Large-scale systems control
ISSN (print): 1819-2440, ISSN (online): 1819-2467
Media registration certificate: ЭЛ № ФС 77 - 44158 от 09.03.2011
Founder
Institute of Control Sciences of the Russian Academy of Sciences
Editor-in-Chief
Novikov D.A., Academician of RAS, Doctor of Sc.
Frequency / Access
6 issues per year / Open
Included in
White list (4th level), Higher Attestation Commission List, RISC, MathNet
Edição corrente
Nº 117 (2025)
Systems analysis
Self-reflexion and learning
Resumo
Self-reflection is the subject's reflection and analysis of their own understanding of reality, the principles and techniques of their activities, the correspondence of their results to their goals, and ways to improve the effectiveness of their activities. This article examines several models of self-reflection, described in terms of learning processes. It is shown that, with a non-decreasing learning curve dependent on the total learning time, the optimal strategy is to maximize the initial learning period. That is, if the conditions for acquiring and mastering experience do not change (during the activity, the subject does not receive qualitatively new information), then self-reflection is meaningless. Self-reflection is justified if, for example, the individual learning rate increases during productive activity. Hierarchical learning models are proposed, in which the achieved values of the learning criterion determine the learning rate at a higher level of the hierarchy, what can also be considered an effect of self-reflection.
Large-scale systems control. 2025;(117):7-18
7-18
A model for representing an individuals' beliefs and attitude using cognitive maps
Resumo
The research is based on the analysis of discourse, which is a common means of socio-political communication. Such an analysis is of particular importance in man-aging socio-political situations in the field of international relations. In the context of the tasks of managing socio-political communication using methods of knowledge representation and processing, a model for representing the beliefs and attitudes of a subject is proposed. The model is based on a cognitive map of the causal relation-ships between factors (concepts) that are significant in the subject's worldview, which verbalizes his belief structure in a given situation. A distinctive feature of the model is the inclusion of valence in it, a parameter for taking into account emotionally colored assessments of the subject (attitudes) in the assessed situation. This model is designed to solve the problems of explication and analysis of the stability and consistency of the structure of beliefs and attitudes of the subject in the assessed situation, analyzing the dynamics of its change through structural analysis. To implement the structural anal-ysis, a system of criteria for emotional consistency and a number of structural indica-tors are proposed to assess the significance and emotional consistency of sequences of statements derived from the model that determine the preferences of the subject when making decisions in a situation. In conclusion, generalizing conclusions are given and the further direction of experimental research is outlined.
Large-scale systems control. 2025;(117):19-51
19-51
Analysis of the three phase queuing system with common buffer
Resumo
This paper examines the non-stationary performance metrics of a three-phase queuing system with a Poisson input flow, exponentially distributed service time across all phases, and a constraint on the total size of the shared buffer. A system of Kolmogorov differential equations is written using specially introduced functions that account for the system's operating principles. A probability translation matrix method is used to solve the system of equations. Expressions are derived for finding the loss probability and system performance. A system with a buffer size of three is considered as an example. The duration of the transient mode is analyzed as a function of the ratios of service rates in individual service phases. It is concluded that the service rates in the first and second phases have the greatest impact on the duration of the transient mode. The dependences of the maximum values of the non-stationary loss probability and the corresponding stationary probabilities for various service rates are analyzed. The analysis of the system performance metrics is conducted for parameters corresponding to modern optical networks. The obtained conclusions are of interest for the design of high-performance computing systems.
