9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

MonteCarlomethod
FromWikipedia,thefreeencyclopedia

NottobeconfusedwithMonteCarloalgorithm.
MonteCarlomethods(orMonteCarloexperiments)areabroadclassofcomputationalalgorithmsthat
relyonrepeatedrandomsamplingtoobtainnumericalresults.Theyareoftenusedinphysicaland
mathematicalproblemsandaremostusefulwhenitisdifficultorimpossibletouseothermathematical
methods.MonteCarlomethodsaremainlyusedinthreedistinctproblemclasses:[1]optimization,numerical
integration,andgeneratingdrawsfromaprobabilitydistribution.
Inphysicsrelatedproblems,MonteCarlomethodsarequiteusefulforsimulatingsystemswithmany
coupleddegreesoffreedom,suchasfluids,disorderedmaterials,stronglycoupledsolids,andcellular
structures(seecellularPottsmodel,interactingparticlesystems,McKeanVlasovprocesses,kineticmodels
ofgases).Otherexamplesincludemodelingphenomenawithsignificantuncertaintyininputssuchasthe
calculationofriskinbusinessand,inmath,evaluationofmultidimensionaldefiniteintegralswith
complicatedboundaryconditions.Inapplicationtospaceandoilexplorationproblems,MonteCarlobased
predictionsoffailure,costoverrunsandscheduleoverrunsareroutinelybetterthanhumanintuitionor
alternative"soft"methods.[2]
Inprinciple,MonteCarlomethodscanbeusedtosolveanyproblemhavingaprobabilisticinterpretation.
Bythelawoflargenumbers,integralsdescribedbytheexpectedvalueofsomerandomvariablecanbe
approximatedbytakingtheempiricalmean(a.k.a.thesamplemean)ofindependentsamplesofthe
variable.Whentheprobabilitydistributionofthevariableistoocomplex,weoftenuseaMarkovChain
MonteCarlo(MCMC)sampler.ThecentralideaistodesignajudiciousMarkovchainmodelwitha
prescribedstationaryprobabilitydistribution.Bytheergodictheorem,thestationaryprobabilitydistribution
isapproximatedbytheempiricalmeasuresoftherandomstatesoftheMCMCsampler.
Inotherimportantproblemsweareinterestedingeneratingdrawsfromasequenceofprobability
distributionssatisfyinganonlinearevolutionequation.Theseflowsofprobabilitydistributionscanalways
beinterpretedasthedistributionsoftherandomstatesofaMarkovprocesswhosetransitionprobabilities
dependsonthedistributionsofthecurrentrandomstates(seeMcKeanVlasovprocesses,nonlinearfiltering
equation).[3][4]Inotherinstanceswearegivenaflowofprobabilitydistributionswithanincreasinglevelof
samplingcomplexity(pathspacesmodelswithanincreasingtimehorizon,BoltzmannBibbsmeasures
associatedwithdecreasingtemperatureparameters,andmanyothers).Thesemodelscanalsobeseenasthe
evolutionofthelawoftherandomstatesofanonlinearMarkovchain.[4][5]Anaturalwaytosimulatethese
sophisticatednonlinearMarkovprocessesistosamplealargenumberofcopiesoftheprocess,replacingin
theevolutionequationtheunknowndistributionsoftherandomstatesbythesampledempiricalmeasures.
IncontrastwithtraditionalMonteCarloandMarkovchainMonteCarlomethodologiesthesemeanfield
particletechniquesrelyonsequentialinteractingsamples.Theterminologymeanfieldreflectsthefactthat
eachofthesamples(a.k.a.particles,individuals,walkers,agents,creatures,orphenotypes)interactswith
theempiricalmeasuresoftheprocess.Whenthesizeofthesystemtendstoinfinity,theserandomempirical
measuresconvergetothedeterministicdistributionoftherandomstatesofthenonlinearMarkovchain,so
thatthestatisticalinteractionbetweenparticlesvanishes.

Contents
1Introduction

https://en.wikipedia.org/wiki/Monte_Carlo_method

1/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

1Introduction
2History
3Definitions
3.1MonteCarloandrandomnumbers
3.2MonteCarlosimulationversus"whatif"scenarios
4Applications
4.1Physicalsciences
4.2Engineering
4.3Computationalbiology
4.4Computergraphics
4.5Appliedstatistics
4.6Artificialintelligenceforgames
4.7Designandvisuals
4.8Financeandbusiness
5Useinmathematics
5.1Integration
5.2Simulationandoptimization
5.3Inverseproblems
5.3.1Petroleumreservoirmanagement
6Inpopularculture
7Seealso
8Notes
9References

Introduction
MonteCarlomethodsvary,buttendtofollowaparticularpattern:
1. Defineadomainofpossibleinputs.
2. Generateinputsrandomlyfromaprobabilitydistributionover
thedomain.
3. Performadeterministiccomputationontheinputs.
4. Aggregatetheresults.
Forexample,consideracircleinscribedinaunitsquare.Giventhat
thecircleandthesquarehavearatioofareasthatis/4,thevalueof
canbeapproximatedusingaMonteCarlomethod:[6]
1. Drawasquareontheground,theninscribeacirclewithinit.
2. Uniformlyscattersomeobjectsofuniformsize(grainsofrice
orsand)overthesquare.
3. Countthenumberofobjectsinsidethecircleandthetotal
numberofobjects.
4. Theratioofthetwocountsisanestimateoftheratioofthe
twoareas,whichis/4.Multiplytheresultby4toestimate.

https://en.wikipedia.org/wiki/Monte_Carlo_method

MonteCarlomethodappliedto
approximatingthevalueof .After
placing30000randompoints,the
estimatefor iswithin0.07%ofthe
actualvalue.Thishappenswithan
approximateprobabilityof20%.

2/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

Inthisprocedurethedomainofinputsisthesquarethatcircumscribesourcircle.Wegeneraterandom
inputsbyscatteringgrainsoverthesquarethenperformacomputationoneachinput(testwhetheritfalls
withinthecircle).Finally,weaggregatetheresultstoobtainourfinalresult,theapproximationof.
Therearetwoimportantpointstoconsiderhere:Firstly,ifthegrainsarenotuniformlydistributed,thenour
approximationwillbepoor.Secondly,thereshouldbealargenumberofinputs.Theapproximationis
generallypoorifonlyafewgrainsarerandomlydroppedintothewholesquare.Onaverage,the
approximationimprovesasmoregrainsaredropped.
UsesofMonteCarlomethodsrequirelargeamountsofrandomnumbers,anditwastheirusethatspurred
thedevelopmentofpseudorandomnumbergenerators,whichwerefarquickertousethanthetablesof
randomnumbersthathadbeenpreviouslyusedforstatisticalsampling.

History
BeforetheMonteCarlomethodwasdeveloped,simulationstestedapreviouslyunderstooddeterministic
problemandstatisticalsamplingwasusedtoestimateuncertaintiesinthesimulations.MonteCarlo
simulationsinvertthisapproach,solvingdeterministicproblemsusingaprobabilisticanalog(seeSimulated
annealing).
AnearlyvariantoftheMonteCarlomethodcanbeseenintheBuffon'sneedleexperiment,inwhichcan
beestimatedbydroppingneedlesonafloormadeofparallelandequidistantstrips.Inthe1930s,Enrico
FermifirstexperimentedwiththeMonteCarlomethodwhilestudyingneutrondiffusion,butdidnot
publishanythingonit.[7]
ThemodernversionoftheMarkovChainMonteCarlomethodwasinventedinthelate1940sbyStanislaw
Ulam,whilehewasworkingonnuclearweaponsprojectsattheLosAlamosNationalLaboratory.
ImmediatelyafterUlam'sbreakthrough,JohnvonNeumannunderstooditsimportanceandprogrammedthe
ENIACcomputertocarryoutMonteCarlocalculations.In1946,physicistsatLosAlamosScientific
Laboratorywereinvestigatingradiationshieldingandthedistancethatneutronswouldlikelytravelthrough
variousmaterials.Despitehavingmostofthenecessarydata,suchastheaveragedistanceaneutronwould
travelinasubstancebeforeitcollidedwithanatomicnucleus,andhowmuchenergytheneutronwaslikely
togiveofffollowingacollision,theLosAlamosphysicistswereunabletosolvetheproblemusing
conventional,deterministicmathematicalmethods.StanislawUlamhadtheideaofusingrandom
experiments.Herecountshisinspirationasfollows:
ThefirstthoughtsandattemptsImadetopractice[theMonteCarloMethod]weresuggested
byaquestionwhichoccurredtomein1946asIwasconvalescingfromanillnessandplaying
solitaires.ThequestionwaswhatarethechancesthataCanfieldsolitairelaidoutwith52cards
willcomeoutsuccessfully?Afterspendingalotoftimetryingtoestimatethembypure
combinatorialcalculations,Iwonderedwhetheramorepracticalmethodthan"abstract
thinking"mightnotbetolayitoutsayonehundredtimesandsimplyobserveandcountthe
numberofsuccessfulplays.Thiswasalreadypossibletoenvisagewiththebeginningofthe
neweraoffastcomputers,andIimmediatelythoughtofproblemsofneutrondiffusionand
otherquestionsofmathematicalphysics,andmoregenerallyhowtochangeprocesses
describedbycertaindifferentialequationsintoanequivalentforminterpretableasasuccession
ofrandomoperations.Later[in1946],IdescribedtheideatoJohnvonNeumann,andwe
begantoplanactualcalculations.
https://en.wikipedia.org/wiki/Monte_Carlo_method

