large dimension of the original data, the joint application of sufficiency heterogeneous mathematical models in terms of complexity and integration of the simulated processes are becoming increasingly important. The more critical may be the external constraints imposed on the value of the optimized functional, and often unattainable within the framework of the constructed model. It is logical to assume that in order to fulfill these restrictions, a purposeful transformation of the original model is necessary, that is, the transition to a mathematical model with a deliberately improved solution. The new model will obviously have a different internal structure (a set of parameters and their interrelations), as well as other formats (areas of definition) of the source data. The possibilities of purposeful change of the initial model investigated by the authors are based on the realization of the idea of strategic reflection. The most difficult in mathematical terms practical implementation of the author's idea is the use of simulation models, for which the algorithms for finding optimal solutions have known limitations, and the study of sensitivity in most cases is very difficult. On the example of consideration of rather standard discrete- event simulation model the article presents typical methodological techniques that allow ranking variable parameters by sensitivity and, in the future, to expand the scope of definition of variable parameter to which the simulation model is most sensitive. In the transition to the “improved” model, it is also possible to simultaneously exclude parameters from it, the influence of which on the optimized functional is insignificant, and vice versa — the introduction of new parameters corresponding to real processes into the model.

The using of determining the measure of interaction between channels when choosing the configuration structure of a control system for complex dynamic objects is considered in the work. The main methods for determining the measure of interaction between subsystems of complex control systems based on the methods RGA (Relative Gain Array), Dynamic RGA, HIIA (Hankel Interaction Index Array), PM (Participation matrix) are presented. When choosing a control configuration, simple configurations are preferable, as they are simple in design, maintenance and more resistant to failures. However, complex configurations provide higher performance control systems. Processes in large dynamic objects are characterized by a high degree of interaction between process variables. For the design of the control structure interaction measures are used, namely, the selection of the control structure and the decision on the configuration of the controller. The choice of control structure is to determine which dynamic connections should be used to design the controller. When a structure is selected, connections can be used to configure the controller. For large systems, it is proposed to pre-group the components of the vectors of input and output signals of the actuators and sensitive elements into sets in which the number of variables decreases significantly in order to select a control structure. A quantitative estimation of the decentralization of the control system based on minimizing the sum of the off-diagonal elements of the PM matrix is given. An example of estimation the measure of interaction between components of strong coupled subsystems and the measure of interaction between components of weak coupled subsystems is given. A quantitative estimation is given of neglecting the interaction of components of weak coupled subsystems. The construction of a weighted graph for visualizing the interaction of the subsystems of a complex system is considered. A method for the formation of the controllability gramian on the vector of output signals that is invariant to state vector transformations is proposed in the paper. An example of the decomposition of the stabilization system of the components of the flying vehicle angular velocity vector is given. The estimation of measures of the mutual influence of processes in the channels of control systems makes it possible to increase the reliability of the systems when accounting for the use of analytical redundancy of information from various devices, which reduces the mass and energy consumption. Methods for assessing measures of the interaction of processes in subsystems of control systems can be used in the design of complex systems, for example, motion control systems, orientation and stabilization systems of vehicles.

Recently, saddle point problems have received much attention due to their powerful modeling capability for a lot of problems from diverse domains. Applications of these problems occur in many applied areas, such as robust optimization, distributed optimization, game theory, and many applications in machine learning such as empirical risk minimization and generative adversarial networks training. Therefore, many researchers have actively worked on developing numerical methods for solving saddle point problems in many different settings. This paper is devoted to developing a numerical method for solving saddle point problems in the nonconvex uniformly-concave setting. We study a general class of saddle point problems with composite structure and H\"older-continuous higher-order derivatives. To solve the problem under consideration, we propose an approach in which we reduce the problem to a combination of two auxiliary optimization problems separately for each group of variables, the outer minimization problem w.r.t. primal variables, and the inner maximization problem w.r.t the dual variables. For solving the outer minimization problem, we use the Adaptive Gradient Method, which is applicable for nonconvex problems and also works with an inexact oracle that is generated by approximately solving the inner problem. For solving the inner maximization problem, we use the Restarted Unified Acceleration Framework, which is a framework that unifies the high-order acceleration methods for minimizing a convex function that has H\"older-continuous higher-order derivatives. Separate complexity bounds are provided for the number of calls to the first-order oracles for the outer minimization problem and higher-order oracles for the inner maximization problem. Moreover, the complexity of the whole proposed approach is then estimated.