Large-scale systems control. 2025;(117):52-73
52-73
Fuzzy optimization of aircraft fleet management taking into account transfer passenger flows
Resumo
A variant of the problem of joint optimization of the structure and number of aircraft fleet and the distribution of aircraft by air lines for an airline planning to perform mass transfer transportation of passengers based on a hub airport has been solved. It is assumed that the task is solved by the airline at the stage preceding the performance of transportation, when the demand for transportation can be known only approximately. In this case, the projected demand levels are determined by expert assessments and should be considered as fuzzy numbers. Joint optimization according to the criterion of economic efficiency is formulated as an integer mathematical programming problem with a fuzzy criterion and crisp constraints. Using the defuzzification technique, the fuzzy problem is reduced to an ordinary mathematical programming problem that can be solved in an acceptable amount of time using available software. Based on the IBM ILOG OPL software package, a solution has been obtained for a series of model examples of a problem with demand levels specified in a fuzzy and "crisp" form. The comparison revealed significant differences between the most significant results of solving the optimization problem with fuzzy and "crisp" initial data. An attempt to replace the solution obtained taking into account the vagueness of the passenger forecast with a "crisp" solution leads to a significant deterioration in the target function. All this indicates that it is advisable to take into account the vague uncertainty of the initial data. The proposed fuzzy model can be used to increase the efficiency of decisions made at the design stage of such promising air transport systems as transfer air transportation systems based on hub airports.
Large-scale systems control. 2025;(117):74-102
74-102
Design of experiment for sensitivity analysis of mathematical models from different classes
Resumo
The article focuses on design of experiment for conducting Sensitivity Analysis of mathematical models used in forecasting and controlling complex systems. Special attention is given to cases where factors are unevenly distributed in space, which is typical for problems with nonlinear dependencies, local features, and high computational complexity. In such situations, the application of mathematical remodeling is justified, whereby models with a complex structure are replaced (remodeled) by objects of a selected remodeling class that have a predefined structure, which allows unificating system research. The purpose of the study is the development and comparison of the design of experiment strategies aimed at improving the efficiency of Sensitivity Analysis. Methods adapted to uneven data distribution are considered. The foundation of sensitivity research is the analysis of finite fluctuations, built upon the application of the Lagrange mean value theorem. Numerical experiments on a test function and its neural network approximation confirmed that the proposed algorithms (central composite design and an adaptive Latin hypercube sampling-based method) enable highly accurate identification of significant factors, aligning with classical methods (Sobol indices, Morris method), while significantly reducing computational costs. It is shown that the remodeling approach refines sensitivity estimates and ensures a unified analysis procedure for models of complex structure.
Large-scale systems control. 2025;(117):103-118
103-118
Mathematical control theory
Branch-and-bound algorithn for solving problem of minimizing fees for external resources
Resumo
We consider a problem of scheduling jobs performed on a single machine. Precedence relations between jobs are established, as well as subsets of jobs that require additional external resources, for which a fee is charged. For each external resource, the extreme (first and last) job to be executed using that resource is uniquely determined. It is necessary to order the jobs without violating the precedence relations and minimizing total rent payments. We have proved that this problem is NP-hard in the strong sense, even if the processing times of all jobs are the same and the prices of all external resources are equal. Based on the properties of the objective function, lower bounds and the branch-and-bound method are proposed to solve the problem. In this method, enumeration is performed according to feasible permutations of extreme jobs using external resources. The computational experiment showed the efficiency of the proposed lower bounds of the objective function, which make it possible to cut off unpromising branches in the search tree. We also determined the type of input data for which the developed method is more successful than another known variant of exact methods that enumerate all jobs. In particular, for large-dimensional problems with fewer than 20 external resources, this method proves to be more efficient than CP Optimizer solver which is based on constraint programming.
Large-scale systems control. 2025;(117):119-140
119-140
Calculation and determination of the stability area of an educational anthropomorphous robot
Resumo
This article discusses the development and study of the educational anthropomorphic robot of the STEM Academy kit. The robot design features increased mass and mass redistribution with heavier feet to lower the center of mass in order to improve stability. The Denavit – Hartenberg technique was used to analyze the kinematic characteristics. A dynamic model wascompiled to calculate the external forces acting on the robot, taking into account the gravity forces acting on the drives based on the Newton – Euler method. A model wasconstructed for the dependence of the zero moment point on the servo rotation angles taking into account the position of the center of mass of each link. An experimental determination of the zero moment point was carried out using strain gauges, the results of which were compared with theoretical data. A new term, the "region of guaranteed static stability" (ROGSS), was introduced, which is a subset of the fundamental concept of the "region of stability", supplemented by taking into account technological factors. An experimental analysis of the stability area of the ROGSS robot was also conducted, which showed that the ROGSS does not correspond to the dimensions of the stability area described by the physical contour of the foot. Comparison of theoretical and experimental results showed that the calculated mathematical model can be integrated into the control system of an anthropomorphic robot, as well as into a simulator of motion programming for assessing static stability. The results obtained allow predicting the stability of the target position in statics and can serve as a basis for further research into dynamic stabilization algorithms.
