|
«INFORMATICS AND APPLICATIONS»
Scientific journal
Volume 18, Issue 4, 2024
Content | About Authors
Abstract and Keywords
- I. N. Sinitsyn 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 development of Pugachev’s conditionally optimal filtering and extrapolation methods for implicit stochastic systems (StS) reducible to explicit continuous and discrete StS. A special review in the field of suboptimal and conditionally optimal filtering and extrapolation for continuous and discrete StS with unsolved derivatives (differences) is given. Mathematical models ofimplicit continuous and discrete Gaussian and non-Gaussian StS reducible to explicit StS are presented. It is supposed that observations do not influence implicit objects of observation and are described by explicit differential (difference) equations. Basic methods for conditionally optimal filtering and extrapolation in implicit StS reducible to explicit StS at Gaussian and non-Gaussian noises are developed. Three examples are discussed. Some generalizations are given.
Keywords: conditionally optimal extrapolation; conditionally optimal filtering; implicit stochastic systems; explicit stochastic systems
- A. V. Borisov Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation, M. V. Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1,Moscow 119991, Russian Federation
- Yu. N. Kurinov M. V. Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1,Moscow 119991, Russian Federation
- R. L. Smeliansky M. V. Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1,Moscow 119991, Russian Federation
Abstract: The paper is devoted to the optimal filtering problem of a class of Markov jump processes. The estimated system state is a Markov jump process with a finite set of possible states representing the probabilistic distributions. The available measurement information includes continuous and counting observations. The continuous observation is a function of the system state corrupted by an independent Wiener process. The counting observation intensity also depends on the state. The filtering problem is to find the conditional mathematical expectation of a scalar function of the state (a signal process) given the available observations. The required estimate represents the solution to a system of the stochastic differential system. The paper also introduces an analog of the Kushner—Stratonovich equation describing the temporal evolution of the state conditional distribution.
A numerical example illustrates the performance of the proposed filtering estimate. It presents the monitoring of the quality state and numerical parameters of a communication channel given the oscillating observations of round-trip time and the flow of the packet losses.
Keywords: Markov jump process; stochastic differential observation system; observations with additive noises; Kushner—Stratonovich equation
- I. N. Sinitsyn Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: The theory of Pugachev conditionally-optimal filtering and extrapolation of stochastic processes described by explicit stochastic differential equations at autocorrelated noise in observations is widely used in modern real-time information processing. The paper is devoted to implicit Gaussian stochastic systems (StS) reducible to explicit StS at observational autocorrelated noises. The main results are: (i) typical mathematical models of observable implicit StS reducible to differential explicit StS; (ii) basic equations for nonlinear conditionally-optimal filters (COF) and conditionally-optimal extrapolators (COE) at noncorrelated and autocorrelated noises; and (iii) illustrative examples for reducible StS. Main conclusions and perspective directions for COF and COE design in the case of implicit differential and functional differential StS are given.
Keywords: autocorrelated observation noise; conditionally-optimal extrapolation; conditionally-optimal filtering; explicit stochastic system; implicit stochastic system
- O. V. Shestakov Department of Mathematical Statistics, Faculty of Computational Mathematics and Cybernetics, M. V Lomonosov Moscow State University, 1-52 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation,
Federal Research Center “Computer Science and Control” of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation, Moscow Center for Fundamental and Applied Mathematics, M.V. Lomonosov Moscow State University, 1 Leninskie Gory, GSP-1, Moscow 119991, Russian Federation
Abstract: Wavelet analysis methods in combination with thresholding procedures are widely used in nonparametric regression problems when estimating a signal function from noisy data. Their popularity is explained by their adaptability to local features of the functions under study, high speed of processing algorithms, and optimality of the estimates obtained. Error analysis of these methods is an important practical task, since it allows one to estimate the quality of both the methods themselves and the equipment used. Sometimes, the nature of the data is such that observations are recorded at random points in time. If the sample points form a variation series of a sample from a uniform distribution over the data recording interval, then the use of standard thresholding procedures is adequate. This paper considers the block thresholding method, in which the wavelet decomposition coefficients are processed in groups that allows one to take into account information about neighboring coefficients. An analysis of the mean square risk estimate of this method is carried out and it is shown that under certain conditions, this estimate turns out to be strongly consistent and asymptotically normal.
Keywords: wavelets; block thresholding; random samples; unbiased risk estimation
- I. V. Uryupin Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- A. A. Sukharev National Research University Higher School of Economics, 20 Myasnitskaya Str., Moscow 101000, Russian Federation
Abstract: The paper is aimed to suggest a model describing the influence of air fares on air transportation demand
in the Russian air transportation system. The authors suggest an approach and a mathematical model for the price
elasticity of air travel demand estimation in the Russian Federation both at the national level and for a single
route. In the context of limited statistical information, the problem of incomplete data on tariffs was solved by an
additional regression model of the dependence of tariffs on the distance of transportation. The results of the study
contribute to airline practitioners and stakeholders by providing a Russian-context-specific allied instrument for
estimating the influence of the air fare change on air transportation demand to solve a range of tasks related to
aircraft design and operation. The article demonstrates the use of the obtained model to assess the potential for changes in demand for transportation when replacing existing types of aircraft with advanced models on a specific
route and in the whole Russian air transportation system.
