Работая с нашим сайтом, вы даете свое согласие на использование файлов cookie. Это необходимо для нормального функционирования сайта, показа целевой рекламы и анализа трафика. Статистика использования сайта отправляется в «Яндекс» и «Google»

Понятно

Научный журнал Моделирование, оптимизация и информационные технологииThe scientific journal Modeling, Optimization and Information Technology

Online media
issn 2310-6018

submit
a manuscript

metadata of articles for the last 2 years

Development of a multi-class classifier algorithm for a federated learning system operating in case of incomplete classes of local classifiers

2024. T.12. № 4. id 1731

Mikhalev P.A. Kutsakin M.A. Karamyhova O.V.

DOI: 10.26102/2310-6018/2024.47.4.021

The relevance of research is due the need to solve the problem of training multi-class classifier models used in federated machine learning system structure operating with a training data set that contains both publicly available data and confidential data that forming hidden classes. A similar problem arises in the context of training a classifier using a training data set, some of which consists of personal information or data of varying degrees of confidentiality. In this regard, this article is aimed at researching the features of the Gaussian mixture model of distributions as a way of representing hidden classes representing confidential data, as well as justifying the choice of an algorithmic method for finding maximum likelihood estimates of its parameters. The main method for solving the problem of identifying the parameters of hidden classes is a reasonably chosen two-stage iterative expectation-maximization procedure (EM-algorithm), which ensures strengthening the relationship between missing (confidential) data and unknown parameters of the data model represented by a Gaussian mixture of distributions. The article presents a diagram of the developed algorithm of a multi-class classifier for federated machine learning system, represented by parallel cycles of forming local learning models and their ensemble into a global learning model.

Keywords: federated machine learning, multi-class classification, confidential training data, gaussian mixture model of distributions, EM-algorithm

Building predictive agent models based on the inclusion of machine learning models in determining the state of agents

2024. T.12. № 4. id 1728

Lisovenko A.S. Tarasov D.A. Shishmareva A.S.

DOI: 10.26102/2310-6018/2024.47.4.017

The field of agent modeling continues to evolve towards the creation of more intelligent agents. This raises the conceptual problem of finding a balance between the determinism of agents' behavior and the ability of these agents to learn and predict their condition. One of the potential ways to solve this problem is to consider the possibility of developing an intermediate approach in the creation of agents, in which agents maintain the determinism of their behavior, but at the same time are able to predict their condition and correct behavior. The article presents a new approach to building intelligent agents, which combines the classical approach of building agents based on a priori set rules and the application of machine learning methods in the rules of agent behavior. A mathematical description of the proposed function for calculating the state of an agent using machine learning models is presented, as well as an algorithm for calculating the states of agents in the model. An example of building an agent model using the proposed approach is also given. The proposed approach makes it possible to develop agent models of complex systems in which agents are reactive but are able to predict their state and take into account the predict in determining their current state.

Keywords: agent modeling, intelligent agents, the approach of building intelligent agents, predicting the state, machine learning

Optimization of management in an organizational system with alternative supplies based on a multivariate choice model and algorithm

2024. T.12. № 4. id 1727

L'vovich Y.E. Shevyreva E.A.

DOI: 10.26102/2310-6018/2024.47.4.018

В статье предлагается оптимизационный подход к принятию решений при управлении в организационной системе с альтернативными поставками на основе модели и алгоритма многовариантного выбора. Охарактеризованы основные особенности, определяющие структуру оптимизационной модели многовариантного выбора: многокритериальность, индивидуальная потребность в поставках по каждой номенклатурной единице и каждому объекту организационной системы, альтернативность выбора поставщика. Показано, что исходной является многокритериальная оптимизационная модель, в которой критерии заданы на множестве альтернативных переменных. Обоснован эквивалентный подход к задаче оптимизации с ограничением по суммарной цели поставок для каждого объекта и целевой функции в виде средневзвешенной свертки остальных показателей, влияющих на эффективность деятельности организационной системы. Для последующей алгоритмизации многовариантного выбора целевая функция и ограничения объединены аддитивной функцией, к которой предъявляется экстремальное требование макс-мина. Алгоритмическая процедура многовариантного выбора управленческих решений сформулирована путем интеграции рандомизированного поиска на основе задачи многоальтернативной оптимизации и генетического алгоритма. Показано преимущество по трудоемкости поиска при сочетании используемых алгоритмов в режиме чередования по сравнению с известным использованием генетического алгоритма только на завершающем этапе выбора окончательного управленческого решения на множестве доминирующих вариантов.

Keywords: management, organizational system, alternative supplies, optimization, randomized search, genetic algorithm

Development of adaptive exponential min sum decoding algorithm

2024. T.12. № 4. id 1725

Zhang W. Mouhamad I. Saklakov V.M.

DOI: 10.26102/2310-6018/2024.47.4.019

This paper presents an optimized min sum (MS) decoding algorithm with low complexity and high decoding performance for LDPC short codes. The MS algorithm has low computational complexity and is simple to deploy. The MS decoding algorithm, while demonstrating a performance gap compared to the belief propagation (BP) and likelihood ratio BP (LLR-BP) decoding algorithms, shows significant potential for optimization. To improve the decoding performance of traditional MS algorithm, secondary external information is introduced into the control node (CNs) update operations of MS algorithm and optimized as adaptive exponential correction factor (AECF). The optimized MS algorithm is named as adaptive exponential exponential MS decoding algorithm (AEMS). The decoding efficiency of the AEMS algorithm for regular, irregular and LDPC codes of the Consultative Committee on Space Data Systems (CCSDS) was extensively tested, then the complexity of the AEMS algorithm was analyzed and compared with other decoding algorithms. The results show that the AEMS algorithm outperforms the offset MS (OMS) and normalized MS (NMS) algorithms in decoding performance, and outperforms the BP algorithm as the signal-to-noise ratio (SNR) gradually increases.

