   |
  |
||
|
|||
 |
|||
“Informatics and Applications” scientific journalVolume 13, Issue 2, 2019Content Abstract and Keywords About Authors PROOF OF THE UNIMODALITY OF THE OBJECTIVE FUNCTION IN M/M/N QUEUE WITH THRESHOLD-BASED CONGESTION CONTROL
Abstract: The problem of limiting the load in the system M/M/N/infinity is considered using a simple threshold strategy. In addition to the service time, each task is characterized by a deadline. Depending on the quality of service, the system receives either a fixed income or a penalty. The quality of control is determined by the marginal average income and the threshold value that maximizes this value is considered as optimal. Usually, it is much easier to find the optimal threshold if the objective function has a single maximum. The experimental results show the unimodality of the objective function for a wide class of arrival flows. However, there is no rigorous proof of this fact and in the paper, this gap is filled up for the Poisson arrivals. The proof is based on the results of the Markov chain theory and queueing theory. Keywords: Markov chains; M/M/N/infinity system; congestion control; threshold control; deadline ON THE CONDITIONALLY MINIMAX NONLINEAR FILTERING CONCEPT DEVELOPMENT: FILTER MODIFICATION AND ANALYSIS
Abstract: The main result of the research is a new suboptimal filter developed from the conditionally minimax nonlinear filtering (CMNF) method for nonlinear stochastic systems in discrete time. The main idea of the proposed modification is to omit the time and resource consuming phase ofa priori CMNF parameter calculation in favor of their online approximation together with the current state estimation. In the original CMNF filter, the simulation study is used in order to approximate dynamic system parameters' unconditional expectation and covariances, while the modified version deals with the conditional moments which are also calculated by means of the Monte-Carlo method. The proposed filter modification is provided with the minimax justification, similar to the underlying CMNF concept. Simulation examples show the proposed algorithm effectiveness and performance gain in comparison with the original conditionally minimax nonlinear filter. Keywords: nonlinear stochastic observation system in discrete time; conditionally minimax nonlinear filtering; Monte-Carlo simulation PROPERTIES OF WAVELET ESTIMATES OF SIGNALS RECORDED AT RANDOM TIME POINTS
Abstract: Wavelet analysis algorithms in combination with threshold processing procedures are widely used in nonparametric regression problems when estimating the signal function from noisy data. The advantages of these methods are their computational efficiency and the ability to adapt to the local features of the function being estimated. The error analysis of threshold processing methods is an important practical task, since it allows assessing the quality ofboth the methods themselves and the equipment used. Sometimes, the nature ofthe data is such that observations are recorded at random times. Ifthe sampling points form a variation series constructed from a sample of a uniform distribution over the data recording interval, then the use of conventional threshold processing procedures is adequate. In this paper, the author analyzes the estimate of the mean square risk of threshold processing and shows that under certain conditions, this estimate is strongly consistent and asymptotically normal. Keywords: wavelets; threshold processing; random samples; mean square risk estimate ARCHITECTURAL DECISIONS IN THE PROBLEM OF IDENTIFICATION OF FRAUD IN THE ANALYSIS OF INFORMATION FLOWS IN DIGITAL ECONOMY
Abstract: An approach to a research of some types of fraud in digital economy with the usage of relationships of cause and effect is formulated. In all types of the considered frauds, the discrepancy between the purposes of financial transactions and actual cost of achievement of these purposes has to be observed. Data on transactions can be collected by observing information flows in which these transactions are reflected. The architecture of data collection and their analysis can be organized by means of the distributed ledgers with the centralized consensus that allows creating an analog of the electronic account book fixing financial and economic activity of subjects of digital economy in the region. The methods of fraud identification considered are based on the contradictions between actions described in transactions and information, which is contained in plans, standards, precedents, etc. The method based on a simplified scheme of implementation of the abstract project is considered. For identification of contradictions, it is necessary to carry out the analysis from the effect to the cause, i. e., to look for anomalies in information describing the generation of the observed effects. It is shown how in implementation of the project it is possible to allocate simple "necessary conditions" of violation of cause and effect relationships, i. e., a set of "necessary conditions" violation of which demonstrates fraud existence. It is possible to call this set of "necessary conditions" by metadata for control of the project for fraud identification. Keywords: digital economy; information flows; relationships of reason and effect; detection of fraudulent schemes ON AUTOMATA MODELS OF BLOCKCHAIN
Abstract: The author considers automata models of blockchain, mostly based on timed automata. The author suggests a new version oftimed automata that avoids some inconveniences that occur in modeling by using standard timed automata where time is represented by real numbers. In the latter case, one should use variables of two types, Boolean and real; when applied to blockchain modeling, this fact causes some difficulties both in obtaining theoretical estimates and in program implementation. The present approach is based on 2-adic analysis since in that case, both time and digital variables are ofone type only; namely, Boolean. Keywords: blockchain; smart contract; timed automaton ON LOCAL AFFINITY BASED METHOD OF SOLVING SYSTEMS OF QUADRATIC BOOLEAN EQUATIONS
Abstract: Solving nonlinear systems of Boolean equations is NP-hard. Nevertheless, certain classes of such systems can be solved by efficient algorithms. There are theoretical and applied reasons for studying these classes and designing corresponding efficient algorithms. The paper proposes an approach to solving the systems of quadratic equations over two-element field. The method makes use of the quadratic functions' representation by their affine normal forms, i. e., in a sense, of their piecewise affine approximation. So, the initial nonlinear instance comes to a number of linear equations systems of the same variables. The paper also discusses possible ways to reduce the complexity of the proposed method. Keywords: Boolean function; system of quadratic Boolean equations; vector space partition; flat; local affinity; affine normal form; algebraic cryptanalysis APPLICATION OF DECISION SUPPORT METHODS FOR THE MULTICRITERIAL SELECTION OF MULTISCALE COMPOSITIONS
