Институт проблем информатики Российской Академии наук
Институт проблем информатики Российской Академии наук
Российская Академия наук

Институт проблем информатики Российской Академии наук




«INFORMATICS AND APPLICATIONS»
Scientific journal
Volume 16, Issue 3, 2022

Content | About  Authors

Abstract and Keywords

A CONDITION FOR NON-EMPTINESS OF THE EPSILON-CORE OF A NONTRANSFERABLE UTILITY FUZZY GAME AND COMPUTATIONAL SCHEMES
  • A. S. Shvedov  National Research University Higher School of Economics, 20 Myasnitskaya Str., Moscow 101000, Russian Federation

Abstract: Zhao (2001) suggested a new condition for non-emptiness of the core of a cooperative game. At first, a necessary and sufficient condition was found by Zhao for a particular case of games with transferable utility. In this case, as was shown by Zhao, the condition can be easily used for construction of a computational scheme which decides if the core of a game is empty or non-empty and gives imputations belonging to the core if the core is non-empty. Then, the condition was generalized by Zhao for games with nontransferable utility. In this paper, the results are generalized in few directions. The problem of non-emptiness of the epsilon-core which is somewhat more general than the problem of non-emptiness of the core is studied. Games with fuzzy payoffs are considered.
For some classes of games with nontransferable utility, possibility of construction of a computational scheme which is similar to the computational scheme for games with transferable utility is established.

Keywords: fuzzy set; cooperative game; epsilon-core; comprehensive set

APPLICATION OF SELF-ORGANIZING NEURAL NETWORKS TO THE PROCESS OF FORMING AN INDIVIDUAL LEARNING PATH
  • A. V. Bosov  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: The problem of dynamic classification of students within the framework of supporting the process of forming an individual trajectory of the user of an electronic learning system is considered. The training model is designed for a mixed form of educational activities with partial independent work and periodic control events in the form of distance tests, partial full-time work with offline tests, and offset tests. The purpose of the classification is to determine the category of the student based on the results of the next control event. The semantics of the categories suggests the possibility of an individual task of different levels of task complexity at the next step of learning. The direction of improving the existing methods of classification is the rejection of the accumulation and use of statistics from previous (other) groups of students. The absence of samples of correct classification justified the use of self-organizing neural networks. For the solution, Kohonen's maps were used, the standard version of which is adapted to the existing learning model and to the task of taking into account the subjective evaluation policy of the teacher. Three variants of the self-learning algorithm are described. Experimental research was carried out, its results are illustrated.

Keywords: electronic learning tool; self-organizing neural network; Kohonen's map; classification; individual learning path

SOME RESULTS OF THE ANALYSIS OF THE PROCESS OF CHANGING THE PRICE OF A DUAL CURRENCY BASKET BASED ON RANDOM PROCESS STATISTICS METHODS
  • P. V. Shnurkov  National Research University Higher School of Economics, 34 Tallinskaya Str., Moscow 123458, Russian Federation
  • M. A. Migulya  National Research University Higher School of Economics, 34 Tallinskaya Str., Moscow 123458, Russian Federation

Abstract: The work is devoted to the study of the process of changing the price of the so-called dual-currency basket in the foreign exchange market of the Russian Federation in the presence of interventions carried out by the Central Bank. The main goal of the work is to verify the compliance of the observed process with a stochastic Markov model with discrete time and a discrete set of states. The theoretical basis of the study is the methods of statistics of Markov random processes. As a result, conditions are established under which the real process can be adequately described by the indicated Markov model.

Keywords: Markov chain with discrete time and discrete set of states; statistics of Markov random processes; stochastic models of the evolution of processes in financial markets; stochastic models of interventions; price of a dual-currency basket

INCREASING FOREX TRADING PROFITABILITY WITH LSTM CANDLESTICK PATTERN RECOGNITION AND TICK VOLUME INDICATOR
  • A. K. Gorshenin  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
  • E. I. Guseynova  M. V Lomonosov Moscow State University, 1 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation

Abstract: The paper introduces the research of the effectiveness of using LSTM (Long-Short Term Memory) for candlestick data and a technical analysis indicator for a large number of the most common currency pairs (27 in total) over a long period in order to build automatic trading strategies. The achieved average total and annual return for 8 years of a model trading were 286% and 15.4%, respectively. It is more than 50 times higher than the values for the classic Buy & Hold trading strategy for the same period. In addition, the paper introduces a new technical indicator based on tick volumes which is an alternative trading strategy (the total and annual returns of LSTM models exceed it by an average of 7.2 and 2.3 times) as well as an additional feature to increase the profitability of the neural network strategy through the use of ensemble learning. For 37% of the analyzed currency pairs, the use of an ensemble of LSTMs allows one to increase further the total return by an average of 17.2%.