Keywords: LDPC, adaptive exponential algorithm, min sum, low complexity, LLR-BP

Architecture of the patient's health predicting system using agent-based modeling and machine learning methods

2024. T.12. № 4. id 1724

Lisovenko A.S. Limanovskaya O.V. Tarasov D.D. Meshchaninov V.N. Gavrilov I.V.

DOI: 10.26102/2310-6018/2024.47.4.016

The introduction of information technology in medical institutions contributes to the development of predictive, preventive and personalized medicine. The task that arises in this case is to create a software analogue of the patient, capable of taking into account his individual indicators and predicting the state of health, is still relevant. The architecture of the patient's health predicting system presented in the work is aimed at solving this problem. A distinctive feature of the system architecture is the combination of the principles of agent modeling and representation of the patient's body in the form of interacting modules, which opens up wide opportunities for modeling the health status of the patient's body. The paper describes the hierarchy of agents in the system architecture, describes the rules of agent interaction and provides a mathematical model for evaluating the effectiveness of therapeutic effects on the patient's body, the solution of which is achieved through the interaction of system agents. The prediction of the patient's health status is performed using downloadable pre-trained machine learning models, while the models are directly involved in determining the behavior of agents. The architecture of the patient's health predicting system, implemented in the form of a software package, is a powerful tool for building agent-based predicting models aimed at modeling physiological and pathological processes and effects on the patient's body.

Keywords: health predicting, patient's health, agent-based modeling, patient's digital double, modular approach

Characteristic function of an actor computing system

2024. T.12. № 4. id 1722

Zekenskii A.A. Gribkov A.A.

DOI: 10.26102/2310-6018/2024.47.4.015

The paper is devoted to the study of the problem of determining a complex feasibility indicator for an actor computing system, which can be expressed as a binary characteristic function. This function depends on the solvability and enumerability of the set of intermediate values of the parameters of the computational problem to be solved, the feasibility of the computational system, i.e. its ability to perform the entire set of necessary computational operations for a given limited time interval (computation cycle), as well as on the degree of confidence in the functional reliability and information security of the computational system, expressed in the form of an integral confidence index. The paper presents a description of the actor model of a computing system in the framework of number theory. The proposed description is based on the representation of a computing system in the form of a composition of actors – function carriers, definitions of computability of these functions, as well as solvability and enumerability of numerical sets of parameter values set for a computing system and arising in it in the process of solving the set tasks. Approaches to ensuring solvability, realisability and trust in the computational system are considered. It is stated that the choice of memory-oriented architecture of computations based on the requirement of realisability is also reasonable from the point of view of providing decidability, enumerability and ensuring trust to the computing system.

Keywords: computing system, actor model, memory-oriented architecture, feasibility, realisability, computability, solvability, enumerability, confidence

Topological skeletons-based correction of volumetric motion artifacts in optical coherence tomography and elastography

2024. T.12. № 4. id 1715

Potlov A.Y.

DOI: 10.26102/2310-6018/2024.47.4.008

An original approach to image stabilization in optical coherence tomography and elastography was presented. The key features of the proposed approach are: I) binarization and application of mathematical morphology digital operations; II) parallel construction of a topological skeleton for each optical image with an emphasis on the equivalent high- and low-level signal; III) complexing of topological skeletons; IV) comparison of a sequence of optical images by combined topological skeletons using «quench» points; V) compensation of volumetric motion artifacts by «reassembling» the original sets of interference signals. The computational efficiency of the proposed method with respect to the dynamics of interference signal acquisition by a specific device was achieved by using sequential and parallel operations. Сomputations using the central and graphical processing units, namely GPU and CPU, were combined for this. High efficiency of volumetric motion artifact correction in optical coherence tomography and elastography is ensured by robustness of topological skeletons constructed with emphasis on high-level equivalent signal to speckle noise corresponding to constructive interference (bright speckles). Topological skeletons for low-level equivalent signal are correspondingly robust to dark speckles (destructive interference result).

Keywords: optical coherence tomography, medical elastography, fiber optic probe, structural images, functional images, topological skeleton, biological tissue, tissue-imitating phantoms, volumetric motion artifacts

Improving the objectivity of the decision-making process in relation to the tasks of sowing crops

2024. T.12. № 4. id 1714

Strygin S.P. Konkina V.V. Pustovarov N.Y. Konovalcev A.S. Dudin M.A.

DOI: 10.26102/2310-6018/2024.47.4.012

The relevance of the study stems from the need to improve the efficiency and economic benefits of crop cultivation. The research in this paper is aimed at developing a decision support system that will improve the method of evaluating the accuracy of seeding units, allow pre-sowing adjustment of row crop seeders and reduce the workload of agronomists. The dotted-nested sowing method was considered, and the total coefficient of variation, dynamic coefficient of variation and accuracy were determined as criteria for assessing the unevenness of seed distribution in the row. As alternatives, soybean varieties “Alaska” and “Lisbon” of different fractions were studied at different design and operating parameters of the seeding unit, namely: the rotation speed of the seeding disk of the unit (15–55 rpm), the position of the seed wiper (from fully open hole to overlapped by 1/3 of the area of the hole), the diameter of the holes of the seeding disk (4–4.5 mm). The paper formulates the problem of decision-making theory within the framework of a specific research area. The problem is solved using the method of hierarchical analysis and complete enumeration. The article's materials are of practical use for agricultural enterprises, including pre-sowing rowed seeder settings.

