|
Research Institute of Intelligent Computer Systems Ternopil National Economic University |
|
2004, Vol. 3, Issue 2 |
|
Contents and abstracts
Xianhong Xu, Simon Jones Faculty of Engineering and Design, University of Bath, BA2 7AY, UK, {x.xu, s.r.jones}@bath.ac.uk, http://www.bath.ac.uk/engineering
Previous code compression research on embedded
systems was based on typical RISC instruction code. THUMB from ARM Ltd
is a compacted 16-bits instruction set showing a density great code
density than its original 32-bits ARM instruction. Our research shows
that THUMB code is compressible and a further 10-15% code size
reduction on THUMB code can be expected using our proposed new
architecture – Code Compressed THUMB Processor. In our proposal, Level
2 cache or additional RAM space is introduced to serve as the temporary
storage for decompressed program blocks. A software implementation of
the architecture is proposed and we have implemented a software
prototype based on ARM922T processor, which runs on the ARMulator.
Valeriy N. Koval, Volodymyr V.
Savyak 40, Prospect Academika Glushkova, 03680, e-mail: icdepval@ln.ua, fax: (044) 2664549
The paper considers the creation of intelligent
solving machines and the arrangement of parallelprogramming in
intelligent distributed multiprocessor systems based on them. There are
proposed some mainconcepts. A system is designed for programming in the
C+Graph high-level language. The ideologyproposed can be considered as
an efficient development of structural high-level language
interpretation whenapplied to multi-microprocessor systems. Equipment
structure of the basic version of intelligent solvingmachines is
considered and some characteristics are discussed.
Miki Sirola Helsinki University of Technology,
Decision making is done in many application areas.
Still most studies are done in such fields as economyand production
planning. In methodologies used there exists more variation. This paper
reviews the decisionconcepts discussed in the literature. Also some
decision models by the author are commented. The field andpractise in
decision science is summarized. Although decision support systems are
the final results of manyprojects, they are mostly based on the
decision concepts behind the studies that deserve also more
detailedexamination.
Decision analysis approach and knowledge-based technologies are
examples of commonlyused concepts.
Arunas Lipnickas 1), Józef Korbicz 2) 1) Department of Control Technology, Kaunas University of Technology, Studentų 48, 3031 Kaunas, Lithuania, E-mail: lipnick@soften.ktu.lt 2) Institute of Control and Computation Engineering, University of Zielona Góra, ul. Podgórna 50, 65-246 Zielona Góra, Poland, E-mail: J.Korbicz@issi.uz.zgora.pl
To improve recognition results, decisions of
multiple neural networks can be aggregated into a committeedecision. In
contrast to the ordinary approach of utilising all neural networks
available to make a committeedecision, we propose creating adaptive
committees, which are specific for each input data point. A
predictionnetwork
is used to identify classification neural networks to be fused for
making a committee decision abouta given input data point. The jth
output value of the prediction network expresses the expectation level
thatthe jth classification neural network will make a correct decision
about the class label of a given input datapoint. The proposed
technique is tested in three aggregation schemes, namely majority vote,
averaging, andaggregation by the median rule and compared with the
ordinary neural networks fusion approach. Theeffectiveness of the
approach is demonstrated on three well known real data sets and also
applied to faultidentification of the actuator valve at one sugar
factory within the DAMADICS RTN.
Wieslaw Winiecki, Piotr Bilski Institute of Radioelectronics, Warsaw University of Technology, Nowowiejska 15/19, Warsaw, Poland, w.winiecki@ire.pw.edu.pl, foxtrot@sgk.com.pl
The problem of the Real-Time virtual instrument is
presented. The requirements for the Real-Time conditions are explained.
The method of the instrument time analysis is proposed. The virtual
spectrumanalyser is examined in order to reveal properties of the
software functions. The optimisation procedure isdescribed and its
results are presented. The conclusions for the future developers have
been articulated.
