Software & Systems

Journal articles №4 2012

21. Complex of the programs and algorithm of calculation of the fractal dimensions and a linear trend of the time series [№4 за 2012 год]
Authors: (mancu@mail.ru) - ; (tsvet@tversu.ru) - , Ph.D;
Abstract: One of the upcoming trends of modeling of complex systems is application of a multifractal dynamics method. The multifractal dynamics models are based on the description of the dynamics of processes represented by multifractal curves. The whole observation time period is got into intervals with defined values of a linear trend characterizing the trend of the value of interest. On each interval, for time series values, a fractal dimension D should be defined. This work offers a new algorithm of computation of a fractal dimension D of the time series envelope in the form of a piece linear function. Computation of the fractal dimension of the time series envelope is based on calculation of the time series envelope length while grouping initial data by different ways. In order to improve computation accuracy, the algorithm uses a trend channel development procedure and rationing of first-calculated fractal dimension values based on the results obtained for the specially generated stochastic time series inside the trend channel.
Keywords: algorithm, programming, normalizing, trend, time series, fractal dimension
Visitors: 15295

22. The experimental cloud computing environment in institute of Mathematics and Mechanics Ural Branch of RAS [№4 за 2012 год]
Author: Ermakov D.G. (Ermak@imm.uran.ru) - Institute of Mathematics and Mechanics Ural Branch of the Russian Federationn Academy of Sciences; (dau@imm.uran.ru) - ;
Abstract: This paper considers the problem of providing a scalable computing infrastructure to consolidate computing equipment used for developing and debugging software as well as for the didactic support of educational courses. Software specifications are formulated and the following existent solutions are reviewed and compared: Eucalyptus, OpenNebula, OpenStack, and Nimbus. A decision is made to perform an experimental deployment of the OpenNebula suite based on the KVM hypervisor and the GNU/Linux operating system. OpenNebula is an open source software designed to control the cloud environment of the Infrastructure-As-a-Service class. The KVM hypervisor was developed by Red Hat and is supported by OpenNebula as a default virtualization solution. The available computational resources of the Institute of Mathematics and Mechanics, Ural Branch of the Russian Academy of Sciences, and dedicated backend nodes for the experimental cloud computing environment are described. The configuration of a typical cloud computing node under the Scientific Linux 6.1 (x86_64) operating system with the LVM, KVM, and libvirt technologies is presented. The cloud environment architecture is presented at both logical and physical levels. Some difficulties concerning the SELinux security system and the fault-tolerance performance are described. A plan of further work in the following directions is suggested: integration with the Active Directory domain system, application of the network-attached storage to keep the images of virtual machines, migration to the MySQL relational database, and the testing and presentation of scalable Web services to cloud computing end users at the IMM UrB RAS.
Keywords: iaas, opennebula, kvm, InfrastractureasaService, cloud computing, virtual machine hypervisor, virtualization
Visitors: 11674

23. Workflow with flexible transitions of stellar type [№4 за 2012 год]
Authors: (svysh@pn.sinp.msu.ru) - , Ph.D; ( jdubenskaya@pn.sinp.msu.ru) - ;
Abstract: A method of building interactive software systems is proposed, which is based on flexible workflow technology. In these systems a violation of the technological discipline is allowed, which comprise unforeseen transitions between the allowed states. During the evolution of the individual workflow instance (token) it may be needed to change the order of the states, to skip some state or to return again to a state in which the token has already been. These actions can be performed in a special operation mode of the information system. To implement this mode an initial description of workflow (which does not allow for violation of the technological discipline) should be in advance modified the addition of special states and transitions. The method is mainly implemented using standard definition of workflow in a high level language. In this case, from the point of view of the operator, unforeseen transition is barely different from the usual transition, and the operator may not be aware of the fact that the interactive system is based on workflow technology. The flexibility of the system is achieved in the course of its operation, and the operator does not need to be a qualified developer or analyst. The method is presented with an example of the open source software called perl-workflow. The proposed method of implementing unforeseen transitions in workflow is applicable, for example, in monitoring systems. These results extend the range of applicability of information systems based on workflow technology.
Keywords: open source, the automated information system, actions, states, flexible routes, graph, information system, technological process
Visitors: 10368