Keywords: multicriteria problem, decision making, method of hierarchy analysis, weight coefficients, objective choice

Analytical modeling of a multicluster special purpose system based on several monitoring scenarios

2024. T.12. № 4. id 1713

Kamil W.A.K. Kochegarov M.V. Mutin D.I.

DOI: 10.26102/2310-6018/2024.47.4.006

The article considers the problem and formulation of the task of modeling the optimal functioning of a multicluster special purpose system (MSPS), based on multi-scenario modeling. The problems associated with the uncertainty of sources and loads in the MSPS in the energy sector are becoming increasingly apparent due to the combination of large-scale renewable energy sources and multi-energy loads. Moreover, such scenarios pose great problems for the optimal functioning of the MSPS. The distributed MSPS in the energy sector is considered as an object of research, and a functioning model based on multi-scenario modeling is proposed to account for forecasting uncertainties arising in the case of distributed electricity generation and multi-energy loads. Traditional models for optimizing the work of the MSPS usually take into account only one deterministic work scenario, which can lead to certain limitations of work strategies. When optimizing, it is necessary to balance the problems with conservative optimization results caused by extreme scenarios and the high complexity of the model caused by the large sample size of the random sample scenario. To solve the above problems, an optimization model based on multi-scenario modeling is proposed for a load-side distributed MSPS in a multicluster system. The optimization model is also applicable to account for the uncertainties associated with distributed wind and solar energy sources and the randomness of load forecasting for cooling, heating and electricity needs.

Keywords: stochastic modeling, integrated system, distributed operation, multicluster system, optimization model, load forecasting

Optimization of resource management in a regional organizational system based on predictive analysis of long-term statistical information

2024. T.12. № 4. id 1710

Gusev P.Y. Lomakov A.V. Lvovich Y.

DOI: 10.26102/2310-6018/2024.47.4.007

The article considers the rationale for the optimization approach to resource provision management in a regional organizational system, which is distinguished by the procedures for integrating the results of predictive analysis of long-term statistical information into the decision-making process. The limitations of methods of expert selection of management actions based on the analysis of the dynamics of changes in the system's performance indicators and the possibility of overcoming this limitation through a formalized representation of the dependence of the integral effect function on additional resource provision, which makes it possible to move on to the search for management solutions through optimization modeling, are shown. The formation of optimization models for the distribution of resource provision in a regional organizational system is considered according to three components: population groups, territorial entities, and time periods. For the first two components, management decisions are determined by setting and solving multi-alternative optimization problems. They allow one to determine promising subsets of population groups and territorial entities for which the need for additional resources determined by experts will give the greatest effect in future periods. Since management decisions contain an expert component along with a formalized choice, they are preliminary in nature. The final formalized decision is achieved by distributing preliminary estimates over time intervals using an optimization model of dynamic programming. It is proposed to use the results of predictive analysis in the form of prognostic models reflecting the data of statistical indicators when forming target functions of optimization models, which allows integrating them into the decision-making process when managing the distribution of resource provision in a regional organizational system.

Keywords: regional organizational system, management, resource provision, predictive analysis, forecasting, optimization, expert assessment

Neural network modeling of street and road network capacity

2024. T.12. № 4. id 1706

Sysoev A.S. Pogodaev A.K. Klyavin V.E. Zhikhoreva S.V. Borovkova G.S.

DOI: 10.26102/2310-6018/2024.47.4.010

The problems of increasing the number of personal transportation vehicles in urban agglomeration as well as the number of cargos lead to applying Intelligent transportation systems based on Machine Learning techniques and Artificial Intelligence models to create strategies to reduce congestion and to prevent accidents. An important parameter of transportation system showing the effectiveness of using existing urban infrastructure is the capacity of the planned route. The paper is devoted to the creating model of urban route capacity based on the capacities of its elements, they are namely stretches and intersections. The approach to create such model is Mathematical Remodeling, where feed-forward neural network is chosen as a unified class to substitute models of different heterogeneous classes during modeling. It is proposed to use index of route capacity to form data sets for model fitting. The given numerical examples show how the proposed approach can be applied. The capacities of three planned routes are estimated and the best route is chosen, the efficiency criterion is traffic flow volume to capacity ratio. The prospective issue of the presented study is analyzing sensitivity of the created model to identify the parameters of route elements affecting the most to the capacity and to control them increasing the total efficiency of the system.

Keywords: neural networks, remodeling, capacity estimation, street and road network, mathematical modeling

Efficient resource allocation in geodistributed heterogeneous dynamic computing environments

2024. T.12. № 4. id 1704

Klimenko A.B.

