|
Research Institute of Intelligent Computer Systems Ternopil National Economic University |
|
2004, Vol. 3, Issue 3 |
|
Contents and abstracts
OPTIMIZATION VIA SIMULATION: SOLUTION
CONCEPTS, ALGORITHMS, PARALLELCOMPUTING STRATEGIES AND COMMERCIAL
SOFTWARE
Gianpaolo
Ghiani1, Pasquale Legato2, Roberto Musmanno2, Francesca
Vocaturo2
1) Dipartimento di
Ingegneria dell’Innovazione Università degli Studi di Lecce,
73100 Lecce, Italy gianpaolo.ghiani@unile.it 2) Dipartimento di Elettronica, Informatica e Sistemistica Università della Calabria, 87030 Rende (CS), Italy {legato,musmanno, vocaturo}@unical.it
Simulation optimization (or optimization via simulation) is defined as
the optimization of performance measures based on outputs from
stochastic simulations. Although several
articles on this topic have been published, the literature on
optimization via simulation is still in
its infancy. In this paper the research in this field is reviewed and
some issues that have not received attention so
far are highlighted. In particular, a survey of solution methodologies
is presented, followed by a critical
review of parallel computing strategies and commercial software
packages. A particular emphasis is put on
problems with discrete decision variables.
László Lasztovicza,
Béla Pataki, Nóra Székely, Norbert Tóth Budapest University of Technology and Economics,
This document presents the computer aided diagnosis
system being
developed to help experts in screening
mammography. It is a very important project because about 8 %
of women develop breast cancer in her lifetime therefore global
screening is necessary. It means that
reliable diagnosis of huge number of images must be solved. The basic
architecture of the system and the
information processing needed is presented. One of the most important
tasks in mammographic diagnosis
systems is microcalcification detection. It is solved by a hierarchical
neural architecture. The
original suggestion of that structure was improved by two ways. The
image features used as inputs to the neural
networks were analyzed and the feature set was extended. The neural
architecture was embedded in a
neural ensemble context for improving the quality of the solution
further. Results of the tests of that
detection procedure show that the false detection ratios are acceptable.
Miticǎ
Craus 1), Laurenţiu Rudeanu 2)
Department of Computer Science and
Engineering
Technical University "Gh. Asachi" 700050 Iaşi, Romania 1) craus@cs.tuiasi.ro 2) lrudeanu@yahoo.com
This article deals with the pyramidal framework designed to be used in
the parallelization of the ant-like algorithms. Such algorithms have
several things in common: they run in
cycles and the process can be divided among different “processing
units”. The parallel implementation
of the Ant Colony Optimization algorithm for the Traveling Salesman
Problem is an application of this
system. The topology of the framework architecture is similar to a
B-tree and contains three types
of processing nodes: a single master (the root), several sub-masters
corresponding to the internal nodes of
the tree and several slaves as leaves. First the master reads the
problem instance, wraps it up in a message
that is sent to all the other processing nodes and initializes the
central data structures. Then, the slaves
take over the control by starting the algorithm while the master and
the sub-masters are waiting for requests
to update the data. The framework has an object-oriented design and was
implemented in C++, using the MPI
library.
Domenico Conforti 1), Domenico Costanzo 2), Rosita Guido 1) 1) D.E.I.S., Università
della Calabria, Ponte Pietro Bucci 41C, 87036 Rende (CS), Italy ,
mimmo.conforti@unical.it,
http://www.deis.unical.it/deis1.0/portale/home/conforti/ 2) Dipartimento di Medicina Sperimentale e Clinica, Università degli Studi “Magna Graecia”, Via Tommaso Campanella 115, 88100 Catanzaro, Italy, domcostanzo@unicz.it
In this paper we considered a very challenging
medical decision making
problem: the short-term prognosis evaluation of breast cancer patients.
In particular, the
oncologist has to predict the more likely outcome of the disease in
terms of survival or recurrence after a given
follow-up period: “good” prognosis if the patient is still alive and
has not recurrence after the follow-up
period, “poor” prognosis if the patient has recurrence or dies within
the follow-up period. This prediction can be
realized on the basis of the execution of specific clinical tests and
patient examinations. The relevant
medical decision making problem has been formulated as a supervised
binary classification problem. By the
framework of generalized Support Vector Machine models, we tested and
validate the behavior of four kernel
based classifiers: Linear, Polynomial, Gaussian and Laplacian. The
overall results demonstrate the
effectiveness and robustness of the proposed approaches for solving the
relevant medical decision making problem.
