系統(tǒng)的結(jié)構(gòu)和系統(tǒng)的行為_第1頁
系統(tǒng)的結(jié)構(gòu)和系統(tǒng)的行為_第2頁
系統(tǒng)的結(jié)構(gòu)和系統(tǒng)的行為_第3頁
系統(tǒng)的結(jié)構(gòu)和系統(tǒng)的行為_第4頁
系統(tǒng)的結(jié)構(gòu)和系統(tǒng)的行為_第5頁
已閱讀5頁,還剩77頁未讀, 繼續(xù)免費閱讀

下載本文檔

版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進行舉報或認領(lǐng)

文檔簡介

系統(tǒng)結(jié)構(gòu)與系統(tǒng)行為

提綱系統(tǒng)動力學(xué)基本原理一階系統(tǒng)二階系統(tǒng)(振蕩)(美國)約翰.D.斯特曼.商務(wù)動態(tài)分析方法[M].北京,清華大學(xué)出版社,2008王其藩.系統(tǒng)動力學(xué)[M],上海財經(jīng)大學(xué)出版社,2009李旭.社會系統(tǒng)動力學(xué)-政策研究的原理、方法和應(yīng)用[M],復(fù)旦大學(xué)出版社,2008[美]Ford環(huán)境模擬-環(huán)境系統(tǒng)的系統(tǒng)動力學(xué)模型導(dǎo)論[M],復(fù)旦大學(xué)出版社,2009

系統(tǒng)動力學(xué)的基本原理

系統(tǒng)的行為由它的結(jié)構(gòu)決定!如何提煉和整理產(chǎn)生復(fù)雜行為的那些簡單結(jié)構(gòu),成為研究復(fù)雜系統(tǒng)的關(guān)鍵問題!系統(tǒng)的結(jié)構(gòu)(1)系統(tǒng)的物理和制度結(jié)構(gòu)-客觀結(jié)構(gòu);(2)系統(tǒng)主體決策結(jié)構(gòu)-主觀結(jié)構(gòu)系統(tǒng)的基本行為:指數(shù)增長;尋的行為;振蕩系統(tǒng)的基本結(jié)構(gòu)?如何設(shè)計?系統(tǒng)的結(jié)構(gòu)系統(tǒng)結(jié)構(gòu):指組成系統(tǒng)的各個要素之間相互依賴關(guān)系的秩序?;芈肥墙M成系統(tǒng)結(jié)構(gòu)的基本單元,決策總是存在于回路之中并通過流率變量具體實施,流率變量調(diào)節(jié)流位變量的取值,為決策提供信息。認識系統(tǒng)結(jié)構(gòu)的三個層次:閉合的邊界,反饋回路和流率方程。系統(tǒng)的結(jié)構(gòu)閉合的邊界:閉合的邊界內(nèi)包含了影響系統(tǒng)行為的主要因素和連接關(guān)系-系統(tǒng)流圖反饋回路:反饋環(huán)及反饋環(huán)的鏈接

閉合的邊界和反饋回路從定性的角度描述了系統(tǒng)結(jié)構(gòu)的整體框架。流率變量方程:決策是運用信息對系統(tǒng)實施控制的過程,流率變量方程描述了過程規(guī)律。

流率變量方程從定量角度描述了系統(tǒng)局部的子系統(tǒng)。目標、觀察值、偏差和控制是流率變量方程的內(nèi)部次級子結(jié)構(gòu)系統(tǒng)行為1.ExponentialGrowth2.ExponentialDecay

3.GoalSeekingBehavior4.S-ShapedGrowth5.SystemOscillation系統(tǒng)結(jié)構(gòu)和系統(tǒng)行為complexsystemBehaviorinteractionbetweenstockandflow

setsofstocks/flowswithinlargerfeedbackloops.