HART PROTOCOL ANALYSER BASED IN LABVIEW J.M. Dias Pereira 1,2),
Octavian Postolache 1,2), P. Silva Girão 2) 1) Escola Superior de
Tecnologia, Instituto Politécnico de Setúbal,
Analysis of Highway Addressable Remote Transducer
(HART) protocol is important in reliability evaluation of HART based
systems. The present work proposes a virtual instrument based solution
for HART signal analysis that includes a digital oscilloscope with GPIB
interface, a HART modem, a PC with a GPIB interface board and a RS232
port, and a software component developed in LabVIEW. The HART analyser
permits the visualisation of signals that correspond to the HART
protocol, gives access to data coming from field instruments and has
the ability to analyse and modify field instrument performance.
Stéphane Raze 1,2), Dominique Dallet 2), Phillipe Marchegay 2) 1) Serma Technologies, 30 av. G. Eiffel, F-33608 PESSAC CEDEX, FRANCE e-mail : s.raze@serma.com, url : http://www.serma.com 2) Laboratoire IXL, Université Bordeaux I, 351 cours de la Libération, F-33405 Talence CEDEX, FRANCE e-mail : dallet@ixl.u-bordeaux.fr, url : http://www.ixl.u-bordeaux.fr
An industrial test bench for Analog-to-Digital
Converters (ADC) is presented in this article. The benchhardware has
been composed with known devices of instrumentation. This hardware had
been identifiedfrom ADCs under test requirements. All of hardware is
handled by a software written in LabVIEWdevelopment system.
S. Sitharama Iyengar 1), Qishi Wu 2), Nageswara S.V. Rao 2) 1) 298 Coates Hall, Louisiana State University, Baton Rouge, LA 70803, iyengar@bit.csc.lsu.edu, http://bit.csc.lsu.edu/~iyengar/ 2) Center for Engineering Science Advanced Research, Computer Science and Mathematics Division, Oak Ridge National Laboratory, Oak Ridge, TN, 37831-6355, {wuqn,raons}@ornl.gov
An increasing number of distributed sensor networks
are being deployed in a wide spectrum of scenarios for both military
and civilian applications. The design of such systems requires the
integration of many methods for solving various problems ranging from
sensor deployment, through data communication, information fusion to
environmental inference and hypothesis testing. This paper is focused
on the network aspects in support of data communication services. We
present some recent developments in networking techniques for multiple
sensor systems in three topics: efficient mobile agent routing in
mobile agent-based distributed sensor networks, bandwidth aggregation
using multiple paths in wide-area wireline networks, and connectivity
enhancement using node movements in small-area wireless mobile ad-hoc
networks.
Oleg Tsymbal 1), Volodymyr Lukin 1), Pertti Koivisto 2), Volodymyr Melnik 2) 1) Dept 504, National Aerospace University (KhAI) 17 Chkalova Street, 61070, Kharkov, Ukraine, Tel/fax +38 0572 441186, E-mail: lukin@xai.kharkov.ua, 2) Institute of Signal Processing, Tampere University of Technology, P.O.Box-553, FIN-33101, Tampere, Finland, E-mail: peko@cs.tut.fi, melnik@cs.tut.fi
Characteristics of impulse bursts in satellite
images are analyzed and methods for burst removal are considered.
Artificial compact burst model is proposed and test images are created.
An advanced multipass algorithm for the detection and
removal of compact bursts in the presence of both additive and
multiplicative noise is proposed. The efficiency of the
algorithm is evaluated quantitatively using the artificial test images
and
visually using the artificial test images and real radar and optical
satellite images. It is shown through experiments that the
proposed method removes impulse bursts efficiently while preserving
information.
Sergiy Lupenko, Mykola Prijmak Ternopil Technical University, Department of Computer Sciences, 54 Ruska street, Ternopil 46001, Ukraine
In work is proved the theorem about periodicity of
higher moment functions of output process of the nonlinear
system, subject to the condition, if on its entrance acts a periodic
white noise. The statistical methods of evaluation of
periodic moment functions of output process of the nonlinear system are
proved.