DOI: 10.26102/2310-6018/2024.47.4.011

Currently, the management of computing resources in geo-distributed heterogeneous dynamic computing environments is a non-trivial scientific problem. Due to the complexity of such systems, the distribution of computing resources becomes a computationally hard problem, usually multi-criteria, with nonlinear constraints, integer or mixed-integer. The solution of such problems produces some additional costs of system exploitation. In addition, the property of geo-distribution also introduces additional resource costs that arise during data transit between computing subtasks in the case when transit sections of the network are involved and the route length is more than one section. The purpose of this study is to implement effective management of computing resources based on the criterion of using computing resources – both in the process of their distribution and in solving a computational task in a computing environment. To achieve the goal of the study, a new formulation of the computational resource distribution problem has been developed, which takes into account the properties of heterogeneity, dynamics and geo-distribution of the computing environment and is distinguished by the presence of controlled parameters that determine the resource costs both for data transmission over the network and for solving the computational resource distribution problem. A method has been developed that allows solving the formulated problem, which includes the stages of developing a metaheuristic repository and its use. The results of the conducted modeling allow us to conclude that the developed method is promising – the computing resource usage for resources distribution has decreased by 28 times with a loss in the quality of the resulting solution of up to 10%.

Keywords: resource allocation, distributed computing, distributed computing management, dynamic computing environment, optimization, metaheuristics

GIS-oriented classification modeling for management in organizational systems with a heterogeneous structure of spatial elements

2024. T.12. № 3. id 1702

Linkina A.V. Ryndin N.A.

DOI: 10.26102/2310-6018/2024.46.3.030

The article presents theoretical approaches to formalizing problems of optimizing the management of complex organizational systems, taking into account GIS-based classification modeling. It is shown that models of complex systems with spatial characteristics can be classified as stochastic due to the wide variability of input parameters and their random distribution (both in space and time). At the same time, it is clarified that spatial characteristics can be considered, in fact, both geographic reference and any other attribute information about the objects of the system under consideration. The problem of presenting a model of a complex organizational system of an agricultural profile is solved, taking into account the hierarchy of characteristics affecting the system. It is clarified that a feature of the system under consideration is the dependence of stability not only on the structure and parameters of the system (as for linear systems), but also on the magnitude of the initial deviation of the system from the equilibrium position, based on the phase space method, widely used in the theory of automatic control. The problem of finding the optimal (equilibrium) state of a complex organizational system of an agricultural profile is formalized, the choice of significant characteristics and their combined influence on the target variable are justified. 3 main types of input variables are defined. It has been studied that, taking into account the Pareto efficiency when predictors influence each other, the constructed model will make it possible to find optimal solutions in a multicriteria system, taking into account the ranking of the significance and weight of features. The possibility of complicating this problem is noted by the fact that with GIS-oriented classification modeling, the heterogeneous structure of spatial elements can solve the inverse problem - finding the system at a minimum in the case where the optimal option is considered to be the absence of influence on the system of individual input parameters when leveled by other input features.

Keywords: optimization of management of complex systems, GIS-oriented approach, classification modeling, formalized information model, spatial features

Comparison of the efficiency of the random forest algorithm and artificial neural networks of the RNN class in the problem of managing the process of structural-parametric synthesis of business process models based on a genetic algorithm

2024. T.12. № 4. id 1701

Petrosov D.A.

DOI: 10.26102/2310-6018/2024.47.4.020

The results presented in this study are relevant for solving the problem of increasing the efficiency of the genetic algorithm in problems related to the use of big data. In the framework of most existing approaches to the application of the evolutionary procedure, efficiency improvement methods are used that are based on classical approaches aimed at pre-setting the operating parameters of the genetic algorithm operators in a specific subject area. At the same time, when working with big data, there is a need to stop and restart the genetic algorithm to obtain the best solutions, since the population of the evolutionary algorithm can be in local extremes and / or the efficiency of the increase in the quality of individuals does not allow finding the required solution in a given time interval. In this case, it becomes relevant to develop new methods that allow you to manage the search process. One of the approaches to solving this problem is the use of the mathematical apparatus of artificial neural networks of the RNN class, which have proven their effectiveness in solving the classification problem and can be used to identify the state of the population of the genetic algorithm. In addition to the approach based on the use of artificial neural networks, it is relevant to assess the possibility of using the "random forest" algorithm to solve the problem of recognizing the state of a population and making decisions on changing the operating parameters of the genetic algorithm operators directly in the process of work, which will allow influencing the trajectory of the population in the solution space. Within the framework of this article, the results of computational experiments on solving the problem of classifying the state of a population of a genetic algorithm by two modern methods will be considered: the "random forest" algorithm and the artificial neural network RNN, the modeling of which is performed using a graph approach based on the theory of Petri nets, which will allow combining the developed models with the model of a genetic algorithm adapted to solving the problem of structural-parametric synthesis using nested Petri nets.

Keywords: mathematical modeling, business processes, systems analysis, petri net theory, genetic algorithm, artificial neural networks, random forest algorithm

Method for solving the problem of planning jobs in an IT project and assigning specialists to it

2024. T.12. № 4. id 1698

Dyatchina A.V. Oleinikova S.A.

DOI: 10.26102/2310-6018/2024.47.4.009

This article is devoted to the development of an iterative approach that provides a simultaneous solution to the problem of planning individual works of an IT project with the assumption of their possible correction and assignment of specialists to these works. At present, the class of project management problems has been studied in sufficient depth, and various methods for forming schedules from the point of view of various criteria (fastest completion, cost, etc.) have been obtained. However, IT projects differ from standard projects in the periodic revision of tasks (error correction, clarification of the result with the customer, etc.), which requires changes in the mathematical apparatus of the problem. In addition, the time of execution of a particular work will depend on its performer. This feature is taken into account extremely rarely, which allows the decision maker to solve the planning problem and the assignment problem separately. However, the analysis of the subject area shows that not only the duration of a specific work will depend on the specialist, but also the probability of its error-free execution the first time. Therefore, it is advisable to take such features into account when simultaneously solving the planning problem and the assignment problem. In this regard, it is necessary to develop a method that allows taking into account these nuances of the problem under study and ensures the best solution to both the planning and assignment problem from the point of view of the objective functions. The method is based on a combination of basic approaches to solving the assignment problem with the critical path method and PERT. As a result, an iterative algorithm for solving the problem of forming a project schedule and assigning performers to all work was obtained, taking into account the stochastic nature of the time of execution of individual works, as well as their possible correction.

