Другие журналы

scientific edition of Bauman MSTUSCIENCE & EDUCATIONBauman Moscow State Technical University. El № FS 77  48211. ISSN 19940408
Informatics, Computer Science and Management
Decomposition method for analysis of closed queuing networks
# 02, February 2014 DOI: 10.7463/0214.0700018 This article deals with the method to estimate the average residence time in nodes of closed queuing networks with priorities and a wide range of conservative disciplines to be served. The method is based on a decomposition of entire closed queuing network into a set of simple basic queuing systems such as MGImN for each node. The unknown average residence times in the network nodes are interrelated through a system of nonlinear equations. The fact that there is a solution of this system has been proved. An iterative procedure based on NewtonKantorovich method is proposed for finding the solution of such system. This procedure provides fast convergence to solution. Today possibilities of proposed method are limited by known analytical solutions for simple basic queuing systems of MGImN type.
Reduction of the reversible circuits gate complexity without using the equivalent replacement tables for the gate compositions
# 03, March 2014 DOI: 10.7463/0314.0699195 This paper studies the problem of reducing the gate complexity of the reversible circuits consisting of NOT and kCNOT gates without using replacement tables. Generalization of kCNOT gate is made for the case of zero value at some of the gate control inputs. For such generalized kCNOT gates the equivalent replacements for the reversible gates compositions are considered. The proof of correctness for such replacements is provided by comparing result transformations of the gate compositions before and after replacement. Application of the equivalent replacements for reduction of the gate complexity of the abstract reversible circuit is shown.
Software complex for finding the path routings in synchronous digital hierarchy networks
# 02, February 2014 DOI: 10.7463/0214.0698921 The paper addresses the communication network constructions based on the synchronous digital hierarchy (SDH) via software complex, which uses Dijkstra algorithm taking into consideration real network conditions. These conditions include dead ends, a standby link path along the fibres and equipment, which differs from equipment with passed main link, etc. The software complex enables us to find optimal main and standby routes for link in SDH network of complicated structure taking into account these restrictions. In addition to the path routings, complex enables saving information on the network nodes, sections, and links in database (DB) i.e., actually, it represents a DB with graphic interface, thereby allowing an operator to have rapid information on its state.
Parameter space normalization in General Hough transform
# 02, February 2014 DOI: 10.7463/0214.0697554 The paper considers an issue of object registration in the image through the generalized Hough transform with uncompensated geometric distortions being available. It proposes the method of geometric distortion correction and the method of parameter space normalization. The paper shows that the proposed algorithms require a relatively small amount of computations, but significantly improve the reliability of object detection. The paper may be of interest to professionals working in the field of pattern recognition as well as to graduate and undergraduate students attending lectures on the pattern recognition.
Reliability assessment for redundant multiagent systems
# 01, January 2014 DOI: 10.7463/0114.0696290 УДК: 004.052.3 The application of logicalandprobabilistic methods to evaluation of reliability indexes of distributed hardwaresoftware multiagent systems is examined in the article. The developed model of multiagent system defines faulttolerance technique, which is based on replication of tasks and actuators of the existing system as well as on introduction of redundant sets of executive containers. The developed model is also applicable for multiagent systems with actuators being aggregated into a set of production lines. Criteria of serviceability and the methodology for assessment of reliability, which defines the procedure of construction of an analytic operability function, were developed on a basis of the introduced model. The developed methodology was validated by the computing experiment described in the article. The model of redundant multiagent system and the introduced reliability assessment methodology could be applied to various problems of synthesis of optimal multiagent system with required reliability indexes.
Research and analyze of physical health using multiple regression analysis
# 01, January 2014 DOI: 10.7463/0114.0696028 УДК: 519.2 This paper represents the research which is trying to create a mathematical model of the "healthy people" using the method of regression analysis. The factors are the physical parameters of the person (such as heart rate, lung capacity, blood pressure, breath holding, weight height coefficient, flexibility of the spine, muscles of the shoulder belt, abdominal muscles, squatting, etc..), and the response variable is an indicator of physical working capacity. After performing multiple regression analysis, obtained useful multiple regression models that can predict the physical performance of boys the aged of fourteen to seventeen years. This paper represents the development of regression model for the sixteen year old boys and analyzed results.
Timed estimates and homomorphisms of asynchronous systems
# 01, January 2014 DOI: 10.7463/0114.0695993 УДК: 519.7 We study asynchronous systems used for mathematical modeling of the parallel computer system. It is considered as a set with partially trace monoid action. This allowed us to introduce homomorphisms of asynchronous systems as corresponding polygonal morphisms. Asynchronous systems with time function are studied using these homomorphisms. An algorithm for computing the minimal execution time of parallel processes in asynchronous systems is constructed. The conditions of homomorphisms of asynchronous systems minimal execution time keeping are found. The algorithm is used for compute of execution time of parallel processes, which consists of Petri net transitions. There are examples of computing minimal execution time for pseudopipeline and wave system. An experiment realized with multithreaded application that is built on a given Petri net is described. The experiment confirms the method of estimating the minimum execution time of the parallel process.