Vladimir Vacic and Tarek M. Sobh Department of Computer Science and Engineering University of Bridgeport, Bridgeport, CT 06604, USA vladimir@vacic.org, sobh@bridgeport.edu
The topic of this paper is a Genetic Algorithm
solution to the Vehicle Routing Problem with Time Windows, a variant of
one of the most common problems in contemporary operations research.
The paper will introduce the problem starting with more general
Traveling Salesman and Vehicle Routing problems and present some of the
prevailing strategies for solving them, focusing on Genetic Algorithms.
At the end, it will summarize the Genetic Algorithm solution proposed
by K.Q. Zhu which was used in the programming part of the project.
SOLVING THE RECTILINEAR STEINER MINIMAL TREE PROBLEM WITH A BRANCH AND CUT ALGORITHM Nahit
Emanet 1), Can Ozturan 2)
1) Computer
Engineering
Department,
Bogazici University, emanetn@boun.edu.tr, asma.cmpe.boun.edu.tr/~emanetn 2) Computer Engineering Department, Bogazici University, ozturaca@boun.edu.tr, asma.cmpe.boun.edu.tr/~ozturan This paper presents a new branch-and-cut algorithm that allows us to reduce the solution time of the concatenation phase of the rectilinear Steiner minimal tree problem in the plane. Our branch-and-cut algorithm is used on an integer programming formulation using what we call cutsec, and sec constraints. We present implementation details of our branch-and-cut program called NEOSteiner and provide computational results on test instances from the SteinLib library. Nickolay
Petrov 1), Tania Pehlivanova 2)
1) U “Prof. As.
Zlatarov”
– Burgas, Bulgaria, Technical College
nicipetrov@lycos.com 2) Trakian University- Stara Zagora, Technical College – Yambol, tania_ipg@abv.bg
The use of microprocessor measurement system gives
the opportunity to improve the precision of the measurements. It is
possible by realization of algorithms, taking into account the changing
conditions of measurement experiment. In the paper it is suggested an
adaptive algorithm for calculation of the mathematical expectation MX
in the conditions of insufficient knowledge about the interval of
correlation.
Volodymyr
Tydnjuk 1), Volodymyr Kravchenko 2), Olena Mul 3),
Mykola Shut 4)
1) International
Scientific
and Educational Center of Information Technologies and Systems,
40 Academician Glushkov avenue, 03680 Kiev, Ukraine 2) Institute of Cybernetics named by V. Glushkov, 40 Academician Glushkov avenue, 03680 Kiev, Ukraine 3) Institute of Computer Information Technologies, Ternopil Academy of National Economy, 3 Peremoga Square, 46004 Ternopil, Ukraine, omu@tanet.edu.te.ua 4) National Pedagogical University named by M. Dragomanov, 9 Pirogov str., 01601 Kiev, Ukraine
The application of general theory of relativity for
building of model of living organisms field is considered. The
information field, which connects living organism with environment,
forms a material object of such field. A non-electromagnetic nature of
such material object is substantiated. The phenomenon of interaction of
thermochemical potentials at distance is investigated.
Mykola Dyvak, Volodymyr Manzhula Department of Computer Sciences Institute of Computer Information Technologies Ternopil Academy of National Economy 3 Peremoga Square, Ternopil 46004, Ukraine e-mail: mdy@tanet.edu.te.ua <>
The task of structural identification of interval
models of the static systems is considered. Algorithm of consequent
complication of structure is proposed. Possibility of receiving of
single solution of task of structural identification is shown for
different initial set structures of models. Adaptive procedure which
eliminates dependence of structural identification result on the value
of interval errors is proposed.
