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

Институт проблем информатики Российской Академии наук



«Systems and Means of Informatics»
Scientific journal
Volume 27, Issue 4, 2017

Content | About  Authors

Abstract and Keywords.

CONTROLLING QUEUE SIZE IN A SINGLE SERVER SYSTEM
  • M. G. Konovalov
  • R. V. Razumchik

Abstract: The authors revisit the well-known problem of active queue management in queueing systems and show numerically that the efficiency of the algorithms based on dropping functions can be improved. Three different approaches are discussed: randomization of the parameters of RED-type dropping functions, use of the arbitrary dropping functions either without or with parallel tracking of the remaining/elapsed service time. The relative efficiency of these three approaches is demonstrated on M/D/1/N queue with implemented RED scheme. For various initial settings, simulation results are given, which show that the new active queue management algorithms outperform RED schemes and the gain increases with the increase of the system load.

Keywords: active queue management; random early detection; drop function

ANALYTICAL MODELING OF NORMAL PROCESSES IN STOCHASTIC SYSTEMS WITH INTEGRAL NONLINEARITIES (III)
  • I. N. Sinitsyn

Abstract: Methodological and algorithmical support for analytical modeling of normal (Gaussian) processes in differential stochastic systems with probabilistic integral nonlinearities (IN) based on Pearson distributions (PD) and stable probabilistic distributions (SPD) is presented. Support is based on the methods of statistical linearization (MSL) and normal approximation (MNA). Probabilistic IN were approximated by power expansions. The MSL and MNA coefficients for probabilistic IN based on PD and SPD are given. The MSL coefficients for incomplete beta-function and F-distribution are considered. Some SPD types are described. Test examples with accuracy estimation are given. Results may be generalized as new types of probabilistic IN (multichannel angular, spherical, etc.) and various numerical approximations (polynomial, rational, fractional rational, orthogonal, asymptotic, and iterative).

Keywords: analytical modeling; x2-distribution; exponential distribution; gamma-distribution; Hermite polynomial and power expansions; method of normal approximation (MNA); method of statistical linearization (MSL); probabilistic integral nonlinearities (PIN)

EFFECTIVE BANDWIDTH ESTIMATION OF HIGHLY RESPONSIBLE SYSTEMS BY THE RANDOMIZED BATCH MEAN METHOD
  • K. A. Kalinina
  • E. V. Morozov

Abstract: The effective bandwidth (EB) estimation of highly responsible systems is considered. A special attention is paid to the reasons causing overestimation of the EB obtained by the regenerative method. Simulations show that the basic reason of the overestimation is the random summation of the input data over regeneration cycles. Numerical results demonstrate an advantage of the randomizing for the EB estimation of highly responsible systems. If in such a system the workload excesses the predetermined threshold (or overloads the finite buffer), then an unacceptable violation of the QoS (quality of service) requirement happens. To estimate the EB in such systems, the modified batch means method with randomized blocks is proposed.

Keywords: effective bandwidth; regenerative estimation; randomizing; QoS; highly responsible system; batch means method

BAYESIAN RECURRENT MODEL OF RELIABILITY GROWTH: A PRIORI DENSITIES OF POLYNOMIAL TYPE
  • A. A. Kudryavtsev
  • S. I. Palionnaia

Abstract: This work is devoted to the Bayesian recurrent model of reliability growth of complex modifiable information systems. It is assumed that the key parameters of the system are unknown but the researcher obtains the information about their a priori distributions. The paper contains formulas for density and mean of marginal system's reliability when indexes of "defectiveness" and "efficiency" of the tool correcting the deficiencies in the system have a priori distributions with polynomial densities. For instance, uniform and parabolic distributions are considered. Likewise, the results for the case of degenerate distribution of one of the parameters are provided. The obtained formulas are illustrated with numerical results and plots.

Keywords: modifiable information systems; reliability theory; Bayesian approach; parabolic distribution; uniform distribution; polynomial densities

RADIO RESOURCE DISTRIBUTION STRATEGIES IN HETEROGENEOUS NETWORKS WITH NARROW-BAND IoT TRAFFIC
  • V. O. Begishev
  • A. K. Samuylov
  • D. A. Molchanov
  • K. E. Samouylov

Abstract: Growth in volumes and types of data that have to be supported in future Internet of Things (IoT) applications is one of the key features in mobile networks transition from 4G+ to 5G. In this regard, IoT service providers have to ensure the massive multimedia data transfers in conjunction with machine-type messages. The need for simultaneous support of several categories of traffic implies the heterogeneity of 5G networks. To cope with this problem, 3GPP proposed and standardized the in-band deployment of Narrow-Band IoT (NB- IoT) technology. The paper presents a set of feasible strategies for sharing resources between multimedia data and data from sensors in a hybrid LTE/NBIoT wireless network. Based on the developed analytical framework, the paper provides a comparison of the proposed strategies, while giving recommendations on the choice of the optimal one.

