Software & Systems

Journal articles №4 2012

11. Controlling in the process management [№4 за 2012 год]
Authors: (IFedorov@mesi.ru) - , Ph.D;
Abstract: The interpretation of process management as pure process model re-design, limit our capability to control the process. Model modification is an essential but not the only way of process management. This article suggests to distinguish between management of an enterprise with the use of business processes and the controlling of the process itself. The first is a holistic management approach focused on aligning all aspects of an organization with the wants and needs of clients. The second is an effort to maintain normal operation of the process. We will distinguish three levels of process controlling. The level of operational control implements the concept of proactive BPM. Its main idea is to exclude managers from the routine duties to manage the process instance. Their involvement happens only in cases where it is really necessary. The proactive management requires special process logic. The level of a tactical controlling is carried out through the manipulation of additional layers and aspects of the process models that are not covered by the process map, modeled with BPMN. Thus it is possible to achieve, for example, the optimal allocation of human resources involved in the process. Finally, a strategic management level involves improvement of the process model and is applicable only if the first two levels are no longer able to achieve their goals or there was a radical change in the business. The analysts and developers of the business processes management systems need to be aware of and possess the techniques of process' controlling. A BPM system developer must know and master the methods of process’s management and include necessary patterns of operational and tactical control in the process model. We assume that this recommendation can help a developer make an executable process model a controllable and can help your organization achieve high productivity and efficiency with a minimal changes of their process model.
Keywords: controlling, reengineering, of organizational management, business process management
Visitors: 11403

12. System of automatic design of intelligent information technologies ensembles For data mining problems [№4 за 2012 год]
Author: Semenkin E.S. (styugin@rambler.ru) - Academician M.F. Reshetnev Siberian State Aerospace University; (shabalovGandrey@mail.ru) - ;
Abstract: An algorithmic core of the program system of intelligent information technologies (IIT) automatic design as well as their ensembles for solving complex weak-defined data mining problems and decision support is considered in the article. Owing to an automation of intelligent systems algorithmic core generation, the process of the design of intelligent decision support systems for end-users is simplified considerably. The generation of the ITT’s structure and parameters is carried out by means of evolutionary algorithms because of their proved reliability and performance in solving optimization problems on complex data structures. A genetic algorithm is used as the basis for the structure design and parameter tuning of connectionist models, fuzzy rule based systems and neuro-fuzzy models. Many real-world problems are large and complex and too difficult for solving it by a single technology. Ensemble techniques allow improving the effectiveness and reliability of the decision support systems. An idea of applying a genetic programming method for building an ensemble decision based on partial individual decisions is proposed in the article. Conducted numerous experiments on real-world problems show efficiency and effectiveness of the proposed techniques of automatic IIT generation algorithms and prospective of the ensemble decision building method based on the genetic programming.
Keywords: ensembles of intelligent systems, neurofuzzy systems, fuzzy systems, artificial neural networks, evolutionary algorithms
Visitors: 9897

13. Software package of adaptive evolutionary algorithms for complex systems modeling and optimization [№4 за 2012 год]
Author: Semenkin E.S. (styugin@rambler.ru) - Academician M.F. Reshetnev Siberian State Aerospace University; (semenkina88@mail.ru) - ;
Abstract: In this paper authors justify relevance and describe results of adaptive evolutionary algorithms development and their application in modeling and optimization of complex systems. Due to the essential evolutionary algorithms efficiency dependence from their settings choice the concept of self-configured evolutionary algorithm is introduced as one of possible adaptive evolutionary methods. Basic ideas of proposed approach to self-configuration of genetic operator different variants are described, results of the verification of the self-configured evolutionary algorithm performance on a set of test problems are discussed, the comparison results of the proposed self-configured evolutionary algorithms with the standard evolutionary algorithms are presented and the way for application of the self-configured genetic programming algorithm for automatic design of various intelligent information technologies are described. In addition, this article describes an approach using genetic programming algorithm with self-configuration for the automatic intelligent information technology ensembles generation, discusses the composition and mission of an adaptive evolutionary algorithms program complex for modeling and optimization of complex systems using self-configured genetic programming algorithms and selfconfigured genetic algorithm. The obtained results allow concluding that the discussed approach has a wide range application areas and high performance in solving practical problems.
Keywords: self configuration, intelligent information technologies, genetic programming algorithm, adaptive evolutionary algorithms
Visitors: 6359