Viktor
Lokazyuk1), Viktor Cheshun2), Vitaliy Chornenkiy3)
1) doctor, professor,
Lvivska str. 11, Ternopil, 46000, Ternopil academy of a national
economy, kism@beta.tup.km.ua
2) doctor, docent, Institutska str. 11, Khmelnitskiy, 29016, Technological university of Podillya, kism@beta.tup.km.ua 3) Institutska str. 11, Khmelnitskiy, 29016, Technological university of Podillya, kism@beta.tup.km.ua
The base principles of a technique of application of 3-layer
feedforward fullconnected artificial neural network for execution of
adaptive algorithms of testing of digital
microprocessor devices are considered. The method of change of weight
coefficients and thresholds of artificial
neurons in the mode of operation of artificial neural network realized
at the hardware level is considered.
The application of this method provides implementation of adaptive
algorithms of testing of the large complexity with the limited hardware
resources of artificial neural network.
Bei Wang, Dung Hoang, Idris Daiz, Chiedu
Okpala, Tarek M. Sobh Department of Computer
Science, University of Bridgeport
The Collective Intelligence Research Tool (CIRT) is
an experimental software and hardware research tool. It provides an
inexpensive and efficient alternative research implementation that
demonstrates simulations of the collective behavior of self-organized
systems, primarily social insects. The software focuses on 2D
simulations of the woodchip-collecting behavior of termites and 3D
simulations of the building behavior of wasps. The hardware simulation
employs a Boe-Bot robot, which has the potential of simulating simple
movements of a social insect, by extending its functionality through
adding sensors and integrating a control chip. Yuri
Kornienko 1), Arkady Borisov 2)
Institute of Information Technology,
Riga Technical University,
1 Kalku Str., Riga LV-1658, Latvia 1) j.kornienko@pf.lv 2) aborisov@egle.sc.rtu.lv
The paper discusses the experiments performed with Machine Learning
algorithms (ID3, C4.5, Bagged-C4.5, Boosted-C4.5 and Naive Bayes) and
an algorithm made on the basis of a combination of genetic algorithms
(GA) and ID3. The latter algorithm is implemented as an extension of
the MLC++ Library of Stanford University. The behaviour of the
algorithm is tested using 24 databases including those with a large
number of attributes. It is shown that owing to “hill-climbing” problem
solving, the characteristics of the classifier made with the help of
the new algorithm became significantly better. The behaviour of the
algorithm is examined when constructing pruned classifiers. The ways to
improve standard Machine Learning algorithms are suggested.
Mykola Dyvak, Petro Stakchiv, Iryna
Kalishchuk Department of Computer
Sciences
The task of interval parameter’s identification of
linear dynamic system is considered. Possibility of application of
calculation procedures of parameter’s interval identification of the
static system for the decision of this task is shown. The method of
parameter’s set identification of the linear dynamic system in form of
multidimensional ellipsoids is proposed.
Oles
Hodych1, Yuriy Shcherbyna1,
Michael Zylan2
1) Ivan Franko Lviv
National University, Ukraine
2) Myklan Systems, Melbourne, Australia
In this article the authors propose an approach to forecasting the
direction of the share price fluctuation, which is based on utilization
of the Feedforward Neural Network in conjunction with Self-Organizing
Map. It is proposed to use the Self-Organizing Map for filtration of
the share price data set, whereas the Feedforward Neural Network is
used to forecast the direction of the share price fluctuation based on
the filtered data set. The comparison results are presented for
filtered and non-filtered share price data sets.
Abdelaziz Hamzaoui1, Artem
Melikhov2, Vadim Tsukanov2, Yuriy Kolokolov2 1) UFR
Sciences Exactes et Naturelles - Moulin de la Housse BP - 51687 The paper is
dedicated to the realization of the structural model of the pulse-phase
control system (PPCS) of the thyristor voltage regulator (TVR) with an
active-inductive (RL) load in the medium of “matLab 6.0”® Describing
the possibilities of “matLab 6.0”® in the dynamic systems simulation,
some problems connected with the use of special instrument means,
included in structure of Toolbox Simulink Library, are examined.
