Другие журналы

scientific edition of Bauman MSTUSCIENCE & EDUCATIONBauman Moscow State Technical University. El № FS 77  48211. ISSN 19940408
Informatics, Computer Science and Management
Global optimization with the use of a biogeographybased method
# 10, October 2013 DOI: 10.7463/1013.0605836 This article deals with a promising global optimization method that is a method of biogeography (BBO). Efficiency of this algorithm and its implementation was demonstrated by the example of the known complicated minimization problem of the multidimensional Rastrigin function. Integration of implementation of the biogeography method developed in MATLAB with PowerFactory software of DlgSilent was considered; PowerFactory software was used to calculate total power losses in the electric circuit being under consideration. Efficiency of the biogeography method for solving optimization problems of IEEE9 and Kuban’s power networks was investigated with the use of MATLAB BBO + PowerFactory. By optimizing a set of compensators and their physical layout, realpower losses in the specified networks were successfully reduced by 9.2% and 2.5% respectively.
Estimation of computational complexity of the fast numerical algorithm for calculating vortex influence in the vortex element method
# 10, October 2013 DOI: 10.7463/1013.0604030 In this paper the authors consider a fast numerical algorithm for calculating convective velocities in the vortex element method. This algorithm is based on derivation of a tree of rectangular regions which contain vortex elements. Influence of parameters of the fast method on computational complexity of the algorithm was analyzed. The number of arithmetic operations required for calculating velocities of all vortex elements was evaluated for modeling an evolution of vortex motion in an ideal incompressible liquid. This theoretical estimation is in good agreement with the results of the numerical experiment. The developed approach could be useful in practice because it allows to select optimal values of the algorithm parameters a priori.
Technology of complex research of energy sector functioning in emergency situations with the use of Bayseian networks
# 08, August 2013 DOI: 10.7463/0813.0603646 In this paper the author proposes a technology of organization and execution of computational experiments in the field of energy security with the purpose of analyzing complex scenarios of emergency situations. This technology is based on application of intellectual research support within integrated ITenvironment and integration with Bayesian networks tool for modeling energy security threats. An example of a naturallyemerging threat of cold snap in the Central interconnected power system was also described. The presented technology gives a new opportunity in describing energy security threats; this is obtained due to the joint use of intellectual technologies for generation of computation scenarios (at qualitative level) and further transition to a traditional quantitative research.
Study of the spectral properties of LiveJournal’s social graph
# 09, September 2013 DOI: 10.7463/0913.0603441 In this paper we compute some characteristics of the spectrum of LiveJournal’s social graph and estimate the vertex and edge expansion ratios of this graph. We use the Power Iteration algorithm to compute first and second elements of the graph’s spectrum. The concept of application of the spectral graph theory methods to problems of social network analysis appeared to be very promising. Methods described in this paper could be applied to analysis of social networks, study of the social interactions between people and many other problems at the intersection of sociology, computer science and information security.
Experiments with teaching of recognition methods for road signs with the use of synthetic data
# 08, August 2013 DOI: 10.7463/0813.0603378 This article considers the task of traffic sign recognition with learning from synthetic data. The method for synthetic dataset generation from one traffic signs pictogram for each class was described. The authors proposed a method for traffic signs segmentation and classification; the method is based on the nearest neighbor search in the space of gradient features derived from synthetic data. The proposed method was evaluated on databases from three different countries – Germany, Belgium and Sweden. A comparative analysis of classification accuracy of the proposed method with training from synthetic and real data was carried out.
A modified cuckoo search method in the global optimization problem
# 09, September 2013 DOI: 10.7463/0913.0603388 In this paper we consider a global optimization problem and a method of Cuckoo Search (CS). CS method could be used for solving the specified problem. This method belongs to a new class of population based algorithms, which have been developed actively in recent years. In contrast with other known population based methods CS has only two free parameters. The goal of this work is to increase the efficiency of the canonical CS method. Several different modifications were presented in this work along with their software implementations; the canonical method was also implemented. Results of comprehensive performance study of proposed modifications by the example of a set of benchmark functions which revealed advantages over the canonical method were presented. One of proposed modifications was used for solving the existent practical problem of minimizing the expenditures on the manufacturing of highpressure vessels. It was shown that proposed modifications provide a better value of the objective function in comparison with results obtained by other researches and also allow one to meet all restrictions.
Parallel algorithm for solving systems of linear algebraic equations with a multidiagonal coefficient matrix
# 07, July 2013 DOI: 10.7463/0713.0590785 This article describes a parallel algorithm for solving systems of linear algebraic equations with a multidiagonal (band) coefficient matrix. A scheme of the algorithm for multiprocessor sharedmemory computing systems is also presented in the article. Data structures for a compact storage of sparse band matrices were implemented. Theoretical and experimental investigations of efficiency of the algorithm were conducted using various problems. Dependences of the program’s speedup on various parameters of the problem and the number of processors were given in this work. The conclusion presents pros and cons of the algorithm.
