简介:Bargraphsareconvenientforshowingcomparisonsamongitems.Thebarsmaybeeitherhorizontalorvertical,andtheyareusedtoshowtheamountsofdifferentitems.Figure1isanexampleofatypicalbargraph.ThegraphshowstheaveragesalespriceofexistinghomesintheNortheastUnitedStatesforthreeyears——1970,1980,and1997.
简介:Basedonthethermaltomographytechnology,thisthesistriestocalculatequantityanddistributionofheatsourceinvivofrombodysurfacetemperature.AsuperpositiontemperatureimageofanumberofpointheatsourcesissurfacefittedtogettheQofheatsources(informationofcancercells)quantitatively.Theresultcanreflectthediseaseareainformationbecausecancercell'sQvalueismuchhigherthanthatofnormalcell.Thisapplicationisanewtryinthediagnosisofbreastcancer,whichhasanimportantvalueontheearlydetectionanddiagnosisofdiseasesource.
简介:一个人工的神经网络(ANN)被用来用鸣温度侧面的很多时间/军刀的手足为中间的空气建模。一个三层的前馈控制网络基于背繁殖(BP)被选算法。当温度是产量向量时,纬度,经度,和高度被选择为网络的输入向量。温度观察在从通过2007年3月16日,它在一样的卫星的1月13日的时期期间偏航,作为样品被拿训练ANN。结果建议网络为为温度的空间变化建模有高质量。在从流行实验NRLMSISE-00模型的ANN产量和那些之间的量的比较说明他们的通常一致的特征和一些特定的差别。NRLMSISE-00模型鈥檚带的吝啬的温度由锝太高?行星的波浪的stratopause附近的K-10K,和振幅和阶段第一号的活动在从在45鈥?上面的ANN模拟的一些方面是不同的0km,建议改进为更多的精确模拟在NRLMSISE-00模型被需要在附近并且在stratopause上面。关键词人工的神经网络-中间的空气-建模-背繁殖算法-NRLMSISE-00模型
简介:ThemaingoalofroutingsolutionsistosatisfytherequirementsoftheQualityofService(QoS)foreveryadmittedconnectionaswellastoachieveaglobalefficiencyinresourceutilization.InthispaperproposesasolutionbasedonHopfieldneuralnetwork(HNN)todealwithoneofrepresentativeroutingproblemsinuni-castrouting,i.e.themulti-constrained(MC)routingproblem.ComputersimulationshowsthatwecanobtaintheoptimalpathveryrapidlywithournewLyapunovenergyfunctions.
简介:Anewefficientcouplingrelationshipdescriptionmethodhasbeendevelopedtoprovideanautomatedandvisualizedwaytomultidisciplinarydesignoptimization(MDO)modelingandsolving.Thedisciplinaryrelationmatrix(DRM)isproposedtodescribethecouplingrelationshipaccordingtodisciplinaryinput/outputvariables,andtheMDOdefinitionhasbeenreformulatedtoadoptthenewinterfaces.Basedonthese,auniversalMDOsolvingprocedureisproposedtoestablishanautomatedandefficientwayforMDOmodelingandsolving.Throughasimpleandconvenientinitialconfiguration,MDOproblemscanbesolvedusinganyofavailableMDOarchitectureswithnofurthereffort.SeveralexamplesareusedtoverifytheproposedMDOmodelingandsolvingprocess.ResultshowsthattheDRMmethodhastheabilitytosimplifyandautomatetheMDOprocedure,andtherelatedMDOframeworkcanevaluatetheMDOproblemautomaticallyandefficiently.
简介:
简介:Usingagentdevelopmenttoolstoconstructanagent-basedsystemisawellappliedapproach.However,thedevelopmenttoolsusuallydonothavethefunctiontocheckthefeasibilityabouttheworkflowoftheagentsystemduringitimplementationstage.Therefore,todevelopanevaluationapproachtoanalyzethefeasibilityofadevelopingagentsystemsuchthattheimproperworkflowofanagentsystemcanbefoundintheearlydesignstageisanecessarytasktoreducetheriskofimplementation.Inthisresearch,aPetriNet(PN)basedthree-stageevaluationapproachwasdeveloped.Intheconceptualstage,thepitfallofthecurrentagentsystemdevelopingprocesswasexaminedandanimprovementanalysisprocesswasspecified.Then,inthesystemdesignstage,anevaluationapproachwhichextractedtheprocesslogfilefromadevelopingagentsystemintoaPNmodelintermsofaprocessminingapproach-αalgorithmwasproposed.ThismodelwassimulatedinaPNsimulationpackage.TheagentsystemperformancewasevaluatedintermsofanalyzingthedeadlockphenomenaofthePNmodel.Finally,intheimplementationstage,theproposedconceptwasimplementedbyusinganagentdevelopingtoolJADEandaPNsimulationtoolCPN.Anagent-basedroboticassemblysystemwasusedtoexaminethepossibledeadlockoftheagentsystem.