Keywords: project management, assignment problem, random service time, PERT, correction of jobs

Developing a software architecture to support decision making when selecting design strategies from multiple alternatives

2024. T.12. № 4. id 1696

Kalach A.V. Borzykh N.Y. Smolentseva T.E.

DOI: 10.26102/2310-6018/2024.47.4.004

The article examines the stages of building software architecture for multi-criteria analysis of design strategies, taking into account the competencies of decision-makers (DMs). The software considered in the work is based on an algorithm for managing the input set of criteria and is aimed at automating the process of selecting the optimal strategy in project organizations. The logical structure of a relational database is described, ensuring efficient storage and processing of information about DMs, criteria, alternatives, and their evaluations. A modular software architecture implemented in C# using the .NET Framework and the MVVM pattern is presented. Special attention is paid to the multi-criteria analysis module, which implements a combination of the Analytic Hierarchy Process, PROMETHEE, and TOPSIS methods, allowing for various aspects of multi-criteria optimization to be taken into account. The software provides flexible tools for managing criteria, considers the interests of various DMs, and easily adapts to changes in preferences. The results of a comparative analysis of the developed product's efficiency are presented, demonstrating a significant reduction in time for strategy analysis compared to manual processing. The proposed software architecture aims to improve the accuracy and validity of decisions made, reduce time and resource costs, and enhance project management quality in conditions of multi-criteria and uncertainty.

Keywords: multi-criteria analysis, decision support, software, DMs, AHP, PROMETHEE, TOPSIS, modular architecture, project organizations

Concept of dynamic positioning system for unmanned small-class underwater vehicles based on visual odometry

2024. T.12. № 3. id 1695

Aliagaev A.R. Azhmukhamedov I.M. Khomenko T.V.

DOI: 10.26102/2310-6018/2024.46.3.029

The article is devoted to the actual problem of underwater robotics - the problem of dynamic positioning of unmanned underwater vehicles of small class. Particular attention is paid to the methods of navigation of unmanned underwater vehicles and methods for creating a dynamic positioning system, including methods for the synthesis of an observer, a regulator and methods for distributing control actions on the propulsion and steering complex of unmanned underwater vehicles. It is revealed that in the existing dynamic positioning systems, expensive hydro acoustic navigation systems and Doppler speed meters are mainly used to generate feedback on the position and speed of unmanned underwater vehicles. Not all unmanned submersibles of the small class of the budget segment are equipped with such systems, while video systems and inertial sensors are present in almost every device. With the development of onboard computing facilities, it becomes possible to use visual odometry algorithms for navigation of unmanned underwater vehicles based on data from a video system as an alternative to hydro acoustic navigation in the task of dynamic positioning. The concept of architecture of the system of dynamic positioning of unmanned underwater vehicles of small class based on visual odometry is proposed, which helps to reduce the cost of navigation equipment and allows to increase the productivity of underwater technical work.

Keywords: dynamic positioning, unmanned underwater vehicle, navigation system, visual odometry, control system

Approximation of an elliptic operator with a singularity in the space of functions specified on the graph

2024. T.12. № 4. id 1671

Приходько И.В. Perova I.V. Gunkina A.S. Part A.A.

DOI: 10.26102/2310-6018/2024.47.3.003

Was proposed an approach to approximation of an elliptic operator used in describing mathematical models of transfer processes of continuum and in problems of controlling elastic vibrations of network-like structures. To ease the problem of studying the presented material, i.e. to simplify the mathematical symbolism of grid functions, the space variable of functions of the domain of definition of the elliptic operator changes on the oriented geometric graph - star, which is not a restrictive circumstance, because an arbitrary graph (in applications – a network) is a collection of stars that differ from each other only in the quantity of edges. An algebraic system and its corresponding finite-dimensional operator are formed, the properties of this operator are established and examples of constructing (and analyzing) difference schemes for the heat transfer equation and the oscillation equation (wave equation) with a space variable changing on a graph (network) are given. In this case, the optimal control problem is reduced to a finite moment problem, which opens the way to obtaining a numerical analysis that does not depend on the dimension of the control vector, it is only necessary to know a limited number of grid eigenfunctions of the finite-difference analogue of the elliptic operator.

Keywords: elliptic operator on a graph, finite-dimensional analog, difference scheme with singularities, optimization of the elliptic operator

Mathematical support for selecting directions for the development program of an organizational system based on a combination of a randomized search algorithm and a genetic algorithm with adaptation

2024. T.12. № 4. id 1668

Ivanov D.V. Lvovich Y.E.

