文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > A Survey on Agent-Based Modelling and Equation-based Modelling
侵权投诉

A Survey on Agent-Based Modelling and Equation-based Modelling

In many domains, agent-based system modeling competes with equation-based modeling approaches that identify system variables and evaluate or integrate sets of equations relating these variables. Both approaches simulate the system how it is executed throug

suchas“novaccination”,“residualimmunity”,“vaccinated”.Theyassumethatthedisease-stagedurationsarethesameacrossgroups.Interestingly,theresultsoftheirequation-basedmodel-differentialequations-aredifferentfrompreviousworkin[4].Basicallythemodelismuchsimpleranddoesnottaketheinteractionamongpersonsintocount.Inaddition,itdoesnotconsidertheinterventionofsociety.However,themodelingprocedureisverysimilarbetweenthem.Identifythesystemvariablesaccordingtothestudyobjective.Thevariablesmayrepresentsystemstates,thenumberofagents,theprobabilityofstateprogression,lifecycles.Abstractorsimplifythestudyobjectbydeclaringsome“reasonable”assumptions.Listthescenarios,andexpressthedynamicprocessindifferentialequations.Oncethevariablesareidenti edandequationsareestablished,theremainingworkistosolvetheequationsandanalyzethesystemdynamicmodels.Thereareseveralintuitivedrag-and-dropsimulationtoolsavailable,tobediscussedinthenextsubsection.

D.Simulationtools

SMILE,“anobjectoriented,equationbased,genericsimulationenvironmentforcomplexsystems.Itiscommonlyusedforthesimulationofcomplexenergysystems,butitisalsoapowerfultoolforsolvingamultitudeofothertimedependentsimulationproblems”.Language:Objective-C,SMILE,Python.doc.guandang.net,“apowerful,interactiveenvironmentformodelingandsolvingscienti candengineeringproblemsbasedonpartialdifferentialequations”.Language:MATLAB.doc.guandang.net

III.AGENT-BASEDMODELLING

Inthissection,wediscussagent-basedmodeling.Whyweneedagent-basedmodeling?Tworeasonsexist:(1)intheexistingmodelingtechniques(equation-basedmodeling)individualpropertiescannotbefullytakenintoaccountinstatevariablemodelsand(2)wewanttounderstandhowindividualpropertiesdeterminethesystem’sproperties.Thusagent-basedmodelingismoreappropriateforproblemslikecoordinationorstrategicinteraction.Thekeyelementinagent-basedmodelingisthepossibilityofcomplexstructuresofsocialinteractions.

A.Concept

Agentbasedmodelisaspeci cindividualbasedcomputationalmodelforcomputersimulationextensivelyrelatedtothethemeincomplexsystems,emergence,MonteCarloMethod,computationalsociology,multiagentsystems,andevolutionaryprogramming[1].Themodelwasdevelopedthroughasimpleconceptualforminthelate1940sbasedontheideaofconstructingagentswithsomepropertiesandthensimulatingtheminparalleltomodeltherealphenomena.ThehistoryoftheagentbasedmodelcanbetracedbacktotheVonNeumannmachineandimprovedbyStanislawUlamandJohnConway.However,thebirthofagentbasedmodelasamodelforsocialsystemswasprimarilybroughtbyCraigReynold.Hetriedtomodeltherealityoflivelybiologicalagents(arti ciallife).Althoughagent-basedmodelisnotnew,themodelingtechniquesisarelativelynewapproachforsystemmodelingandsimulation.Itisalsocalledindividual-basedmodeling.Agent-basedmodelingisamethodtostudymacroworldthroughde ningmicrolevelofthesocialsystem–aindividualwithspeci cproperties.Sincetheearly1990s,agent-basedmodelinghasincreasinglybeenusedinmostofthesocialsciences,suchasecologicaleconomics,andecology.

B.Howtobuildanduseagent-basedmodel

Accordingto[2],theredoesnotexistawell-recognisedmethodofbuildingagentsforalargesystem.Basically,everyagentdesignhastoincludemechanismsforreceivinginput,forstoringhistory,fordevisingnextactionandcarryingouttheactionfollowedbyoutputs.Theauthorssummarizethetechniquesforconstructingagents.Oneapproachofdesignforanagentistouseproductionsystems.Suchasystemconsistsofasetofrules,aworkingmemoryandaruleinterpreter.Therulesmightcontainasetofifandelsestatements.Forexample,incrawl shpredator-preysystem,somerulesmaybelikethefollowing:IF

IF

IF

IFIIIIamhungryTHENleavesheltertolookforfoodamnexttosomefoodTHENpickitupamholdingfoodTHENIeatitseeapotentialdangerTHENreturntoshelter

Animprovementtotheproductionsystemistoallowagentstohavethecapabilityoflearning,thusprocessingtheabilityofadaptingtochangingenvironment.Thetechniquescommonlyusedforthelearningareneuralnetworksandevolutionaryalgorithm(e.g,geneticalgorithm).Thismeanstheagentsde nednotonlyusetherules,butalsoalterandde nenewrules.Comparedwithpreviousapproach,itmaybringsomeuncertainty,unpredictibilityandcomplexitytosystems.Howeverthesystemsimulatedmatchesmorecloselytothecounterpartinrealitywhenitisusedforthestudyofintelligentworldsorevents.

上一页第4页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email