系統(tǒng)行為:ExponentialGrowthExactAnalyticalSolutionThefirstorderpositivefeedbackloopsystemisalinearsystem.Asaresult,inadditiontoitssimulatedsolution,itpossessesanexactanalyticalsolution.Thissolutionis:Stockt=Stock0*eCoef*t

InitializationPhase(初始值)Createalistofallequationsinrequiredorderofevaluation.Calculateinitialvaluesforallstocks,flowsandauxiliaries(inorderofevaluation).IterationPhase(迭代)Euler'smethod

EstimatethechangeinstocksovertheintervalDT.Calculatenewvaluesforstocksbasedonthisestimate.Usenewvaluesofstockstocalculatenewvaluesforflowsandauxiliaries.AddDTtosimulationtime.StopiteratingwhenTime>=StopTime

TimeConstant(時間常數(shù))Thetimeconstantofthissystem,Tc,measuredinunitsoftime,isdefinedtobe1/Coef.(描述什么?)

TimeConstantGiventhedefinitionofthesystemtimeconstant,thefollowingquestioncanbeasked:Howmuchexponentialgrowthoccursduringonetimeconstant'sworthoftime?TimeConstantStockt=Stock0*eCoef*tStockt=Stock0*eCoef*TcStockt=Stock0*eCoef*(1/Coef)Stockt=Stock0*e1Stockt=Stock0*2.71828(什么意思?)TimeConstantInotherwords,theamountofexponentialgrowththatoccursduringonetimeconstantis2.71828timestheinitialamountinthesystem'sstock.Pushingthislogictothenextstep,theamountofexponentialgrowththatoccursduringtwotimeconstantsis7.3890(i.e.,2.71828*2.71828)timestheinitialamountinthesystem'sstock.TimeConstantDoublingTime(倍增時間)Thesecondmeasureofhowquicklyexponentialgrowthisproceedinginafirstorder,linear,positivefeedbacksystemisitsdoublingtime.Asystem'sdoublingtimeissimplytheamountoftimeittakesforthecontentsofitsstocktodouble.DoublingTimeThedoublingtimeofafirstorder,linear,positivefeebackloopsystem,Td,isequaltoseventypercentofitstimeconstant.Thatis:Td=.7*Tc.DoublingTime2*Stock0=Stock0*eCoef*TdSincethesystem'stimeconstant,Tc,is1/Coef,thenthesystem'sCoef=1/Tc.Thus:2*Stock0=Stock0*e1/Tc*TdDividingbothsidesbyStock0yields:2=eTd/TcTakingthenaturallogarithmofbothsidesyields:ln(2)=ln(eTd/Tc)Or:.69=Td/TcOr:.7=Td/TcOr:Td=.7*TcDoublingTimeDoublingTimeTheconceptofdoublingtimeleadsdirectlytoaheuristicknownastheRuleof70.DoublingTimetheRuleof70saysthatthenominalsizeofU.S.economywilldoublein(70/.07055)*100=9.86years.ExponentialGrowth小結(jié)要點:基本結(jié)構(gòu);Positiveloopsgenerateexponentialgrowth(firstorderpositivefeedbackloop)增長速度:CoefTheTimeConstant:TCDoubleTime:Td關(guān)系:TC=1∕Coef;Td=0.7TCExponentialDecayMeasuresofExponentialDecay:HalfLife(半衰期)Onemeasureofhowlongitwilltakeforafirstorder,linear,negativefeedbackloopsystemtoreachitsgoalofzeroisitshalflife.Thehalflifeofsuchasystemisthetimeittakestodrainhalfofthecontentsfromthesystem'sstock.Analogoustothecaseofthedoublingtimeforafirstorderpositivefeedbackloopsystemabove,thesystem'stimeconstantmustbedeterminedbeforeitshalflifecanbefound.ExponentialDecayAswaspreviouslyshown,theanalyticalsolutionstoEquations1ofExponentialGrowthandEquation2ofGoalSeekingBehaviordifferonlybysign.