3/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

StanislawUlam[8]
Beingsecret,theworkofvonNeumannandUlamrequiredacodename.AcolleagueofvonNeumannand
Ulam,NicholasMetropolis,suggestedusingthenameMonteCarlo,whichreferstotheMonteCarlo
CasinoinMonacowhereUlam'sunclewouldborrowmoneyfromrelativestogamble.[7]Usinglistsof
"trulyrandom"randomnumberswasextremelyslow,butvonNeumanndevelopedawaytocalculate
pseudorandomnumbers,usingthemiddlesquaremethod.Thoughthismethodhasbeencriticizedascrude,
vonNeumannwasawareofthis:hejustifieditasbeingfasterthananyothermethodathisdisposal,and
alsonotedthatwhenitwentawryitdidsoobviously,unlikemethodsthatcouldbesubtlyincorrect.
MonteCarlomethodswerecentraltothesimulationsrequiredfortheManhattanProject,thoughseverely
limitedbythecomputationaltoolsatthetime.Inthe1950stheywereusedatLosAlamosforearlywork
relatingtothedevelopmentofthehydrogenbomb,andbecamepopularizedinthefieldsofphysics,
physicalchemistry,andoperationsresearch.TheRandCorporationandtheU.S.AirForceweretwoofthe
majororganizationsresponsibleforfundinganddisseminatinginformationonMonteCarlomethodsduring
thistime,andtheybegantofindawideapplicationinmanydifferentfields.
ThetheoryofmoresophisticatedmeanfieldtypeparticleMonteCarlomethodshadcertainlystartedbythe
mid1960s,withtheworkofHenryP.McKeanJr.onMarkovinterpretationsofaclassofnonlinear
parabolicpartialdifferentialequationsarisinginfluidmechanics.[9][10]Wealsoquoteanearlierpioneering
articlebyTheodoreE.HarrisandHermanKahn,publishedin1951,usingmeanfieldgenetictypeMonte
Carlomethodsforestimatingparticletransmissionenergies.[11]MeanfieldgenetictypeMonteCarlo
methodologiesarealsousedasheuristicnaturalsearchalgorithms(a.k.a.Metaheuristic)inevolutionary
computing.Theoriginsofthesemeanfieldcomputationaltechniquescanbetracedto1950and1954with
theworkofAlanTuringongenetictypemutationselectionlearningmachines[12]andthearticlesbyNils
AallBarricelliattheInstituteforAdvancedStudyinPrinceton,NewJersey.[13][14]
QuantumMonteCarlo,andmorespecificallyDiffusionMonteCarlomethodscanalsobeinterpretedasa
meanfieldparticleMonteCarloapproximationofFeynmanKacpathintegrals.[15][16][17][18][19][20][21]The
originsofQuantumMonteCarlomethodsareoftenattributedtoEnricoFermiandRobertRichtmyerwho
developedin1948ameanfieldparticleinterpretationofneutronchainreactions,[22]butthefirstheuristic
likeandgenetictypeparticlealgorithm(a.k.a.ResampledorReconfigurationMonteCarlomethods)for
estimatinggroundstateenergiesofquantumsystems(inreducedmatrixmodels)isduetoJackH.
Hetheringtonin1984[21]Inmolecularchemistry,theuseofgeneticheuristiclikeparticlemethodologies
(a.k.a.pruningandenrichmentstrategies)canbetracedbackto1955withtheseminalworkofMarshall.N.
RosenbluthandArianna.W.Rosenbluth.[23]
TheuseofSequentialMonteCarloinadvancedSignalprocessingandBayesianinferenceismorerecent.It
wasin1993,thatGordonetal.,publishedintheirseminalwork[24]thefirstapplicationofaMonteCarlo
resamplingalgorithminBayesianstatisticalinference.Theauthorsnamedtheiralgorithm'thebootstrap
filter',anddemonstratedthatcomparedtootherfilteringmethods,theirbootstrapalgorithmdoesnotrequire
anyassumptionaboutthatstatespaceorthenoiseofthesystem.Wealsoquoteanotherpioneeringarticle
inthisfieldofGenshiroKitagawaonarelated"MonteCarlofilter",[25]andtheonesbyPierreDelMoral[26]
andHimilconCarvalho,PierreDelMoral,AndrMoninandGrardSalut[27]onparticlefilterspublishedin
themid1990s.Particlefilterswerealsodevelopedinsignalprocessingintheearly19891992byP.Del
Moral,J.C.Noyer,G.Rigal,andG.SalutintheLAASCNRSinaseriesofrestrictedandclassified
https://en.wikipedia.org/wiki/Monte_Carlo_method

4/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

researchreportswithSTCAN(ServiceTechniquedesConstructionsetArmesNavales),theITcompany
DIGILOG,andtheLAASCNRS(https://www.laas.fr/public/en)(theLaboratoryforAnalysisand
ArchitectureofSystems)onRADAR/SONARandGPSsignalprocessingproblems.[28][29][30][31][32][33]
TheseSequentialMonteCarlomethodologiescanbeinterpretedasanacceptancerejectionsampler
equippedwithaninteractingrecyclingmechanism.
From1950to1996,allthepublicationsonSequentialMonteCarlomethodologiesincludingthepruning
andresampleMonteCarlomethodsintroducedincomputationalphysicsandmolecularchemistry,present
naturalandheuristiclikealgorithmsappliedtodifferentsituationswithoutasingleproofoftheir
consistency,noradiscussiononthebiasoftheestimatesandongenealogicalandancestraltreebased
algorithms.Themathematicalfoundationsandthefirstrigorousanalysisoftheseparticlealgorithmsare
duetoPierreDelMoral[26][34]in1996.Branchingtypeparticlemethodologieswithvaryingpopulation
sizeswerealsodevelopedintheendofthe1990sbyDanCrisan,JessicaGainesandTerryLyons,[35][36][37]
andbyDanCrisan,PierreDelMoralandTerryLyons.[38]Furtherdevelopmentsinthisfieldwere
developedin2000byP.DelMoral,A.GuionnetandL.Miclo.[16][39][40]

Definitions
ThereisnoconsensusonhowMonteCarloshouldbedefined.Forexample,Ripley[41]definesmost
probabilisticmodelingasstochasticsimulation,withMonteCarlobeingreservedforMonteCarlo
integrationandMonteCarlostatisticaltests.Sawilowsky[42]distinguishesbetweenasimulation,aMonte
Carlomethod,andaMonteCarlosimulation:asimulationisafictitiousrepresentationofreality,aMonte
Carlomethodisatechniquethatcanbeusedtosolveamathematicalorstatisticalproblem,andaMonte
Carlosimulationusesrepeatedsamplingtodeterminethepropertiesofsomephenomenon(orbehavior).
Examples:
Simulation:Drawingonepseudorandomuniformvariablefromtheinterval[0,1]canbeusedto
simulatethetossingofacoin:Ifthevalueislessthanorequalto0.50designatetheoutcomeas
heads,butifthevalueisgreaterthan0.50designatetheoutcomeastails.Thisisasimulation,butnot
aMonteCarlosimulation.
MonteCarlomethod:Pouringoutaboxofcoinsonatable,andthencomputingtheratioofcoinsthat
landheadsversustailsisaMonteCarlomethodofdeterminingthebehaviorofrepeatedcointosses,
butitisnotasimulation.
MonteCarlosimulation:Drawingalargenumberofpseudorandomuniformvariablesfromthe
interval[0,1],andassigningvalueslessthanorequalto0.50asheadsandgreaterthan0.50astails,is
aMonteCarlosimulationofthebehaviorofrepeatedlytossingacoin.
KalosandWhitlock[6]pointoutthatsuchdistinctionsarenotalwayseasytomaintain.Forexample,the
emissionofradiationfromatomsisanaturalstochasticprocess.Itcanbesimulateddirectly,oritsaverage
behaviorcanbedescribedbystochasticequationsthatcanthemselvesbesolvedusingMonteCarlo
methods."Indeed,thesamecomputercodecanbeviewedsimultaneouslyasa'naturalsimulation'orasa
solutionoftheequationsbynaturalsampling."

