|
RUSSIAN ACADEMY OF SCIENCES
THE INSTITUTE OF INFORMATICS PROBLEMS (IPI RAN)
The Systems and Means of Informatics
Special Issue, 2008
Оглавление
ON A SIMULATION APROACH TO CLUSTER STABILTY VALIDATION
- Zeev Barzily Software Engineering Department, ORT Braude College
of Engineering, Karmiel, Israel
- Mati Golani Software Engineering Department, ORT Braude College
of Engineering, Karmiel, Israel
- Zeev Volkovich Software Engineering Department, ORT Braude College
of Engineering, Karmiel, Israel
Abstract: In the current paper we outline a new approach to the “true
number of clusters” determination problem. Our method combines
both the stability and density concentration approaches. In the
spirit of the density estimation methodology, we consider each
cluster as an island of “high” density of items in a sea of “low”
density. In addition, following the cluster steadiness concept, we
suggest that these islands are “resistant” to a random noise. In
other words, we believe that adding noise to the attributes of the
data elements does not change the clusters structure. A second
novelty of our approach is the proposition to measure the similarity
between source-data clusters and noisy-data clusters by means
of two sample test statistics, represented by probability metricsdistances.
Such a pair seems as an appropriate database for the
true number of clusters determination. As a consequence of the
high resemblance between these samples, within the partitions,
the similarity is expected to be amplified under the true number
of clusters. According to our model, the true number of clusters
corresponds to the empirical distance distribution which is most
concentrated at zero. Thus, our procedure can be considered as the
creation of an empirical normalized distance distribution, followed
by testing its concentration at zero. This test is carried out by
means of the sample mean and the size of the sample first quartile.
AN APPROACH TO CLASSIFICATION OF COMPUTER SYSTEMS FAULTS LOCALIZATION MODELS
- Sergey Frenkel Institute of Informatics Problems, Russian Academy of Sciences, Moscow, Russia
- Eugene Levner Holon Institute of Technology, Holon, Israel
- Victor Zakharov Institute of Informatics Problems, Russian Academy
of Sciences, Moscow, Russia
Abstract: This paper suggests a characterization of fault localization
strategies and a classification of probability-based search algorithms,
which are widely used in the current practice of complex
computer systems and networks systems maintenance and service.
The characterization is performed in terms of some tuples
including: characteristics of fault models, localization procedures,
cost functions, and other factors. Such characterization (and an
induced classification) can be used for a rational choice of search
algorithms at early system design stages, for development of
strategies of possible fault localization during the target systems
maintenance and service. Search algorithms classification is based
on the special notation like those used in the queuing theory.
COMPLEXITY AND CONSISTENCY OF STATISTICAL CRITERIA
- Alexander Grusho
Lomonosov Moscow State University, Moscow, Russia
- Nick Grusho
Russian State University for the Humanities, Moscow,
Russia
- Elena Timonina
Russian State University for the Humanities, Moscow,
Russia
Abstract: We investigate relations between consistency of statistical
criteria sequences in finite probability spaces and asymptotic complexity
of test calculations. It is proved that for every consistent
test sequence it is possible to construct another consistent sequence
for which complexity of calculations of membership functions
of new criteria critical sets can be made asymptotically small
in comparison with similar complexity for initial criteria. However
such simplification appears to be fictitious as a matter of fact.
To prevent fictitious simplification of calculations of membership
functions of critical sets in a sequence of criteria it is necessary
to set additional restrictions on classes of considered criteria. It is
shown, that in case of such natural restrictions any simplification
of calculations can lead to a failure of a consistency of criteria
sequence.
Estimations of efficiency of two-level criteria are resulted,
when simple but not consistent in the given class of alternatives
criteria work at first, and only in case of nonacceptance of a hypothesis
hard computable but consistent criteria are applied.
BAYESIAN QUEUEING AND RELIABILITY MODELS
- Alexey Kudryavtsev
Faculty of Computational Mathematics and Cybernetics,
Lomonosov Moscow State University, Moscow, Russia
- Sergey Shorgin
Institute of Informatics Problems, Russian Academy
of Sciences, Moscow, Russia
- Vsevolod Shorgin
Institute of Informatics Problems, Russian Academy
of Sciences, Moscow, Russia
- Valery Chentsov
Institute of Informatics Problems, Russian Academy
of Sciences, Moscow, Russia
Abstract: The Bayesian approach for certain tasks of queueing systems
theory and reliability theory is investigated. The method provides
the randomization of system characteristics with regard of a priori
distributions of input parameters. This approach could be used,
for instance, for calculating average values and for construction
of confidential intervals applicable for performance and reliability
characteristics of large groups of systems or devices. The results
for certain models of input flow and service time parameters are presented.