24. Fuzzy classifier base rule collective forming [№4 за 2012 год]
Author: (romaserg@list.ru) - ;
Abstract: An approach of fuzzy classifier base rule collective forming with generated with self-tuning coevolutionary algorithms fuzzy classifier ensemble is observed. Results of the approach applying for practical classification problems are presented. The main result of our work is collective fuzzy classifier forming method. Having generated some fuzzy classifiers we are able to construct more effective classifier from previous classifiers using again cooperative-competitive coevolutionary algorithm. A number of using fuzzy rules isn’t increasing with this method. The approach of multistep fuzzy classifier forming has the following features: the method improves classification performance without increasing number of rules, the method reduces diversity of performance values for multiple algorithm runs, i.e. the method has higher statistical stability, the method is more effective for more complicated classification problems (more attributes and classes).Fuzzy classifier forming methods comparison with alternative classification methods by performance value demonstrates that both fuzzy classifier forming methods have better efficiency that present-day classification algorithms.
Keywords: coevolutionary algorithm, algorithm ensemble, collective decision, fuzzy classifier
Visitors: 6305

25. Noise combating layered space block code [№4 за 2012 год]
Author: (maxgof@gmail.com) - ;
Abstract: The transmitter is equipped with multiple transmit antennas can transmit multiple signals simultaneously, and the receiver uses multiple receiving antennas, allows more reliable reception. Creating space code, which will be used in the multi-antenna communication systems, we have to choose between the speed of transmission and noise combating. This article presents noise combating layered space code. Parameters allow you to get the code of its implementation, focused on a particular channel with fading. Feature of the code is that it does not have symbolic redundancy, and the fact that each of the transmit antennas are always involved. The proposed code uses the spatial, temporal and frequency types of diversity.
Keywords: generator matrix, positioning function, channel with fading, noise combating layered space code, mimo communication system
Visitors: 5823

26. A fuzzy cognitive map as a tool to model innovation at the regional level [№4 за 2012 год]
Authors: Palyukh B.V. (pboris@tstu.tver.ru) - Tver State Technical University, Ph.D; (tatjank@yandex.ru) - , Ph.D;
Abstract: For developing strategic innovative programs for regional industry and improve implementation of major investment projects, the technique of modeling innovation using multilevel fuzzy cognitive model to determine the relative impact of performance strategies of socio-economic development and innovation strategies of the regional industry. Two basic features of modeling innovation using fuzzy cognitive maps as a special directed graphs, nodes (concepts) which are the characteristics of the socio-economic development and regional industry. The first feature modeling innovation necessitates the definition of traditional indicators of system interaction graph nodes: consonance, dissonance, mediated by the mutual influence of nodes on each other, etc. The second feature is caused by the weight of arcs between nodes in the graph (the concept) changes over time, so that can change not only the value but also the nature of influence. A new kind of fuzzy cognitive maps with the uncertainty of system characteristics, wherein the relationship between the concepts of the influence of the graph are fuzzy sets. Accounting for the negative impact of concepts achieved by broadening the base set for these fuzzy sets on negative values. A simplified two-level cognitive model of management of innovation processes in the regional industrial complex of the Smolensk region. It is shown that the use of the changing relationships between concepts of cognitive maps will extend the scope of the unit of cognitive modeling for management innovation and improve the validity of the provisions of innovative strategies and programs for the development of the region.
Keywords: the regional industrial complex, the region, innovation, innovation, modeling, fuzzy sets, fuzzy cognitive model
Visitors: 16170

27. Genetic algorithm for nfa state minimization problem [№4 за 2012 год]
Authors: Tsyganov A.V. (andrew.tsyganov@gmail.com) - Ulyanovsk State Pedagogical University named after I.N. Ulyanov (Professor), Ph.D;
Abstract: The state minimization problem for nondeterministic finite automata is a well-known computationally hard combinatorial optimization problem. A lot of exact and approximate methods were proposed for it. All known exact algorithms for this problem are exhaustive and often become impractical even for relatively small automata. In the present paper we discuss a new heuristic algorithm for NFA state minimization problem which is based on the classical Kameda– Weiner algorithm and genetic algorithm. The main idea of the proposed method is to replace the exhaustive search for legitimate covers of the RAM (Reduced Automaton Matrix) with the fast but incomplete search for covers by the means of genetic algorithm. The implementation of the proposed method with the usage of parallel computing techniques is described and the results of computational experiments are provided.
Keywords: mpi, OpenMP, parallel computing, generic algorithm, kameda–weiner algorithm, state minimization, nondeterministic finite automata
Visitors: 10629