Keywords: Internet of Things; NB-IoT (Narrow-Band IoT); LTE (Long-Term Evolution); analytical model; resource allocation strategies

OPTIMAL INVENTORY CONTROL OF CONTINUOUS PRODUCT IN REGENERATION THEORY WITH DETERMINATE DELAY OF THE DELIVERY AND THE PERIOD OF REAL REPLENISHMENT
  • P. V. Shnurkov
  • E. Yu. Pimenova

Abstract: The problem of inventory control in the stochastic model of regeneration is investigated. The purpose is to find a theoretical solution of the optimal control problem. To solve the problem, the authors have found the distribution function of a random time from the replenishment moment till the next order which determines the optimal replenishing product plan. It is proved that the initial problem solution exists and it is attained on the degenerate probability distribution. In addition, the optimal deterministic value of the control parameter is the global extreme point of a function which was expressed in an analytical representation. The existence of deterministic optimal control assertion was proved. The authors have specified the analytical conditions on the initial model characteristics under which the optimal control exists.

Keywords: optimal control in stochastic systems; inventory management; controlled regenerating process

ARCHITECTURE OF THE TOOL FOR DEVELOPING THE FUNCTIONAL HYBRID INTELLIGENT SYSTEMS WITH A HETEROGENEOUS VISUAL FIELD
  • A. V. Kolesnikov
  • S. V. Listopad

Abstract: Hybrid intelligent systems integrating experts' knowledge models solve heterogeneous tasks (problems). At the same time, elements of hybrid intelligent systems interact using symbolic-logical models of knowledge, which significantly limit their capabilities in comparison with teams of experts who successfully argue both logically and figuratively. A typical architecture of hybrid intelligent systems that synthesize integrated methods for solving problems over a heterogeneous visual field and an architecture of the tool for developing such systems are proposed. Hybrid intelligent systems with a heterogeneous visual field alternately simulate collective verbal-symbolic and visual-figurative reasoning depending on the uncertainty of the decision-making situation. Visual-figurative reasoning allows the system to "see" an approximate solution of the problem which can be refined later by the methods of verbal-symbolic reasoning.

Keywords: hybrid intelligent system; heterogeneous visual field; visual reasoning; visual management

EFFECTIVE USE OF PROGRAMMABLE GRAPHICS PROCESSING UNITS IN PROBLEMS OF MOLECULAR DYNAMICS SIMULATION
  • S. A. Semenov
  • D. L. Reviznikov

Abstract: Graphics processing units (GPU) make it possible to expand significantly the capabilities of computing systems. The article discusses the use of graphics processors in problems of molecular dynamics modeling with a complex particle interaction potential. In order to improve the performance of calculations on the GPU, the following methods are implemented: reduction of the number of requests to the global memory; reduction of the number of branches; selection of the optimal load of multiprocessors; and use of equivalent mathematical expressions that perform functions for faster code execution. Parallel execution of the program is realized by partitioning the space into cells, compiling and updating the list of neighboring atoms in order to minimize memory collisions, distribution of operations over computational flows, and allocation of additional memory for creating copies of the coordinates of interacting atoms. Several examples illustrating the computational complexity of the algorithms are considered.

Keywords: high-performance computing; graphics processors; shared memory; molecular dynamics simulation; nanomaterials

TECHNOLOGICAL SERVICE PREPARATION OF INFORMATION AND ANALYTICAL PRODUCTS BY MEANS OF A SEGMENTED SITUATIONAL CENTER
  • A. A. Zatsarinny
  • V. I. Korolev

Abstract: The situational center is presented as an object in the protected execution, where the information security is supported by segmentation of the resources on the basis of information-technology (IT) infrastructure security contour. The article considers the matters of integration and consolidation of information resources to create an information-analytical product. It provides general description of the IT-landscape of the situational center. It proposes the algorithm of information support to make the information-analytical product of situational management in IT-infrastructure with dedicated security contour.

Keywords: situational center; information security; information resources; information-technology infrastructure; information and analytical product

CLASSIFICATION OF VULNERABILITIES OF INTEGRATED CONTROL SYSTEMS AT THE EARLY STAGES OF LIFE CYCLE
  • A. P. Suchkov

Abstract: The article deals with problematic issues of creating integrated management systems by combining the existing and future heterogeneous automated control systems. It is important to identify and analyze potential vulnerabilities in the implementation of these processes. Systematization of the threats and vulnerabilities based on their proper classification will provide a complete and coherent picture of the emerging situation in a large volume of work scheduled.
The article describes an attempt to create such a classification.