ExponentialDecayStockt=Stock0*e-(1/TimeCoef)*TcStockt=Stock0*e-(1/TimeCoef)*TimeCoefStockt=Stock0*e-1Stockt=Stock0*(1/e)Stockt=Stock0*.37ExponentialDecaythesystemloses63%ofthecontentsofitsstockafteronetimeconstanthaselapsed.thesystemloses86%ofthecontentsofitsstockaftertwotimeconstantshaveelasped(Stockt=Stock0*.37*.37=Stock0*.14)and95%ofthecontentsofitsstockafterthreetimeconstantshaveelapsed(Stockt=Stock0*.37*.37*.37=Stock0*.05).ExponentialDecayAsinthecaseofafirstorderpositivefeedbackloopsystem,thenegativeloopsystem'shalflifeis70%ofitstimeconstant.Thatis:Th=.7*Tc.ExponentialDecayExponentialDecayTheanalysisofthefirstordernegativefeedbackloopsystemwithanimplicitgoalofzerocanbeextendedtothemoregeneralnegativefeedackloopmodeloriginallypresentedinFigure2ofGoalSeekingBehavior,givingthefollowinginsight:Afirstorder,negative,feedbackloopsystemreachesapproximately95%ofitsgoalafterthreetimeconstantshaveelapsed.ExponentialDecay小結(jié)要點:結(jié)構(gòu);TIMECONSTANT;HALFLIFE.Th=.7*Tcapproximately95%ofitsgoalafterthreetimeconstantsGoalSeekingBehaviorthetwotypesoffeedbackloopsthatexistintheworld:positiveloopsandnegativeloops.Positiveloops“generate”exponentialgrowth(orrapiddecline)negativeloopsproducegoal-seekingbehavior.GenericFirstOrderNegativeFeedbackLoopSystemGoalSeekingBehaviorFigure1presentsageneric,firstorder,linear,negativefeedbackloopsystem.Themodelisinitializedinequilibriumbysettingtheinitialvalueofthestockequaltoitsgoal,andthen"shocked"outofequilibrium(attime=1)byastepfunction

thatchangesthesystem'sgoalfromtentofifteen.GoalSeekingBehaviorAsystemcontainingnegativefeedbackloopswillbeinequilibriumonlywhenallofitsstocksareequaltoallofitsgoalssimultaneously.GoalSeekingBehaviorTheimplicationofthisideaisveryimportant.Intherealworlditisraretofindactualsystems(particularlyactualsocialsystems)thatexistinstatesofequilibriumbecauseactualsystemscontainmanynegativefeedbackloopsthatdonotreachtheirgoalssimultaneously.Thus,systemdynamicistsoftenstarttheirmodelsinstatesofequilibrium,notbecausetheyfeelthatthisreflectsreality,butratherbecauseitisausefulreferencepointforanalyzingthe"pure"behaviorsoftheirmodels

GoalSeekingBehaviorGoalSeekingBehaviorThetimeittakesthesystemtoreachitsnewgoalisdeterminedbythesizeofTimeCoef.ThebiggerTimeCoefis,thelongerittakesforthesystemtoadjusttoitsnewgoal.GoalSeekingBehavior