28. Studying the statistical properties of finite automata minimization algorithms using ReFaM [№4 за 2012 год]
Authors: Tsyganov A.V. (andrew.tsyganov@gmail.com) - Ulyanovsk State Pedagogical University named after I.N. Ulyanov (Professor), Ph.D; (PhoenixDragonViSta@yandex.ru) - ; (alexlumen@rambler.ru) - ;
Abstract: In the present paper we consider the minimization of nondeterminisitc finite automata using ReFaM. This experimental open source software tool provides several exact and approximate state minimization algorithms, such as classical Kameda–Weiner algorithm and some heuristic algorithms based on it, which are implemented using OpenMP and MPI parallel programming techniques. Usually software products that deal with finite automata and related structures do not provide algorithms for nondeterministic finite automata minimization due to their computational complexity therefore the considered program can be used for research and educational purposes. One of the distinguishing features of this software is that it explains the minimization process in details and collects statistics of all minimization steps. The description of the program features as well as some experimental results is provided.
Keywords: mpi, OpenMP, parallel computing, heuristic algorithms, state minimization, nondeterministic finite automata
Visitors: 10654

29. Implementation method for interactive monitoring systems based on flexible workflows [№4 за 2012 год]
Authors: (svysh@pn.sinp.msu.ru) - , Ph.D; ( jdubenskaya@pn.sinp.msu.ru) - ; (peter@kapella.gpi.ru) - , Ph.D;
Abstract: Implementation method is proposed for flexible interactive monitoring systems applicable to practically significant cases where the information about the original system goes into monitoring system irregularly, intermittently, with delays or in violation of chronological order. The method also addresses a situation when the original system consciously violates the technological discipline. The monitoring system is constructed using workflow technology. A violation of the original workflow is described with a supplementary workflow. The monitoring system creates a tracking token, which is coupled to the token of the original system. The workflow for the tracking token is specified in a high-level language, and is a simple function of the workflow of the original system. To allow for flexible transitions the system has a special mode of operation. The proposed system is able to reconstruct the history of each instance of the source workflow, analyzing scattered messages from the original system and has some tolerance to the loss of these messages.
Keywords: flexible routes, graph, information system, the automated information system, interactive system, technological process
Visitors: 10663

30. Intelligent control of multilinked objects implemented in the situational subprograms [№4 за 2012 год]
Authors: (muraveva_ea@mail.ru) - , Ph.D;
Abstract: A novel technique has been suggested for increasing multidimensional precise logic controller (MPLC) speed of action based on arrangement of production rules in situation subprograms, initial points of which are registered in special register by fuzzification procedure according to equality to the logical unit of the control parameters precise terms. It enabled to process but one subprogram selected during fuzzification rather than the entire system of the controller production rules without damaging control adequacy in each scanning cycle. According to their logical nature, each situation subprogram is a software implementation of a production rule, which must be executed when one of the precise terms of each MPLC controlled parameter is equal to a logical unit. The correct assignment of MPLC controlled output variables ranges the microprocessor executes one rule for each MPLC loop. And to find this rule, the standard procedure for applying to the subprogram is used, which is now well minimized by microprocessor time and storage. It should be noted that in the standard multivariate fuzzy logic controllers the entire system is executed, comprising out of tens or even hundreds of production rules in similar situations in each loop. Two fundamental properties (at any given instant both only one term and only an antecedent of one rule is equal to a logical unit) of precise terms set and production rules systems under consideration are the theoretical basis for improving MPLC speed and reducing its error. This article describes the results of a quantitative estimation of scannig time reduction of the program implementing MPLC in comparison with a typical fuzzy controller. Proposed logic controller used in the distiller control system. Using of multi-dimensional logic controller has reduced the average error of regulation by 42 %, and the liquid pH value at the output of distiller has reduced by 57 %.
Keywords: fuzzification procedure, scanning cycle, precise term, speed, production rule, multidimensional precise logic controller, situational subprogram, multilinked object, intelligent control
Visitors: 8632

◄ ← Preview | 1 | 2 | 3 | 4 | 5 | 6 | Next → ►