A model of algorithmic representation of a business process
# 01, January 2014 DOI: 10.7463/0114.0687808 УДК: 61; 338.2; 338.3 This article presents and justifies the possibility of developing a method for estimation and optimization of an enterprise business processes; the proposed method is based on identity of two notions – an algorithm and a business process. The described method relies on extraction of a recursive model from the business process, based on the example of one process automated by the BPM system and further estimation and optimization of that process in accordance with estimation and optimization techniques applied to algorithms. The results of this investigation could be used by experts working in the field of reengineering of enterprise business processes, automation of business processes along with development of enterprise informational systems.
Synthesis methods of the fast VilenkinChrestenson transforms
# 01, January 2014 DOI: 10.7463/0114.0687462 УДК: 519.216.1/2 This paper deals with the synthesis methods of the high efficiency fast VilenkinChrestensontransforms, using their scalar representation form and distinctive decomposition methods of multipoint transforms. Analytical expressions for the fast VilenkinChrestenson algorithms were obtained at various levels of decimation of signals’ original samples along with analytical estimations of their computational complexity. Tools for graphical representation of this computational process were presented in the form of signal graphs. Presented theoretical results were illustrated by specific examples which confirmed their applied practicability and efficiency.
Statistical modeling of programs' performance
# 01, January 2014 DOI: 10.7463/0114.0679688 A task of evaluation of program's performance occurs often in the process of design of computer systems or during the iterative compilation. A traditional way to solve this problem is an emulation of program's execution on the target system. A modern alternative approach to evaluation of program's performance is based on the statistical modeling of program's performance on a computer under investigation. This statistical method of modeling programs' performance called Velocitas was introduced in this work. The method and its implementation in the Adaptor framework were presented. Investigation of the method's effectiveness showed high adequacy of program's performance prediction.
FPGA implementation of general cellular automata based cryptographic hash functions: performance and effectiveness
# 01, January 2014 DOI: 10.7463/0114.0675812 In the paper the author considers hardware implementation of the GRACEH family general cellular automata based cryptographic hash functions. VHDL is used as a language and Altera FPGA as a platform for hardware implementation. Performance and effectiveness of the FPGA implementations of GRACEH hash functions were compared with Keccak (SHA3), SHA256, BLAKE, Groestl, JH, Skein hash functions. According to the performed tests, performance of the hardware implementation of GRACE–H family hash functions significantly (up to 12 times) exceeded performance of the hardware implementation of previously known hash functions, and effectiveness of that hardware implementation was also better (up to 4 times).
Promising structures with memory based on inorganic materials
# 12, December 2013 DOI: 10.7463/1213.0669629 At present time many research groups working in the field of microelectronics are trying to develop a technology for producing a memristor also investigating possibilities of its integration into the CMOS process. Main classes of materials with memory effect such as metallic oxides, chalcogenides, solid electrolytes, polymers etc. were found. All of these classes have resistive switching properties; but their switching mechanisms are different. There are not fully studied elements in literature, but for some of them a stable memristive effect was achieved. Also an operating model for these elements was developed. In this article the authors present a review of such structures with a memory effect based on inorganic materials and analysis of their characteristics along with possible application domains.
Automatic synthesis of the helicopter programmed motion along the horizontal line
# 10, October 2013 DOI: 10.7463/1013.0660675 For fourdimensional model describing the helicopter motion along a given horizontal line the problem of automated synthesis of programmed motion is solved. This solution provides a helicopter motion from a given state of rest in the given state of rest. Time to perform the maneuver is not set. The considered helicopter model is a control dynamic system, which is not a flat system. For such systems general approaches to the terminal problem solution are presently unknown. To solve the terminal problem two approaches are applied. The first approach is based on the use of finite symmetry, which converts the initial conditions of the problem in the final conditions. The use of such symmetry can reduce the number of final conditions. The second approach is based on the use of covers and consists in the construction of a special mapping, which for two given dynamic systems surjectively maps the set of solutions of the first system in the set of solutions of the second system. The programmed motion in this case can be found as the solution of two related specifically posed Cauchy problems for these dynamic systems. The resulting programmed control is a piecewise continuous function of time.
GPU realization of a neural network face recognition algorithm
# 12, December 2013 DOI: 10.7463/1213.0659387 In this paper the author considers implementation of a neural network face recognition algorithm on a graphical processing unit. The algorithm is based on a classic multilayer perceptron. The training method is backward propagation of errors. This implementation has a seven times performance boost compared with traditional CPU implementation. Error rate of recognition approximately equals to 1%.