Keywords: LSTM; ensemble learning; candlestick; technical indicator; FOREX; currency pairs

ON EXTREMUM SUFFICIENT CONDITIONS IN MULTIDIMENSIONAL VARIATION CALCULUS PROBLEMS
  • N. S. Vasilyev  N. E. Bauman Moscow State Technical University, 5-1,2nd Baumanskaya Str., Moscow 105005, Russian Federation

Abstract: Variation principles give formalization and general approach to construct and study the models from different fields of knowledge. They provide system presentations about theories origins. In the models, sought-for solution is a stationary point of a criterion. Its search on the basis of necessary conditions should not accomplish problem investigation. Sufficient conditions are needed to assert its optimality. In natural sciences, such results substantiate principles of energy or Hamilton's action minimization. In the paper, invariant surface integrals discovery gave possibility to prove minimum availability in multidimensional variation calculus problems. The functional in the problems may depend on several unknown functions of many variables and their high-order derivatives. Classical theorems are generalized.

Keywords: extremal; extreme hypersurface; field of normals; divergence and flow of a vector field; differential form; external differentiation; integral invariance; Lagrange multiplier

SEQUENTIAL ANALYSIS AND METRIC ESTIMATES OF PEAK LOAD FLOWS IN THE MULTIUSER NETWORK
  • Yu. E. Malashenko  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: The set of vectors of internodal flows in a multiuser communication network under peak load is analyzed. Within the framework of the multicommodity model, the throughput capacities of edges are considered as the components of a vector of resources of various types that are required for the transmission of various kinds of flows. When conducting computational experiments, at each iteration, the norms of vectors of jointly permissible internodal flows are calculated, during the transmission of which the capacity of all network edges is fully used.
The proposed method and computational experiments have shown that the equalizing phased distribution leads to an uneven distribution of flows for different groups of correspondents. Metric values obtained during experiments indicate deformation of the sets of accessible flows. Within the framework of the model, all correspondents are tantamount and the flows are noninterchangeable; however, in the case of an equalizing peak load distribution, adjacent pairs of nodes are in a privileged position when using residual capacity. The obtained metric values can be used to evaluate the functional characteristics of the transmission network in the finite capacity loading mode.

Keywords: multicommodity flow network model; set of achievable internodal flows; peak load distribution

MODEL SELECTION FOR MATRIX FACTORIZATION WITH MISSING COMPONENTS
  • M. P. Krivenko  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: The work is dedicated to the problem of factorizing a matrix with missing components into a product of two lower-rank matrices. The influence of the intensity of missing on the choice of the factorization model is studied. Two algorithms for parameter estimation are considered: alternating least squares (ALS) and Wiberg - for two factorization models: with and without means. It is substantiated that there is no need to use a model with averages: it is a special case of another model and, in some cases, leads to ambiguous solutions. During the experiments, the preference was given to a more stable ALS algorithm. The advantages of the insertion method over random filling in the initial settings of iterative algorithms for estimating model parameters are demonstrated.
The reasons for the negative properties of the existing version of the Wiberg algorithm are revealed. Based on the experiments, it was found that with an increase in the probability of missing, the accuracy of the presentation of the available data increases which leads to an underestimation of the true value of the model dimension.