MonteCarloandrandomnumbers

https://en.wikipedia.org/wiki/Monte_Carlo_method

5/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

MonteCarlosimulationmethodsdonotalwaysrequiretrulyrandomnumberstobeusefulwhilefor
someapplications,suchasprimalitytesting,unpredictabilityisvital.[43]Manyofthemostusefultechniques
usedeterministic,pseudorandomsequences,makingiteasytotestandrerunsimulations.Theonlyquality
usuallynecessarytomakegoodsimulationsisforthepseudorandomsequencetoappear"randomenough"
inacertainsense.
Whatthismeansdependsontheapplication,buttypicallytheyshouldpassaseriesofstatisticaltests.
Testingthatthenumbersareuniformlydistributedorfollowanotherdesireddistributionwhenalarge
enoughnumberofelementsofthesequenceareconsideredisoneofthesimplest,andmostcommonones.
Weakcorrelationsbetweensuccessivesamplesisalsooftendesirable/necessary.
SawilowskyliststhecharacteristicsofahighqualityMonteCarlosimulation:[42]
the(pseudorandom)numbergeneratorhascertaincharacteristics(e.g.,along"period"beforethe
sequencerepeats)
the(pseudorandom)numbergeneratorproducesvaluesthatpasstestsforrandomness
thereareenoughsamplestoensureaccurateresults
thepropersamplingtechniqueisused
thealgorithmusedisvalidforwhatisbeingmodeled
itsimulatesthephenomenoninquestion.
Pseudorandomnumbersamplingalgorithmsareusedtotransformuniformlydistributedpseudorandom
numbersintonumbersthataredistributedaccordingtoagivenprobabilitydistribution.
Lowdiscrepancysequencesareoftenusedinsteadofrandomsamplingfromaspaceastheyensureeven
coverageandnormallyhaveafasterorderofconvergencethanMonteCarlosimulationsusingrandomor
pseudorandomsequences.MethodsbasedontheirusearecalledquasiMonteCarlomethods.

MonteCarlosimulationversus"whatif"scenarios
TherearewaysofusingprobabilitiesthataredefinitelynotMonteCarlosimulationsforexample,
deterministicmodelingusingsinglepointestimates.Eachuncertainvariablewithinamodelisassigneda
bestguessestimate.Scenarios(suchasbest,worst,ormostlikelycase)foreachinputvariablearechosen
andtheresultsrecorded.[44]
Bycontrast,MonteCarlosimulationssamplefromaprobabilitydistributionforeachvariabletoproduce
hundredsorthousandsofpossibleoutcomes.Theresultsareanalyzedtogetprobabilitiesofdifferent
outcomesoccurring.[45]Forexample,acomparisonofaspreadsheetcostconstructionmodelrunusing
traditionalwhatifscenarios,andthenrunningthecomparisonagainwithMonteCarlosimulationand
triangularprobabilitydistributionsshowsthattheMonteCarloanalysishasanarrowerrangethanthewhat
ifanalysis.Thisisbecausethewhatifanalysisgivesequalweighttoallscenarios(seequantifying
uncertaintyincorporatefinance),whiletheMonteCarlomethodhardlysamplesintheverylowprobability
regions.Thesamplesinsuchregionsarecalled"rareevents".

Applications
MonteCarlomethodsareespeciallyusefulforsimulatingphenomenawithsignificantuncertaintyininputs
andsystemswithalargenumberofcoupleddegreesoffreedom.Areasofapplicationinclude:
https://en.wikipedia.org/wiki/Monte_Carlo_method

6/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

Physicalsciences
Seealso:MonteCarlomethodinstatisticalphysics
MonteCarlomethodsareveryimportantincomputationalphysics,physicalchemistry,andrelatedapplied
fields,andhavediverseapplicationsfromcomplicatedquantumchromodynamicscalculationstodesigning
heatshieldsandaerodynamicformsaswellasinmodelingradiationtransportforradiationdosimetry
calculations.[46][47][48]InstatisticalphysicsMonteCarlomolecularmodelingisanalternativeto
computationalmoleculardynamics,andMonteCarlomethodsareusedtocomputestatisticalfieldtheories
ofsimpleparticleandpolymersystems.[23][49]QuantumMonteCarlomethodssolvethemanybody
problemforquantumsystems.[3][4][15]Inexperimentalparticlephysics,MonteCarlomethodsareusedfor
designingdetectors,understandingtheirbehaviorandcomparingexperimentaldatatotheory.In
astrophysics,theyareusedinsuchdiversemannersastomodelbothgalaxyevolution[50]andmicrowave
radiationtransmissionthrougharoughplanetarysurface.[51]MonteCarlomethodsarealsousedinthe
ensemblemodelsthatformthebasisofmodernweatherforecasting.

Engineering
MonteCarlomethodsarewidelyusedinengineeringforsensitivityanalysisandquantitativeprobabilistic
analysisinprocessdesign.Theneedarisesfromtheinteractive,colinearandnonlinearbehavioroftypical
processsimulations.Forexample,
Inmicroelectronicsengineering,MonteCarlomethodsareappliedtoanalyzecorrelatedand
uncorrelatedvariationsinanaloganddigitalintegratedcircuits.
Ingeostatisticsandgeometallurgy,MonteCarlomethodsunderpinthedesignofmineralprocessing
flowsheetsandcontributetoquantitativeriskanalysis.
Inwindenergyyieldanalysis,thepredictedenergyoutputofawindfarmduringitslifetimeis
calculatedgivingdifferentlevelsofuncertainty(P90,P50,etc.)
impactsofpollutionaresimulated[52]anddieselcomparedwithpetrol.[53]
InFluidDynamics,inparticularRarefiedGasDynamics,wheretheBoltzmannequationissolvedfor
finiteKnudsennumberfluidflowsusingtheDirectSimulationMonteCarlo[54]methodin
combinationwithhighlyefficientcomputationalalgorithms.[55]
Inautonomousrobotics,MonteCarlolocalizationcandeterminethepositionofarobot.Itisoften
appliedtostochasticfilterssuchastheKalmanfilterorParticlefilterthatformstheheartofthe
SLAM(SimultaneousLocalizationandMapping)algorithm.
Intelecommunications,whenplanningawirelessnetwork,designmustbeprovedtoworkforawide
varietyofscenariosthatdependmainlyonthenumberofusers,theirlocationsandtheservicesthey
wanttouse.MonteCarlomethodsaretypicallyusedtogeneratetheseusersandtheirstates.The
networkperformanceisthenevaluatedand,ifresultsarenotsatisfactory,thenetworkdesigngoes
throughanoptimizationprocess.
Inreliabilityengineering,onecanuseMonteCarlosimulationtogeneratemeantimebetween
failuresandmeantimetorepairforcomponents.
InsignalprocessingandBayesianinference,particlefiltersandsequentialMonteCarlotechniques
areaclassofmeanfieldparticlemethodsforsamplingandcomputingtheposteriordistributionofa
signalprocessgivensomenoisyandpartialobservationsusinginteractingempiricalmeasures.