Large-scale systems control. 2025;(117):141-170
141-170
Control systems analysis and design
Analysis of motion separation by control for two identical pendulums
Resumo
The paper studies the possibility of separating the motions of two identical dynamic systems using a single scalar action generated by a feedback mechanism. Such a problem may arise with underactuated control of oscillatory mechanical systems, with selective control of individual molecule systems, etc. A typical case is studied when separation must be performed based on the energy level of subsystems. To solve the problem, it is proposed to synthesize feedback control. The control algorithm is based on the speed gradient method extended to account for constraints. To design control under constraints the inner penalty function is used. Its operability in this case follows from the invariance of the subsystem energies in the absence of control. It is shown that even a small difference in the initial states of the controlled systems allows the required separation to be achieved, and the result depends weakly on significant changes in the parameters of the controlled system and on the presence of disturbances. Finally, the result depends weakly on the parameters of the controller (gain gamma) and the strength of the penalty alpha.
Large-scale systems control. 2025;(117):171-187
171-187
Synthesis of static output feedback control laws subject to feasibility the kimura condition
Resumo
A lot of control tasks, such as, for example, the synthesis of static output feedback, are expressed in the form of bilinear matrix inequalities. Solving such tasks based on iterative algorithms attended with considerable time, especially in the case of large-scale systems. If no solution is found for the initial values, then repeating calculations with different initial values does not guarantee success. The reason is the non-convexity of the feasible sets. The article investigates the possibility of bilinear matrix inequalities reducing to linear matrix inequalities by replacing of Lyapunov function matrix arbitrary to block-diagonal matrix. A sufficient condition for such a replacing is Kimura condition feasibility. It is proved that the necessary conditions for the linear matrix inequality feasibility are satisfied by two linear non-generate transformations of the system basis. To the synthesis problem investigating within the framework of linear matrix inequalities computational experiments were performed, in which 1000 linear systems were randomly generated. Based on the results of computational experiments, a hypothesis is proposed according to which the Kimura condition is a sufficient condition for bilinear matrix inequalities reducing to linear matrix inequalities with nonempty feasible sets.
Large-scale systems control. 2025;(117):188-199
188-199
Information technologies in control
Optimization approach to solving pnp problem based on parameterization by rodrigues vector
Resumo
In many cases, the task of positioning a moving object using camera images can be reduced to well-known PnP (perspective-n-point) problem. For solving it, optimization algorithm is considered. The reprojection error using the parameterization of the rotation matrix by the Rodrigues vector is considered. Analytical partial derivatives were obtained for it. A feature of the task of positioning a moving object is that the new solution in most cases is in the neighborhood of the previous one. The main disadvantage of the optimization approach is the need to calculate the initial solution. It can be eliminated using the solution from the previous iteration. This also helps to eliminate the problem with the Rodrigues vector, which increases indefinitely as the rotation angle reaches 180 degrees. The BFGS numerical algorithm was used to solve the optimization problem. A heuristic for selecting the initial approximation of the inverse Hessian and the first step of the linear search is proposed. This made it possible to speed up the algorithm. The proposed approach is compared with known implementations from the OpnCV library on synthetic data. The experiment showed that the proposed approach has good performance in terms of accuracy and execution speed.
Large-scale systems control. 2025;(117):200-219
200-219
Identification of rfid tags from a sensor field using uavs
Resumo
Radio Frequency Identification (RFID) technology is widely used in scientific and engineering applications, including integration with Unmanned Aerial Vehicles (UAVs) for object identification in hard-to-reach environments. Two key performance characteristics of such systems are the probability of successful tag reading and the reading time, both of which can be quantitatively described using semi-Markov process models. This paper considers two fundamentally different scenarios. In the first scenario, tags are sparsely distributed, which eliminates collisions. An analytical model is proposed to describe the interaction between the reader and a single tag. In the second scenario, densely placed tags lead to collisions. To evaluate system performance in this case, a discrete-event simulation model is developed. The model takes into account the specifics of the EPC Gen2 protocol, radio channel parameters, spatial configuration of tags, and data reading strategies. The paper compares analytical and simulation results and investigates how tag density, data volume, and UAV altitude affect reading performance metrics.
