PAGE1PAGE1PAGE1中英文對照外文翻譯文獻(文檔含英文原文和中文翻譯)原文：ProjectRiskAnalysisChapter1Introduction1.1AboutthiscompendiumThiscoursecompendiumistobeusedinthecourse“Risikostyringisprojector”.Thefocuswillbeonthefollowingtopics:?Riskidentification?Riskstructuring?Riskmodelinginthelightofatimescheduleandacostmodel?RiskfollowsupWewillalsodiscusselementsrelatedtodecisionanalysiswhereriskisinvolved,anduseoflifecyclecostandlifecycleprofitmodels.Thecoursecompendiumcomprisesalargenumberofexercises,anditisrecommendedtodomostoftheexercisesinordertogetagoodunderstandingofthetopicsandmethodsdescribed.AseparateMSExcelprogram,pRisk.xlshasbeendevelopedinordertoassistnumericalcalculationsandtoconductMonteCarlosimulation.1.2DefinitionsAleatoryuncertaintyVariationofquantitiesinapopulation.Wesometimesusethewordvariabilityratherthanaleatoryuncertainty.EpistemicuncertaintyLackofknowledgeaboutthe“world”,andobservablequantitiesinparticular.DependencyTherelationbetweenthesequencesoftheactivitiesinaproject.ObservablequantityAquantityexpressingastateofthe“world”,i.e.aquantityofthephysicalrealityornature,thatisunknownatthetimeoftheanalysisbutwill,ifthesystembeinganalyzedisactuallyimplemented,takesomevalueinthefuture,andpossiblybecomeknown.ParameterWeusethetermparameterintwowaysinthisreport.Themainuseofaparameteristhatitisaquantitythatisapartoftheriskanalysismodels,andforwhichweassignnumericalvalues.Themoreacademicdefinitionofaparameterusedinaprobabilitystatementaboutanobservablequantity,X,isthataparameterisaconstructwherethevalueoftheparameteristhelimitingvaluewherewearenotabletosaturateourunderstandingabouttheobservablequantityXwhatsoevernewinformationwecouldgetholdof.ParameterestimateThenumericvalueweassesstoaparameter.ProbabilityAmeasureofuncertaintyofanevent.RiskRiskisdefinedastheanswertothethreequestions[14]:i)whatcangowrong?ii)Howlikelyisit?Andifitgoeswrong,iii)whataretheconsequences?TodescribetheriskisascenarioRiskacceptanceAdecisiontoacceptarisk.RiskacceptancecriterionAreferencebywhichriskisassessedtobeacceptableorunacceptable.ScheduleAplanwhichspecifiesthestartandfinalizationpointoftimesfortheactivitiesinaproject.StochasticdependencyTwoormorestochasticvariablesare(stochastically)dependentiftheexpectationofonestochasticvariabledependsonthevalueofoneormoreoftheotherstochasticvariables.StochasticvariableAstochasticvariable,orrandomquantity,isaquantityforwhichwedonotknowthevalueitwilltake.However,wecouldstatestatisticalpropertiesofthevariableormakeprobabilitystatementaboutthevalueofthequantity.1.3DEFINITIONSUncertaintyLackofknowledgeabouttheperformanceofasystem,andobservablequantitiesinparticular.Chapter2RiskManagementGenerally,riskmanagementisdefined(IEC60300-3-9)asa“systematicapplicationofmanagementpolicies,proceduresandpracticestothetasksofanalyzing,evaluatingandcontrollingrisk”.Itwillcomprise(IECdefinitionsinparentheses):?Riskassessment,i.e.