Computationalbiology
https://en.wikipedia.org/wiki/Monte_Carlo_method

7/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

MonteCarlomethodsareusedinvariousfieldsofcomputationalbiology,forexampleforBayesian
inferenceinphylogeny,orforstudyingbiologicalsystemssuchasgenomes,proteins,[56]ormembranes.[57]
Thesystemscanbestudiedinthecoarsegrainedorabinitioframeworksdependingonthedesired
accuracy.Computersimulationsallowustomonitorthelocalenvironmentofaparticularmoleculetoseeif
somechemicalreactionishappeningforinstance.Incaseswhereitisnotfeasibletoconductaphysical
experiment,thoughtexperimentscanbeconducted(forinstance:breakingbonds,introducingimpuritiesat
specificsites,changingthelocal/globalstructure,orintroducingexternalfields).

Computergraphics
PathTracing,occasionallyreferredtoasMonteCarloRayTracing,rendersa3Dscenebyrandomlytracing
samplesofpossiblelightpaths.Repeatedsamplingofanygivenpixelwilleventuallycausetheaverageof
thesamplestoconvergeonthecorrectsolutionoftherenderingequation,makingitoneofthemost
physicallyaccurate3Dgraphicsrenderingmethodsinexistence.

Appliedstatistics
Inappliedstatistics,MonteCarlomethodsaregenerallyusedfortwopurposes:
1. Tocomparecompetingstatisticsforsmallsamplesunderrealisticdataconditions.AlthoughTypeI
errorandpowerpropertiesofstatisticscanbecalculatedfordatadrawnfromclassicaltheoretical
distributions(e.g.,normalcurve,Cauchydistribution)forasymptoticconditions(i.e,infinitesample
sizeandinfinitesimallysmalltreatmenteffect),realdataoftendonothavesuchdistributions.[58]
2. Toprovideimplementationsofhypothesisteststhataremoreefficientthanexacttestssuchas
permutationtests(whichareoftenimpossibletocompute)whilebeingmoreaccuratethancritical
valuesforasymptoticdistributions.
MonteCarlomethodsarealsoacompromisebetweenapproximaterandomizationandpermutationtests.
Anapproximaterandomizationtestisbasedonaspecifiedsubsetofallpermutations(whichentails
potentiallyenormoushousekeepingofwhichpermutationshavebeenconsidered).TheMonteCarlo
approachisbasedonaspecifiednumberofrandomlydrawnpermutations(exchangingaminorlossin
precisionifapermutationisdrawntwiceormorefrequentlyfortheefficiencyofnothavingtotrack
whichpermutationshavealreadybeenselected).

Artificialintelligenceforgames
Mainarticle:MonteCarlotreesearch
MonteCarlomethodshavebeendevelopedintoatechniquecalledMonteCarlotreesearchthatisuseful
forsearchingforthebestmoveinagame.Possiblemovesareorganizedinasearchtreeandalargenumber
ofrandomsimulationsareusedtoestimatethelongtermpotentialofeachmove.Ablackboxsimulator
representstheopponent'smoves.[59]
TheMonteCarloTreeSearch(MCTS)methodhasfoursteps:[60]
1. Startingatrootnodeofthetree,selectoptimalchildnodesuntilaleafnodeisreached.
2. Expandtheleafnodeandchooseoneofitschildren.
3. Playasimulatedgamestartingwiththatnode.
https://en.wikipedia.org/wiki/Monte_Carlo_method

8/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

4. Usetheresultsofthatsimulatedgametoupdatethenodeanditsancestors.
Theneteffect,overthecourseofmanysimulatedgames,isthatthevalueofanoderepresentingamove
willgoupordown,hopefullycorrespondingtowhetherornotthatnoderepresentsagoodmove.
MonteCarloTreeSearchhasbeenusedsuccessfullytoplaygamessuchasGo,[61]Tantrix,[62]
Battleship,[63]Havannah,[64]andArimaa.[65]
Seealso:ComputerGo

Designandvisuals
MonteCarlomethodsarealsoefficientinsolvingcoupledintegraldifferentialequationsofradiationfields
andenergytransport,andthusthesemethodshavebeenusedinglobalilluminationcomputationsthat
producephotorealisticimagesofvirtual3Dmodels,withapplicationsinvideogames,architecture,design,
computergeneratedfilms,andcinematicspecialeffects.[66]

Financeandbusiness
Seealso:MonteCarlomethodsinfinance,QuasiMonteCarlomethodsinfinance,MonteCarlo
methodsforoptionpricing,Stochasticmodelling(insurance)andStochasticassetmodel
MonteCarlomethodsinfinanceareoftenusedtoevaluateinvestmentsinprojectsatabusinessunitor
corporatelevel,ortoevaluatefinancialderivatives.Theycanbeusedtomodelprojectschedules,where
simulationsaggregateestimatesforworstcase,bestcase,andmostlikelydurationsforeachtaskto
determineoutcomesfortheoverallproject.MonteCarlomethodsarealsousedinoptionpricing,default
riskanalysis.[67][68][69]

Useinmathematics
Ingeneral,MonteCarlomethodsareusedinmathematicstosolvevariousproblemsbygeneratingsuitable
randomnumbers(seealsoRandomnumbergeneration)andobservingthatfractionofthenumbersthat
obeyssomepropertyorproperties.Themethodisusefulforobtainingnumericalsolutionstoproblemstoo
complicatedtosolveanalytically.ThemostcommonapplicationoftheMonteCarlomethodisMonteCarlo
integration.

Integration
Mainarticle:MonteCarlointegration
Deterministicnumericalintegrationalgorithmsworkwellinasmallnumberofdimensions,butencounter
twoproblemswhenthefunctionshavemanyvariables.First,thenumberoffunctionevaluationsneeded
increasesrapidlywiththenumberofdimensions.Forexample,if10evaluationsprovideadequateaccuracy
inonedimension,then10100pointsareneededfor100dimensionsfartoomanytobecomputed.Thisis
calledthecurseofdimensionality.Second,theboundaryofamultidimensionalregionmaybevery
complicated,soitmaynotbefeasibletoreducetheproblemtoaniteratedintegral.[70]100dimensionsisby
nomeansunusual,sinceinmanyphysicalproblems,a"dimension"isequivalenttoadegreeoffreedom.
https://en.wikipedia.org/wiki/Monte_Carlo_method

9/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

MonteCarlomethodsprovideawayoutofthisexponentialincreaseincomputationtime.Aslongasthe
functioninquestionisreasonablywellbehaved,itcanbeestimatedbyrandomlyselectingpointsin100
dimensionalspace,andtakingsomekindofaverageofthefunctionvaluesatthesepoints.Bythecentral
limittheorem,thismethoddisplays
convergencei.e.,quadruplingthenumberofsampledpoints
halvestheerror,regardlessofthenumberofdimensions.[70]
Arefinementofthismethod,knownasimportancesamplinginstatistics,involvessamplingthepoints
randomly,butmorefrequentlywheretheintegrandislarge.Todo
thispreciselyonewouldhavetoalreadyknowtheintegral,butone
canapproximatetheintegralbyanintegralofasimilarfunctionor
useadaptiveroutinessuchasstratifiedsampling,recursivestratified
sampling,adaptiveumbrellasampling[71][72]ortheVEGAS
algorithm.
Asimilarapproach,thequasiMonteCarlomethod,useslow
discrepancysequences.Thesesequences"fill"theareabetterand
samplethemostimportantpointsmorefrequently,soquasiMonte
Carlomethodscanoftenconvergeontheintegralmorequickly.
Anotherclassofmethodsforsamplingpointsinavolumeisto
simulaterandomwalksoverit(MarkovchainMonteCarlo).Such
methodsincludetheMetropolisHastingsalgorithm,Gibbs
sampling,WangandLandaualgorithm,andinteractingtypeMCMC
methodologiessuchastheSequentialMonteCarlosamplers.[73]

