701
1 INTEGRATEDEMERGENCYCONTROL
ENVIRONMENT
The purpose of the experimental studies conducted
during development of the emergency computing
center (ECC) is connected with the study of the
characteristics of the vessel’s interaction with the
externalenvironmentinemergencysituations(Fig.1)
based on the on‐board intelligent system (IS) of
the
new generation operating in the emergency
computing mode (Urgent Computing).‐UC) [1]‐
[17]. The integrated environment of IP is organized
within the framework of conceptual solutions based
on the dynamic theory of catastrophes [6] and is
characterized by the interaction of functional
components: fuzzy (FUZZY), neuro‐fuzzy (NF) and
neuroevolutionary(NE)systems.Thereal‐timemode
problem is provided by the service‐oriented
architecture[13],[14]inUCmode,andtheproblemof
uncertainty is within the concept of the minimum
description length (Minimal Decision Length (MDL)
[4]andcomplexitytheory[10].
In the practice of operation the situations take
place:duetouncertaintythevessel’scharacteristicsof
interest are not available for direct observation and
measurement,atthesametimethedataofaphysical
experiment can be quite complex and involve large
financialcosts.Inthiscase,someindirectinformation
abouttheinteractiondynamics is acquiredbased on
the interpretation of the results of a computational
experiment [6]‐[8]. Thus, the tasks of on‐board IE
andtheintegratedcomplexoftheECCareconsidered
to determine the causes according to the results
obtained as a result of measurements. The tasks of
this type are usually called inverse [11].
The causal
inverse tasks are individual and are used in the
construction of mathematical models of interaction
basedonthedynamictheoryofcatastrophes[6].
Operational Control of Marine Catastrophes Based on
Competitive Computing Technologies
V.A.Bondarev,Yu.I.Nechaev&P.Yu.Kovalishin
KaliningradStateTechnicalUniversity,Kaliningrad,Russian Federation
ABSTRACT:Thepaperconsidersthestructureandfunctionalelementsoftheemergencycomputingcenterthat
carries out operational control of sea catastrophes of vessels of the fishing fleet based on the integration of
intelligentsystems ofnewgenerationsandhigh‐performancecomputingwithinthe
problem‐orienteddynamic
environment of the virtua l testing area. The real time operation of the computer center is provided by the
systemintegration of information, algorithmicand software supportbased on dynamic measurement data
and a structured knowledge base. The focus is on providing decision support in complex dynamic
environments
usingmoderncatastrophetheory.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 12
Number 4
December 2018
DOI:10.12716/1001.12.04.08
702
SYSTEM OF EMERGENCY SITUATIONS
MODELING
Intellectual
technologies
FUZZY-
systems
NF-
systems
NE-
systems
High performance
computing
SOA- systems
UC- systems
Uncertainty- systems
Functional state
Behavior modeling
Training simulation
Figure1.Informationprocessingparadigmofemergencymodeling
2 METHODSANDMODELSUSINGFOR
EMERGENCYSITUATIONSINTERPRETATION
The procedure for solving tasks in the treatment of
causal relationships is often associated with
overcoming serious mathematical difficulties. The
successofthesolutionstronglydependsbothonthe
qualityandquantityoftheinformationobtainedfrom
theexperiment,andonthe
methodofitsprocessing
[2]. That is why the dynamic theory of catastrophes
providescreation of a multi‐model program
complex (MMPC) using a geometric and analytical
interpretation of the vessel’s behavior in emergency
situations. Based on this representation, the solution
is carried out within the framework of the Soft
Computing [17] and Data Mining [1] procedures
using mathematical models based on the modified
MathieuandDuffing [6]differentialequations.Inthis
case,thesolutionoftheinverseproblemispreceded
bythestudyofthepropertiesofthedirectproblemon
the basis of a formal conceptual analysis [16]. It
is
assumedthattheoriginaldatahavelargedimensions
(many factors and conditions of the vessel), do not
alwaysobeythenormaldistribution,areincomplete,
inaccurate and noisy. It is believed that in non‐
standard situations in conditions of considerable
uncertainty, the initial data is extremely difficult to
establish as
a result of specially organized physical
modelingandtheonlywaytoobtaininformationisa
computational experiment. This characteristic of the
sourcedataallowsustoformulatethefollowingbasic
requirements for a mathematical model of the
interactionoffunctionalelementsoftheEVC:
informative interpretability using the concept
of
Soft Computing and Data Mining based on the
geometric and analytical components of the
dynamiccatastrophemodel[6];
effective computability based on parallel
information processing algorithms in a
multiprocessorcomputingenvironment[6]‐[8].
It is these two requirements that determine the
construction of the interaction model in the
development of the information environment and
algorithms for controlling emergencies, as well as
testingtheknowledgebaseofIE(Fig.2).
A
B
C
D
E
F
G
Modeling environment Modeling tasks
Risk assessment of decisions
Analysis of the physical
features of the situation
Discovery and analysis of the
interaction model
Emergency Detection and
Analysis
Analysis of alternatives in a
fuzzy environment
Emergency Information Modeling Environment
Figure2. Information environment of interpretation of
decisions when modeling emergency situations: A‐
numerical analysis; B‐phase plane method; C‐graph
theory;(D‐G)‐dynamicenvironmentsoffuzzy(D),neural
network(E);F‐cognitive(F)andphysical(G)modeling
The functional analysis of the information
environment modeling ensures that the tasks and
modelingtoolsareconsistentwiththeECCcomputer
complexused.Astheinitialtimepointt0tomonitor
thecurrentsituationintheinformationenvironment,
the set of necessary tasks Task (t0) and their set of
functionsOpt
(t0)areallocated:

,,,
210
210 iiitp
CCpiiCtCet
 (1)
i.e.thefullsetoftasksСet(t0)isdividedintoclasses
Сiinaccordancewiththeirpurposeintheoperation
oftheECC.
AtoolfordescribingtasksCandtheorderoftheir
distributionistheinteractionmatrix.Theelementsof
the matrix denote the intensity of
data transmission
during the operation of tasks that provide
identification, approximation and forecast of the
vesseldynamicsbymovingfromthej–thstageofthe
selectedsequencetothei–th.
3 THEGENERATIONOFALTERNATIVESAND
THECHOICEOFDECISIONSINTHE
EMERGENCYSITUATIONSCONTROL
The construction of the multi‐model program
complex(MMPC) and the generation of alternatives
when choosing a solution that includes various
703
scenarios of interaction in emergency situations is
carriedoutintheformof a situational modelofthe
gamewithadynamicallychangingclassofstrategies
and a controlled scenario (Fig. 3). To solve this
problem,thescenariodescribedbyafinitegraph[6]‐
[8]isformulated:
G=(S
S,WS), (2)
wheretheSsstructureistheunionofallconsidered
S
S
tj
emergencies taking into account the moments
of time defining controls on the implementation
interval at a given moments t
j, j=1,…,N, and the
structureW
SSSSS describes transitions between
situationsbydisplayingasetoftheIEoperatortactics
asadecisionmaker(DM).
Figure3. Information Processing Model for Emergency
Control
Theproblemunderconsiderationisadjacenttothe
problem of controlling an ensemble of trajectories
under conditions of uncertainty [5]. However, the
approach to the generation of alternatives involves
thestudyofthebehaviorofnonlinearnonstationary
systems (NN‐systems) for isolatedtrajectories which
isequivalenttothetaskofcontrolling
anensemble.
Axiom 1. When the conditions of irregularity and
control with geometric constraints are met, an ensemble
predictionoftrajectoriesunderuncertaintycan beachieved
by building solutions generated by elements minimizing
thefunction of interpretingthe behavior space at a given
stage of system evolution within the framework
of a
dynamictheoryofcatastrophes[6].
Let us consider the application of this axiom in
constructinganensembleoftrajectoriesofadynamic
catastrophemodelbygeneratingsolutionsofanNN
systemwithintheframeworkoftheUCconcept[15].
Suppose that the initial state of the system is not
known in advance and the restriction on admissible
values of the quantity under investigation is given
х
0
Х
0
,гдеХ
0
isagivencompactsetinRn.
Thenateachmomentoftimetmanyalternatives
areknown


,,,,,,
0000
XxxutxUXutXutX
(3)
uniting all trajectories of the system with a known
controlu=u(t)andwithallpossiblevectorsх
0
Х
0
.
Thus,wehaveatrajectoryensemble[5]:
k
tttutX
0
,,  (4)
generatedbytheХ
0
andthecontrolu(•)foragiven
perturbation w (t).By choosing the function u (t),
onecancontrolthesetoftheensembleinthecontrol
space. The control goal on the implementation
intervalis tobring the ensemble X (t, u (•))into a
Euclidean –
neighborhood,atagiventime,and
the optimality requirement is that the quantity be
minimal.
Let control be defined in a given class of
functionsu()U, and–beasetinR
n
. Then
the Euclidean distance of a point x to in the
formofasymbold(x,)hastheform:
,min, zzxxd (5)
andthecorresponding–neighborhood–set
()isdefinedbytheinequality
.,
xdx
(6)
Foragivenfunctionu(•)ofthecontrolspace, the
setX(t,u(•))willcompletelyliein(u(•))‐(u())‐
neighborhoods