–Riskanalysis(“Systematicuseofavailableinformationtoidentifyhazardsandtoestimatetherisktoindividualsorpopulations,propertyortheenvironment”)–Riskevaluation(“Processinwhichjudgmentsaremadeonthetolerabilityoftheriskonthebasisofriskanalysisandtakingintoaccountfactorssuchassocio-economicandenvironmentalaspects”)?Riskreduction/control(Decisionmaking,implementationandriskmonitoring).Thereexistsnocommondefinitionofrisk,butforinstanceIEC60300-3-9definesriskasa“combinationofthefrequency,orprobability,ofoccurrenceandtheconsequenceofaspecifiedhazardousevents”.Mostdefinitionscomprisetheelementsofprobabilitiesandconsequences.However,someasKlinkeandRennsuggestaverywidedefinition,stating:“Riskreferstothepossibilitythathumanactionsoreventsleadtoconsequencesthataffectaspectsofwhathumansvalue”.Sothetotalriskcomprisesthepossibilityofnumber(“all”)unwanted/hazardousevents.Itispartoftheriskanalysistodelimitwhichhazardstoinclude.Further,riskusuallyreferstothreatsinthefuture,involvinga(high)degreeofuncertainty.Inthefollowingwewillpresentthebasicelementsofriskmanagementasitisproposedtobeanintegralpartofprojectmanagement.2.1ProjectobjectivesandcriteriaInclassicalriskanalysisofindustrialsystemstheuseofso-calledriskacceptancecriteriahasplayedacentralroleinthelasttwoortreedecades.Basicallyuseofriskacceptancecriteriameansthatsomesevereconsequencesaredefined,e.g.accidentwithfatalities.Thenwetrytosetanupperlimitfortheprobabilityoftheseconsequencesthatcouldbeaccepted,i.e.wecouldnotaccepthigherprobabilitiesinanysituations.Furthertheseprobabilitiescouldonlybeacceptedifriskreductionisnotpossible,orthecostofriskreductionisveryhigh.Inrecentyearsithasbeenadiscussionintheriskanalysissocietywhetheritisfruitfulornottouseriskacceptancecriteriaaccordingtotheprinciplesabove.Itisarguedthatveryoftenriskacceptancecriteriaaresetarbitrary,andthesedonotnecessarilysupporttheoverallbestsolutions.Therefore,itcouldbemorefruitfultousesomekindofriskevaluationcriteria,ratherthanstrictacceptancecriteria.Inprojectriskmanagementwecouldestablishacceptancecriteriarelatedtotwotypesofevents:?Eventswithsevereconsequencesrelatedtohealth,environmentandsafety.?Eventswithsevereconsequencesrelatedtoprojectcosts,projectquality,projectduration,oreventerminationoftheproject.Inthiscoursewewillhavemainfocusontheprojectcostsandthedurationoftheproject.Notethatbothprojectcostandprojectdurationarestochasticvariablesandnotevents.Thusitisnotpossibletoestablishacceptancecriteriatoprojectcostordurationdirectly.Basically,therearethreetypesofnumericvalueswecouldintroduceinrelationtosuchstochasticvariablesdescribingtheproject:1.Target.Thetargetexpressesourambitionsintheproject.Thetargetshallbesomethingwearestrivingat,anditshouldbepossibletoreachthetarget.Itispossibletointroduce(internal)bonuses,orotherrewardsinordertoreachthetargetsinaproject.2.Expectation.Theexpectationsarethevaluethestochasticvariableswillachieveinthelongrun,orourexpectationabouttheoutcome.Theexpectationislessambitiousthanthetarget.Theexpectationwillinarealisticwayaccountforhazards,andthreatsandconditionswhichoftencontributetothefactthatthetargetsarenotmet.3.Commitment.Thecommitmentsarevaluesrelatedtothestochasticvariableswhichareregulatedinagreementsandcontracts.Forexampleitcouldbestatedinthecontractthatanewbridgeshallbecompletedwithinagivendate.Ifwearenotabletofulfillthecommitments,thiswillusuallyresultineconomicalconsequences,forexamplepenaltiesfordefaults,orintheworstcasecancelingofthecontract.2.2RiskidentificationAscenarioisadescriptionofaimaginedsequenceorchainofevents,e.g.wehaveawaterleakage,andwearenotabletostopthisleakagewithordinarytighteningmediumduetothepossibleenvironmentalaspectswhichisnotclarifiedatthemoment.Furtherthegreenmovementisalsolikelytoenterthesceneinthiscase.Ahazardistypicallyrelatedtoenergies,poisonousmediaetc,andiftheyarereleasedthiswillresultinanaccidentorasevereevent.Athreatisawidertermthanhazard,andweincludealsoaspectsas“wrong”methodapplied,“l(fā)ackofcompetenceandexperience”.Thetermthreatisalsoveryoftenusedinconnectionwithsecurityproblems,e.g.sabotage,terrorism,andvandalism.2.3StructuringandmodelingofriskInSection2.2wehaveidentifiedmethodstoidentifyeventsandthreats.Wenowwanttorelatetheseeventsandthreatstotheexplicitmodelswehaveforprojectcostsandprojectduration.2.3.1Modelforprojectexecutiontime/schedulemodelingWhenanalyzingtheexecutiontimeforaprojectwewillhaveaprojectplanandtypicallyaGanttdiagramasastartingpoint.TheGanttdiagramistransformedintoaso-calledflownetworkwheretheconnectionsbetweentheactivitiesareexplicitlydescribed.Suchaflownetworkalsocomprisesdescriptionofdurationoftheactivitiesintermsofprobabilitystatements.ThedurationofeachactivityisstochasticVariables,whichwedenoteTiforactivityinaflownetworkwemightalsohaveuncertainactivitieswhichwillbecarriedoutonlyunderspecialconditions.Theseconditionscouldbedescribedintermsofevents,andweneedtodescribetheprobabilityofoccurrenceofsuchevents.Thus,thereisasetofquantities,i.e.timevariablesandeventsinthemodel.TheobjectiveisnowtolinktheundesiredeventsandthreatsdiscussedinSection2.2tothesetimevariablesandevents.Timevariablesaredescribedbyaprobabilitydistributionfunction.Suchadistributionfunctioncomprisesparametersthatcharacterizethetimevariable.OftenaparametricprobabilitydistributionisdescribedbythethreequantitiesL(low),M(mostlikely)andHhigh.Ifanundesiredeventoccur,itislikelythatthevaluesofL,MandHwillbehigherthanincasethiseventdoesnotoccur.AwaytoincludetheresultfromtheriskidentificationprocessisthentoexpressthedifferentvaluesofL,MandHdependingonwhetherthecriticaleventoccursornot.Ifweinadditionareabletoassesstheprobabilityofoccurrenceofthecriticalevent,theknowledgeaboutthiscriticaleventhasbeencompletelyincludedintotheriskmodel.Basedonsuchanexplicitmodelingofthecriticalevent,wecouldalsoeasilyupdatethemodelincaseofnewinformationaboutthecriticaleventisobtained,forexamplenewinformationcouldbeavailableatalaterstageintheprocessandchangesoftheplancouldstillbepossibleinlightofthenewinformation.2.3.2CostmodelingThecostmodelisusuallybasedonthecostbreakdownstructure,andthecostelementswillagainbefunctionsoflaborcost,overtimecost,purchaseprice,hourcostofrentingequipment,materialcost,amountofmaterialetc.Theprobabilisticmodelingofcostisusuallyeasierthanformodelingprojectexecutiontime.Theprincipleisjusttoaddalotofcostterms,whereeachcosttermistheproductoftheunitpriceandthenumberofunits.Weintroducepriceandvolumeasstochasticvariablestodescribetheunitpriceandthenumberofunits.ThepriceandvolumevariablesshouldalsobelinkedtotheundesiredeventsandthreatswehaveidentifiedinSection2.2.Oftenitisnecessarytolinkthecostmodeltotheschedulemodel.Forexampleincaseofdelaysitmightbenecessarytoputmoreeffortintotheprojecttocatchupwiththeproblems,andtheseeffortscouldbeverycostly.Also,iftheprojectisdelayedwemayneedtopayextracosttosub-contractorsthathavetopostponetheirsupportintotheproject.2.3.3UncertaintyinscheduleandcostmodelingAsindicatedabovewewillestablishprobabilisticmodelstodescribethedurationandcostofaproject.Theresultofsuchaprobabilisticmodelingisthatwetreatthedurationandcostasstochasticvariables.Sincedurationandcostsarestochasticvariables,thismeansthatthereisuncertaintyregardingthevaluestheywilltakeintherealprojectweareevaluating.Sometimeswesplitthisuncertaintyintothreedifferentcategories,i)Aleatoryuncertainty(variabilityduetoe.g.weatherconditions,laborconflicts,breakdownofmachinesetc.),ii)parameterorepistemicuncertaintyduetolackofknowledgeabout“true”parametervalues,andiii)modeluncertaintyduetolackofdetailed,orwrongmodeling.Undersuchthinking,thealeatoryuncertaintycouldnotbereduced;itisbelievedtobetheresultofthevariabilityintheworldwhichwecannotcontrol.Uncertaintyintheparametersis,however,believedtobereduciblebycollectingmoreinformation.Alsouncertaintyinthemodelsisbelievedtobereduciblebymoredetailedmodeling,anddecompositionofthevariouselementsthatgointothemodel.Itisappealingtohaveamentalmodelwheretheuncertaintycouldbesplitintoonepartwhichwemightnotreduce(variability),andonepartwhichwemightreducebythoroughanalysisandmoreinvestigation(increasedknowledge).Ifweareabletodemonstratethatthepartoftheuncertaintyrelatedtolackofknowledgeandunderstandinghasbeenreducedtoasufficientdegree,wecouldthenclaimhighconfidenceintheanalysis.Insomesituationtheownerortheauthoritiesputforwardrequirements.Whichcouldbeinterpretedasconfidenceregardingthequalityoftheanalysis?Itisthoughnotalwaysclearwhatismeantbysuchaconfidencelevel.Asanexample,letE(C)betheexpectedcostofaproject.Aconfidencestatementcouldnowbeformulatedas“TheprobabilitythattheactualprojectcostiswithinanintervalE(C)±10%shouldatleastbe70%”.Itis,however,notstraightforwardtodocumentsuchaconfidencelevelinarealanalysis.The“Successiveprocess(trinnvisprosessen)”[4]isanattempttodemonstratehowtoreducethe“uncertainty”intheresulttoacertainlevelofconfidence.WealsomentionthatEven[12]hasrecentlyquestionedsuchanapproachwherethereexistmodeluncertaintyandparameteruncertainty,andemphasizesthatweintheanalysisshouldfocusontheobservablequantitieswhichwillbecomeevidentforusiftheprojectisexecuted,e.g.thecosts,andthatuncertaintyinthesequantitiesrepresentthelackofknowledgeaboutwhichvaluestheywilltakeinthefuture.Thisdiscussionisnotpursuitanymoreinthispresentation.2.4Riskelementsforfollowup:RiskandopportunityregisterAsriskelementsandthreatsareidentifiedinSection2.2thesehavetobecontrolledasfaraspossible.Itisnotsufficienttoidentifytheseconditionsandmodeltheminthescheduleandcostmodels,wealsohavetomitigatetheriskelementsandthreats.Inordertoensureasystematicfollowupofriskelementsandthreatsitisrecommendedtoestablishaso-calledthreatlog.Theterms’RiskRegister‘a(chǎn)nd’Risk&OpportunityRegister‘(R&OR)issometimesusedratherthantheterm’threatlog.‘AR&ORisbestmanagedbyadatabasesolution,forexampleanMS-AccessDatabase.Eachrowinthedatabaserepresentsoneriskelementorthreat.Thefieldsinsuchadatabasecouldvary,butthefollowingfieldsseemsreasonable:?ID.Anidentifierisrequiredinordertokeeptrackofthethreatinrelationtothequantitativeriskmodels,tofollowupactionsET.?Description.Adescriptionofthethreatisnecessaryinordertounderstandthecontentoftheproblem.Itcouldbenecessarytostatetheimmediateconsequences(e.g.occupationalaccident),butalsoconsequencesintermsofthemainobjectivesoftheproject,e.g.timeandcosts.?Likelihoodorprobability.Ajudgmentregardinghowprobableitisthatthethreatortheriskconditionwillbereleasedintermsofe.g.undesiredorcriticalevents.?Impact.Ifpossible,giveadirectimpactoncostandscheduleiftheeventoccurs,eitherbyanexpectedimpact,orbyL,MandHvalues.?Referencestocostandschedule.InordertoupdatethescheduleandcostmodelsitisconvenienttogiveanexplicitreferencefromtheR&ORintothescheduleandcostmodels.?Manageability.Hereitisdescriedhowthethreatcouldbeinfluenced,eitherbyimplementingmeasurestoeliminatethethreatpriortoitrevealsitself,ormeasuresinordertoreducetheconsequencesincaseofthethreatwillmaterialize.?Alertinformation.Itisimportanttobeawareofinformationthatcouldindicatethedevelopmentofthethreatbeforeiteventuallywillmaterialize.Ifsuchinformationisavailablewecouldimplementrelevantmeasuresifnecessary.Forexampleitcouldbepossibletotakegroundsamplesatacertaincost,bututilizingtheinformationfromsuchsamplescouldenableustochooseappropriatemethodsfortunnelpenetration.?Measures.Listofmeasuresthatcouldbeimplementedtoreducetherisk.?Deadlineandresponsible.Identificationofwhoisresponsibleforimplementingandfollowupofthemeasureorthreat,andanydeadlines.?Status.Bothwithrespecttothethreatandanymeasureitisvaluabletospecifythedevelopment,i.e.didthetreatrevealitselfintoundesiredeventswithunwantedconsequences,didthemeasureplayanypositiveeffectetc.2.5CorrectionandcontrolAstheprojectdevelopstheR&ORistheprimarycontroltoolforriskfollowup.Byfollowingthestatusofthevariousthreats,riskelementsandmeasureswecouldmonitortheriskintheproject.Thisinformationshouldofcoursebelinkedtothetimeandcostplans.Ifagiventhreatdoesnotrevealintermsofundesiredevents,thetimeandcostestimatescouldbeloweredandthisgaincouldbeutilizedinotherpartoftheproject,orinotherprojects.Intheoppositesituationitisnecessarytoincreasethetimeandcostestimates,andweneedtoconsidernewmeasures,andmaybespendsomeofthereservestocatchupincaseofanexpecteddelay.Duringthelifecycleofaprojectitwilloccurnewthreatsandriskelementswhichwedidnotforeseeintheinitialriskidentificationprocess.SuchthreatsmustcontinuouslybeenteredintotheR&OR,andmeasuresneedtobeconsidered.