DOI: 10.26102/2310-6018/2024.47.4.013

The article is devoted to the development of an optimization approach to the selection of directions for the optimization system development program. It is shown that the formalization of the process of optimal selection of a management decision when forming a development program leads to a model of multi-alternative optimization. It is advisable to implement the solution of the optimization problem using a directed randomized search. However, in this case it is only possible to form a set of dominant options, which requires the use of expert assessment to select the final option for distributing organizational system objects between the directions of the development program. Another approach is proposed based on a combination of a randomized search algorithm and a genetic algorithm with adaptation. In order to integrate these algorithms into a single iterative scheme for searching for an optimal solution, first of all, the condition for the transition from the first iterative process of a randomized search to the formation of a genetic algorithm population with elements corresponding to random values of alternative variables is substantiated. Parents are selected from this population and a transition to the second iterative process of probabilistic selection of the best option for combining crossbreeding and reproduction schemes is carried out. It is shown that a two-level adaptive algorithm using the values of the fitness function corresponding to the structure of the original optimization problem is acceptable for correcting the probability characteristics from one iteration process. The third iteration process is aimed at including seven mutation options in the selection of genetic algorithm elements. It is shown by what condition the listed search processes are stopped for the subsequent selection of the optimal management solution.

Keywords: organizational system, development program, multi-alternative optimization, randomized search, genetic algorithm, adaptation

Comparison of optimization methods in simulation models of complex organizational and technical systems

2024. T.12. № 3. id 1665

Beketov S.M. Zubkova D.A. Redko S.G.

DOI: 10.26102/2310-6018/2024.46.3.027

The relevance of the study is due to the need to improve the effectiveness of management decisions in complex organizational and technical systems. The problem of this study is to choose the most appropriate optimization method for specific tasks of organizational systems. The purpose of the article is to compare modern methods of optimization of complex organizational and technical systems, in particular, in the model of the transport system. Special attention is paid to minimizing the target function, which takes into account such parameters as passenger traffic, passenger waiting time, vehicle loading and the impact on the traffic situation. The study analyzed suitable optimization methods and implemented software implementation of optimization approaches for the transport system in the Python programming language. The practical part allows evaluating the effectiveness of each method in terms of the results of the objective function, the adequacy of the selected model parameters and the execution time of the algorithm. The results showed that the methods of particle swarm and differential evolution provide the best minimization of the objective function with optimally selected parameters of the range of motion, speed and capacity of the vehicle, however, these optimization methods require a lot of time for calculations. The materials of the article are of practical value for specialists in the field of process optimization and transport planning, offering recommendations on the choice of optimization methods depending on the goals and conditions of the task.

Keywords: optimization methods, organizational and technical system, simplex method, annealing method, double annealing method, differential evolution method, particle swarm method

Algorithm of formation of training and test samples for data character analysis

2024. T.12. № 4. id 1663

Chirkov A.V.

DOI: 10.26102/2310-6018/2024.47.4.014

The article presents an adaptive algorithm for forming training and test datasets for the ANFIS system, used to diagnose the technical condition of electrical equipment. A key feature of the proposed approach is the consideration of temporal dependencies and anomalous data, which enhances the accuracy and completeness of identifying faulty equipment states. The process of testing the algorithm on synthetic data, including vibration, temperature, current, and voltage parameters, is described. The conducted analysis shows that adaptive data partitioning improves the system's ability to identify anomalies compared to the classical method of dataset partitioning. The algorithm is highly applicable for equipment diagnostics in industries where it is crucial to account for dynamic changes in parameters and rare anomalous events.To assess the algorithm's efficiency, it was compared with traditional dataset partitioning methods. The experiment demonstrated that the proposed method enhances the accuracy of classifying anomalous equipment states. Additionally, the algorithm reduces the likelihood of false positives when detecting faults. A notable feature of the development is its ability to adapt to various types of equipment, making it a universal solution for diagnostics in different industrial sectors. The algorithm's future applications are related to its integration into predictive maintenance and monitoring systems, which will increase equipment reliability and reduce repair and maintenance costs.

Keywords: ANFIS, neuro-fuzzy model, adaptive dataset formation, equipment diagnostics, time series, anomalous data, industrial diagnostics, electrical equipment

Mathematical modeling of relations between agents of an organizational system

2024. T.12. № 4. id 1661

Rossikhina L.V. Betskov A.V. Makarov V.F. Kondratiev V.D.

DOI: 10.26102/2310-6018/2024.47.4.001

The article discusses the main types of relationships (conflict, assistance and independence) active agents, the manifestation of which is possible when they interact in the organizational system. The agent's activity is understood as the possibility of independent goal-setting, according to which he chooses actions and his unscrupulous behavior. To characterize active agents, the concept of a utility function is introduced, which determines the agent's choice of actions that allow its usefulness to be maximized, as a rule, this is profit. The mathematical formalization of the relations of active agents is given for the option of achieving the common goal of the organizational system, as well as taking into account the achievement of local goals by active agents. To describe the interaction of active agents in the process of achieving a common goal, a matrix of the state of the organizational system is proposed, which allows to identify the existing cores of conflict, independence and assistance between active agents. The elements of the matrix are quantitative estimates of the set of agent relationships. To determine quantitative estimates of the set of agent relationships, an algorithm based on the calculation of the relative discrepancy of utility functions has been developed, which allows determining the nature and degree of agent relationships. The author's classification of agent relations according to the degree of their manifestation is proposed. An example illustrating the practical implementation of the algorithm is given.

Keywords: agent, multiple relationships, conflict, assistance, independence, utility function, quantitative assessment of relationships, matrix of the state of the organizational system

Development of intelligent models for proactive protection of critical infrastructure of the financial sector using the example of information support for contract systems

2024. T.12. № 4. id 1652

Korchagin S. Rubtsov D. Bespalova N. Serdechny D.