.,,max utXxxdu
(7)
Consequently, the optimal control u0, formed on
the basis of the evaluation of the function of
interpretation of the space of behavior, must satisfy
thecondition


 

,,,,maxmin
,,min
000
utXxPuxd
utXxxdu
х
u
(8)
,TttPtuuP
(9)
whereР(t)–convexcompactsetinspaceR
P
.
Generation of scenarios for the interaction of the
vesselwiththeexternalenvironment on the basisof
relations (3)‐(7) is carried out on the basis of the
modelof the“essence‐relationship” type [6]‐[8].In
accordance with this structure, alternatives are
formed (emergency scenarios) comparing and
choosing the
preferred alternative. A generalized
criterionforthe transformation of informationwhen
comparing alternatives can serve as an indicator of
informationefficiency[3].Withregardtotheproblem
ofchoosingasolutionforcontrollingthedynamicsof
ashipinanemergency,thiscriteriontakestheform:
704
,1)(
x
EEECR
(10)
where

dtd
XE
(11)
relative change of the current parameter (Х
t)
characterizedbythedynamicsoftheinteractionof
the emergency vessel with the external
environment (equilibrium landing parameters,
restrictionsontheemergencystabilityparameters)
fromthemaximumallowable
d;

XXXE
tx
(12)
relative change of the current value of the
determining parameter (factor)Х
t (ХtХ) from
selectedvalueХinthe specified interval(a , b).
HereXcanbearegular,randomandfuzzyvalue.
Wheninterpretingsolutions,differentmodelscan
be used. The simplest model is associated with the
constructionandanalysisoftheselectionfunction[7],
[8],whichallowsforthe
targetednarrowing ofmany
alternativesinthedecisionsupportsystem(Decision
Making Process) when choosing and justifying the
best decision based on the principle of competition
[6]:
Ф(Q)={q
iQ|U}, (13)
where q
i – object from the setQ, selected by
conditionU.
Theselectionconditionisrepresentedasatuple
U=<(t),(R)>, (14)
where (t) – current vessel’s condition information;
(R)=<R,T>–setofselectionrules.R–relationship
between elementsх
i,i, T – type of choice
(equivalence,match,preference).
4 EVALUATIONOFTHEEFFECTIVENESSOF
DECISIONSINTHEEMERGENCYSITUATIONS
CONTROL
Weformalizethetaskofevaluatingtheeffectiveness
of the DMP system in controlling emergency
situationsintheECCmodels.Theimplementationof
thecontrolfunctiondependsontheparameter
vector
ХR
n
determiningthedynamicsofthevesselandthe
state vector of the environmentWR
m
in an
emergency. If [Х, W]A, then the solution with the
parameter vectorX maintains the condition of the
vessel in a dynamic environmentcharacterized by
vectorW.If [X, W]Bthen the generated
solution leads to inefficient system operation DMP.
The above conditions determine the solution
to the
problemofchoosingasolution:

AW)(X,0,,* WXх ; (15)
ВW)(X,0,,*
WXх
, *Х*х , (16)
wherex*–selectedclassofseparatingfunctions.
When conditions (15), (16) are implementedthe
rangeofpossiblevaluesofthecontrolledparameters
ofthevesselisestablishedinthemodelsoftheECC
computer complex, which is limited by various
factors, including the features of operation and the
level of
intelligent technologies used. Each specific
implementationofthetechnicalsolutioncorresponds
to certain values of the parameters that satisfy the
conditions[8]:
.,,1,
maxmin1
niXXX
ni
 (17)
Thus, in the n‐dimensional parameter space for
each implementation, you the vector of parameters
canberepresented
,,,
1
T
n
xxX (18)
which belongs to the parameter space defined by
inequalities(17).
Thevectorofparameters(18)uniquelydefinesthe
characteristics of the vessel, the set of which we
denoteby(Сh)
j,j=1,…,m.
Thenumberofcharacteristicsisdeterminedbythe
characteristics of the vesselʹs dynamics in various
operatingconditions.
Letusassociateeachsetofcharacteristicswiththe
vector
T
m
ChChH ,,
1
 (19)
m–dimensionalspace.
Inthiscase,theshipcanbeconsideredasasystem
withninputsforparametersх
iХandmoutputsfor
characteristics(Ch)
j.
ToeachvectorXoftheparameterspace(18)sucha
system associates the space vector of the
characteristics of an emergency situation defined by
(19). The considered model allows us to construct a
geometric interpretation of various versions of the
tasks,theiranalysisandoptimalmappingintheECC
system.
Thechoiceofsolutiondependsonthecomplexity
of the situation being controlled, in emergency
situations, especially in conditions of considerable
uncertainty, a collective decision strategy is used by
the experts of the ECC. When developing this
strategy, you can use the approach [9], the
developmentofwhichin
thisapplicationisrelatedto
the following features. Consider the feedback
characterizing the implementation of the choice of
solutions in the presence of two choices, which we
denote и. The degree of reliability of these
features is characterized by the numbersА
иА
.
ThenumberofexpertsECCwhomadeachoice
orisdenotedbyХ
иХ
.Atthesametime,inthe
course of the situation development it is possible to
change the decisions of experts. Relative number of
705
expertswillingtochangethedecisionintheprocess
ofdevelopmentofthesituationisinproportiontothe
number of those experts who have already made a
choice . In this case, the preference of choice is
defined asА
(А
+А
). Similarly, the number of
expertswillingtochangetheselectiontowill
be proportionalХ
in accordance with the formula 
А
/(А
+А
). This leads to a system of equations for
Х
AAAXAAAXXdtdX
(
(20)
or considering thatХ
= N –Х
, where N – total
numberofexperts


dX dt X NA A A X


(21)
Thus, various choices affect the efficiency of the
DMSsystemwhichisafunctionoftheinstantaneous
state of the vessel due to its dependence on the
variables characterizing the emergency. These
considerations can be generalized to the case of an
arbitrary number of elections K, taking into account
the real situation, when the preference of the i ‐th
option depends on the number of the expert group,
whichmustmakeachoice:

1
1 ( ) , 1, ..., .
k
iiijijij
ij
dX dt X X N A A i K





(22)
Here the group of experts is heterogeneous and
falls into several subgroups, each of which has its
ownideaoftherelativepreferenceofthischoice.
Riskassessmentofdecisionsmadeonthebasisof
the developed strategy is carried out using various
interpretations[2],[6]‐[8].Thetheory,methodsand
technologiesfordevelopingvariousclassesoftasksin
ariskassessmentsystemcovervariousproblemareas
[4]‐[6]. The complexity and interrelation of these
areas bring to the fore the problem of assessing the
quality of models, analyzing and streamlining the
choice of the most preferable models for solving
applied
problems. The urgency of the problem is
exacerbated if the dynamics of the vessel are
described by a multi‐model computing complex [2]
which may include heterogeneous and combined
modelseachofwhichisevaluatedbyitsownsystem
ofindicators.
5 CRITERIONBASISOFEMERGENCYCONTROL
Aconceptualmodelof
aship’sbehaviorinemergency
situations formalizes the processes of building
applied tasks and criterial functions for interpreting
interaction processes in the implementation interval.
Figure 4 presents the sequence of operations that
determinesthecriterialbasisforassessingthesafety
of a vessel in the form of the main stages
of
determining the parameters of the environment, the
dynamics of interaction, as well as the stage of
evolutioninpredictingthebehaviorofthevessel.
CONCEPTUAL MODEL OF FUNCTIONING OF THE EMERGENCY SITUATION
CONTROL SYSTEM
Identification
Restoration of the spectra of external
disturbances (wave and wind parameters)
Approximation
Assessment of vessel dynamic characteristics
(interaction parameters)
Prediction
Prediction of a ship’s behavior in an
emergency (evolutionary stage)
Figure4. Criteria basis for assessing the dynamics of the
vesselatthestagesofoperationofthebehaviorspace
Improvement of the theoretical, methodological
and technical support of operational control of
emergencysituationsisassociatedwiththeuseoftwo
systems of criteria‐based assessments (Fig. 5). The
first (local) criteria system is related to ensuring the
safetyconditionandcanbeimplementedonthebasis
of the developed standards
in the form of an
embedded procedure of inference rules. The second
(global) system includes national and international
requirements, which are ensured regardless of the
particulardynamicsofthevessel.
The local system is developed in the process of
creating a dynamic knowledge base and takes into
accountthecharacteristic
featuresofthevesselunder
study.Improvementofthelocalsystemiscarriedout
in the direction of creating a fuzzy criteria system
based on the methods of formalizing information
with regard to its incompleteness and uncertainty
within the framework of the concept of “soft
calculations”[17].
Thetransitionfroma
localsystemtoaglobalone
is carried out through a description (a conceptual
modelofthesystem),whichfixes informationabout
the vessel being modeled and the process of
interaction in terms of typical competing
mathematical models and knowledge structures.
When choosing a simulation scheme for an
emergency,the
conceptofafunctioningenvironment
is introduced, which makes it possible to use
information of an applied nature on the purpose of
modeling, the laws ofsystem evolution,the existing
mathematical apparatus for studying methods and
algorithms for making decisions on ship
management. Thus, the object of the considered
appliedtheory
ofrationingofvesselcharacteristicsis
themodelingprocess.
Figure5. Criterion basis for the rationing of extreme
situations
The practical implementation of the formulated
approach is associated with the creation of a fuzzy
706
systemoflogicalrulesfortheknowledgebaseofthe
IE [2], [6]‐[8]. One group of rules allows you to
defineanassessmentofthedangerofasituation,the
other is aimed at preventing it and can be directly
managedtomakedecisionsonreducingthespeedof
DOandchangingthecourseangle
6 ASSESSMENTOFTHEADEQUACYOFMODELS
OFTHEVESSEL’SBEHAVIORINEMERGENCY
MANAGEMENT
The strategy of assessing the adequacy of the ECC
computingcomplex (Fig.6)determinesthe
formalization of the procedure based on the
consideration of factors characterizing a priori
information [2],
the concept ofМDL [4], and the
problem of complexity [10]. As follows from this
figure,theadequacyproblemissolvedbyintegrating
aprioriinformationwhichdeterminesthechoiceofa
solutioninaccordancewiththeconceptualmodelof
the dynamic theory of catastrophes [6] which is
adaptedtothe
probleminquestion.
Prior Information
Dynamic
measurement data
Physical simulation
results
The results of
mathematical
modeling
Concept MDL
Formation of an
information array
Selecting data
structures
Evaluation error
description
Difficulty problems
Formation of a set of
models
Model selection
based on criteria
Evaluation of the
adequacy of the
model
Figure6. The flow of information that determines the
strategy for assessing the adequacy of the model of
interactioninanemergency
Theassessmentofadequacyiscarriedoutonthe
basis of the modified O. Balchi scheme [12] for a
specificapplicationinorder to takeinto account the
dataofphysical,neuro‐fuzzyandneuro‐evolutionary
modeling (Fig. 7). At the same time, improvements
weremadeinconsideringtheinteractionmodel
asan
integralpartofthepracticalapplicationbasedonit‐
thetaskofmodelingemergenciesintheoperationofa
software package based on the principle of
competition.
Thefirstcycleisassociatedwiththedevelopment