Keywords: integrated management system; stage of life cycle; threats; vulnerability classification

BANS OF PROBABILITY MEASURES IN THE PROBLEM OF INSIDER DETECTION
  • E. A. Martyanov

Abstract: A data model for a common work of a finite set of analysts is considered. Among analysts, there can be an insider who due to access to forbidden data can cause damage to the organization. Operation of each analyst is described by a random series of samples of data, admissible for him. The process of addressing the forbidden data is described by bans of probability measures defining honest operation of analysts. The method of definition of the set of bans which allows to use the theory of bans correctly is constructed. It helps to identify the insider by a finite number of observations over operations of analysts.

Keywords: insider; information security; theory of bans of probability measures; mathematical statistics

THE INFLUENCE OF USER ERRORS ON SEMANTIC NET QUALITY DYNAMICS
  • I. M. Adamovich
  • O. I. Volkov

Abstract: The article is devoted to the description and rationale of the approach to the modeling of errors of independent users in forming of a semantic net which is the basis for the distributed technology of concrete historical investigation support and to the quantitative assessment of degree of influence of these errors over the net structure. The proposed approach is to expand the semantic net model which is based on the concepts of graphodynamics and the Barabasi-Albert model by including of the new mechanism of parallel fixation of mistaken and appropriate nonmistaken (ideal) user actions and also the mechanism of imitation of search queries which perform parallely in distorted and nondistorted subnets. Within this approach, the classification of errors affecting the net structure was made and the integral index of semantic net quality was also developed. The experimental studies of user errors impact on semantic net quality dynamics have been undertaken using this model.

Keywords: semantic net; model; user errors; search query; net quality

METHODS OF FREQUENCY ANALYSIS OF CONNECTIVES TRANSLATIONS AND REVERSIBILITY OF STATISTICAL DATA GENERALIZATION
  • I. M. Zatsman
  • M. G. Kruzhkov
  • E. Ju. Loshchilova

Abstract: The methods of Russian connectives frequency analysis are examined, including analysis of their translation models in Russian-French parallel texts.
The parallel texts are integrated into a supracorpora database (SCDB) which also includes bilingual annotations of translation correspondences. The annotations include properties of the examined linguistic items (Russian connectives) and, at the same time, properties of the corresponding linguistic items found in the translation. These properties are organized as a faceted classification in the SCDB describing the translation models found in the SCDB from various perspectives. A characteristic feature of the connectives translations frequency analysis methods implemented in the SCDB is the reversibility of the calculated statistical data, meaning that the calculated frequency values act as hyperlinks to the lists of the annotations those values are based on, which represent occurrences of the corresponding connectives in the parallel texts of the SCDB. The use of faceted classifications in the SCDB allows for multidimensional statistical analysis of the annotated connectives and translation models. The calculated statistical data are verifiable because they allow tracing the given values directly to the annotations they are based on. The main goal of this paper is to describe methods of frequency analysis of connectives translation models, including those that support the reversibility of the calculated statistical data on different generalization levels.

Keywords: supracorpora database; translation models; annotation of translation models; faceted classifications; corpus linguistics; generalization; reversibility of generalization process

FEATURES OF REALIZATION OF THE "PERSONAL TELEMEDICINE" SERVICE
  • G. Y. Ilushin
  • V. I. Limansky

Abstract: The possibility of application of mobile telemedicine units (MTU) and services of personal telemedicine (PTM) for increase in availability and improvement of quality of the medical care provided to inhabitants of remote and difficult to access settlements of the Russian Federation is considered. The article describes the appointment and the main functions of the PTM service, the use of which allows organizing remote monitoring of health of patients and reducing the number of visits to doctors of a medical institution by patients in the course of outpatient treatment. The approaches to realization of the PTM service on the basis of services and applications of a patient's personal cabinet are defined.
A possible scheme of organization of the process of providing medical services to the population of remote and difficult to access settlements with the use of MTU and PTM services is suggested.

Keywords: telemedicine centers; mobile telemedicine units; medical institution; telecommunication center; patient's personal cabinet

SITUATIONAL INFORMATIZATION OF POPULATION ACTIVITIES: THE MODEL OF THE OBJECT
  • A. V. Ilyin
  • V. D Ilyin

Abstract: The article describes the model of the population activities considered as an aggregate of education, production, and other kinds of activity. The model is represented by the predefined set of "state spaces" describing the components of population activities regulated by state. Improvement of the object through informatization is studied as a phased process of planned transitions from the starting situation to the target one. The object state is assessed by comparing the portraits of the target and achieved situations. Information technologies, based on the situational informatization methodology, are implemented in the human-machine environment for problem solving ("s-environment"), which serves as an infrastructural base of online services (banking, logistics, etc.). The concept and the methodology of situational informatization are considered as contribution to implementation of the state program "Digital Economy of the Russian Federation."

Keywords: situational informatization; human-machine environment for problem solving; spaces of statuses of the population activities; country's potential; digital economy