Roman
Kochan1, Orest Kochan2
1) Lecturer of SCS
Department,
Institute of Computer Information Technologies, Ternopil Academy of National Economy, 3 Peremoga Square, Ternopil 46004, Ukraine 2) Student of physical faculty of I. Franko Lviv National University, Lviv, Ukraine
In the article there is presented the developed precision block of data
acquisition on the base of the 24-bit ADC type AD7712. It is designed
for the usage consisting of the distributed information-measurement
systems and networks. Designed block of data acquisition provides
controlled distance reprogramming – replacement of the program of its
work through a serial interface during work (after authentication of
the proper command). The developed block of data acquisition provides
automatic execution of the following procedures, intended for
increasing of measuring result accuracy: (i) setting of zero; (ii)
calibration on the standard voltage source; (iii) correction of
temperature coefficient of standard voltage source; (iv) correction of
quadratic component error of ADC nonlinearity. Execution of accuracy
improvement procedures provides the constant voltage measuring
allowable error ±0.01% in working conditions and the nonlinearity of
conversion characteristics less than ±0.0003%.
Cheng-Fa Tsai, Feng-Cheng Lin Department of
Management Information Systems,
This paper proposes a new and efficient hybrid heuristic scheme for
solving job-shop scheduling problems (JSP). A new and efficient
population initialization and local search concept, based on genetic
algorithms, is introduced to search the solution space and to determine
the global minimum solution to the JSP problem. Simulated results imply
that the proposed novel JSP method (called the PLGA algorithm)
outperforms several currently used approaches. This investigation also
considers a real-life job-shop scheduling system design, which
optimizes the performance of the job-shop scheduling system subject to
a required service level. Simulation results demonstrate that the
proposed method is very efficient and potentially useful in solving
job-shop scheduling problems.
Tymoor
Korkishko1, Ruslan Shevchuk2
1) Senior lecturer of
Computer Science department , tko@tanet.edu.te.ua
2) Assistant of Computer Science department, rsh@tanet.edu.te.ua Institute of Computer Information Technologies Ternopil Academy of National Economy Peremoga Square 3, Ternopil 46004, Ukraine
In the paper the operating device of IPSec protocol optimized for
treatment media of packages is investigated. Analytical expressions
which describe time of media packages processing depending on the
parameters of operating device structure are offered, the mathematical
model of operating device of IPSec processor is developed. On the basis
of mathematical model, with the purpose of reduction of delay and
jitter, which rise up during the generation of media package, software
is developed for optimization of structures descriptions of processor
IPSec operating device. The row of the optimized structures of
operating device is got for different services of given IPSec treatment
at different technological descriptions of component base. The analysis
of results allowed to set that in most cases, the least time of
treatment media of packages is observed at iterative and
iterative-conveyer realization of IPSec operating device.
Zoran Perić1,
Srdjan Bogosavljević2, Aleksandra Jovanović3 1,3
“Faculty of Electronic Engineering”, University of Nis, Serbia; e-mail:
peric@elfak.ni.ac.yu
In this paper we consider quantization of complex
variables and mean-square error (MSE). The best polar quantizer is
Wilson’s unrestricted polar quantizer (UPQ) [1]. The MSE minimization
is constrained only by the total number of quantization points, N. Our
method is different from Wilson’s algorithm [1] that has predetermined
number of points Mi at each magnitude level i, 1≤i≤L, which makes it
impractical for large number of points. In our approach, we consider
MSE as a function of the vector M= L i i M ≤ ≤ 1 ) ( whose elements are
numbers of phase quantization levels at each magnitude level. The
Wilson’s method finds the optimal quantizer in such a way that the
decision and reconstruction levels r, m are iterative found for each
combination M, while the optimal combination is found by searching all
combinations. Wilson’s algorithm cannot be applied for middle and great
N. The asymptotic analysis of the polar quantizers with circular
symmetric densities is given in [2]. This analysis is approximate and
cannot be applied for any number of points and for great N, which will
be shown in this paper. We define the extension of the MSE over RL
(denoted by MSE(P)). We prove the convexity of this function and show
an efficient way to find M= L i i M ≤ ≤ 1 ) ( by Popt. Our algorithm
consists of two main iterative processes. The first iterative process
finds Popt, ropt, mopt with ε accuracy, while the second iterative
process determines Mopt, mopt, ropt using Popt as the starting value.
