适性化多代理人网际网路环境资讯侦搜.ppt
上传人:王秋****哥哥 上传时间:2024-09-10 格式:PPT 页数:20 大小:424KB 金币:10 举报 版权申诉
预览加载中,请您耐心等待几秒...

适性化多代理人网际网路环境资讯侦搜.ppt

适性化多代理人网际网路环境资讯侦搜.ppt

预览

免费试读已结束,剩余 10 页请下载文档后查看

10 金币

下载此文档

如果您无法下载资料,请参考说明:

1、部分资料下载需要金币,请确保您的账户上有足够的金币

2、已购买过的文档,再次下载不重复扣费

3、资料包下载后请先用软件解压,在使用对应软件打开

適性化多代理人網際網路環境資訊偵搜CollaborativeMultiagentAdaptationforBusinessEnvironmentalScanningthroughtheInternetOutlineIntroductionUCRES:User-centered,Continuous,andResource-boundedESGoal&clallengesFindingIOIcontinuouslyTimeliness(TES):minimizingtheaveragetimedelayoffindingIOICompleteness(CES):maximizingthepercentageofIOIfoundControllingtheresourceconsumedEffectiveness(EES):maximizingthepossibilityoffindingIOIineachinquiryofdataConsideringthepreferencesoftheuser(i.e.addingimportanceweightstoTES,CES,andEES)WeightedTESWeightedCESWeightedEESRelatedmultiagenttechnologyforUCRESInformationgatheringagentsAimingtosatisfyusers'"one-shot"needs,e.g.Intelligentlylocatingtheinformationwiththeadditionalconsiderationoftime/cost/qualitytradeoffs,andFilteroutirrelevantinformationButwhenandhowfrequentlytoscanforIOI?InformationmonitoringagentsAimingtomonitorapredefinedsetoftargetsperiodicallyoradaptivelyButwheretofindtheIOItomonitor?AdaptiveagentsfordecisionsupportAimingtoprovidetailoredinformationforsupportingdecisionmakingAdaptingtowhattomonitorbyobservingtheuser’spreferenceand/orproblemsolvingstrategiesButwhenandhowfrequentlytoscanforIOI?Multiagentcoordination&biddingAimingtoresolveconflictsandbuildconsensusamongtheagentsButhowtodefineacoordinationand/orbiddingprotocolforUCRES?Learningformultiagentcoordination,bidding,andorganizationAimingtolearnHowtheactionsaffecteachother,Whatinformationisrequiredforcoordination,Whentotriggeragentcoordination,Usageofwhattheagentsarebiddingfor,RestructuringoftheorganizationforcoordinationandcollaborationButhowtosimultaneouslyadapttoUser’spreferenceIOI’sdistributionintheInternetUpdatebehavioroftheIOILimitedamountofresourceMultiagentAdaptationforUCRESOverviewofAESABehaviorofeachScanningAgentBehavioroftheControllingAgentExperimentSimulatingusers’dynamicpreferencesThesystemsevaluatedAESABasicsettingForthecontrollingagent:1=2=20,=1inquirypersecond,=1inquiryper200secondsVariantsAESA-1Forthescanningagents,=0.5,=0.5,