Abstract: The article discusses the use of decision-making support methods for the task of selecting multiscale compositions (MC) - computational analogues of multiscale physical and mathematical models created for analyzing various heterogeneous processes associated with the formation of new composite materials with predetermined properties. When solving specific problems, different multiscale models and their corresponding MC can be constructed. The question arises of comparing these models and assessing their "effectiveness" for specific problem. On the stage of predictive modeling, the authors propose a methodology for comparison of multiscale models through evaluation and selection of appropriate MC using methods of decision-making support under multiple criteria. As an illustration of the possibility of choosing the best alternative in the presence of additional information on evaluation criteria of MC, a model example associated with the study of electronic and structural properties of thin films InN (GaN) on silicon substrates is considered. Keywords: multiscale modeling; decision theory; quality criteria; alternative; decision support methods; multiple criteria; value function RESEARCH OF THE OPTIMAL CONTROL PROBLEM OF INVENTORY OF A DISCRETE PRODUCT IN THE STOCHASTIC REGENERATION MODEL WITH CONTINUOUSLY OCCURING CONSUMPTION AND RANDOM DELIVERY DELAY
Abstract: The paper considers the optimal control problem of inventory of a discrete product in a regeneration scheme with a Poisson flow of customer requirements. In the system, deferred demand is allowed, the volume of which is limited by a given value. The control parameter is the level of the stock, at which achievement it is necessary to make an order for replenishment. The indicator of management effectiveness is the average specific profit received in one regeneration period. The optimal control problem is solved on the basis of the statement about the extremum of a fractional-linear integral functional on the set of discrete probability distributions. Keywords: inventory management of a discrete product; controlled regenerative process; extremal problem for a fractional-linear integral functional ESTIMATION OF THE RELEVANCE OF THE NEURAL NETWORK PARAMETERS
Abstract: The paper investigates a method for optimizing the structure of a neural network. It is assumed that the number of neural network parameters can be reduced without significant loss of quality and without significant increase in the variance of the loss function. The paper proposes a method for automatic estimation of the relevance of parameters to prune a neural network. This method analyzes the covariance matrix of the posteriori distribution of the model parameters and removes the least relevant and multicorrelate parameters. It uses the Belsly method to search for multicorrelation in the neural network. The proposed method was tested on the Boston Housing data set, the Wine data set, and synthetic data. Keywords: neural network; hyperparameters optimization; Belsly method; relevance of parameters; neural network pruning BAYESIAN MODELS OF FACTORS BALANCE WITH A PRIORI WEIBULL AND NAKAGAMI DISTRIBUTIONS
Abstract: Bayesian balance models are considered. Within this approach, it is assumed that the parameters affecting a system can be divided into positive, which support system functioning, and negative, which interfere with the functioning. Thus, the ratio of negative to positive factors - balance index - is considered as an indication of system's functioning effectiveness. The study is carried out assuming that the factors depend on the environment state and their exact value cannot be obtained due to external reasons, e. g., equipment faults, lack of resources, etc. It is also assumed that the principles of factors' changes are known a priori and remain invariable. Considering these assumptions, it is natural to use the Bayesian method, which implies randomization of the initial parameters supposing that their a priori distributions are known. As a result, the balance index becomes a random variable as well. This paper continues a series of studies by the authors devoted to the application of Bayesian methods in the problems of queuing and reliability. In this work, the obtained probability characteristics of the factor balance index in the case of a priori Weibull and Nakagami distributions are presented. The results are formulated using gamma-exponential function. Keywords: Bayesian approach; balance models; mixed distributions; Weibull distribution; Nakagami distribution; gamma-exponential function HETEROGENEOUS THINKING PROTOCOL OF HYBRID INTELLIGENT MULTIAGENT SYSTEM FOR SOLVING DISTRIBUTIONAL POWER GRID RECOVERY PROBLEM
Abstract: The problem of power supply restoration in the regional distributional power grid after large-scale accidents is characterized by high combinatorial complexity, heterogeneity, underdetermination, inaccuracy, and ambiguity. The use of the collective problem solving mechanism to overcome the listed non-factors in the sense of A. S. Narinyani is impossible due to time constraints. To solve this problem, a new class of intelligent systems that model collective decision-making under the guidance of a facilitator is proposed, namely, hybrid intelligent multiagent systems of heterogeneous thinking. Unlike traditional hybrid intelligent systems that integrate models of expert knowledge, they additionally model group processes and effects arising from collective problem solving, adapting to the dynamic nature of the problem of restoring the regional distribution grid. The paper discusses one of the components of such systems, namely, the protocol for organizing collective heterogeneous thinking of agents. Keywords: heterogeneous thinking; hybrid intelligent multiagent system; distributional power grid recovery COMPATIBILITY OF LOGICAL SEMANTIC RELATIONS: METHODS OF QUANTITATIVE ANALYSIS
Abstract: The paper deals with logical semantic relations (LSR) that ensure text coherence and examines their explicit markers in text, i. e., connectives. An overview of existing approaches to definition and classification of LSRs is presented; different LSR combination patterns are examined. To make quantitative assessment of LSR compatibility, one should be able to annotate LSRs and their markers in texts. To support such annotation, a supracorpora database was developed, including customizable faceted classifications for LSRs and connectives.