<> USING BOUNDARY SCAN IN DIGITAL SYSTEMS Marina Kaminskaya Kharkiv National University of Radioelectronics, Department of computing automation and design, Lenin Prospect 14, Kharkiv 61726, Ukraine
Fast fault simulation method of stuck-at-faults
evaluation of digital circuit represented on gate description level is
offered. This method allows minimizing supervisory test and raising his
quality.
THE IMPLEMENTATION OF THE NEURAL NETWORK FOR THE CLASSIFICATION PROBLEM Ulyana Lisovik, Oleksandr Lipchanskiy Kharkiv National University of Radioelectronics, Department of computing automation and design, Lenin Prospect 14, Kharkiv 61726, Ukraine
The example of NN realization is considered. Also
description of all its design stages from NN function model description
to its timing and hardware characteristics estimation is considered. NN
structural model is presented in VHDL code. Through SynplifyPro 7.0
package from Synplicity® the system synthesis with the orientation on
Virtex-II XC2V6000 family is made out. The estimation of the optimality
of the synthesized NN model utilization is accomplished. NN structures
are shown; hardware costs are taken to the table.
Veselina Ivanova Nedeva Technical College, Yambol, Trakia University, Stara Zagora, Bulgaria, Technical college, 8600 Yambol, Bulgaria, Gr.Ignatiev Str. 38, tel. 046 66-91-78, E-mail: vnedeva@yahoo.com
The article presents the results of a research in
the field of the marketing information technologies. The subjects of
the research are the marketing information systems, mainly the
information systems used in the marketing business activities of the
studied companies. 25 marketing information systems were covered. Data
were collected in the period 1999-2002. A study was conducted on the
basis of criteria and indices preliminarily worked out as well.
Analysis was made, results were obtained and a concept for the building
of a marketing information system of an integrated kind was proposed.
SIMULATION OF DC-DC CONVERTER SYMBOLIC DYNAMICS:DETECTION AND IDENTIFICATION OF BIFURCATION TYPES Abdelaziz
Hamzaoui 1), Pavel Ustinov 2), Andrey Sholonik 3),
Yury Kolokolov 4)
1) LAM – UFR
Sciences
Exactes et Naturelles – Moulin de la Housse
BP 1039 – 51687 REIMS Cedex 2 – France Tel: +33 (0) 3 25 42 46 43, Fax: +33 (0) 3 26 91 31 06 e-mail: a.hamzaoui@iut-troyes.univ-reims.fr 2, 3, 4) Department of Design and Technology of Electronic and Computer Systems, Orel State Technical University, 29 Naugorskoye Shosse, 302020 Orel, Russia Tel: 2, 3) +7 0862 419879, 4) +7 0862 421661, Fax: 2, 3, 4) +7 0862 416684 e-mail: 2) p_s_ustinov@orel.ru, 3) sholonik@ostu.ru, 4) kolokolov@ostu.ru
In this paper continuous-time model to discrete-time
symbolic model transformation for an algorithmization of the DC-DC buck
converter dynamics investigation process is derived. The proposed
concept of symbolic index allows evaluate quantitatively the symbolic
model change and makes possible the “nonsmooth” bifurcations revealing.
The numerical simulation that had been carried out verifies utilizing
efficiency of symbolic modeling both as for possible bifurcation
pattern forecasting, as for bifurcation type identification.
Iryna
Turchenko 1), Volodymyr Kochan 2), Anatoly
Sachenko 3)
Institute of Computer
Information
Technologies
Ternopil Academy of National Economy 3 Peremoga Square, Ternopil 46004 Ukraine 1) vtu@tanet.edu.te.ua, 2) vk@tanet.edu.te.ua, 3) as@tanet.edu.te.ua
The possibility of artificial neural network usage
for recognition of a signal of a multi-parameter sensor is described in
this paper. The general structure of data acquisition channel with
usage of neural networks as well as mathematical model of output signal
of a multi-parameter sensor is studied in this article. The model of
neural network, training algorithm and achieved results of simulation
modeling of a multi-parameter sensor signal recognition using MATLAB
software are presented at the end of this paper.
|