DOI: 10.26102/2310-6018/2024.47.4.005

The paper proposes an approach to developing intelligent models of proactive protection focused on information support of contract systems in the financial sector. A methodology for developing intelligent models is presented, which includes components for monitoring, forecasting and preventing cyberattacks. The proposed methodology formed the basis for practical implementation in Python using the Numpy and Scirket Learn libraries. Particular attention is paid to the use of advanced machine learning and artificial intelligence algorithms to identify and prevent potential threats in real time. As a practical example, the application of the developed intelligent models to protect the information support of contract systems used in the financial sector is considered. Key vulnerabilities, potential attacks and methods for their proactive detection and blocking are analyzed. The results of the study are confirmed by the data of a computational experiment and demonstrate the high efficiency of the proposed approach in increasing the resilience of the critical information infrastructure of the financial sector to cyberattacks. The implementation of intelligent models of proactive protection allows us to significantly reduce the risks of compromising the integrity and availability of key data, minimize financial and reputational losses, and predict and prevent potential threats.

Keywords: mathematical modeling, cybersecurity, intelligent models, proactive defense, financial sector, government contracts, critical information infrastructure

Fractal approach to Monte Carlo based numerical simulation of photon transport in biological tissues

2024. T.12. № 3. id 1648

Potlov A.Y.

DOI: 10.26102/2310-6018/2024.46.3.022

The paper presents a computationally efficient approach to mathematical modeling of the photon migration process in biological tissues. In this case, the tissues of living organisms are described as strongly scattering media with pronounced anisotropy and a relative refractive index higher than that of air. The proposed approach is a modified version of the Monte Carlo statistical testing method, in connection with which the calculation of the photon mean free path, the probability of an absorption or scattering act, energy loss during an absorption act, a new direction of motion in the case of an act of scattering and the behavior of a photon at the boundary of the modeled object or its separate relatively isolated section are performed according to classical formulas. The main distinctive feature of the proposed solution is the description of a photon packet as a tree-like fractal. In this case, the reference trajectory is calculated in the classical way, and the rest are completed according to the principle of self-similarity, adjusted for the presence or absence of areas of abrupt change in optical properties. This approach allows increasing the computing performance by reducing the number of photons in a packet with a proportional increase in the number of packets under consideration. The proposed solution is intended for use in the development of new and improvement of known methods of optical tomography and elastography.

Keywords: mathematical modeling, high-performance computing, biological tissues, optical tomography, optical elastography, monte Carlo method, photon trajectories, fractals

Ensuring functional reliability of telecommunication systems based on topological resource

2024. T.12. № 3. id 1647

Gvozdev V.E. Guzairov M.B. Rakipova A.S. Galimov R.R. Prykhodko V.E. Teplyashyn P.N.

DOI: 10.26102/2310-6018/2024.46.3.024

Modern special-purpose communication and computing systems perform tasks, first of all, to deliver information between spatially distributed bodies involved in solving network-centric control problems. Modern communication and computing systems are characterized by a transition to a hybrid structure, a decentralized network architecture, which predetermines the formation of a single information space based on the integration of different departmental affiliations information systems, and created on the basis of various methodological and technological platforms. In this work, topological and resource approaches are used as approaches that allow us to study the properties of local information systems from a unified methodological position. The conceptual basis was the proposition that a promising approach to routing in conditions of dynamic changes in the state of a telecommunication system is the formation of a backup message delivery paths set, which will increase the reliability and stability of the system. The features of the backup paths formation are determined, limiting the possibility of mechanical transfer of backup methods developed for technical systems to the TCS area. A metric has been proposed that allows one to analyze possible paths for transmitting messages between the source node and the destination node based on a set of static and dynamic characteristics.

Keywords: communication and computing systems, functional reliability, telecommunication systems, topology, routing, dynamic structure

The forecast of the prevalence of cancer among residents of the Moscow region based on a regression model

2024. T.12. № 3. id 1644

Stepanov V.S.

DOI: 10.26102/2310-6018/2024.46.3.023

The article makes an attempt to identify the relationship between cancer prevalence in urban areas and several environmental factors, taking into account a demographic indicator. The regression dependence of the prevalence of oncologic diseases in the territories of urban districts of the Moscow region and several districts of the capital with the proportion of elderly residents and a number of sanitary and hygienic indicators of the territories has been established. The complex of factor explanatory variables included the indicator of atmospheric air pollution of the territory, two variables with the concentration of surface ozone and benz(a)pyrene on it, qualitative variables in terms of the level of its man-made pollution and the volumes of polluted water discharge, the proportion of elderly population. Daily cigarette smoking by adults is also taken into account. On this basis, a regression model with a variable structure is constructed, which has a determination coefficient of 98.5% and an approximation error below 2%. The model parameters were estimated using the least squares method based on data for 51 urban districts of the region and 5 districts of Moscow. The presence of lags in the factors makes it possible to make a forecast of the number of people suffering from tumors of any localization, in the municipal context and with a planning horizon of 1 year. Based on the created model, it is possible to plan primary prevention measures more effectively and allocate medical resources.

Keywords: regression model, atmospheric air pollution, discharge of polluted waste water, benz(a)pyrene, surface ozone, suspended particles, technogenic pollution, malignant neoplasm, city district, municipality

Development of a mivar expert system for planning shop resources and analysis of deviations

2024. T.12. № 3. id 1641

Varlamov O.O. Zhang X. Baldin A.V. Myshenkov K.S. Sidorenko E.V.

