Chistiakova ô. ÷., Ivanov á. ÷., Kolert ë.
Systems of Automatized Design of Three-Dimensional Geometric Model of Overtiming Production of Polimeric Films
      Techniques of the analysis of multi-assortmental recustomized polymer films manufacture realization and synthesis of three-dimensional geometrical model of projected and modernized recustomized chemical-engineering manufactures, structure and the characteristic of the software, structure of functional maintenance of projecting CAD subsystem of three-dimensional geometrical model of recustomized polymer films manufacture are considered. The example of use of developed CAD system at designing of new calender line at factory Klockner Pentaplast Russia is shown.

Bojko A. N., Betin E. A.
On a Contractability Property of Hypergraphs
      We describe the question of hypergraph's Contractability, which describe the assembling of products. In this article we examine some class of hypergraphs, especially hypergraps with maximum hyperedge degree as four. We lay down sufficient conditions of theirs Contractability.

Filippova A. S.
Some Problems of Rectangular Packing Decoding: a Short Review of Up-to-Date Technologies
      Under consideration there are: rectangular packing problems of different dimensions and the basic trends of development of coding-decoding methods for different packings. The first part of the given article is devoted to short characteristics of the main papers concerning the matter. Then, the block coding method and block structure decoders are shown, the latter being worked out by the team to which the author belongs. The algorithm reorganizing a pair of sequences into a pair of block-structures has been described for the first time. It represents a new decoder for packing construction.

Safronov V. V.
Hypervector Translation of Complex System in Number of Leaders
      Statement of a hypervector problem of a conclusion of complex system in number of leaders is considered. Generally she is reduced to polycriterial to a problem of discrete programming. The method of its decision is offered.

Malykhina G. F., Merkusheva A. V.
Information and Entropy Conceptions in Neuronet Methods for Signal Analysis by Means of Information-Measuremental System
      Information and entropy conception elements and using-method are represented for signal processing by means of neural networks. It is shown how the system-bounded notions of information, entropy, Kullback-Leibler's divergence measure for distribution of analysed signals, form analytical framework on optimization criteria for net learning and on system self-organization principles. The correlations are given for mutual-information maximization principle and applications on image processing and on reconstructing the of signal tipes by means of data measurements for sensors registering only linear combinations of analysed signals totality.

Devyatisil’nyi A. S., Dorozhko V. M.
Neural Control System of a Car Longitudinal Motion
      The problem of a car autonomous informational system construction enhancing safety and comfort level while driving in a transport stream is considered; the solutions based on a modern ontology of neural structure concept, situational control and tuning (training) are given. The results of the numerical analysis and experiments are provided.

Knyazeva M. A.
Optimizing Compilers Controlled by Knowledge Bases
      The problems that we face while constructing optimizing compilers are analyzed. The concept of constructing optimizing compilers controlled by knowledge bases is given. The medium in which the optimizing compilers are constructed is described.

Ivanova G. S.
Control Flow Graph Vertex Numbering Algorithm for its Structuring
      In this paper, requirements for numbering are determined, rules are formulated and control flow graph vertex labeling algorithm is proposed. Purpose of this labeling is revealing of its structured features.

Beskhlebnova G. A., Gorbatkov S. A.
The Iteration Algorithm of "Streams" Formation in Configuration Intended for a Neural Network Forecast Model of Corrosion Damage Processes of Trunk Pipe
      New neural network technology for modeling corrosion damage dynamical processes of trunk pipe is proposed. It is worked out a base of common cybernetic system laws.

Zedginidze I. G., Beraya N.O.
The Comparative Analysis of Composite Designs of the Second Order at Presence of Errors of Means of Measurement
      The comparative analysis of symmetric and economic asymmetrical composite designs of the second order for two, three and four variable errors is lead{carried out} at the account of the measuring means used at installation of levels of factors according to the design. Advantage of use in some cases of asymmetrical specially focused composite designs of the second order being more economic is shown.

Avedian E. D.
Historical Aspects of Development of Multilayer Neural Networks Theory
      The comparative analysis of multilayer neural networks (MNN) learning algorithms published by the Soviet and American scientists in 70th and 80th years of the last century is resulted. It is shown that the basic difference of the learning algorithms of the MNN consists in the way of gradient calculation of the minimizing functional. In the Soviet scientific literature (1973) is used the direct method of gradient calculation, in American — the recurrent method (the back-propagation algorithm, 1986) with the same end results. From the given analysis follows the priority of the Soviet scientific school in the field of development of the MNN learning theory.