Large-scale systems control. 2025;(117):220-245
220-245
The use of llm models on single-board computers for the implementation of autonomous uav flight
Resumo
This paper examines the use of large language models (LLM) to control unmanned aerial vehicles (UAVs) using natural language commands. The research is aimed at solving a key problem – the discrepancy between the high computing requirements of LLM and the limited resources of on-board computers. The main focus is on optimizing LLM for operation on energy-efficient single-board computers with neuroprocessors, such as OrangePi 5B based on Rockchip RK3588S. The paper presents the results of testing Qwen2.5-Coder quantized models, demonstrating the preservation of code generation quality at processing speeds of up to 17.8 tokens/s. A specialized test (benchmark) was developed to evaluate the effectiveness of LLM integration into UAV autonomous control architecture and the correctness of code generation, including 125 scenarios. The results confirm the feasibility of LLM in autonomous drone control systems, although they reveal typical errors associated with sensor data processing and coordinate systems. The study proposes a promising direction for the development of intelligent UAV control systems with a natural language interface (NLP). The study included both a scientific approach (development of a specialized test) and a technological innovation (performance analysis on single-board computers) aimed at integrating LLM into UAV autonomous control architecture.
Large-scale systems control. 2025;(117):246-264
246-264
Networking in control sciences
Model of encoding time intervals by active agents
Resumo
Interval time coding, i.e. the ability to perceive and estimate duration in the range from seconds to minutes, is one of the key cognitive processes underlying the adaptive behavior of biological species. This ability allows individuals to extract temporal patterns from the environment, optimize resource extraction strategies, coordinate communication, and form forecasts about future events. The paper describes a model of an active agent with an internal structure represented by an ensemble capable of generating rhythmic activity in given time intervals. The ensemble consists of three nodes: a half-center oscillator (two nodes exciting each other in antiphase) and a trigger node with memory. The half-center oscillator excites a trigger that accumulates excitation in memory. When excitation reaches a threshold value, the trigger is activated and, thus, transmits a signal to the agent. The trigger activation frequency depends on its parameters: receptor weights, discount coefficient, threshold value. The proposed model of temporal coding by such agents demonstrates properties inherent in biological systems: compliance with Weber's law (direct dependence of the variation of the estimate on the duration of the signal), memory fading in the absence of a stimulus, a return to homeostatic parameters, and a shift in estimates to average values. Research in the field of biologically inspired interval coding not only deepens our understanding of time perception, evaluation, and prediction, but also stimulates the development of adaptive AI systems, robotics, and human-machine interfaces.
Large-scale systems control. 2025;(117):265-285
265-285
Control of social-economic systems
Mechanisms of sustainable development of railway transport by means of capital repairs of diesel locomotives
Resumo
An organizational mechanism for planning and stimulating sustainable development of the organizational system has been developed. Sufficient conditions for the optimality of this mechanism have been found. Based on the obtained results, the problem of sustainable development of rail transport through major repairs of diesel locomotives has been considered. An organizational mechanism has been developed to reduce emissions of harmful substances by a diesel locomotive as a result of major repairs. This mechanism includes an algorithm for standardizing and categorizing the indicators of emissions of the main harmful substances, an algorithm for collapsing the categories of emission indicators and forming a category of environmental friendliness of a diesel locomotive, as well as a procedure for stimulating for major repairs of a diesel locomotive. Emission standards for harmful substances are formed on the basis of domestic standards, as well as national and regional requirements abroad. In this case, the category of emissions of harmful substances of a certain type is recognized as satisfactory if the indicator of these emissions does not exceed the norm specified in the corresponding domestic environmental standard, good - if this indicator also meets the requirements of national or regional environmental standards of the countries of advanced diesel locomotive building, and excellent - if this indicator meets the best world standards. The algorithm of these categories’ convolution defines a comprehensive assessment of the environmental friendliness of a diesel locomotive after a major overhaul. Depending on this assessment, incentives are provided for interested parties. Examples of calculations of environmental friendliness categories for domestic and foreign diesel locomotives are given. The developed standards for emission indicators of the main harmful substances can be used as guidelines for major overhauls of diesel locomotives. The results obtained were used in developing a concept of requirements for diesel locomotives of JSC Russian Railways in the field of environmental protection.
