Computer Methods in Applied Mechanics and Engineering.docx

Computer Methods in Applied Mechanics and Engineering.docx

《Computer Methods in Applied Mechanics and Engineering.docx》由会员分享，可在线阅读，更多相关《Computer Methods in Applied Mechanics and Engineering.docx（31页珍藏版）》请在冰豆网上搜索。

ComputerMethodsinAppliedMechanicsandEngineering

Topologicalclusteringforwaterdistributionsystemsanalysis  

EnvironmentalModelling&Software,InPress,CorrectedProof,Availableonline15February2011

LinaPerelman,AviOstfeld

 Showpreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

Purchase

$19.95

716

HyphArea—Automatedanalysisofspatiotemporalfungalpatterns  OriginalResearchArticle

JournalofPlantPhysiology,Volume168,Issue1,1January2011,Pages72-78

TobiasBaum,AuraNavarro-Quezada,WolfgangKnogge,DimitarDouchkov,PatrickSchweizer,UdoSeiffert

 Closepreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

AbstractAbstract|Figures/TablesFigures/Tables|ReferencesReferences

Abstract

Inphytopathologyquantitativemeasurementsarerarelyusedtoassesscropplantdiseasesymptoms.Instead,aqualitativevaluationbyeyeisoftenthemethodofchoice.Inordertoclosethegapbetweensubjectivehumaninspectionandobjectivequantitativeresults,thedevelopmentofanautomatedanalysissystemthatiscapableofrecognizingandcharacterizingthegrowthpatternsoffungalhyphaeinmicrographimageswasdeveloped.Thissystemshouldenabletheefficientscreeningofdifferenthost–pathogencombinations（e.g.,barley—Blumeriagraminis,barley—Rhynchosporiumsecalis）usingdifferentmicroscopytechnologies（e.g.,brightfield,fluorescence）.Animagesegmentationalgorithmwasdevelopedforgray-scaleimagedatathatachievedgoodresultswithseveralmicroscopeimagingprotocols.Furthermore,adaptabilitytowardsdifferenthost–pathogensystemswasobtainedbyusingaclassificationthatisbasedonageneticalgorithm.ThedevelopedsoftwaresystemwasnamedHyphArea,sincethequantificationoftheareacoveredbyahyphalcolonyisthebasictaskandprerequisiteforallfurthermorphologicalandstatisticalanalysesinthiscontext.BymeansofatypicalusecasetheutilizationandbasicpropertiesofHyphAreacouldbedemonstrated.ItwaspossibletodetectstatisticallysignificantdifferencesbetweenthegrowthofanR.secaliswild-typestrainandavirulencemutant.

ArticleOutline

Introduction

Materialandmethods

Experimentalset-up

Thehostandpathogens

Fluorescenceimageacquisitionprotocol

Brightfieldimageacquisitionprotocol

Segmentationofhyphalcolonies

Achievingadaptivebehavior

Imagecorpus

Results

R.secalisinfectedmaterialandsegmentation

B.graminisinfectedmaterialandsegmentation

Discussion

Acknowledgements

References

Purchase

$41.95

717

AutomatedgenerationofcontrapuntalmusicalcompositionsusingprobabilisticlogicinDerive  OriginalResearchArticle

MathematicsandComputersinSimulation,Volume80,Issue6,February2010,Pages1200-1211

GabrielAguilera,JoséLuisGalán,RafaelMadrid,AntonioManuelMartínez,YolandaPadilla,PedroRodríguez

 Closepreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

AbstractAbstract|Figures/TablesFigures/Tables|ReferencesReferences

Abstract

Inthiswork,wepresentanewapplicationdevelopedinDerive6tocomposecounterpointforagivenmelody（“cantusfirmus”）.Theresultisnon-deterministic,sodifferentcounterpointscanbegeneratedforafixedmelody,allofthemobeyingclassicalrulesofcounterpoint.Inthecasewherethecounterpointcannotbegeneratedinafirststep,backtrackingtechniqueshavebeenimplementedinordertoimprovethelikelihoodofobtainingaresult.ThecontrapuntalrulesarespecifiedinDeriveusingprobabilisticrulesofaprobabilisticlogic,andtheresultcanbegeneratedforbothvoices（aboveandbelow）offirstspeciescounterpoint.

Themaingoalofthisworkisnottoobtaina“professional”counterpointgeneratorbuttoshowanapplicationofaprobabilisticlogicusingaCAStool.Thus,thealgorithmdevelopeddoesnottakeintoaccountstylisticmelodiccharacteristicsofspeciescounterpoint,butratherfocusesontheharmonicaspect.

Theworkdevelopedcanbesummarizedinthefollowingsteps:

（1）Developmentofaprobabilisticalgorithminordertoobtainanon-deterministiccounterpointforagivenmelody.

（2）ImplementationofthealgorithminDerive6usingprobabilisticLogic.

（3）ImplementationinJavaofaprogramtodealwiththeinput（“cantusfirmus”）andwiththeoutput（counterpoint）throughinter-communicationwiththemoduledevelopedinDerive.Thisprogramalsoallowsuserstolistentotheresultobtained.

