python離散事件模擬和simpy簡(jiǎn)介

1、IntroductiontoDiscreteEventSimulationtheSimPyLanguageNmMatloffFebruary132008c?20062008N.S.MatloffContents1WhatIsDiscreteEventSimulation(DES)32WldViewsinDESProgramming32.1TheActivityientedParadigm.........................

2、.......32.2TheEventientedParadigm.................................42.3TheProcessientedParadigm................................63IntroductiontotheSimPySimulationLanguage73.1SimPyOverview...................................

3、.....83.2IntroductiontoSimPyProgramming..............................93.2.1MachRep1.py:OurFirstSimPyProgram.......................103.2.2MachRep2.py:IntroducingtheResourceClass....................143.2.3MachRep3.py:Intro

4、ducingPassivateReactivateOperations.............163.2.4MMk.py:“DoItYourself”QueueManagement....................183.2.5SMP.py:SimultaneousPossessionofResources....................203.2.6Cell.py:DynamicCreationofThreads...

5、......................223.3NoteTheseRestrictionsonPEMs...............................253.4SimPyDataCollectionDisplay..............................253.4.1IntroductiontoMonits................................253.4.2TimeAver

6、ages.....................................263.4.3TheFunctionMonit.timeAverage().........................273.4.4ButIRecommendThatYouNotUseThisFunction..................2711WhatIsDiscreteEventSimulation(DES)Considersimulati

7、onofsomesystemwhichevolvesthroughtime.Thereisahugevarietyofsuchapplications.Onecansimulateaweathersystemfinstance.Akeypointthoughisthatinthatsettingtheeventsbeingsimulatedwouldbecontinuousmeaningfexamplethatifweweretogra

8、phtemperatureagainsttimethecurvewouldbecontinuousnobreaks.Bycontrastsupposewesimulatetheoperationofawarehouse.Purchasederscomeinarefilledreducinginventybutinventyisreplenishedfromtimetotime.Hereatypicalvariablewouldbethe

9、inventyitselfi.e.thenumberofitemscurrentlyinstockfagivenproduct.Ifweweretographthatnumberagainsttimewewouldgetwhatmathematicianscallastepfunctioni.e.asetofflatlinesegmentswithbreaksbetweenthem.Theeventshere—decreasesincr

10、easesintheinventy—arediscretevariablesnotcontinuousones.DESinvolvessimulatingsuchsystems.2WldViewsinDESProgrammingSimulationprogrammingcanoftenbedifficult—difficulttowritethecodedifficulttodebug.Thereasonfthisisthatitrea

11、llyisafmofparallelprogrammingwithmanydifferentactivitiesinprogresssimultaneouslyparallelprogrammingcanbechallenging.Fthisreasonmanypeoplehavetriedtodevelopseparatesimulationlanguagesatleastsimulationparadigms(i.e.program

12、mingstyles)whichenabletoprogrammertoachieveclarityinsimulationcode.SpecialsimulationlanguageshavebeeninventedinthepastnotablySIMULAwhichwasinventedinthe1960shassignificancetodayinthatitwasthelanguagewhichinventedtheconce

13、ptofobjectientedprogrammgthatissopopulartoday.HoweverthetrendtodayistosimplydevelopsimulationlibrarieswhichcanbecalledfromdinarylanguagessuchasCinsteadofinventingentirenewlanguages.1Sothecentralfocustodayisontheprogrammi

14、ngparadigmsnotonlanguage.Inthissectionwewillpresentanoverviewofthethreemajdiscreteeventsimulationparadigms.SeveralwldviewshavebeendevelopedfDESprogrammingasseeninthenextfewsections.2.1TheActivityientedParadigmLetusthinko

15、fsimulatingaqueuingsystem.Jobsarriveatromtimesthejobservertakesarandomtimefeachservice.Thetimebetweenarrivalsofjobsthetimeneededtoserveajobwillbecontinuousromvariablespossiblyhavingexponentialothercontinuousdistributions

16、.FconcretenessthinkofanexampleinwhichtheserverisanATMcashmachinethejobsarecustomerswaitinginline.Undertheactivityientedparadigmwewouldbreaktimeintotinyincrements.Iffinstancethemeaninterarrivaltimeweresay20secondswemightb

17、reaktimeintoincrementsofsize0.001.Ateachtimepointourcodewouldlookaroundatalltheactivitiese.g.currentlyactivejobservicingcheckfthepossibleoccurrenceofeventse.g.completionofservice.Ourgoalistofindthelongrunaveragejobwait1T