Dodecapod’s movement algorithm within a linear cylindrical tube of variable crosssection
# 08, August 2013 DOI: 10.7463/0813.0587740 A dodecapod is a onesection parallel manipulator which could be considered as an evolution of the wideknown hexapod, or the Stewart platform. The idea of a dodecapod was proposed by prof. S.N. Sayapin in 2012. In this article the authors consider one of the dodecapod’s functional capabilities: movement within a linear cylindrical tube of uniform and variable crosssections. A design model of the dodecapod was presented in this paper along with propagation conditions for dodecapod’s movement within the tubes and the algorithms of changing the lengths of dodecapod’s rods, which make its movement possible. The presented results allow to draw a conclusion that a onesectional parallel manipulator based on a dodecapod could be applied to creation of mechanisms for execution of technological operations in linear cylindrical tubes of uniform and variable crosssections.
Analysis of University’s corporative network traffic by the methods of nonlinear dynamics
# 08, August 2013 DOI: 10.7463/0813.0587054 Results of calculations of incoming and outgoing traffic’s dynamic characteristics were presented in this paper along with the distribution of hardware capacities by the example of the university’s corporate network server. Lyapunov and Hurst exponents, which characterize chaotic and fractal properties of the processes, were evaluated for network traffic. Traffic’s phase diagrams were investigated and, for the first time, appearance of attractors was revealed; these attractors allow one to conduct a more detailed analysis of the impact of load on the network’s traffic capacity. A selfsimilarity was found for all the processes; this fact confirms the possibility of using fractal models for working with data, including solving urgent problems of predicting behavior of the time series.
A statistical model of the market sector of macroeconomic management
# 06, June 2013 DOI: 10.7463/0613.0586255 The author considers a statistical model of the market sector of the macroeconomic management system, where supply and demand lack any regular pattern due to the statistical spread of customers’ incomes and sellers’ profitability. Probability density distribution of customers’ incomes and sellers’ profitability was determined on the basis of studying market statistics. Using these distributions it was possible to obtain distribution density of possible supply and demand over the market price. Natural and monetary sales volumes, mean price, supply and demand dynamics were also determined. A feature of this model is simultaneous existence of an unsatisfied demand and a stock in trade, which, for example, is typical for the labor market.
Special AWS for designing software for navigation complexes and aerial surveillance
# 08, August 2013 DOI: 10.7463/0813.0585918 This paper is based on the results of investigations of computeraided design facilities for systems software and software for terrestrial and onboard facilities of aerial surveillance and Earth remote sensing. The article presents a review of typical airplanes used for carrying navigation equipment and aerial surveillance along with the investigation of intelligence equipment, key factors of airplane navigation and hardware properties which determine the quality of intelligence data obtained in different modes of aerial surveillance, special features of intelligence data preprocessing inflight, peculiarities of ground information system for processing intelligence data and preparing intelligence reports. Configuration of a special AWS was justified.
Numerical solving static equations in a charged gas flow model with the use of Haar wavelet method
# 08, August 2013 DOI: 10.7463/0813.0583405 This article considers the Cauchy problem for the second order automodeling nonlinear ordinary differential equation in the problem of twofluid plasma motion. An algorithm for solving this problem with the use of Haar wavelet method was described. A numerical experiment showed that Haar wavelet method is more accurate than the fourth order RungeKutta method for few collocation points in problem 1 with a known analytical solution. In problem 2, where the exact solution is unknown, relative error between numerical solutions by Haar wavelet method and RungeKutta method tends to decrease with the growth of the number of collocation points.
Estimation of working precision of the method of automated identification of lexical items by using virtual texts
# 08, August 2013 DOI: 10.7463/0813.0583389 This article describes a method for analyzing working precision of algorithms of automated identification of lexical items of natural language word forms on the basis of virtual texts. Estimation of working precision of this complex automated method was carried out with lexical items of Russian word forms. Expediency of applying this comprehensive solution to the problem was proved. The nature of identification errors was revealed, and on the basis of the results of the conducted analysis recommendations were given on modification of the proposed method in order to increase accuracy of automated identification of Russian lexical items.
The University’s knowledge management system
# 06, June 2013 DOI: 10.7463/0613.0581872 In 2012 at MSAL named after Kutafin a scientific research was carried out; the aim of the research was to determine prospects for further development of the University’s knowledge management system. In recent years many researchers paid significant attention to development of this branch of IT penetration into the field of education. After analyzing the existent experience, a group of authors defined a check list of the most promising and successful projects which had been implemented, in a varying degree, with the use of knowledge management systems. The authors made special reference to the system for generating educational content developed at Bauman MSTU under the auspices of prof. I.P. Norenkov. This particular direction that allows to increase efficiency of development and actualization of electronic educational resources, in the future, will help to create unified subject areas in universities, which, in their turn, will clear the way to new learning technologies based on project and crossdisciplinary forms of the teaching process. This article describes problems of electronic educational resources in the university and approaches to their solving. A set of scenarios of formation, visualization and application of knowledge bases was also considered.