Keywords: lower-rank matrix approximation; singular value decomposition; missing data; ALS algorithm; Wiberg algorithm

INFORMATICS' MEDIUM MODELS OF INFORMATION TECHNOLOGY: THEORETICAL FOUNDATIONS FOR THEIR CREATING
  • I. M. Zatsman  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: The variants of the information technology-oriented (ITO) model are considered which are used in the design of information technologies (IT) for discovering linguistic and medical knowledge from texts within the framework of two projects under the grants from the Russian Foundation for Basic Research (RFBR) and the Natural Science Foundation of China (NSFC). The variants of the ITO model were created within the paradigm of dividing the subject domain of informatics into media of different nature. The aim of the paper is to describe the initial data and the theoretical foundations for the generalization of the ITO model. These foundations are planned to be used to create a generalized model of IT, its types, and particular cases as constituents of the new class of models that are proposed to be called the informatics' medium ones. The principal idea of generalization is to distribute the stages of the IT being designed, their inputs, and outputs according to media of the subject domain of informatics and the boundaries between them in the interests of IT modeling. The boundaries locate entities of multiple nature (semiotic signs on the boundary between the mental and information media, Unicode coding tables on the boundary between the information and digital media, etc.). Such IT is proposed to be called the multimedium one. Medium models are focused, first of all, on solving the topical problem of creating unified tables for coding meanings (= senses) of ambiguous words in semantic web knowledge bases when processing large volumes of texts (by analogy with Unicode tables for encoding letters and symbols).

Keywords: medium models in informatics; media of informatics subject domain; ITO model variants; theoretical foundations; multimedium information technologies

COMPUTER-ASSISTED TEXTUAL ANALYSIS IN TRANSLATION: REDUCING THE SPECTRUM OF TRANSLATION MODELS IN SUPRACORPORA DATABASES
  • V. A. Nuriev  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: The paper refines the approach aimed at reducing the spectrum of translation models in supracorpora databases (SDBs). Being an information resource of broad potential application, SDBs can be used to research on problems in the field of information science, computer linguistics, medicine, etc. Here, SDBs are regarded from the perspective of the corpus-based translation studies. It is shown how this automated instrument can be applied in 'close and distant reading' - an approach that advocates the idea of using modern information resources in literary translation. The special focus is on opportunities that SDBs could offer for reducing the spectrum of translation models. Due to the synonymic potential, characteristic of natural languages, in translation, instead of the only possible solution, one has to choose between relatively interchangeable alternatives (words, collocations, syntactic constructions, etc.). Choosing the only one output equivalent, a translator seeks to narrow the choice set. Hence, the goal of the paper is to refine the approach that would allow using SDBs for narrowing the choice set of relevant translation models.

Keywords: corpus-based translation studies; digital humanities; computer-assisted textual analysis; distant reading; parallel texts; translation; translation models; supracorpora database; multiple choice

ON AN APPROACH FOR ESTIMATING THE RATE OF CONVERGENCE FOR NONSTATIONARY MARKOV MODELS OF QUEUEING SYSTEMS
  • I. A. Kovalev  Department of Applied Mathematics, Vologda State University, 15 Lenin Str., Vologda 160000, Russian Federation, Moscow Center for Fundamental and Applied Mathematics, M.V. Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation
  • Y. A. Satin  Department of Applied Mathematics, Vologda State University, 15 Lenin Str., Vologda 160000, Russian Federation
  • A. V. Sinitcina  P. G. Demidov Yaroslavl State University, 14 Sovetskaya Str., Yaroslavl 150003, Russian Federation
  • A. I. Zeifman  Department of Applied Mathematics, Vologda State University, 15 Lenin Str., Vologda 160000, Russian Federation, Moscow Center for Fundamental and Applied Mathematics, M.V. Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation, P. G. Demidov Yaroslavl State University, 14 Sovetskaya Str., Yaroslavl 150003, Russian Federation, Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: The transformation of the forward Kolmogorov system is considered which allows one to obtain simple estimates on the rate of convergence for Markov chains with continuous time describing queuing systems. In addition, the concept of the logarithmic norm of the operator function and the associated estimates of the norm of the Cauchy matrix are used. The results obtained make it possible to estimate the rate of convergence for new classes of models in which the matrix is not significantly nonnegative and the use of the logarithmic norm method does not guarantee the possibility of obtaining estimates of the rate of convergence. Previously, a rather laborious more general method of inequalities was used for such situations. A theorem is formulated on obtaining the rate of convergence when the intensities of the matrix change. An estimate was obtained for the process of birth and death with constant intensities. As an example, a special nonstationary model with group service of requirements (service in pairs) is investigated.