MonteCarlointegrationworksby
comparingrandompointswiththe
valueofthefunction

Simulationandoptimization
Mainarticle:Stochasticoptimization
Anotherpowerfulandverypopularapplicationforrandomnumbers
innumericalsimulationisinnumericaloptimization.Theproblem
istominimize(ormaximize)functionsofsomevectorthatoftenhas
alargenumberofdimensions.Manyproblemscanbephrasedin
thisway:forexample,acomputerchessprogramcouldbeseenas
tryingtofindthesetof,say,10movesthatproducesthebest
Errorsreducebyafactorof
evaluationfunctionattheend.Inthetravelingsalesmanproblemthe
goalistominimizedistancetraveled.Therearealsoapplicationsto
engineeringdesign,suchasmultidisciplinarydesignoptimization.Ithasbeenappliedwithquasione
dimensionalmodelstosolveparticledynamicsproblemsbyefficientlyexploringlargeconfigurationspace.
Thetravelingsalesmanproblemiswhatiscalledaconventionaloptimizationproblem.Thatis,allthefacts
(distancesbetweeneachdestinationpoint)neededtodeterminetheoptimalpathtofollowareknownwith
certaintyandthegoalistorunthroughthepossibletravelchoicestocomeupwiththeonewiththelowest
totaldistance.However,let'sassumethatinsteadofwantingtominimizethetotaldistancetraveledtovisit
eachdesireddestination,wewantedtominimizethetotaltimeneededtoreacheachdestination.Thisgoes
beyondconventionaloptimizationsincetraveltimeisinherentlyuncertain(trafficjams,timeofday,etc.).
Asaresult,todetermineouroptimalpathwewouldwanttousesimulationoptimizationtofirst
https://en.wikipedia.org/wiki/Monte_Carlo_method

10/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

understandtherangeofpotentialtimesitcouldtaketogofromonepointtoanother(representedbya
probabilitydistributioninthiscaseratherthanaspecificdistance)andthenoptimizeourtraveldecisionsto
identifythebestpathtofollowtakingthatuncertaintyintoaccount.

Inverseproblems
Probabilisticformulationofinverseproblemsleadstothedefinitionofaprobabilitydistributioninthe
modelspace.Thisprobabilitydistributioncombinespriorinformationwithnewinformationobtainedby
measuringsomeobservableparameters(data).As,inthegeneralcase,thetheorylinkingdatawithmodel
parametersisnonlinear,theposteriorprobabilityinthemodelspacemaynotbeeasytodescribe(itmaybe
multimodal,somemomentsmaynotbedefined,etc.).
Whenanalyzinganinverseproblem,obtainingamaximumlikelihoodmodelisusuallynotsufficient,aswe
normallyalsowishtohaveinformationontheresolutionpowerofthedata.Inthegeneralcasewemay
havealargenumberofmodelparameters,andaninspectionofthemarginalprobabilitydensitiesofinterest
maybeimpractical,orevenuseless.Butitispossibletopseudorandomlygeneratealargecollectionof
modelsaccordingtotheposteriorprobabilitydistributionandtoanalyzeanddisplaythemodelsinsucha
waythatinformationontherelativelikelihoodsofmodelpropertiesisconveyedtothespectator.Thiscan
beaccomplishedbymeansofanefficientMonteCarlomethod,evenincaseswherenoexplicitformulafor
theaprioridistributionisavailable.
Thebestknownimportancesamplingmethod,theMetropolisalgorithm,canbegeneralized,andthisgives
amethodthatallowsanalysisof(possiblyhighlynonlinear)inverseproblemswithcomplexapriori
informationanddatawithanarbitrarynoisedistribution.[74][75]
Petroleumreservoirmanagement
MonteCarlomethodsareverypopularinhydrocarbonreservoirmanagementinthecontextofnonlinear
inverseproblems.Thisincludesgeneratingcomputationalmodelsofoilandgasreservoirsforconsistency
withobservedproductiondata.Forthegoalofdecisionmakinganduncertaintyassessment,MonteCarlo
methodsareusedforgeneratingmultiplegeologicalrealizations.[76]

Inpopularculture
TheMonteCarloMethod,the1998albumbythesouthernCaliforniaindierockbandNothing
PaintedBlue.(Scat.1998).

Seealso
AuxiliaryfieldMonteCarlo
BiologyMonteCarlomethod
ComparisonofriskanalysisMicrosoftExceladdins
DirectsimulationMonteCarlo
DynamicMonteCarlomethod
KineticMonteCarlo
Meanfieldparticlemethods
Particlefilter
https://en.wikipedia.org/wiki/Monte_Carlo_method

11/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

ListofsoftwareforMonteCarlomolecularmodeling
MonteCarlomethodforphotontransport
MonteCarlomethodsforelectrontransport
Morrismethod
Geneticalgorithms
QuasiMonteCarlomethod
Sobolsequence

Notes
1. Kroese,D.P.Brereton,T.Taimre,T.Botev,Z.I.(2014)."WhytheMonteCarlomethodissoimportant
today".WIREsComputStat6:386392.doi:10.1002/wics.1314(https://dx.doi.org/10.1002%2Fwics.1314).
2. Hubbard2009
3. Kolokoltsov,Vassili(2010).NonlinearMarkovprocesses.CambridgeUniv.Press.p.375.
4. DelMoral,Pierre(2013).MeanfieldsimulationforMonteCarlointegration
(http://www.crcpress.com/product/isbn/9781466504059).Chapman&Hall/CRCPress.p.626."Monographson
Statistics&AppliedProbability"
5. "SequentialMonteCarlosamplersP.DelMoralA.DoucetA.Jasra2006JournaloftheRoyalStatistical
Society:SeriesB(StatisticalMethodology)WileyOnlineLibrary"
(http://onlinelibrary.wiley.com/doi/10.1111/j.14679868.2006.00553.x/abstract).onlinelibrary.wiley.com.
Retrieved20150611.
6. Kalos&Whitlock2008
7. Metropolis1987
8. Eckhardt1987
9. McKean,Henry,P.(1967)."Propagationofchaosforaclassofnonlinearparabolicequations".LectureSeriesin
DifferentialEquations,CatholicUniv.7:4157.
10. McKean,Henry,P.(1966)."AclassofMarkovprocessesassociatedwithnonlinearparabolicequations"
(http://www.ncbi.nlm.nih.gov/pmc/articles/PMC220210/pdf/pnas001670283.pdf)(PDF).Proc.Nat.Acad.Sci.
USA56(6):19071911.Bibcode:1966PNAS...56.1907M(http://adsabs.harvard.edu/abs/1966PNAS...56.1907M).
doi:10.1073/pnas.56.6.1907(https://dx.doi.org/10.1073%2Fpnas.56.6.1907).PMC220210
(https://www.ncbi.nlm.nih.gov/pmc/articles/PMC220210).PMID16591437
(https://www.ncbi.nlm.nih.gov/pubmed/16591437).
11. Herman,KahnHarris,Theodore,E.(1951)."Estimationofparticletransmissionbyrandomsampling"
(https://dornsifecms.usc.edu/assets/sites/520/docs/kahnharris.pdf)(PDF).Natl.Bur.Stand.Appl.Math.Ser.12:
2730.
12. Turing,AlanM."Computingmachineryandintelligence"(http://mind.oxfordjournals.org/content/LIX/236/433).
MindLIX(238):433460.doi:10.1093/mind/LIX.236.433
(https://dx.doi.org/10.1093%2Fmind%2FLIX.236.433).
13. Barricelli,NilsAall(1954)."Esempinumericidiprocessidievoluzione".Methodos:4568.
14. Barricelli,NilsAall(1957)."Symbiogeneticevolutionprocessesrealizedbyartificialmethods".Methodos:143
182.
15. DelMoral,Pierre(2004).FeynmanKacformulae.Genealogicalandinteractingparticleapproximations
(http://www.springer.com/mathematics/probability/book/9780387202686).Springer.p.575."Series:
ProbabilityandApplications"
16. DelMoral,PierreMiclo,Laurent(2000).BranchingandInteractingParticleSystemsApproximationsof
FeynmanKacFormulaewithApplicationstoNonLinearFiltering.
(http://archive.numdam.org/ARCHIVE/SPS/SPS_2000__34_/SPS_2000__34__1_0/SPS_2000__34__1_0.pdf)
(PDF).LectureNotesinMathematics1729.pp.1145.doi:10.1007/bfb0103798
(https://dx.doi.org/10.1007%2Fbfb0103798).