ThreeSimulationsoftheGenericFirstOrderNegative

FeedbackLoopSystemgiveastepreductionattime=1

GoalSeekingBehavior討論:GoalSeekingBehavior產(chǎn)生的本質(zhì)原因是什么?SystemOscillationSystemOscillationOscillationreferstoabehaviorinwhichthevaluesofthestocksvaryaroundsomeaveragevalueinarepeatingpattern.SystemOscillationSustainedoscillation:the“peak“isthesameeachcycle.Expandingoscillation:the“peak“becomesgreatereachperiod.Dampedoscillation:the“peak“becomessmallereachperiod.SystemOscillationThemathematicsfunctionofoscillationStock=A?sin(??time+?)WhereA,?,and?areconstantAcademicperformancemodel二級反饋環(huán)與系統(tǒng)振蕩AcademicperformancemodelWhythesystemdemonstratesthebehaviorofsustainedoscillation?Thestockstheflowsgap26hoursaweekDebriefoftheacademicperformancemodelAnoscillatingsystemrequirestwostocks.Oscillatingsystemhavesomemomentum-gap.Gap---increaseheramountofstudying--….---currentgrades:desiredlevel流位變量的積累作用:延遲Whysecond-ordersystemcanoscillrteThesystemmustbeanegativefeedbackloop:negativefeedbackloopsalwaysattempttoclosethegapbetweensomedesiredstateofthesystemandtheactualstateofthestate,knowasgoal-seekingbehaviorThesystemmustbeatleastsecond-order:thesecondstockcancontinuetochangethesystemevenifoneofthestockisatitsgoalvalue.SystemOscillationSystemoscillationisthecharacteristicsymptomofnegativefeedbackstructuresinwhichtheinformationusedtotakegoal-seekingactionisdelayed.Insuchcases,acontrolactionisnotbasedonthecurrentstateofasystembutonsomepreviousstateorvalue.Asyoucanimagine,usingdatedinformationtocontroltheapproachtoatargetislikelytocausethesystemtomissorovershootitsgoal.Specificstructuralcharacteristicsofasystem,suchasdelaydurationanddelayorder,willdeterminewhethertheresultingoscillationwilldiminishovertime,reachsomesustainedamplitudeandfrequency,orexperienceanincreasingamplitude.SystemOscillationSystemOscillationSystemOscillationGiventhisstructure,Figure3showswhathappensifthegoal,DesiredSystemLevel,isincreasedby5units.Becauseofthedelay,thesystemovershootsitstargetatabouttime=6,reachesamaximumof22unitsattime=10,thenundershootsatabouttime=12,andsoon.Eventually,thesystemdoesfinditstarget50timeunitsafterthegoalchanged!SystemOscillationTypesofOscillationsInformationdelaysareprevalentintherealworldsoit'snosurprisethatoscillationsarethemostcommondynamicbehaviors.Therearefourbasictypes:SustainedOscillation;DampedOscillation;ExplodedOscillation;ChaosOscillation.SustainedOscillation·CharacterizedbyperiodicityofoneConstantamplitudeNostablefixedequilibriumisreachedDampedOscillation·AmplitudeofoscillationdissipatesovertimeStableequilibriumisreachedSystemOscillationExplodedOscillation·NotcommonintherealworldAmplitudeincreasesovertimeGrowtheitherleadstoasustainedpatternordeteriorationofthesystemChaosOscillation·RandomtimepaththatneverrepeatsIrregularamplitudeandinfiniteperiodChaososcillationsmaybefoundinmanynaturalsystemssuchasweatherpatternsSystemOscillation

Systemoscillationiscreatedwhendecisioninformationinnegativefeedbackstructuresisdelayed.SystemOscillation的原因S-ShapedGrowth

S-shapedgrowthisthecharacteristicbehaviorofasysteminwhichapositiveandnegativefeedbackstructurefightfordominancebutresultinlong-runequilibrium.thepositivefeedbackexponentialgrowthloop(showninblue)initiallydominatesthesystem,causingthesystemvariabletoincreaseatanincreasingrate.However,asthesystemapproachesitslimitor"carryingcapacity",thegoal-seekingnegativefeedbackloopbecomesthedominantloop.Inthisnewmode,thesystemapproachesstasisasymptotically.This"shiftingdominance"eventfromonelooptoanotherischaracteristicofnonlinearbehaviorandisfoundinavarietyofphysical,social,andeconomicsystems.S-ShapedGrowthS-ShapedGrowthS-ShapedGrowthS-ShapedGrowthThepositiveBirthRateloop,showninblue,isresponsiblefortheexponentialgrowthexperiencedduringtheearlystagesofherdgrowth.Asthepopulationoftheherdincreases,thebirthrate-givensomeFecundity(fertility)constant-increases.Increasingthebirthratefurtherpromotesthegrowthoftheherd.Inbalance,thenegativeDeathRateloop,showninred,isresponsiblefordrawingdowntheherdpopulation.S-ShapedGrowthS-ShapedGrowthO

溫馨提示

  • 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
  • 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
  • 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負責(zé)。
  • 6. 下載文件中如有侵權(quán)或不適當內(nèi)容,請與我們聯(lián)系,我們立即糾正。
  • 7. 本站不保證下載資源的準確性、安全性和完整性, 同時也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。

評論

0/150

提交評論