14. Development of journal partition for academic site based on CMS NetCat [№4 за 2012 год]
Authors: (nad@imach.uran.ru) - ; (raul@imach.uran.ru) - , Ph.D; () - ;
Abstract: The question is discussed how to create Internet version of periodic science edition in accordance with requirements of VAC. To optimize time required for development and filling of content we use approach based on modernization of existing site of organization. For development of academic sites we use CMS NetCat, which is the most applicable in Russia site constructor. Site is represented as an external interface to internal database that contains information about organization, its goals, plans and results of activity. Data is extracted from the database by requests that are created in accordance with actual demands for representation of information. To represent partition of periodic science edition we create special requests to base of authors and publications.
Keywords: vac, periodic science edition, database of publications, NetCat, site modernization, site development, web programming, content managements systems
Visitors: 8412

15. Qualitative and quantitative characteristics of the open information systems [№4 за 2012 год]
Authors: (chti@chti.ru) - , Ph.D; (chti@chti.ru) - , Ph.D;
Abstract: This article deals with the method of the formation of the virtual laboratory complexes and its location through the use of an automated system in the open information space to carry out the remote experiments. As such, the virtual laboratory complex and the automated system are the effective tools to integrate the intellectual assets, user-friendly and often indispensable tool to solve the scientific, educational and business problems. The research resources include the intellectual assets (object models and modes of their research, computer programs of research, experiments results). We determined the main processes of the formation of the virtual laboratory complexes: problem statement, preparation for the formalization, formalization, algorithm elaboration, building of the multitemporal models or problem solution using the hardware and software system; operation with the simulation model of the object; and analysis of the results and search for knowledge retrieval. The automated system and the virtual laboratory complexes are developed in accordance with the principles of the open systems and ensure the following characteristics achievement: ergonomics, it’s handy and practical tool for the scientific, educational and business researchers; scalability, it’s expandability by users and virtual laboratories; it supports the tasks of the several basic groups of users (science, education, business); it has a mechanism to expand according to the types and number of the virtual laboratories; interoperability, it integrates and manages by diverse software and hardware, by process-oriented research; mobility, it supports the virtual laboratories, regardless of their hardware.
Keywords: regulations, software, quality, virtual experiment, system analysis, information technologies, integration
Visitors: 14630

16. Knowledge base editor for CLIPS [№4 за 2012 год]
Authors: Yurin A.Yu. (iskander@irk.ru) - Institute of system dynamics and control theory SB RAS, National Research Irkutsk State Technical University, Ph.D; Grishchenko M.A. (makcmg@icc.ru) - Institute of System Dynamics and Control Theory SB RAS;
Abstract: The paper describes the knowledge base editor designed for non-programmers. Application of the editor automates the formalization and language-specific representation of knowledge. High universality of the editor is achieved by the use of a universal data model – an abstract (meta) level of the rules’ representation. The description of the requirements for the editor, tasks, editor architecture, editor basic functions and a universal model are presented. The model includes the concepts: knowledge base, template, fact, slot, rule, condition, precondition, action, function, variable, argument. The feature of the editor is the ability to integrate with domain models built using CASE-tools (for example, Poseidon for UML, IBM Rational Rose), which use the XMI format for the description of the models. The editor has an extensible architecture that is implemented the ability to connect the dynamic libraries (modules) for supporting different knowledge representation languages. Now the CLIPS representation support module is implemented. This module provides the translation of knowledge from the universal model to CLIPS and back. The editor used for development of knowledge bases for solving the problems of technical state prognosis and calculating of remaining life of parts of unique machines and equipment in the petrochemical industry and for automated design of fault and event trees on the basis of a model of the dynamics of technical states.
Keywords: metalevel, editor, rules, knowledge base, expert systems
Visitors: 12052

17. Extracting meaning from sentences of natural language [№4 за 2012 год]
Authors: (a_vavilenkova@mail.ru) - , Ph.D;
Abstract: It was proposed a unified form of logical-linguistic model, which is based on the syntactic structure of a sentence of natural language, the meaning of which is to determine reference between the subject (the subject) and the predicate (predicate). The research of various types of sentences was carried out by means of constructing their logicallinguistic models. This proves that the author’s proposed unified form is a kind of template owing by we can formalize all types of sentences. Semantic references between words can be traced in the context formed phrases. The construction of these models allows to compare the texts in the context and to extract knowledge from text information, as well as to find out among them the controversy. Verification of the proposed ideas to represent text information in the form of logical-linguistic models of unified form is an intellectual system CSOFLLM (computer-aided system of forming logical-linguistic models), which is based on the principles of expert systems.
Keywords: formalization, structural linguistics, unified form, a predicate, the extraction of meaning, logicallinguistic model, semantics
Visitors: 7120