The article presents the results of numerical modeling of the vortex spatial non-stationary motion of the medium arising near the lateral and bottom surfaces of the descent module during its movement in the atmosphere of Mars. The numerical study was performed for the high-speed streamline regime at various angles of attack. Mathematical modeling was carried out on the basis of the Navier – Stokes model and the model of equilibrium chemical reactions for the Martian atmosphere gas. The simulation results showed that under the considered conditions of the descent module motion, a non-stationary flow with a pronounced vortex character is realized near its lateral and bottom surfaces. Numerical calculations indicate that, depending on the angle of attack, the nonstationarity and vortex nature of the flow can manifest itself both on the entire lateral and bottom surfaces of the module, and, partially, on their leeward side. For various angles of attack, pictures of the vortex structure of the flow near the surface of the descent vehicle and in its near wake are presented, as well as pictures of the gas-dynamic parameters fields. The non-stationary nature of the flow is confirmed by the presented time dependences of the gas-dynamic parameters of the flow at various points on the module surface. The carried out parametric calculations made it possible to determine the dependence of the aerodynamic characteristics of the descent module on the angle of attack. Mathematical modeling is carried out on the basis of the conservative numerical method of fluxes, which is a finitevolume method based on a finite-difference writing of the conservation laws of additive characteristics of the medium using «upwind» approximations of stream variables. To simulate the complex vortex structure of the flow over descent module, the nonuniform computational grids are used, including up to 30 million finite volumes with exponential thickening to the surface, which made it possible to reveal small-scale vortex formations. Numerical investigations were carried out on the basis of the developed software package based on parallel algorithms of the used numerical method and implemented on modern multiprocessor computer systems. The results of numerical simulation presented in the article were obtained using up to two thousand computing cores of a multiprocessor complex.

Sensor fusion is one of the important solutions for the perception problem in self-driving cars, where the main aim is to enhance the perception of the system without losing real-time performance. Therefore, it is a trade-off problem and its often observed that most models that have a high environment perception cannot perform in a real-time manner. Our article is concerned with camera and Lidar data fusion for better environment perception in self-driving cars, considering 3 main classes which are cars, cyclists and pedestrians. We fuse output from the 3D detector model that takes its input from Lidar as well as the output from the 2D detector that take its input from the camera, to give better perception output than any of them separately, ensuring that it is able to work in real-time. We addressed our problem using a 3D detector model (Complex-Yolov3) and a 2D detector model (Yolo-v3), wherein we applied the image-based fusion method that could make a fusion between Lidar and camera information with a fast and efficient late fusion technique that is discussed in detail in this article. We used the mean average precision (mAP) metric in order to evaluate our object detection model and to compare the proposed approach with them as well. At the end, we showed the results on the KITTI dataset as well as our real hardware setup, which consists of Lidar velodyne 16 and Leopard USB cameras. We used Python to develop our algorithm and then validated it on the KITTI dataset. We used ros2 along with C++ to verify the algorithm on our dataset obtained from our hardware configurations which proved that our proposed approach could give good results and work efficiently in practical situations in a real-time manner.

The features of the component circuits method (MCC) in modeling chemical-technological systems (CTS) are considered, taking into account its practical significance. The software and algorithmic implementation of which is currently a set of computer modeling programs MARS (Modeling and Automatic Research of Systems). MARS allows the development and analysis of mathematical models with specified experimental parameters. Research and calculations were carried out using a specialized software and hardware complex MARS, which allows the development of mathematical models with specified experimental parameters. In the course of this work, the model of a perfect-mixing reactor was developed in the MARS modeling environment taking into account the physicochemical features of the uranium extraction process in the presence of nitric acid and tributyl phosphate. As results, the curves of changes of the concentration of uranium extracted into the organic phase are presented. The possibility of using MCC for the description and analysis of CTS, including extraction processes, has been confirmed. The use of the obtained results is planned to be used in the development of a virtual laboratory, which will include the main apparatus of the chemical industry, as well as complex technical controlled systems (CTСS) based on them and will allow one to acquire a wide range of professional competencies in working with “digital twins” of real control objects, including gaining initial experience working with the main equipment of the nuclear industry. In addition to the direct applied benefits, it is also assumed that the successful implementation of the domestic complex of computer modeling programs and technologies based on the obtained results will make it possible to find solutions to the problems of organizing national technological sovereignty and import substitution.