Large-scale systems control. 2025;(117):286-319
286-319
Vehicle control and navigation
Development trends of inertial navigation systems of unmanned aerial vehicles based on the results of patent research
Resumo
A comprehensive analysis of current trends in the development of inertial navigation systems for unmanned aerial vehicles has been carried out, based on an extensive patent study based on the analysis of a large number of patents covering the period from 2003 to 2024, which allows us to obtain a reliable picture of technological progress in this area. The main focus of the work is on four key areas of improvement of inertial navigation systems: improving the accuracy of determining coordinates, increasing reliability and resistance to external influences, reducing weight and cost, as well as expanding the functionality of the systems. A special place in the study is occupied by the analysis of the development of autonomy and noise immunity as priority areas of modernization. In the course of the study, a detailed analysis of the element base of foreign inertial navigation systems of tactical class was carried out, and the main technical characteristics of modern systems were identified. Based on the patent analysis, the leading developers in this field have been identified and a picture of the geographical distribution of innovative developments has been compiled. The methodological basis of the research was an integrated approach, including patent analysis, comparative evaluation of technical characteristics and identification of trends in the development of inertial navigation systems. The results of the study make it possible not only to assess the current state of technology, but also to form forecasts for the further development of autonomous navigation systems for unmanned aircraft.
Large-scale systems control. 2025;(117):320-334
320-334
Testing of the characteristics of an stellar inertial navigation system on a ship in continuous stellar calculation mode
Resumo
This work is devoted to the study of the functioning of an stellar inertial navigation system of the projection type in the mode of determining the location of a marine object by observing celestial bodies. The advantages of astronomical inertial navigation systems for marine and aviation objects are shown. The stages of development of astronomical navigation tools are considered. The principles of building stellar navigation systems of the projection type are described, and the design features of the system are considered. A mathematical model of the errors of a stellar inertial system of the projection type is constructed. Based on the results of the mathematical model, the errors of the stellar system are ranked according to their impact on its output parameters. The article considers the method of algorithmic conjugation of the star sensor and inertial unit bases based on the use of the gradient descent method with a constant step. The conditions for conducting the experiment are specified. The article also provides a diagram of the geodetic reference of the stellar system relative to the ship's coordinate system. A qualitative experimental assessment of the error in determining the vessel's location using an experimental model of a projection-type SINS has been performed. It has been concluded that SINS can be used to solve maritime navigation tasks. It has been concluded that the methodology of algorithmic coupling of the star sensor and inertial unit bases should be used as an integral part of the technology for the development and production of stellar inertial systems.
Large-scale systems control. 2025;(117):335-350
335-350
Simulation tools
A program for checking the compliance of fire breaks regulatory and technical requirements on the master plan of petrochemical complexes
Resumo
The general plan of the enterprise is a tool for the effective use of the territory, ensuring the placement of facilities, taking into account technological connections and functional zoning, which allows to increase the productivity of the plant and reduce logistical costs. Automatization of the master plan can help improve the quality of design, both by eliminating errors related to the human factor, and by sorting through more options to find the optimal solution. The software package will allow you to instantly generate new layouts of objects depending on changes in current conditions, offering several options to choose from. For automated placement of objects on the master plan, it is necessary to take into account many factors. The program, among other things, should perform the following functions: determine the actual minimum distances between objects, contain data on regulatory fire distances, compare actual minimum distances with regulatory ones, identify cases of non-compliance with the requirements of regulatory and technical documentation, and generate a report on actual and regulatory distances. To achieve these objectives, a specialized program has been developed to verify the compliance of fire-fighting areas with the regulatory and technical requirements on the master plan. The module is the first stage for automating the design of the master plan.
Large-scale systems control. 2025;(117):351-363
351-363