简介:为了研究有限理性假设下出行者的自适应调整行为对交通网络分流的影响,利用累积前景理论结合演化元胞自动机建立了具有个体交互机制的多主体路径选择模型.在模型中将出行者划分为风险追求者与风险厌恶者,基于出行时间可靠性并借鉴元胞遗传算法的思想设计了具有异质特点的出行者动态参照点及其演化规则,使出行者个体能够依据决策环境的变化动态地调整自身的出行时间预算,更加符合出行者的实际行为特征.最后将多主体参照点演化规则与传统的相继平均算法相结合,求解路网配流.研究发现:演化模型较好地继承了传统模型中的路径分流特点;不同的出行者类型比例及出行者的信息接收程度是影响路网分流结构的重要因素.
简介:Althoughthegreyforecastingmodelhasbeensuccessfullyadoptedinvariousfieldsanddemonstratedpromisingresults,theliteraturesshowitsperformancecouldbefurtherimproved,suchasfortheDGM(1,1)model,basedonaconcavesequence,themodelingerrorwillbelarger.Inthispaper,firstlythedefinitionofsequenceconvexityisgivenout,anditisprovedthattheoutputsequenceofDGM(1,1)modelisaconvexsequence.Next,theresidualchangelawofDGM(1,1)modelbasedontheconcavesequenceisdiscussed,andthenon-equidistanceDGM(1,1)modelisproposed.Finally,byintroducingthesymmetrytransformation,aconcavesequenceistransformedintoaconvexsequence,calledthesymmetricsequenceoftheconcavesequence,andthenconstructthenon-equidistanceDGM(1,1)modelbasedontheconvexsequence.Theexampleresultsshowthatthenovelmethodismoreaccuratethanthedirectmodelingforaconcavesequence.
简介:Inthispaper,westudytheexistenceof0-1universalminimaltotaldominatingfunctionsinagraph.Weestablishaformulationoflinearinequalitiestocharacterizeuniversalminimaltotaldominatingfunctionsandshowthatforakindofgraphswhoseadjacentmatricesarebalanced,theexistenceofuniversalminimaltotaldominatingfunctionscoincideswiththatof0-1ones.Itisalsoprovedthatforgeneralgraphs,theproblemoftestingtheexistenceof0-1universalminimaltotaldominatingfunctionsisNP-hard.
简介:Inthispaper,weanalyzeabulkinputM[X]/M/1queuewithmultipleworkingvacations.Aquasiuppertriangletransitionprobabilitymatrixoftwo-dimensionalMarkovchaininthismodelisobtained,andwiththematrixanalysismethod,highlycomplicatedprobabilitygeneratingfunction(PGF)ofthestationaryqueuelengthisfirstlyderived,fromwhichwegotthestochasticdecompositionresultforthestationaryqueuelengthwhichindicatestheevidentrelationshipwiththatoftheclassicalM[X]/M/1queuewithoutvacation.ItisimportantthatwefindtheupperandthelowerboundsofthestationarywaitingtimeintheLaplacetransformorderusingthepropertiesoftheconditionalErlangdistribution.Furthermore,wegainthemeanqueuelengthandtheupperandthelowerboundsofthemeanwaitingtime.
简介:IntheexistingliteratureofRepairableQueueingSystems(RQS),i.e.,queueingsystemswithserverbreakdowns,itisalmostallassumedthatinterarrivaltimesofsuccessivecustomersareindependent,identicallyexponentiallydistributed.Inthispaper,wedealwithmoregenericsystemGI/PH/1withserver'sexponentialuptimeandphase-typerepairtime.Withmatrixanalysistheory,weestablishtheequilibriumconditionandthecharacteristicsofthesystem,derivethetransientandstationaryavailabilitybehaviorofthesystem.
简介:根据项目组织环境特点,构建了以PMO为知识中介的有限理性知识转移多群体动态演化博弈模型,通过计算分析得到演化稳定策略和演化稳定相图。研究表明,知识溢出效应、知识协同效应、知识转移成本以及知识转移的奖惩措施,是知识转移行为主要影响因素,并决定项目型组织知识转移的整体演化趋势。