Computer-aided assembly planning of complex products is an important area of modern information technology. The sequence of assembly and decomposition of the product into assembly units largely depend on the mechanical structure of a technical system (machine, mechanical device, etc.). In most modern research, the mechanical structure of products is modeled using a graph of connections and its various modifications. The coordination of parts during assembly can be achieved by implementing several connections at the same time. This generates a $k$-ary basing relation on a set of product parts, which cannot be correctly described by graph means. A hypergraph model of the mechanical structure of a product is proposed. Modern discrete manufacturing uses sequential coherent assembly operations. The mathematical description of such operations is the normal contraction of edges of the hypergraph model. The sequence of contractions that transform the hypergraph into a point is a description of the assembly plan. Hypergraphs for which such a transformation exists are called $s$-hypergraphs. $S$-hypergraphs are correct mathematical models of the mechanical structures of any assembled products. A theorem on necessary conditions for the contractibility of $s$-hypergraphs is given. It is shown that the necessary conditions are not sufficient. An example of a noncontractible hypergraph for which the necessary conditions are satisfied is given. This means that the design of a complex technical system may contain hidden structural errors that make assembly of the product impossible. Therefore, finding sufficient conditions for contractibility is an important task. Two theorems on sufficient conditions for contractibility are proved. They provide a theoretical basis for developing an efficient computational procedure for finding all $s$-subgraphs of an $s$-hypergraph. An $s$-subgraph is a model of any part of a product that can be assembled independently. These are, first of all, assembly units of various levels of hierarchy. The set of all $s$-subgraphs of an $s$-hypergraph, ordered by inclusion, is a lattice. This model can be used to synthesize all possible sequences of assembly and disassembly of a product and its components. The lattice model of the product allows you to analyze geometric obstacles during assembly using algebraic means.

Searching optimal trajectory of motion is a complex problem that is investigated in many research studies. Most of the studies investigate methods that are applicable to such a problem in general, regardless of the model of the object. With such general approach, only numerical solution can be found. However, in some cases it is possible to find an optimal trajectory in a closed form. Current article considers a time optimal problem with state limitations for a wheeled mobile differential robot that moves on a horizontal plane. The mathematical model of motion is kinematic. The state constraints correspond to the obstacles on the plane defined as circles that need to be avoided during motion. The independent control inputs are the wheel speeds that are limited in absolute value. Such model is commonly used in problems where the transients are considered insignificant, for example, when controlling tracked or wheeled devices that move slowly, prioritizing traction power over speed. In the article it is shown that the optimal trajectory from the starting point to the finishing point in such kinematic approach is a sequence of straight segments of tangents to the obstacles and arcs of the circles that limit the obstacles. The geometrically shortest path between the start and the finish is also a sequence of straight lines and arcs, therefore the time-optimal trajectory corresponds to one of the local minima when searching for the shortest path. The article proposes a method of search for the time-optimal trajectory based on building a graph of possible trajectories, where the edges are the possible segments of the tajectory, and the vertices are the connections between them. The optimal path is sought using Dijkstra’s algorithm. The theoretical foundation of the method is given, and the results of computer investigation of the algorithm are provided.

We investigate machine learning methods with a certain kind of decision rule. In particular, inverse-distance method of interpolation, method of interpolation by radial basis functions, the method of multidimensional interpolation and approximation, based on the theory of random functions, the last method of interpolation is kriging. This paper shows a method of rapid retraining “model” when adding new data to the existing ones. The term “model” means interpolating or approximating function constructed from the training data. This approach reduces the computational complexity of constructing an updated “model” from $O(n^3)$ to $O(n^2)$. We also investigate the possibility of a rapid assessment of generalizing opportunities “model” on the training set using the method of cross-validation leave-one-out cross-validation, eliminating the major drawback of this approach — the necessity to build a new “model” for each element which is removed from the training set.

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.