Approach to calculation of weighting coefficients of experts’ rank assessments when selecting a development option for an information system
# 08, August 2013 DOI: 10.7463/0813.0580272 An approach to calculation of weighting coefficients of experts’ rank assessments, which is aimed at improvement of processing results of an expert survey, was proposed. This approach was designed to increase the concurrence level of experts’ opinions within a formed working group. Weighting coefficients are calculated either considering known numerical values of experts’ experience levels or on the basis of experts’ rank assessments produced from a test assignment. Recommendations on formation of a competent working group of experts who belong to the same school of sciences, as well as rules for using the normalized rank system, are at the heart of this approach. An example that illustrates efficiency of the proposed approach was presented in this paper.
Measuring the density of conduction current without distorting the field being investigated
# 07, July 2013 DOI: 10.7463/0713.0578057 Problems of development and design of sensors of the electric field or the density of conduction current in conducting media, without distorting the field being investigated, are considered in this article. Construction of a sensor which doesn’t distort the picture of the field when one brings it into a specified volume of a conductive area was considered on condition that the field is homogeneous. This condition is met when the sensor’s geometrical dimensions are much smaller than the scale of its heterogeneity. Original designs of real electric field sensors with electrolytic water channels which do not distort the field being investigated and allow one to measure both one component of the tension and three components, were considered. Recommendations on selecting geometrical dimensions of water channels were given. Analytical dependences for determining metrological characteristics of such sensors were also obtained.
Synthesis of orientation mechanisms for the space observatory “Millimetron”. 4. The concept of intelligence system of active vibration protection and very precise pointing of the space observatory “Millimetron”
# 06, June 2013 DOI: 10.7463/0613.0574243 This is the fourth paper in the series of four works dedicated to design of parallel mechanisms for orientation of the space observatory “Millimetron”. In this work the authors consider problems of providing highprecision geometry of the space observatory “Millimetron” released on the orbit along with its very precise pointing at the observable object. These problems are caused by very high sensitivity of the observatory and its operation under conditions of deep vacuum and extralow temperatures. The concept of intelligence system of active vibration protection and very precise pointing of the space observatory “Millimetron” was introduced. This is a unique system; it allows to increase efficiency of vibration protection and precise pointing of the observatory significantly.
Investigation of strength of block ciphers based on generalized cellular automata against linear cryptanalysis
# 05, May 2013 DOI: 10.7463/0513.0574231 In this paper the authors investigates strength of block ciphers based on generalized cellular automata, against classical linear cryptanalysis. A linear cryptanalysis is one of the main standard techniques for cracking block ciphers. Sufficient condition for strength of this family of ciphers against linear cryptanalysis is the main result of this paper. This condition was used to demonstrate that with an arbitrary graph of a cellular automaton in case of correct selection of a local connecting function and other parameters in order to create ciphers with key length of 128 and 256 bits and block length of 128 bits which couldn’t be cracked by linear cryptanalysis when using 4 rounds and 7 steps of a cellular automaton per round.
Approach to increasing the level of consistency of expert opinion when selecting the variant of development of the data processing system
# 06, June 2013 DOI: 10.7463/0613.0574220 The authors propose a new approach aimed at formation of an expert group and improvement of polling organization for this expert group in order to obtain relevant information when selecting the best variant of development of the data processing system. The specified approach is based on the recommendations on the formation of a competent working expert group; rules for using a standardized rank system; recommendations on organization of expertpolls; rules for processing results of these polls; recommendations on result adjustments for certain experts of the working group aimed at the increase in the level of consistency of expert opinions. An example that illustrates efficiency of the proposed approach was presented in the article.
Synthesis of orientation mechanisms for the space observatory “Millimetron”. 3. Synthesis of parallelstructure mechanisms for orientation of the space observatory’s antenna
# 05, May 2013 DOI: 10.7463/0513.0571127 This is the third paper in the series of four works dedicated to design of parallel mechanisms for orientation of the space observatory “Millimetron”». In this work, a method of synthesis of plain sixlink mechanisms with three degrees of freedom was proposed. This method takes into account restrictions imposed by requirements to orientation of the space observatory’s antenna. A distinctive feature of the proposed method is decomposition of the synthesized mechanism into two more simple mechanisms – a double pendulum mechanism and a twobeam mechanism. The first mechanism has two degrees of freedom and is designed for maintaining a point in a constant position; the second mechanism’s output link rotates around the first one by a specific angle. All unknown sizes of the mechanism’s links and turning angles were determined in this work. The authors defined a dependence of the turning angle of the output link on the turning angle of the input link of the twobeam mechanism that is a transfer function. Besides, the problem of maintaining an invariable position of the center of mass of the space observatory’s antenna was considered when changing its orientation.