Keywords: rate of convergence; ergodicity bounds; logarithmic norm; queuing systems

HYSTERETIC CONGESTION CONTROL IN WIRELESS CLOUD SENSOR NETWORKS
  • B. Aliyu  Peoples' Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya Str., Moscow 117198, Russian Federation
  • E. A. Machnev  Peoples' Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya Str., Moscow 117198, Russian Federation
  • E. V. Mokrov  Peoples' Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya Str., Moscow 117198, Russian Federation

Abstract: The paper considers the problem of congestion analysis in a wireless sensor network. A hysteresis control mechanism is proposed which screens out part of the load in the event of an overload of the system in order to stabilize its operation. A mathematical model is constructed in the form of a Markov process with a finite state space and a formula is obtained for analyzing the performance indicators of a wireless sensor network. The probabilistic-temporal characteristics of hysteresis control, the packet drop probability, and the average queue length are studied numerically. The hysteresis mechanism is compared with the IRED (improved random early detection) algorithm with different control window sizes. It is shown that the hysteresis mechanism provides an advantage in the region of high loads compared to the IRED mechanism.

Keywords: wireless sensor networks; load control; hysteresis control; Markov process; queuing system

CONTROLLABLE QUEUING SYSTEM WITH ELASTIC TRAFFIC AND SIGNALS FOR ANALYZING NETWORK SLICING
  • A. S. Vlaskina  Peoples' Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya Str., Moscow 117198, Russian Federation
  • S. A. Burtseva  Peoples' Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya Str., Moscow 117198, Russian Federation
  • I. A. Kochetkova  Peoples' Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya Str., Moscow 117198, Russian Federation, Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
  • S. Ya. Shorgin  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: A controllable queuing system with impatient elastic traffic and minimum bit rate guarantee transmission is built to analyze the dynamic network slicing in wireless network. Resource management of allocation between two segments is carried out by controller signal and in the case of an idle resource of one segment and waiting users of another. Depending on the state of the system, a decision is made how to allocate the resource. The decision-making is based on the efficiency indicators of slicing taking into account the resource usage of each of the segments, deviation from the recommended "initial" slicing, and additional signal load from the controller.
The corresponding objective control function is proposed and an iterative algorithm for calculating the parameters of the optimal resource allocation policy is applied.

Keywords: 5G; network slicing; controllable queuing system; Markov decision process; elastic traffic

METADATA IN SECURE ELECTRONIC DOCUMENT MANAGEMENT
  • A. A. Grusho  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
  • N. A. Grusho  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
  • E. E. Timonina  Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation

Abstract: The paper is devoted to the study of information security problems of network-centric electronic document management systems in the distributed information system of the enterprise. Electronic document management is a secure distributed system of information technologies, each of which is generated by part of the organizational structure of the enterprise, i. e., there is a secure connection that provides the functionality of document management. To ensure information security, the concept of metadata defined earlier in the works of the authors was used. Unlike earlier works that used the concept of metadata, electronic document management uses many variants of information technologies determined by a subset of users participating in the work on the document and the structure of orders generated by the execution of documents. New in the work is the modular design of metadata which allows solving the problem of synthesis and management of a variety of emerging variants of information technologies of electronic document management and ensuring their information security

Keywords: information security; metadata; electronic document management

KINEMATIC MODELS OF PURSUIT PROBLEMS ON THE PLANE BY THE METHODS OF PARALLEL APPROACH AND PURSUIT
  • A. A. Dubanov  Banzarov Buryat State University, 6a Ranzhurov Str., Ulan-Ude 670000, Russian Federation
  • V. A. Nefedova  Banzarov Buryat State University, 6a Ranzhurov Str., Ulan-Ude 670000, Russian Federation

Abstract: This article provides accurate pursuit models based on the parallel approach and chase methods. This article is a modification of the methods of parallel rapprochement and chasing what happens when the pursuit begins. It cannot instantly change the direction of movement. This should be a less affordable option. The proposed method is based on the fact that the pursuer chooses a step at the iteration stage and will try to follow the predicted trajectories. Based on the materials of the article, test programs have been written that calculate the trajectories of the pursuer taking into account the stated conditions. Execution of animated images visualizes the change in the coordinates of the pursuer, target, and predicted time trajectories.

Keywords: target; pursuer; trajectory; convergence; modeling