DOI: 10.26102/2310-6018/2024.46.3.017

To create mechanical engineering artificial intelligence, mivar technologies of logical artificial intelligence are used. The production process is often accompanied by a large number of events, and various types of deviations and interference directly or indirectly affect the stable and efficient operation of production, and also lead to a decrease in product quality. Predicting variances and disturbances in production planning is a research problem that is the basis of resource planning for production systems. There is a known approach to solving optimization problems of resource allocation of production systems based on the construction of logical inference in a mivar knowledge base, which represents a resource allocation plan. This paper analyzes the deviations and/or disturbances caused by production interference on the shop floor, namely materials, personnel, equipment, processes, and so on, and proposes a definition of production interference in the shop floor production environment. A significant degree of interference results in delays in product deliveries, reductions in quality levels and other deviations from the planned production plan. A mivar expert system has been developed to predict deviations in production processes after planning workshop resources. The expert system was developed using the software package KESMI Wi!Mi "Razumator". Deviations in the production environment were analyzed, a system of factors influencing deviations was established, and a corresponding mivar model for predicting production deviations in the workshop was built. The use of a mivar expert system effectively and quickly solves the problem of decision support based on flexible complex calculations when calculating weights. Therefore, the mivar expert system plays a critical role in predicting interference in the planning of workshop operations, significantly increasing the efficiency of the entire enterprise management system.

Keywords: mivar networks, mivar expert system, decision support system, KESMI, razumator, big knowledge, optimization, distribution of production resources of the workshop, deviations in production processes

Features of applying deep learning methods to detect small objects in video in rainy conditions

2024. T.12. № 3. id 1640

Shtekhin S.E. Stadnik A.V.

DOI: 10.26102/2310-6018/2024.46.3.019

This paper discusses methods for detecting small objects in video when recognizing manual labor operations that take place outdoors, in the open air, and are affected by weather conditions. Approaches to improve the accuracy of detecting such objects in adverse weather conditions, such as rain, are considered. This paper explores a two-stage approach. At the first stage, computer vision methods and deep learning methods such as convolutional neural networks are used to identify and classify various weather conditions in video. At the second stage, when adverse weather conditions are detected, a study is conducted of various deep learning methods for filtering weather conditions in video. The main focus is on assessing the impact of various filtering methods on the accuracy of detecting small objects. The paper considers the applicability of this approach to detecting small tools in video data when recognizing manual labor operations performed during repair and maintenance of a railway track. The obtained results can be useful in the study of labor processes occurring outdoors, in algorithms for recognizing manual labor operations in video data.

Keywords: deep learning, transformer, object detection, recognition of weather conditions on video, filtering of weather conditions, filtering of noise in the image, neural networks, technological operations

Artificial intelligence technologies in the rehabilitation of people with disabilities: analysis of the publication flow

2024. T.12. № 3. id 1638

Sufelfa A.R. Petrishcheva K.N. Shcherbina K.K. Ponomarenko G.N.

DOI: 10.26102/2310-6018/2024.46.3.026

Artificial intelligence technologies are actively used in medicine, which significantly expands the possibilities of disease prevention, diagnosis, treatment and monitoring. Rehabilitation of the disabled, located at the intersection of medicine and the social sphere, traditionally adopts innovative development approaches from the healthcare sector. The issues of using artificial intelligence technologies in the rehabilitation of the disabled, taking into account the specifics of rehabilitation measures for different patients, require study. The purpose of the work is to analyze the foreign studies on the topic of using artificial intelligence technologies in the rehabilitation of the disabled and to identify the most used artificial intelligence methods for subsequent implementation in practice. Publications from the international medical database PubMed over the past 5 years (from January 2019 to May 2024) were analyzed. According to the analysis among artificial intelligence technologies broken down by information processing method, some of the main ones were machine learning, deep learning and neural networks, with different ways of combining all three methods. Most often, these methods are used to create health monitoring and prediction systems (based on machine learning) and (medical) decision support systems (based on neural networks). They have a high potential for use in the rehabilitation of people with disabilities in the areas of medical and social examination, developing individual rehabilitation programmes and monitoring the effectiveness of rehabilitation measures.

Keywords: artificial intelligence, data processing methods, machine learning, rehabilitation, people with disabilities, publication analysis, decision support system, health indicators monitoring

Language models and ontologies, security threats in distributed system

2024. T.12. № 3. id 1634

Donskikh N.I.

DOI: 10.26102/2310-6018/2024.46.3.016

Research in the field of large language models and natural language processing systems has intensified due to the emergence of new, latent and serious risks, for example, violations of the output generation processes, malicious requests in automatic mode. Synergistic scenarios for large language models are being developed. The main hypothesis taken into account in this study is the possibility of insurance (with a given probability) from the generation of prohibited content and its "mixing" with the user query, taking into account ontological properties and connections to improve the quality of search in practical tasks, for example, using an ontology library. Methods of analysis-synthesis, modeling-forecasting, expert-heuristic, probability theory and decision-making were used. The main results of the article: 1) analytics on the problems of applying large language models in achieving stability in the system infrastructure (a table of key methods was proposed); 2) a language model of network infrastructure stability based on estimates of distributions when mixing words is proposed, which uses the Bayesian method; 3) a similar language model was proposed and studied on the basis of an expert-heuristic approach to assessing risks (uncertainties in the system), in particular, using an information-entropy approach. Research can be developed by complicating models (hypotheses) and the "depth" of risk accounting.

Keywords: large language models, resilience, risks, information security, governance