ArticleOutline

1.Introduction

1.1.Historicalbackground

1.2.“CantusFirmus”andcounterpoint

1.3.Workdeveloped

1.4.Derive6andJava

2.Descriptionofthealgorithm

2.1.Theprocess

2.2.Rules

2.3.Example

2.4.Backtracking

3.Descriptionoftheenvironment

3.1.Menubar

3.2.Realtimemodifications

3.3.Inter-communicationwithDerive

3.4.Playingthecomposition

4.Results

4.1.Example1:

Abovevoiceagainst“Cantusfirmus”

4.2.Example2:

Belowvoiceagainst“Cantusfirmus”

4.3.Example3:

Aboveandbelowvoicesagainst“Cantusfirmus”

5.Conclusionsandfuturework

References

Purchase

$31.50

718

StudyofpharmaceuticalsamplesbyNIRchemical-imageandmultivariateanalysis  OriginalResearchArticle

TrACTrendsinAnalyticalChemistry,Volume27,Issue8,September2008,Pages696-713

JoséManuelAmigo,JordiCruz,ManelBautista,SantiagoMaspoch,JordiCoello,MarceloBlanco

 Closepreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

AbstractAbstract|Figures/TablesFigures/Tables|ReferencesReferences

Abstract

Near-infraredspectroscopychemicalimaging（NIR-CI）isapowerfultoolforprovidingagreatdealofinformationonpharmaceuticalsamples,sincetheNIRspectrumcanbemeasuredforeachpixeloftheimageoverawiderangeofwavelengths.

JoiningNIR-CIwithchemometricalgorithms（e.g.,PrincipalComponentAnalysis,PCA）andusingcorrelationcoefficients,clusteranalysis,classicalleast-squareregression（CLS）andmultivariatecurveresolution-alternatingleastsquares（MCR-ALS）areofincreasinginterest,duetothegreatamountofinformationthatcanbeextractedfromoneimage.Despitethis,investigationoftheirpotentialusefulnessmustbedonetoestablishtheirbenefitsandpotentiallimitations.

Weexploredthepossibilitiesofdifferentalgorithmsintheglobalstudy（qualitativeandquantitativeinformation）ofhomogeneityinpharmaceuticalsamplesthatmayconfirmdifferentstagesinablendingprocess.Forthispurpose,westudiedfourexamples,involvingfourbinarymixturesindifferentconcentrations.

Inthisway,westudiedthebenefitsandthedrawbacksofPCA,clusteranalysis（K-meansandFuzzyC-meansclustering）andcorrelationcoefficientsforqualitativepurposesandCLSandMCR-ALSforquantitativepurposes.

WepresentnewpossibilitiesinclusteranalysisandMCR-ALSinimageanalysis,andweintroduceandtestnewBACRAsoftwareformappingcorrelation-coefficientsurfaces.

ArticleOutline

1.Introduction

2.Structureofhyperspectraldata

3.Preprocessingthehyperspectralimage

4.Techniquesforexploratoryanalysis

4.1.PrincipalComponentAnalysis（PCA）

4.2.Clusteranalysis

4.2.1.K-meansalgorithm

4.2.2.FuzzyC-meansalgorithm

4.2.3.Numberofclusters

4.2.3.1.Silhouetteindex

4.2.3.2.PartitionEntropyindex

4.3.Similarityusingcorrelationcoefficients

5.Techniquesforestimatinganalyteconcentrationineachpixel

5.1.ClassicalLeastSquares

5.2.MultivariateCurveResolution-AlternatingLeastSquares

5.3.AugmentedMCR-ALSforhomogeneoussamples

6.Experimentalanddatatreatment

6.1.Reagentsandinstruments

6.2.Experimental

6.3.Datatreatment

7.Resultsanddiscussion

7.1.PCAanalysis

7.2.Clusteranalysis

7.2.1.K-meansresultsforheterogeneoussamples

7.2.2.FCMresults

7.3.Correlation-coefficientmaps–BACRAresults

7.4.CLSresults

7.5.MCR-ALSandaugmented-MCR-ALSresults

8.Conclusionsandperspectives

Acknowledgements

References

Purchase

$31.50

719

ValidationandautomatictestgenerationonUMLmodels:

theAGATHAapproach  OriginalResearchArticle

ElectronicNotesinTheoreticalComputerScience,Volume66,Issue2,December2002,Pages33-49

DavidLugato,CélineBigot,YannickValot

 Closepreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

AbstractAbstract|ReferencesReferences

Abstract

Therelatedeconomicgoalsoftestgenerationarequiteimportantforsoftwareindustry.Manufacturerseverseekingtoincreasetheirproductivityneedtoavoidmalfunctionsatthetimeofsystemspecification:

thelaterthedefaultsaredetected,thegreaterthecostis.Consequently,thedevelopmentoftechniquesandtoolsabletoefficientlysupportengineerswhoareinchargeofelaboratingthespecificationconstitutesamajorchallengewhosefalloutconcernsnotonlysectorsofcriticalapplicationsbutalsoallthosewherepoorconceptioncouldbeextremelyharmfultothebrandimageofaproduct.

ThisarticledescribesthedesignandimplementationofasetoftoolsallowingsoftwaredeveloperstovalidateUML（theUnifiedModelingLanguage）specifications.ThistoolsetbelongstotheAGATHAenvironment,whichisanautomatedtestgenerator,developedatCEA/LIST.

TheAGATHAtoolsetisdesignedtovalidatespecificationsofcommunicatingconcurrentunitsdescribedusinganEIOLTSformalism（ExtendedInputOutputLabeledTransitionSystem）.ThegoaloftheworkdescribedinthispaperistoprovideaninterfacebetweenUMLandanEIOLTSformalismgivingthepossibilitytouseAGATHAonUMLspecifications.

InthispaperwedescribefirstthetranslationofUMLmodelsintotheEIOLTSformalis