https://en.wikipedia.org/wiki/Monte_Carlo_method

12/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

17. DelMoral,PierreMiclo,Laurent(2000)."AMoranparticlesystemapproximationofFeynmanKacformulae."
(http://www.sciencedirect.com/science/article/pii/S0304414999000940).StochasticProcessesandtheir
Applications86(2):193216.doi:10.1016/S03044149(99)000940(https://dx.doi.org/10.1016%2FS0304
4149%2899%29000940).
18. DelMoral,Pierre(2003)."ParticleapproximationsofLyapunovexponentsconnectedtoSchrdingeroperators
andFeynmanKacsemigroups"(http://journals.cambridge.org/download.php?
file=%2FPSS%2FPSS7%2FS1292810003000016a.pdf&code=a0dbaa7ffca871126dc05fe2f918880a)(PDF).
ESAIMProbability&Statistics7:171208.doi:10.1051/ps:2003001
(https://dx.doi.org/10.1051%2Fps%3A2003001).
19. Assaraf,RolandCaffarel,MichelKhelif,Anatole(2000)."DiffusionMonteCarloMethodswithafixednumber
ofwalkers"(http://qmcchem.upstlse.fr/files/caffarel/31.pdf)(PDF).Phys.Rev.E61:45664575.
Bibcode:2000PhRvE..61.4566A(http://adsabs.harvard.edu/abs/2000PhRvE..61.4566A).
doi:10.1103/physreve.61.4566(https://dx.doi.org/10.1103%2Fphysreve.61.4566).
20. Caffarel,MichelCeperley,DavidKalos,Malvin(1993)."CommentonFeynmanKacPathIntegralCalculation
oftheGroundStateEnergiesofAtoms".Phys.Rev.Lett.71:2159.Bibcode:1993PhRvL..71.2159C
(http://adsabs.harvard.edu/abs/1993PhRvL..71.2159C).doi:10.1103/physrevlett.71.2159
(https://dx.doi.org/10.1103%2Fphysrevlett.71.2159).
21. Hetherington,Jack,H.(1984)."Observationsonthestatisticaliterationofmatrices"
(http://journals.aps.org/pra/abstract/10.1103/PhysRevA.30.2713).Phys.Rev.A.30(2713):27132719.
Bibcode:1984PhRvA..30.2713H(http://adsabs.harvard.edu/abs/1984PhRvA..30.2713H).
doi:10.1103/PhysRevA.30.2713(https://dx.doi.org/10.1103%2FPhysRevA.30.2713).
22. Fermi,EnriqueRichtmyer,Robert,D.(1948)."NoteoncensustakinginMonteCarlocalculations"
(http://scienzecomo.uninsubria.it/bressanini/montecarlohistory/fermi1948.pdf)(PDF).LAM805(A).
"DeclassifiedreportLosAlamosArchive"
23. Rosenbluth,Marshall,N.Rosenbluth,Arianna,W.(1955)."MonteCarlocalculationsoftheaverageextension
ofmacromolecularchains".J.Chem.Phys23:356359.Bibcode:1955JChPh..23..356R
(http://adsabs.harvard.edu/abs/1955JChPh..23..356R).doi:10.1063/1.1741967
(https://dx.doi.org/10.1063%2F1.1741967).
24. Gordon,N.J.Salmond,D.J.Smith,A.F.M.(April1993)."Novelapproachtononlinear/nonGaussianBayesian
stateestimation"(http://ieeexplore.ieee.org/xpl/login.jsp?
tp=&arnumber=210672&url=http%25253A%25252F%25252Fieeexplore.ieee.org%25252Fxpls%25252Fabs_all.js
p%25253Farnumber%25253D210672).RadarandSignalProcessing,IEEProceedingsF140(2):107113.
doi:10.1049/ipf2.1993.0015(https://dx.doi.org/10.1049%2Fipf2.1993.0015).ISSN0956375X
(https://www.worldcat.org/issn/0956375X).
25. Kitagawa,G.(1996)."MontecarlofilterandsmootherfornonGaussiannonlinearstatespacemodels".Journal
ofComputationalandGraphicalStatistics5(1):125.doi:10.2307/1390750
(https://dx.doi.org/10.2307%2F1390750).JSTOR1390750(https://www.jstor.org/stable/1390750).
26. DelMoral,Pierre(1996)."NonLinearFiltering:InteractingParticleSolution."
(http://web.maths.unsw.edu.au/~peterdelmoral/mprfs.pdf)(PDF).MarkovProcessesandRelatedFields2(4):
555580.
27. Carvalho,HimilconDelMoral,PierreMonin,AndrSalut,Grard(July1997)."OptimalNonlinearFiltering
inGPS/INSIntegration."(http://homepages.laas.fr/monin/Version_anglaise/Publications_files/GPS.pdf)(PDF).
IEEETrans.onAerospaceandelectronicsystems33(3).
28. P.DelMoral,G.Rigal,andG.Salut.Estimationandnonlinearoptimalcontrol:Anunifiedframeworkfor
particlesolutions
LAASCNRS,Toulouse,ResearchReportno.91137,DRETDIGILOGLAAS/CNRScontract,April(1991).
29. P.DelMoral,G.Rigal,andG.Salut.NonlinearandnonGaussianparticlefiltersappliedtoinertialplatform
repositioning.
LAASCNRS,Toulouse,ResearchReportno.92207,STCAN/DIGILOGLAAS/CNRSConventionSTCANno.
A.91.77.013,(94p.)September(1991).
30. P.DelMoral,G.Rigal,andG.Salut.Estimationandnonlinearoptimalcontrol:Particleresolutioninfiltering
andestimation.Experimentalresults.
ConventionDRETno.89.34.553.00.470.75.01,Researchreportno.2(54p.),January(1992).
https://en.wikipedia.org/wiki/Monte_Carlo_method

13/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

31. P.DelMoral,G.Rigal,andG.Salut.Estimationandnonlinearoptimalcontrol:Particleresolutioninfiltering
andestimation.Theoreticalresults
ConventionDRETno.89.34.553.00.470.75.01,Researchreportno.3(123p.),October(1992).
32. P.DelMoral,J.Ch.Noyer,G.Rigal,andG.Salut.Particlefiltersinradarsignalprocessing:detection,
estimationandairtargetsrecognition.
LAASCNRS,Toulouse,Researchreportno.92495,December(1992).
33. P.DelMoral,G.Rigal,andG.Salut.Estimationandnonlinearoptimalcontrol:Particleresolutioninfiltering
andestimation.
Studieson:Filtering,optimalcontrol,andmaximumlikelihoodestimation.ConventionDRETno.
89.34.553.00.470.75.01.Researchreportno.4(210p.),January(1993).
34. DelMoral,Pierre(1998)."MeasureValuedProcessesandInteractingParticleSystems.ApplicationtoNon
LinearFilteringProblems"(http://projecteuclid.org/download/pdf_1/euclid.aoap/1028903535).AnnalsofApplied
Probability(PublicationsduLaboratoiredeStatistiqueetProbabilits,9615(1996)ed.)8(2):438495.
doi:10.1214/aoap/1028903535(https://dx.doi.org/10.1214%2Faoap%2F1028903535).
35. Crisan,DanGaines,JessicaLyons,Terry(1998)."Convergenceofabranchingparticlemethodtothesolution
oftheZakai".SIAMJournalonAppliedMathematics58(5):15681590.doi:10.1137/s0036139996307371
(https://dx.doi.org/10.1137%2Fs0036139996307371).
36. Crisan,DanLyons,Terry(1997)."Nonlinearfilteringandmeasurevaluedprocesses".ProbabilityTheoryand
RelatedFields109(2):217244.doi:10.1007/s004400050131(https://dx.doi.org/10.1007%2Fs004400050131).
37. Crisan,DanLyons,Terry(1999)."AparticleapproximationofthesolutionoftheKushnerStratonovitch
equation".ProbabilityTheoryandRelatedFields115(4):549578.doi:10.1007/s004400050249
(https://dx.doi.org/10.1007%2Fs004400050249).
38. Crisan,DanDelMoral,PierreLyons,Terry(1999)."Discretefilteringusingbranchingandinteractingparticle
systems"(http://web.maths.unsw.edu.au/~peterdelmoral/crisan98discrete.pdf)(PDF).MarkovProcessesand
RelatedFields5(3):293318.
39. DelMoral,PierreGuionnet,Alice(1999)."OnthestabilityofMeasureValuedProcesseswithApplicationsto
filtering".C.R.Acad.Sci.Paris39(1):429434.
40. DelMoral,PierreGuionnet,Alice(2001)."Onthestabilityofinteractingprocesseswithapplicationstofiltering
andgeneticalgorithms"(http://web.maths.unsw.edu.au/~peterdelmoral/ihp.ps).Annalesdel'InstitutHenri
Poincar,37(2):155194.doi:10.1016/s02460203(00)010645(https://dx.doi.org/10.1016%2Fs0246
0203%2800%29010645).
41. Ripley1987
42. Sawilowsky2003
43. Davenport1992
44. Vose2000,p.13
45. Vose2000,p.16
46. "GPUbasedhighperformancecomputingforradiationtherapy".PhysicsinMedicineandBiology59:R151
R182.Bibcode:2014PMB....59R.151J(http://adsabs.harvard.edu/abs/2014PMB....59R.151J).doi:10.1088/0031
9155/59/4/R151(https://dx.doi.org/10.1088%2F00319155%2F59%2F4%2FR151).
47. "Advancesinkilovoltagexraybeamdosimetry".PhysicsinMedicineandBiology59:R183R231.
Bibcode:2014PMB....59R.183H(http://adsabs.harvard.edu/abs/2014PMB....59R.183H).doi:10.1088/0031
9155/59/6/R183(https://dx.doi.org/10.1088%2F00319155%2F59%2F6%2FR183).
48. "FiftyyearsofMonteCarlosimulationsformedicalphysics".PhysicsinMedicineandBiology51:R287R301.
Bibcode:2006PMB....51R.287R(http://adsabs.harvard.edu/abs/2006PMB....51R.287R).doi:10.1088/0031
9155/51/13/R17(https://dx.doi.org/10.1088%2F00319155%2F51%2F13%2FR17).
49. Baeurle2009
50. MacGillivray&Dodd1982
51. Golden1979
52. IntPanisetal.2001
53. IntPanisetal.2002
54. G.A.Bird,MolecularGasDynamics,Clarendon,Oxford(1976)