Keywords: databases; quantitative analysis; connectives; logical semantic relations; annotation of relations; generalization of information objects MODELING THE PROCESS OF NETWORK PLANNING OF A PORTFOLIO OF PROJECTS WITH HETEROGENEOUS RESOURCES UNDER FUZZINESS
Abstract: The paper discusses the problem of project portfolio scheduling subject to resource constraints and fuzzy activity durations. Employees ofvarious specializations are considered as the only type ofresource. Fuzzy W-algebra was used to avoid such difficulties and limitations of traditional fuzzy arithmetics as the support size extension, the need of fuzzy number comparison, and some others. The proposed model provides a fuzzy estimation of project execution times and optimal resource allocation. To solve the problem, the genetic algorithm based on activity list representation was implemented. The paper also provides a numerical example which demonstrates advantages of applying the approach to organizations' workflow in the case of making estimates for decision making under uncertainty. Keywords: project management; W-algebra; fuzzy arithmetic; combinatorial optimization; genetic ON THE GENESIS OF THE INFORMATION SOCIETY: INFORMATICS-CYBERNETIC MODEL REPRESENTATION
Abstract: The concept of the information society genesis is introduced, which is viewed from the standpoint of informatics-cybernetic modeling of the development of Humankind as a self-controlling hierarchical-networking system. On this basis, the author obtained quantitative assessments of its typical spatial-temporal characteristics, representing geometric progressions with the denominator "e to the degree e" (15.15426...), as well as coordinated with them in time and space of the psychoanthropological, educational, and informational communication parameters and possibilities of a person who becomes complicated in this process and his communities of various sizes. This allowed us to push the framework of the information society for the entire historical and even archaeological epoch of such development. The resulting sequence of information technologies "signal poses/sounds/movements - mimics/gestures - speech/language - writing - replicating texts - computers - telecommunications - information nanotechnology - ..." allows us to consider the information society genesis in the broad context of a unified historical retrospective and perspective. Keywords: information society; information technologies; informatics-cybernetic model; self-controlling hierarchical-networking system of Humankind; archaeological epoch A GAUSSIAN APPROXIMATION OF THE DISTRIBUTED COMPUTING PROCESS
Abstract: The authors propose a refinement of the stochastic model describing the dynamics of the Desktop Grid (DG) project with many hosts and many workunits to be performed, originally proposed by Morozov et al. in 2017. The target performance measure is the mean duration of the runtime of the project. To this end, the authors derive an asymptotic expression for the amount of the accumulated work to be done by means of limit theorems for superposed on-off sources that lead to a Gaussian approximation. In more detail, depending on the distribution of active and idle periods, Brownian or fractional Brownian processes are obtained. The authors present the analytic results related to the hitting time of the considered processes (including the case in which the overall amount of work is only known in a probabilistic way), and highlight how the runtime tail distribution could be estimated by simulation. Taking advantage of the properties of Gaussian processes and the Conditional Monte-Carlo (CMC) approach, the authors present a theoretical framework for evaluating the runtime tail distribution. Keywords: Gaussian approximation; distributed computing; fractional Brownian motion VIRTUAL EXPERIMENTS IN DATA INTENSIVE RESEARCH
Abstract: Organization and management of virtual experiments (VE) in data intensive research (DIR) has been widely studied in the several past years. The authors survey existing approaches to deal with VEs and hypotheses and analyze VE management in a real use-case from the astronomy domain. A review of existing systems that can act as platforms for conducting a VE has been carried out. Requirements for a system to organize VEs in data intensive domains have been gathered and overall structure and functionality for system running VEs are presented. The relationships between hypotheses and models in VE are discussed. The authors also illustrate how to model conceptually VEs, respective hypotheses, and models. Potential benefits and drawbacks of such approach are discussed, including maintenance of experiment consistency and reduction of potential number of experiments. Keywords: virtual experiment; hypothesis; conceptual modeling; data intensive research
|
 | |
 Phone of the Center: +7 (499) 135-62-60E-mail of the Center: ipiran@ipiran.ru | 
RUS
|