一、介紹（一）關(guān)于本綱要本課程綱要過程中研究的是“風險也是一種項目”。重點將是就以下議題：?風險識別?風險結(jié)構(gòu)?風險模型中光的時間表和成本模型?風險跟進我們也將討論相關(guān)的決策分析，風險涉及的元素，并使用生命周期成本和生命周期的盈利模式來介紹風險管理的具體內(nèi)容。（二）定義1.偶然的不確定性。我們有時會使用這個詞表示數(shù)量的變化。2.認知的不確定性，缺乏知識的“世界”，特別是觀察到的數(shù)量。3.依據(jù)，在一個項目中的活動的序列之間的關(guān)系。4.觀察到的數(shù)量，一定量的需要表達的狀態(tài)，即對物理現(xiàn)實的數(shù)量或性質(zhì)，是未知的時間分析，在未來，如果被分析的系統(tǒng)實施，可能成為眾所周知的。5.參數(shù)，我們使用的術(shù)語以兩種方式在本報告的參數(shù)。參數(shù)的主要用途是，它是一個量是風險分析模型的一部分并且是我們指定數(shù)值。更多學術(shù)有關(guān)可觀察到的數(shù)量。6.參數(shù)估計，我們評估的一個參數(shù)的數(shù)值。7.可能性，一種衡量事件的不確定性。8.風險，風險是指三個問題：I）什么可以去做了嗎？II）項目的可能性有多大呢？，III）如果它出了問題，后果是什么？9.風險接受，決定接受的風險。10.風險接受準則，風險被評估為可接受或不可接受的參考。11.時間表，該計劃規(guī)定的活動時間在一個項目的開始和完成點。12.隨機依賴，兩個或更多個隨機變量（隨機）依賴的期望的一個隨機變量，如果依賴于其他的一個或多個隨機變量的值。二、風險管理一般來說，風險管理（IEC60300-3-9）定義為“系統(tǒng)應(yīng)用程序的管理政策，程序和做法的分析，評估和控制風險”的任務(wù)。這將包括（IEC定義在括號中）：——風險分析，系統(tǒng)利用現(xiàn)有的信息，以識別危險，并估計個體或群體，財產(chǎn)或環(huán)境的風險。——風險評估，評估過程中，判斷風險的耐受性是在風險分析的基礎(chǔ)上，考慮的因素，如社會，經(jīng)濟和環(huán)境方面的?！獪p少風險/控制（決策，執(zhí)行和風險監(jiān)控）。存在的風險，但并無統(tǒng)一的定義，例如IEC60300-3-9風險定義為“相結(jié)合的頻率或概率，發(fā)生的后果某一特定危險事件。大多數(shù)定義包括概率和后果的元素。雷恩提出一個很廣泛的定義，他表示：“風險是指可能人的行動或事件導(dǎo)致的后果，影響人類重視的方面因此，總的風險包括（“全部”）不想要的/危險事件的可能性。這是界定哪些危害進行風險分析的一部分。此外，風險通常是指在未來的威脅，涉及的不確定性。在下面的建議是項目管理的一個組成部分，我們將介紹風險管理的基本要素。（一）項目的目標和標準在古典風險分析的工業(yè)系統(tǒng)使用所謂的風險接受準則發(fā)揮了核心的作用，在過去的兩年或幾十年。基本上使用的風險接受準則定義了一些嚴重的后果，例如死亡事故。然后，我們嘗試設(shè)置一個上限可以接受這些后果的概率，即較高的概率在任何情況下，我們不能接受。此外，這些可能性只能接受降低風險是不可能的，或降低風險的成本是非常高的。在最近幾年中，它一直是討論在風險分析的方面是富有成果的，或不根據(jù)上述原則，使用風險接受準則。有人認為，很多時候風險接受準則任意設(shè)定，而這些不一定是支持風險接受的最佳整體解決方案。因此，它可能是使用起來更加豐碩的成果。一些風險評估標準，而不是嚴格的驗收標準。在項目風險管理中，我們可以建立兩種類型的事件相關(guān)的驗收標準：?關(guān)系到健康，環(huán)境和安全的嚴重后果的事件。?事件相關(guān)的項目成本，項目質(zhì)量，工期，甚至終止項目的嚴重后果。在這個過程中，我們將主要集中在項目成本和項目的持續(xù)時間。需要注意的是這兩個項目的成本和項目時間是隨機變量，而不是事件。因此，它是無法確定驗收標準的項目直接成本或持續(xù)時間?；旧?，有三種類型的數(shù)值，我們可以引入在關(guān)系到項目的隨機變量來描述：1．目標，在該項目的目標應(yīng)是我們正在努力，應(yīng)該是可以達到的目標。引入獎金或其他獎勵以達到項目的目標是可能的。2．期望，期望是隨機變量，將實現(xiàn)從長遠來看，我們期望的結(jié)果。預(yù)期是做出這樣那樣雄心勃勃的目標。預(yù)期將是克服危害的現(xiàn)實途徑，威脅和經(jīng)常作出貢獻的事實，各項指標均滿足的條件。3．承諾，承諾有關(guān)的協(xié)議和合同監(jiān)管的隨機變量，這些變量的值。例如，它可以在合同中注明之內(nèi)完成一個給定的日期，一種新的聯(lián)系。如果我們不能夠履行的承諾，這通常會導(dǎo)致經(jīng)濟后果，例如違約處罰，或在最壞的情況下取消該合同。（二）風險識別不期望的事件是可能發(fā)生的，例如大的水泄漏在隧道里。情景是一個想象的事件的序列或鏈的描述，如我們有一個漏水的，我們是不能夠阻止這與普通緊縮可能對環(huán)境方面的不明確的時刻，由于介質(zhì)泄漏。此外，綠色運動在這種情況下，也有可能進入現(xiàn)場。通常是一個危險的能量，有毒介質(zhì)等，如果他們被釋放，這將導(dǎo)致事故或嚴重的事件。威脅是一個更廣泛的長期危險，我們還包括“錯誤”的方法應(yīng)用的方面，“缺乏能力和經(jīng)驗”。長期威脅也很常使用的安全問題，例如：破壞，恐怖主義和破壞行為。（三）構(gòu)建與風險建模在（二）節(jié)中，我們已經(jīng)確定的方法來識別事件和威脅?，F(xiàn)在我們要與這些事件和威脅，我們有明確的模式，項目成本和項目工期。1.模型項目執(zhí)行時間/日程建模當分析一個項目的執(zhí)行時間，我們將有一個典型的項目計劃和甘特圖作為出發(fā)點。被變換成一個所謂的流動之間的活動連接的網(wǎng)絡(luò)中明確描述的甘特圖。這樣的流網(wǎng)絡(luò)還包括持續(xù)時間在概率語