https://en.wikipedia.org/wiki/Monte_Carlo_method

14/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

55. Dietrich,S.Boyd,I.(1996)."AScalaroptimizedparallelimplementationoftheDSMCtechnique".Journalof
ComputationalPhysics126:32842.Bibcode:1996JCoPh.126..328D
(http://adsabs.harvard.edu/abs/1996JCoPh.126..328D).doi:10.1006/jcph.1996.0141
(https://dx.doi.org/10.1006%2Fjcph.1996.0141).
56. Ojeda&etal.2009,
57. Milik&Skolnick1993
58. Sawilowsky&Fahoome2003
59. http://sander.landofsand.com/publications/MonteCarlo_Tree_Search__A_New_Framework_for_Game_AI.pdf
60. MonteCarloTreeSearchAbout(http://mcts.ai/about/index.html)
61. ParallelMonteCarloTreeSearchSpringer(http://link.springer.com/chapter/10.1007/9783540876083_6)
62. http://www.tantrix.com:4321/Tantrix/TRobot/MCTS%20Final%20Report.pdf
63. http://www0.cs.ucl.ac.uk/staff/D.Silver/web/Publications_files/pomcp.pdf
64. ImprovingMonteCarloTreeSearchinHavannahSpringer(http://link.springer.com/chapter/10.1007/9783
642179280_10)
65. http://www.arimaa.com/arimaa/papers/ThomasJakl/bcthesis.pdf
66. SzirmayKalos2008
67. Carmona,RenDelMoral,PierreHu,PengOudjane,Nadia(2012).Carmona,RenA.Moral,PierreDelHu,
Pengetal.,eds."AnIntroductiontoParticleMethodswithFinancialApplications"
(http://link.springer.com/chapter/10.1007/9783642257469_1).NumericalMethodsinFinance.Springer
ProceedingsinMathematics(SpringerBerlinHeidelberg)12:349.doi:10.1007/9783642257469_1
(https://dx.doi.org/10.1007%2F9783642257469_1).ISBN9783642257452.
68. "NumericalMethodsinFinanceSpringer"(http://link.springer.com/10.1007/9783642257469).
link.springer.com.doi:10.1007/9783642257469(https://dx.doi.org/10.1007%2F9783642257469).
69. Kroese,D.P.Taimre,T.Botev,Z.I.(2011).HandbookofMonteCarloMethods.JohnWiley&Sons.
70. Pressetal.1996
71. MEZEI,M(31December1986)."Adaptiveumbrellasampling:Selfconsistentdeterminationofthenon
Boltzmannbias".JournalofComputationalPhysics68(1):237248.Bibcode:1987JCoPh..68..237M
(http://adsabs.harvard.edu/abs/1987JCoPh..68..237M).doi:10.1016/00219991(87)900544
(https://dx.doi.org/10.1016%2F00219991%2887%29900544).
72. Bartels,ChristianKarplus,Martin(31December1997)."ProbabilityDistributionsforComplexSystems:
AdaptiveUmbrellaSamplingofthePotentialEnergy".TheJournalofPhysicalChemistryB102(5):865880.
doi:10.1021/jp972280j(https://dx.doi.org/10.1021%2Fjp972280j).
73. "SequentialMonteCarlosamplersDelMoralDoucetJasra2006JournaloftheRoyalStatisticalSociety:
SeriesB(StatisticalMethodology)WileyOnlineLibrary"(http://doi.wiley.com/10.1111/j.1467
9868.2006.00553.x).JournaloftheRoyalStatisticalSociety:SeriesB(StatisticalMethodology)68:411436.
doi:10.1111/j.14679868.2006.00553.x(https://dx.doi.org/10.1111%2Fj.14679868.2006.00553.x).
74. Mosegaard&Tarantola1995
75. Tarantola2005
76. Shirangi,M.G.,HistorymatchingproductiondataanduncertaintyassessmentwithanefficientTSVD
parameterizationalgorithm,JournalofPetroleumScienceandEngineering,
http://www.sciencedirect.com/science/article/pii/S0920410513003227

References
Anderson,HerbertL.(1986)."Metropolis,MonteCarloandtheMANIAC"
(http://library.lanl.gov/cgibin/getfile?00326886.pdf)(PDF).LosAlamosScience14:96108.
Baeurle,StephanA.(2009)."Multiscalemodelingofpolymermaterialsusingfieldtheoretic
methodologies:Asurveyaboutrecentdevelopments".JournalofMathematicalChemistry46(2):
363426.doi:10.1007/s1091000894673(https://dx.doi.org/10.1007%2Fs1091000894673).
Berg,BerndA.(2004).MarkovChainMonteCarloSimulationsandTheirStatisticalAnalysis(With
WebBasedFortranCode).Hackensack,NJ:WorldScientific.ISBN9812389350.
Binder,Kurt(1995).TheMonteCarloMethodinCondensedMatterPhysics.NewYork:Springer.
https://en.wikipedia.org/wiki/Monte_Carlo_method

