基于mas的列车运行智能调整方法研究

《基于mas的列车运行智能调整方法研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、北京交通大学博士学位论文基于MAS的列车运行智能调整方法研究姓名：杜鹏申请学位级别：博士专业：交通运输规划与管理指导教师：杨肇夏20050601AbstractTrainscheduIingisoneofthemostimponantpartsofrailwaytransportationdispatchingM蚰yresearcheshavebeencarnedoutinthisareaaswenaslotsofachievemems，whichhavebeenpIayingimportantrolesinrailwaytransponation．Asthenationaleco

2、nomyiskeepingprogressingandthemarket·orientedsystemisc。ntinuallyimprOVing，thedemandofrailwaytransponationismuchhigherandthepressureontrajnschedulingismuchharderthanbefore．HoweveLitjsdi踊cultfortraditionaImodels，la唱eIybasedoncentral—controllingmethods，toc。pewiththeproblem’seVerincreasingscaIe，co

3、mpIexity，anddemandondynamicfeature．HowtoimprovetheinteIllgenceandreducethelabourintensionofdispatchersisstillahotspotintheprocessofrailwaymodemjzation．Byapplylngmulti—agenttheoryintotheresearchofthisarea，MAS_basedinte川gentschedulingmethod，ispresented．ThemethodisbasedOndist“buteddecisjon·makjn昏

4、whjchproVidesanewideatothesoIVingofla唱e-scalenetwork-basedtrainschedulingproblem，whichisdifnculIfortraditlonalcenlralizeddecjsion-makingmethods．HerethetrainschedulingprobIemisabstractedintOaprobIemofa盯0cdfi略矗m打ed理s0“rcesrom“打fp把“se髂．ThefUnctionofcemraIlzeddecision-Ⅱlakingisdistributedintos协tiO

5、ns／如eldsandtrains，whicharethenlocaldecisionmakerS，respOnsibleforlocalplans，thecOmbinatiOnOfwhichisthetoalplan．LOcaldecisiOnmakersmakedecisionsonitsowntosomeextent，whichistheself-regulatingmechanism；forthosepIanswhichareoutofslngledecision-maker’sabilityreIeVantlocaldecision—makersnegotiatesoaS

6、toreachagreement，whichisnegotiatjonmechanism；andthenegotiationprocessmaybeadjustedbyamedjatoLwhichismediatiOnmechanism．Therefore，inthedistr．buteddecisionmOdeI，thediagramOftrainschedulingisnOtdIanedcentrally'butinstead，producedbythecoIIectiOnOfagroupOf”Iativeautonomousdecisionmakers．Thetrainsch

7、edulingproblemisagentizedusingBDImodeI．RegardingtheshonageoffbnTlalizatiOn，ZlanguageisappliedforagentmOdeldesriptjon．Statjonsa11dneldsareconsidered髂resourceagents，whoseresOurcesaresidingsinstatiOnand仃acksinneld．Itjsde矗nedinthemodelBeIie