G-NETWORK WITH THE ROUTE CHANGE
- Rosanna Manzo
Department of Information Engineering and Applied
Mathematics University of Salerno, Fisciano, Italy
- Alexander Pechinkin
Institute of Informatics Problems, Russian Academy
of Sciences, Moscow, Russia
Abstract: Queueing networks with negative customers (G-networks),
Poisson flow of positive customers, non-exponential nodes, and
dependent service at the different nodes are under consideration.
Every customer arriving at the network is defined by a set of
random parameters: customer route, the length of customer route,
customer volume and its service time at each route stage as well.
The arrival of a negative customer to a queuing system causes
one of the ordinary (or “positive”) customers to be removed (or
“killed”) if any is present. The “killed” customer continues its way
along the new random route. For such G-networks, the multidimensional
stationary distribution of the network state probabilities
is shown to be representable in product form.
FAN-BEAM STOCHASTIC TOMOGRAPHY
- Oleg Shestakov
Faculty of Computational Mathematics and Cybernetics,
Lomonosov Moscow State University, Moscow, Russia
Abstract: In fan-beam tomography settings an object is illuminated by
the divergent pencil of rays emitted from the source which moves
around it. This scheme works much faster than traditional parallel
beam tomography. In some biological and physical applications an
abject under study is described by random function. This paper
considers the problem of recovering probabilistic characteristics of
random function from fan-beam projections.
ESTIMATION OF DELAY DISTRIBUTION IN HIV DYNAMICS
- Anastasia Ushakova
Department of Mathematical Sciences, Norwegian
University of Science and Technology, Trondheim,
Norway
Abstract: In this paper we present two methods of estimating a delay
distribution in biological dynamical systems. The model of HIV
infection serves as an example of such systems. The first method
is based on parametric approach and on approximation of the delay
density by a gamma-density. The second method is nonparametric
and is based on solution of a convolution equation with selection
of the regularization parameter from a parametric start.
ON LINGUISTIC CLASSIFICATION OF BACTERIAL GENOMES
- Zeev Volkovich
Software Engineering Department, ORT Braude College
of Engineering, Karmiel, Israel
- Valery Kirzhner
Institute of Evolution, University of Haifa, Haifa, Israel
- Zeev Barzily
Software Engineering Department, ORT Braude College
of Engineering, Karmiel, Israel
Abstract: The paper is devoted to classification of 185 full prokaryote
genomes using a modification of the compositional spectra
method. This modification suggests separate calculation of the
compositional spectra for coding and non-coding subsequences
of the genome. For each subsequence, the corresponding vector,
in Euclidian space, can be obtained using certain manipulations
of the compositional spectra. This allows analyzing the structure
of genome and determining the most probable number of
genome clusters without any additional information. Our clustering
method is based on the application of the external indexes of
partitions agreement and the number of the misclassified items
within repeated partitions. A biological justification, for the four
and the two letters alphabets, substantiates the appropriateness of
the outcomes acquired.
EXPERIMENTALLY OBTAINED MEASUREMENTS FOR RECONSTRUCTION OF ORTHOGRAPHIC VIEWS
- Miri Weiss-Cohen
Ort Braude College of Engineering, Karmiel, Israel
- Alina Bondarenko
Rafael, Haifa, Israel
- Yoram Halevi
Technion-Israel Institute of Technology, Haifa, Israel
Abstract: The orthographic views are constructed from a moving sensor,
either a distancemeter or a camera, in an automatic fashion
using estimation techniques. The raw distance measurements are
processed via a filter which generates estimates of the part dimensions
and position. The fact that the outcome of the estimation
(measurement) process is a set of explicit contour equations is
suitable for creating orthographic views using variational geometry
representation.
ON INTRUSION IN A LINEARLY-STRATIFIED
AMBIENT: THE ASYMMETRIC STEADY-STATE
MODEL
- Tamar Zemach
Department of Computer Science, Technion, Haifa
- Marius Ungarish
Software Engineering Department, ORT Braude College
of Engineering, Karmiel, Israel
Abstract: The behavior of the steady intrusive gravity currents spreading
into a stratified ambient fluid is investigated. The intrusive
gravity current of thicknessh and density?c which propagates
with speedU at the neutral buoyancy level of a long horizontal
channel of heightH into a stratified ambient fluid whose density
increases linearly from ?o to ?b is investigated. The intrusive
and the ambient fluids are assumed to be asymmetric due to
axis passing the stagnation point of the system. The Boussinesq,
high-Reynolds number two-dimensional configuration is discussed.
The Long’s model combined with the flow-force balance over the
width of the channel and the pressure balances over a density
current are used to obtain the desired results. It is shown that
the intrusion velocity decreases with decreasing the asymmetry
of the system and approaches its minimum for the symmetric
configuration. In additional, the comparison between asymmetric
and symmetric configurations shows no significant differences
between the models.
|
|