Formalization of implementing a release cycle of electronic periodicals
# 11, November 2013 DOI: 10.7463/1113.0658430 This article presents analysis of a typical process related to implementation of a release cycle of electronic periodicals in a publishing house of a higher educational establishment. A set of main components that represent various parts of the publisher’s organizing structure and affect the operating behavior was determined. A process of maintaining queries from users by editors, reviewers, and bibliographers was considered. As a result of the conducted investigations, it was shown that the law of inflow of publication queries in an electronic periodical was the simplest one. Parameters of that simplest inflow of users’ queries were also determined. A formalized scheme represented in the form of a closedloop querying system was proposed for description of a typical process of the release cycle of electronic periodicals.
Method of automated construction of a digital relief model
# 12, December 2013 DOI: 10.7463/1213.0657404 The author considers a problem of effective construction of a digital relief model (DRM) for a certain region on the basis of a large number of aerial photographs. Main stages of DRM construction were described. Special attention was paid to increase efficiency of search for the “nearest” (in terms of content) images in the database; it was proposed to use a search scheme based on inverse indexing of images’ unique characteristics and the database organized in the form of a tree of “visual words”. Construction of a lexical tree with the use of hierarchical clusterization of the attribute space, obtained on the basis of SIFT descriptors, was shown; weighing coefficients of “visual words”, calculated on the principle of TFIDF (Term Frequency – Inverse Document Frequency), were taken into account.
Abstractions and base operations of a specialized description language of algorithms for solving problems of structural analysis and synthesis
# 12, December 2013 DOI: 10.7463/1213.0656686 This article justifies the need for a specialized programming language that would allow to describe solving procedures for tasks of structural analysis and synthesis in terms of mathematical models being used. Compared with the existing languages and subroutine libraries such language would not only allow to reduce program development time significantly but also provide an opportunity to develop more efficient programs in terms of computational complexity of algorithms and runtime. According to the analysis of algorithms and representations of graph models a set of the necessary abstractions and operations on them was determined. The abstractions such as "set", "multiset", "sequence", "vector" and "matrix" were proposed for basic abstractions. Operations on these abstractions were divided into mathematical and related to specific implementation features of the algorithm descriptions. A set of elementary operations of the indicated types was defined for each abstraction.
AORTA software system for automation of investigation in the field of estimating patient’s predisposition for atherosclerosis on the basis of genetic and phenotypic markers
# 01, January 2014 DOI: 10.7463/0114.0649691 The authors propose a method for automated investigation of aorta areas affected by atherosclerosis, in order to find a relation between somatic mutations of the mitochondrial genome in the cells of aortic walls and a degree of manifestation of arterial sclerotic disease in these areas. This method consists of several basic steps; they are aorta image segmentation, marking of connected components of the image, plotting an adjacency graph of image areas and decomposition of this adjacency graph. A structure of software system was proposed; AORTA software system which implements the proposed method was designed and developed. Tests and approbation of the system in the SRI of general pathology and pathophysiology of RAMS demonstrated its working capacity and usability.
Synthesis of a quasioptimal trajectory of an unmanned aerial vehicle
# 12, December 2013 DOI: 10.7463/1213.0646471 A problem of constructing spatial trajectories for unmanned aerial vehicles (UAV) and synthesis of program controls were considered in this paper. A nonlinear mathematical model of UAV movement as a material point along the trajectory was described. A kinematic trajectory is based on the timedependent polynomials. Program control implementing this trajectory is based on the concept of inverse dynamic problems. Кинематическая траектория строится на основе полиномов по времени, а реализующее ее программное управление – на основе концепции обратных задач динамики. The purpose of this paper is to determine the minimal time of the UAV maneuver so that the constructed kinematic trajectory and control would satisfy the limits on state variables and control. Numerical optimization methods were used to solve this problem. Results of numerical simulation were also presented.
Twocriterion identification of kinetic parameters of olefin hydroalumination by HAlBui2 and ClAlBui2
# 12, December 2013 DOI: 10.7463/1213.0645511 This work was carried out in the context of investigation and development of original twocomponent neutral catalyst systems made of mixedvalence metal compounds (titanium and zirconium) and organoaluminum compounds. A problem of identification of kinetic parameters of olefin hydroalumination, which belongs to a class of inverse problems of chemical kinetics, was considered in this paper. As a rule, identification problem is formulated as a singlecriterion optimization problem for residuals between calculations and experimental data. The disadvantage of such a formulation consists in the fact that important prior information about kinetic features of a chemical reaction under investigation is ignored during identification. A distinct feature and novelty of this work is a new formulation of the identification problem; the problem was considered as a twocriterion optimization task which allowed to take into account both experimental data on the kinetics of the chemical reaction under investigation and prior information about this reaction. Solution to this problem represents the Pareto set. In order to approximate the Pareto set a modified Adaptive Weighted Sum method was used. One of the purposes of this work is to test the modified AWS method while solving real world twocriterion inverse problems of chemical kinetics. Twocriterion identification problem of kinetics of olefin hydroalumination by halbui2 and clalbui2 was formulated in this work. The algorithms and software which were used are described in this paper. The obtained calculation results were analyzed and discussed.