ofcompetingmodels(modeling‐M),implementedon
the basis of the neurodynamic system
(ND‐system)
andmethodsofclassicalmathematics.Intheprocess
ofperformingthiscycle,thestructuralandparametric
synthesis of the neural network in the NF and NE
modeling tasks is implemented, and for the
competing model, the assessment of the overall
structure and components is carried out within the
framework
of sequential statistical analysis
procedures.
The second cycle refers to the implementation of
the corresponding mathematical (imitation)
experiments performed with competing models
(simulation‐S) for given initial conditions and
elementsoftheinputvector.HereareformedtheNF
andNEmodelsofdatabankanalysisonthebasisof
whichthecomputationalexperiment,generation and
analysis of alternatives and the assessment of
adequacy are implemented,. The construction and
analysis of the competing model is carried out in
accordance with the formalization of the problem
under conditions of significant uncertainty in
accordancewiththealgorithm[6].
Figure7.Modifiedmethodologicalschemeforassessingthe
adequacy of neural network and standard (traditional)
models:M,S,P‐informationconversioncycles
The third cycle is the most important. It is in
carrying out physical (physical‐P) experiments, on
thebasisofwhichthemodelsareformedthatprovide
an assessment of adequacy under conditions of
complete uncertainty. On this cycle, in the ND
system,thecomponentsoftheNFandNEmodels
are
formed using physical modeling data, a
computational experiment and an adequacy
assessmentareimplemented.
Intellectual support of the procedures M, S, P is
provided by the control system of calculations and
visualizationofsimulationresults.Asassessmentsof
the adequacy of fuzzy, neural network, and
competing models, one should
follow the
recommendationssetoutinthepaper[8].
7 CONCLUSION
The considered conceptual solutions implement the
strategyofthedynamictheoryofcatastropheswhich
determines the evolution of the vessel under
conditions of continuous changes in the behavior of
the object and the external environment. The
managementoftheinteraction
processisprovidedby
thelevelsoffunctioningofthestructuralcomponents
of the ECC using an interactive control system that
providesanalysisofthestateofthecontrolledvessel,
development of management decisions and their
implementation, based on the problem‐oriented
onboardIE.
TheEECmulti‐modelcomplexcontains
interaction
algorithms and operating environments as well as
visualization models that implement computation
707
and data presentation using elements of modern
cognitivecomputergraphics.Operationsthat
determine the flow of information based on the
dynamic theory of catastrophes and intelligent
technologies are implemented within the framework
ofconceptualsolutionsforcontrollingemergenciesin
anon‐stationarydynamicenvironment.
Theimplementationofoperationalandlong
‐term
planning of emergency scenarios is organized. as a
model with a dynamically changing class of
strategies. The generation of scenarios is carried out
onthe basis ofan entity‐relationship type model. In
accordance with this structure, alternative scenarios
andthechoiceofthepreferredcomputingtechnology
basedon
theprinciple of competitionareformed. A
generalizedcriterionwhencomparing alternativesis
the indicator of information efficiency and
intelligenceofthesystem.
LITERATURE
[1]Barsegyan A.A., Kupriyanov M.S.. Stepanenko V.V.,
Kholod I.I. Metody i modeli analiza dannykh: OLAP i
DataMining.–Sankt‐Peterburg.BKHV‐Peterburg,2004.
[2]Bondarev V.A., Nechayev YU.I. Iskusstvennyy intellekt
v chrezvychaynykh situatsiyakh moreplavaniya. –
Sankt‐Peterburg,2017.
[3]Vasilʹyev V.A. Prinyatiye resheniy pri proyektirovanii
preobrazovateley informatsii // Avtomatizatsiya
i
sovremennyyetekhnologii.№11.2003,s.30‐33.
[4]Kolmogorov A.N. Teoriya informatsii i teoriya
algoritmov.–M.:Nauka,1987.
[5]Kurzhanskiy A.B. Upravleniye i nablyudeniye v
usloviyakhneopredelennosti.–M.:Nauka,1977.
[6]NechayevYU.I.Teoriya katastrof:sovremennyy
podkhodpriprinyatii resheniy.–Sankt‐Peterburg:Art‐
Ekspress,2011.
[7]Nechayev
YU.I.,PetrovO.N.Nepotoplyayemostʹsudov:
podkhod na osnove sovremennoy teorii katastrof. –
Sankt‐Peterburg:Art‐Ekspress,2014.
[8]Nechayev YU.I. Topologiya nelineynykh
nestatsionarnykhsistem:teoriyaiprilozheniya.–Sankt‐
Peterburg:Art‐Ekspress,2015.
[9]Nikolis Dzh. Dinamika iyerarkhicheskikh sistem. – M.:
Mir,1980.
[10]SolodovnikovV.V.,TumarkinV.I.Teoriyaslozhnostii
proyektirovaniyesistemupravleniya.–M.:Nauka,1990.
[11]Tikhonov A.N., Arsenin V.YA. Metody resheniya
nekorrektnykhzadach.M.:Nauka,1986.
[12]Balci O. Verification, validation and accreditation //
Proceedingsofthe 1998WinterSimulation Conference.
—1998,p.p.41–48.
[13]Gallopoulos S., Housts E., Rice J. Problem Solving
Environment // IEEE Computational
Science and
Engineering.Summer,1994.
[14]Lublinsky B. Defining SOA as an architectural style. 9
January 2007. [Электронный ресурс]:
http://www.ibm.com/developerworks/architecture/libra
ry/ar–soastyle/
[15]UrgentComputingWorkshop2007. Argonne National
Lab, University of Chicago, April 25‐26, 2007.
[Электронный ресурс]:
<http://spruce.teragrid.org/workshop/urgent07.php>.
[16]WilleR.Restructuringlatticetheoryanapproachbased
on
hierarchiesofconcept/orderedsets/EditorI.Rival.
—Dordrecht‐Boston,1982.
[17]Zadeh L. Fuzzy logic, neural networks and soft
computing //Соmmutation on the ASM‐1994. Vol.37.
№3,р.р.77–84.