Keywords: air transportation system; mathematical modeling; price elasticity; airlines; air transportation
- Ya. G. Martyushova Moscow Aviation Institute (National Research University), 4 Volokolamskoe Shosse, Moscow 125933, Russian Federation
- A. V. Naumov Moscow Aviation Institute (National Research University), 4 Volokolamskoe Shosse, Moscow 125933, Russian Federation
- A. E. Stepanov Moscow Aviation Institute (National Research University), 4 Volokolamskoe Shosse, Moscow 125933, Russian Federation
Abstract: The problem of constructing an optimal strategy for passing a time-limited test in the form of a stochastic programming problem with probabilistic constraints is considered. The strategy is a set of test tasks that maximizes the number of points scored for the test, the excess of which, while simultaneously fulfilling the limit on the time of the test, is guaranteed with a preselected confidence level, acting as a task parameter. The random parameters of the task are the user’s response time to each test task and the correctness of the user’s response to the task modeled by a random variable with a Bernoulli distribution. The resulting stochastic programming problem with probabilistic constraints is reduced to a deterministic integer mathematical programming problem. An algorithm to solve the initial problem is presented.
Keywords: time-limited test; problem with probabilistic constraints; integer mathematical programming
- M. G. Konovalov Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- R. V. Razumchik Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: Consideration is given to the dispatching system with a single dispatcher without a queue for storing incoming jobs. There is a finite number of infinite capacity queues running in parallel, each having a single server for serving jobs one-by-one in FIFO (first in, first out) manner. It is assumed that the dispatcher has perfect information about the system upon making a routing decision about the arrived job. Once the decision is made, it is irrevocable but is executed by the job with a random delay. The system is modeled by a two-phase tandem queue, with an infinite-server queue at the first phase and a fixed number of single-server queues at the second phase. The method to construct simple dispatching policies by mixing is proposed, which can result in robust rules, having better performance than conventional static and dynamic policies. Simulations show significant reductions in mean response times (as well as its percentiles) in large-scale systems.
Keywords: parallel service systems; dispatching; load balancing; random delay
- A. V. Daraseliya Peoples’ Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya Str., Moscow 117198, Russian Federation
- E. S. Sopin 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
- K. E. Samuylov 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
- E. A. Koucheryavy National Research University Higher School of Economics, 20 Myasnitskaya Str., Moscow 101000, Russian Federation
Abstract: Future Cellular Internet of Things (CIoT) 5G/6Gtechnologies are expected to be able to handle various types of traffic with different statistical characteristics, resource requirements, and transmission directions. In order to achieve this, the random access and data transmission phases in these technologies need to be optimized. This study proposes a mathematical model of the service procedure, which takes into account the sequential phases of random access and data transition, and different types of traffic in the uplink and downlink directions. The results show that modern CIoT systems such as NB-IoT (Narrow Band Internet of Things) are not optimized for a wide range of load conditions. Numerical results show that the optimal allocation of resources between the random access and data transmission phases can vary significantly depending on the load in the uplink and downlink directions. The proposed model allows for optimal configuration of the random access and data transition phases in future CIoT systems.
Keywords: 5G; 6G; mMTC; CIoT; random access; latency; stability criterion; resource allocation
- S. L. Frenkel Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
- V. N. Zakharov 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 predicting the degradation rate of technical system characteristics (life time, LT) is usually solved within the framework of the accelerating testing (AT) paradigm under stress effects. However, statistical methods for evaluating the results may be ineffective for AT when the system performance depends on a very large number of factors. Accelerating testing also has its own specifics at the early stages of experimental design work when, having only a small number of device copies, it is necessary to estimate their potential service life to assess the feasibility of continuing the development. The article analyzes the extent to which modern mathematical and statistical models that formthe ATmethodology, namely, survival analysis, extreme value theory, allow obtaining forecasts of the service life of the designed devices under real operating conditions at the early stages of development/design. The authors indicate the problems of solving the LT forecasting problem using known machine learning tools, consider and propose a heuristic method for solving the LT forecasting problem in real conditions. As an example, the authors consider the prediction of performance degradation for new solar cell designs whose performance tends to degrade. This heuristic refers to the extraction of the Hondrick–Prescott trend from a nonstationary time series that represents the degradation of a quality characteristic. The applicability of the proposed heuristic to predict degradation in other technical applications, particularly, in computer networks, is discussed and justified.
Keywords: accelerating testing; machine learning
- A. M. Dostovalova Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences, 44-2 Vavilov Str., Moscow 119333, Russian Federation
Abstract: The paper considers the neural ensemble neural network architecture that uses a quadtree model for SAR (Synthetic Aperture Radar) image segmentation under the lack of training data. The Neural Quadtree network (NQN) consists of segmentation network forming the image pixels features and a graph convolution network with the special branch pruning block establishing the spatial and hierarchical connections between pixels. The NQN is used for segmenting of several SAR images that differ a lot both in presented surfaces and characteristics (Sentinel-1, ESAR (Experimental SAR), HRSID (High-Resolution SAR Images Dataset)). A comparison was made of the results of processing images of NQN and a conventional quad-tree using a common U-Net network segmentor. The NQN demonstrates the higher quality in target detection in comparison with a conventional quadtree. The difference in Recall values for such objects classes between NQN and quadtree ranges from 2.13% to 11.63%.
Keywords: quadtree; graph convolution network; SAR images; target detection
|