15/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

ISBN0387543694.
Caflisch,R.E.(1998).MonteCarloandquasiMonteCarlomethods.ActaNumerica7.Cambridge
UniversityPress.pp.149.
Davenport,J.H."Primalitytestingrevisited".ProceedingISSAC'92Papersfromtheinternational
symposiumonSymbolicandalgebraiccomputation:123129.doi:10.1145/143242.143290
(https://dx.doi.org/10.1145%2F143242.143290).ISBN0897914899.
Doucet,ArnaudFreitas,NandodeGordon,Neil(2001).SequentialMonteCarlomethodsin
practice.NewYork:Springer.ISBN0387951466.
Eckhardt,Roger(1987)."StanUlam,JohnvonNeumann,andtheMonteCarlomethod"
(http://library.lanl.gov/cgibin/getfile?1513.pdf)(PDF).LosAlamosScience,SpecialIssue(15):131
137.
Fishman,G.S.(1995).MonteCarlo:Concepts,Algorithms,andApplications.NewYork:Springer.
ISBN038794527X.
C.ForasteroandL.ZamoraandD.GuiradoandA.Lallena(2010)."AMonteCarlotooltosimulate
breastcancerscreeningprogrammes".Phys.InMed.AndBiol.55(17):52135229.
Bibcode:2010PMB....55.5213F(http://adsabs.harvard.edu/abs/2010PMB....55.5213F).
doi:10.1088/00319155/55/17/021(https://dx.doi.org/10.1088%2F00319155%2F55%2F17%2F021).
Golden,LeslieM.(1979)."TheEffectofSurfaceRoughnessontheTransmissionofMicrowave
RadiationThroughaPlanetarySurface".Icarus38(3):451455.Bibcode:1979Icar...38..451G
(http://adsabs.harvard.edu/abs/1979Icar...38..451G).doi:10.1016/00191035(79)901994
(https://dx.doi.org/10.1016%2F00191035%2879%29901994).
Gould,HarveyTobochnik,Jan(1988).AnIntroductiontoComputerSimulationMethods,Part2,
ApplicationstoPhysicalSystems.Reading:AddisonWesley.ISBN020116504X.
Grinstead,CharlesSnell,J.Laurie(1997).IntroductiontoProbability.AmericanMathematical
Society.pp.1011.
Hammersley,J.M.Handscomb,D.C.(1975).MonteCarloMethods.London:Methuen.ISBN0
416523404.
Hartmann,A.K.(2009).PracticalGuidetoComputerSimulations
(http://www.worldscibooks.com/physics/6988.html).WorldScientific.ISBN9789812834157.
Hubbard,Douglas(2007).HowtoMeasureAnything:FindingtheValueofIntangiblesinBusiness.
JohnWiley&Sons.p.46.
Hubbard,Douglas(2009).TheFailureofRiskManagement:WhyIt'sBrokenandHowtoFixIt.John
Wiley&Sons.
Kahneman,D.Tversky,A.(1982).JudgementunderUncertainty:HeuristicsandBiases.Cambridge
UniversityPress.
Kalos,MalvinH.Whitlock,PaulaA.(2008).MonteCarloMethods.WileyVCH.ISBN9783527
407606.
Kroese,D.P.Taimre,T.Botev,Z.I.(2011).HandbookofMonteCarloMethods
(http://www.montecarlohandbook.org).NewYork:JohnWiley&Sons.p.772.ISBN0470177934.
MacGillivray,H.T.Dodd,R.J.(1982)."MonteCarlosimulationsofgalaxysystems"
(http://www.springerlink.com/content/rp3g1q05j176r108/fulltext.pdf)(PDF).AstrophysicsandSpace
Science(SpringerNetherlands)86(2).
MacKeown,P.Kevin(1997).StochasticSimulationinPhysics.NewYork:Springer.ISBN981
3083263.
Metropolis,N.(1987)."ThebeginningoftheMonteCarlomethod"(http://library.lanl.gov/la
pubs/00326866.pdf)(PDF).LosAlamosScience(1987SpecialIssuededicatedtoStanislawUlam):
125130.
Metropolis,NicholasRosenbluth,AriannaW.Rosenbluth,MarshallN.Teller,AugustaH.Teller,
Edward(1953)."EquationofStateCalculationsbyFastComputingMachines".JournalofChemical
Physics21(6):1087.Bibcode:1953JChPh..21.1087M
https://en.wikipedia.org/wiki/Monte_Carlo_method

16/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

(http://adsabs.harvard.edu/abs/1953JChPh..21.1087M).doi:10.1063/1.1699114
(https://dx.doi.org/10.1063%2F1.1699114).
Metropolis,N.Ulam,S.(1949)."TheMonteCarloMethod".JournaloftheAmericanStatistical
Association(AmericanStatisticalAssociation)44(247):335341.doi:10.2307/2280232
(https://dx.doi.org/10.2307%2F2280232).JSTOR2280232(https://www.jstor.org/stable/2280232).
PMID18139350(https://www.ncbi.nlm.nih.gov/pubmed/18139350).
M.MilikandJ.Skolnick(Jan1993)."Insertionofpeptidechainsintolipidmembranes:anofflattice
MonteCarlodynamicsmodel".Proteins15(1):1025.doi:10.1002/prot.340150104
(https://dx.doi.org/10.1002%2Fprot.340150104).PMID8451235
(https://www.ncbi.nlm.nih.gov/pubmed/8451235).
Mosegaard,KlausTarantola,Albert(1995)."MonteCarlosamplingofsolutionstoinverse
problems".J.Geophys.Res.100(B7):1243112447.Bibcode:1995JGR...10012431M
(http://adsabs.harvard.edu/abs/1995JGR...10012431M).doi:10.1029/94JB03097
(https://dx.doi.org/10.1029%2F94JB03097).
P.OjedaandM.GarciaandA.LondonoandN.Y.Chen(Feb2009)."MonteCarloSimulationsof
ProteinsinCages:InfluenceofConfinementontheStabilityofIntermediateStates".Biophys.Jour.
(BiophysicalSociety)96(3):10761082.Bibcode:2009BpJ....96.1076O
(http://adsabs.harvard.edu/abs/2009BpJ....96.1076O).doi:10.1529/biophysj.107.125369
(https://dx.doi.org/10.1529%2Fbiophysj.107.125369).
IntPanisLDeNockerL,DeVliegerI,TorfsR(2001)."Trendsanduncertaintyinairpollution
impactsandexternalcostsofBelgianpassengercartrafficInternational".JournalofVehicleDesign
27(14):183194.doi:10.1504/IJVD.2001.001963
(https://dx.doi.org/10.1504%2FIJVD.2001.001963).
IntPanisL,RablA,DeNockerL,TorfsR(2002).P.Sturm,ed."DieselorPetrol?Anenvironmental
comparisonhamperedbyuncertainty".MitteilungenInstitutfrVerbrennungskraftmaschinenund
Thermodynamik(TechnischeUniversittGrazAustria).Heft81Vol1:4854.
Press,WilliamH.Teukolsky,SaulA.Vetterling,WilliamT.Flannery,BrianP.(1996)[1986].
NumericalRecipesinFortran77:TheArtofScientificComputing.FortranNumericalRecipes1
(Seconded.).CambridgeUniversityPress.ISBN052143064X.
Ripley,B.D.(1987).StochasticSimulation.Wiley&Sons.
Robert,C.P.Casella,G.(2004).MonteCarloStatisticalMethods(2nded.).NewYork:Springer.
ISBN0387212396.
Rubinstein,R.Y.Kroese,D.P.(2007).SimulationandtheMonteCarloMethod(2nded.).New
York:JohnWiley&Sons.ISBN9780470177938.
Savvides,SavvakisC.(1994)."RiskAnalysisinInvestmentAppraisal".ProjectAppraisalJournal9
(1).doi:10.2139/ssrn.265905(https://dx.doi.org/10.2139%2Fssrn.265905).
Sawilowsky,ShlomoS.Fahoome,GailC.(2003).StatisticsviaMonteCarloSimulationwith
Fortran.RochesterHills,MI:JMASM.ISBN0974023604.
Sawilowsky,ShlomoS.(2003)."Youthinkyou'vegottrivials?"
(http://education.wayne.edu/jmasm/sawilowsky_effect_size_debate.pdf)(PDF).JournalofModern
AppliedStatisticalMethods2(1):218225.
Silver,DavidVeness,Joel(2010)."MonteCarloPlanninginLargePOMDPs"
(http://books.nips.cc/papers/files/nips23/NIPS2010_0740.pdf)(PDF).InLafferty,J.Williams,C.K.
I.ShaweTaylor,J.Zemel,R.S.Culotta,A.AdvancesinNeuralInformationProcessingSystems
23.NeuralInformationProcessingSystemsFoundation.
SzirmayKalos,Lszl(2008).MonteCarloMethodsinGlobalIlluminationPhotorealistic
RenderingwithRandomization.VDMVerlagDr.Muellere.K.ISBN9783836479196.
Tarantola,Albert(2005).InverseProblemTheory
(http://www.ipgp.jussieu.fr/~tarantola/Files/Professional/SIAM/index.html).Philadelphia:Societyfor
IndustrialandAppliedMathematics.ISBN0898715725.
https://en.wikipedia.org/wiki/Monte_Carlo_method

17/18

9/22/2015

MonteCarlomethodWikipedia,thefreeencyclopedia

Vose,David(2008).RiskAnalysis,AQuantitativeGuide(Thirded.).JohnWiley&Sons.
Retrievedfrom"https://en.wikipedia.org/w/index.php?
title=Monte_Carlo_method&oldid=681893137"

WikimediaCommonshas
mediarelatedtoMonte
Carlomethod.

Categories: MonteCarlomethods Numericalanalysis

https://en.wikipedia.org/wiki/Monte_Carlo_method

18/18