18. Neutral networks and regression models in S&P 500 index forecasting [№4 за 2012 год]
Author: (oskolkovama@gmail.com) - ; (parshakov.petr@gmail.com) - ;
Abstract: The article presents a comparative analysis of neural network modeling and regression analysis for forecasting the S&P 500 index. Initially, the forecast of the absolute value of the index is provided, then we justify the use of stationery data, that is, the return of S&P 500. The comparison of two methods is carried out in two stages. Firstly methods are compared by the coefficient of determination on the periods of three and twelve months, and by the quality of trend predictions. Note that the choice of model and its testing is performed at different time intervals (the so-called in-sample and out-of-sample periods). Taking into account the fact that the primary desire of a typical trader is to gain a profit at the second stage we have chosen such trading criteria as profit and profit, weighted on risk (drawdown). On a longer time interval (12 months) regression shows the best results, but in terms of economic gains neural network win. When we consider a shorter period (3 months) neural network has better results. Thus, neural networks are able to assess the dynamics of the stock due to its flexibility and ability to find non-linear patterns.
Keywords: artificial intellect, index yield forecasting, s&p 500, neural network modeling
Visitors: 11851

19. The algorithm and software implementation of the hybrid training method of artificial neural networks [№4 за 2012 год]
Authors: Belyavsky G.I. (gbelyavski@sfedu.ru) - Scientific Reseach Institute of Mechanics and Applied Mathematics Southern Federal University, Ph.D; Lila V.B. (lila@i-intellect.ru) - Rostov State University of Civil Engineering, Ph.D; Puchkov E.V. (puchkoff@i-intellect.ru) - Rostov State University of Civil Engineering, Ph.D;
Abstract: The task of training an artificial neural network can be considered as a task of optimisation, and the main problem is to choose among a variety of optimization methods most suitable. The choice in favor of gradient methods based on the fact that, as a rule, in the problems of network teaching objective function can be expressed as a differentiable function of all weighting coefficients. However, the complicated nature of dependence between weighting coefficients leads to the fact that the objective function has local extremums and saddle points, making the use of gradient methods are not always justified. For solving optimization problems with multiextremal criterion the methods of random search, which include genetic algorithms, are used. Genetic algorithms, as a rule, are notable for their slow convergence. For the comparative performance of gradient methods and genetic algorithm a software with a web interface has been created. In the capacity of the task of training the neural network a problem of approximation of two-dimensional Rosenbrock function was used. The results of this research showed that gradient methods have fast convergence only at the beginning of training, and the genetic algorithm – in the end. Thus, a hybrid algorithm based on the sequential use of gradient methods and genetic algorithm has been proposed.
Keywords: web application, rosenbrock function, genetic algorithm, hybrid training method, multilayer perceptron, artificial neural networks
Visitors: 13647

20. Evolutionary three-stage approach for designing of neural networks ensembles for classification problems [№4 за 2012 год]
Authors: Bukhtoyarov V.V. (vladber@list.ru) - Academician M.F. Reshetnev Siberian State Aerospace University, Ph.D;
Abstract: The use of neural networks ensemble approach for solving classification problems is discussed. Methods for forming ensembles of neural networks and methods for combining solutions in ensembles of classifiers are reviewed briefly. The main ideas of comprehensive evolutionary approach for automatic design of neural networks ensembles are described. A new variant of a three-level evolutionary approach to decision making in ensembles of neural networks is proposed for classification problems. Steps to produce solutions using proposed method are described. Comparative experiments on a number of known test problems (benchmarks) were carried out to evaluate the efficiency of proposed approach. A brief description of the used test problems is given. The conditions and results of the comparative statistical study of different classification methods efficiency are also described. The results of the numerical experiments which were carried out show that the proposed approach is not less effective on used test problems than the most of methods which were taken into consideration. And it also allows to obtain best results on a number of test problems. The directions for further investigations concerning evaluating of the efficiency of the proposed approach on wide range of large-scale data sets and prospects of applying approach in real-world applications are briefly marked in the conclusion.
Keywords: combining strategies, ensembles, neural network, classifiers
Visitors: 10061

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