This paper eliminates incompleteness from [1] and [2]. We also give an
example of the quantizer construction for a Gaussian source. The
authors see their work as a contribution in knowing the best possible
solution in these classes of problems and also a possibility of
applying the technique exposed inhere on other classes of problems and
on larger dimensions. Igor
Voytovych 1), Myhailo Primin 2),
Valery Vasyliev 3), Pavlo Sutkovyy 3),
Mykola Budnyk 3), Igor Nedayvoda 4), Anatoly Rusanov 4), Tetyana Ryzhenko 4) Glushkov Institute for Cybernetics,
prospect Academika Glushkova 40, 03680 Kyiv-187, Ukraine
Phone: 380-44-2660128, 266-1267, Fax: 380-44-2663348, E-mail: d220@public.icyb.kiev.ua 1) PhD, Prof., Corr. Member of the NAS of the Ukraine 2) PhD, (contact author) 3) Candidates of Sciences 4) Researchers
Purpose of the work was to present algorithms and software developed
for working with multi-channel magneto-cardiograph. Such software is
intended for control of operation and computer processing of
magnetocardiographic (MCG) data obtained from the human heart.
Magnetocardiograph is controlled as “virtual device” from PC
mouse/keyboard, so as manually from electronic unit and all that
through the control microprocessors embedded into hardware units.
Software processing is performed both in on-line mode during process of
data acquisition and in off-line manner during post-processing.
Software allows preliminary processing, reconstruction and analysis of
magnetic maps, and also inverse problem solution. Package is intended
both for scientific studying of the heart electric activity and
studying of MCG informative indexes for clinical diagnostic of
cardiology diseases. Above software is planned to use at Strazhesko
Institute of Cardiology (Kyiv) within framework of project supported by
Science and Technology Center in Ukraine (STCU).
Konstantyn Spasokukotskiy 1),
Hans-Rolf Tränkler 1), Kateryna Lukasheva 2) 1) Universität der
Bundeswehr München, Institut für Meß- und Automatisierungstechnik,
Werner-Heisenberg-Weg 39. D-85577 Neubiberg 2) National Technical University of Ukraine (KPI), Chair of Acoustic and Acoustoelectronic Ukraine, 03056 Kiev, Prosp. Peremogy 37, Korp. 12., spasokukotskiy@yahoo.com, lukasheva1@yandex.ru, ima@unibw-muenchen.de http://smarthome.unibw-muenchen.de, http://www.unibw-muenchen.de/ima/
This paper describes a practical method of measurement a HVAC control
new variable. The method is based upon model-based estimation of
thermal comfort. The thermal comfort is the only physical value, that
truly corresponds to the changed (due to dynamic processing)
environment conditions in buildings. The dynamic processing is a
consequence of a modern demand-driven decentralized room climate
control, that has been presented earlier, or a consequence of
improvement of wall thermal insulation, that is beyond the limits of
the actual insulation standards (for example 2002 – Energy saving
regulations in Germany). The differences between various model types
will be discussed. Some results will be shown for the realized model
type.
Yury
Kolokolov, Anna Monovskaya
JSC “Electrotex”
5, Naygorskoye Shosse, 302020, Orel, Russia, Department of Design and Technology of Electronic Systems State Technical University of Orel, 29, Naygorskoye Shosse, 302020, Orel, Russia tel: 7 0862 421661, fax: 7 0862 416684, e-mail: anna@ostu.ru
Practical realization of model-based emergency forecasting in pulse
systems for the case of multiparameter variation in a possible wide
range associates with the problem of a-priori forming of
singlestatement information on system dynamics. The fractal diagram is
proposed to use for the purpose of system dynamics presentation. This
diagram allows one-to-one correspondence between the both, parameter
and phase,spaces establishing by line up stages.
Nataliya Yatskiv Institute of Computer
Information Technologies
In this article data compression method on the basis
of Galois basic functions of zero and first rank is proposed,
efficiency of coding with the usage of this method is investigated.
Dmitry
Malyuk 1), Georgy Boyarintsev 2)
1) Brest State Technical
University, 224017, Republic of Belarus, Brest,
Moskovskaya st., 267, rodjko.o.v@mail.ru 2) Brest State Technical University, 224017, Republic of Belarus, Brest, Moskovskaya st., 267
An approach to detect sharp changes in economical systems was developed
in this paper. We will give brief introduction in current methods of
prediction of nonlinear and chaotic time-series and give definition of
local Lyapunov exponents (LLE). Then author’s approach will be
described. Also some numerical results and discussions will be given.
|