欢迎来到天天文库
浏览记录
ID:36536574
大小:3.29 MB
页数:98页
时间:2019-05-11
《分布式网络管理模型及建模方法》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、东南大学博士学位论文分布式网络管理模型及建模方法姓名:李伟申请学位级别:博士专业:计算机应用技术指导教师:罗军舟20070101ABSTRACTAtpresent,networkmanagementisdevelopingfromthecentralizedmanagementtothedistributedmanagement.Applyingthemulti—agentsystemtonetworkmanagementisanoveltrainofthoughtofconstructingthenewgenerationnetworkmanagementsystem.Ho
2、wever,networkmanagementbasedonmulti·agentisstillontheprobingstageandmostoftheresearchesfocusonhowtou∞mobileagenttoconstructdistributednetworkmanagementsystem.Thoughcurrentresearcheseffectivelyimprovethenetworkmanagementsystemintheaspectofdistributedmanagement,dynamicupgradeandfunctionexpans
3、ion,theylacksubstantialsupportsfornetworkmanagementintelligence.Furthermore,mostofthenetworkmanagementmodelsaredescribedbytheinformalmethod,whichcallnotremovetheinconformityandthedifferentmeaningsinthedescription.Thisleadstothedifferentmeaningsandthearbitrarinessthatthepeopleunderstandthesp
4、ecificationanddevelopthesystem.Inordertosolvetheabove-mentionedproblems,basedonthethoughtofCompumrSupportedCooperativeWork,thedissertationresearchesthedistribtedintelligentnetworkmanagementmodelforthelarge—scalenetworkanditsformalmodeling.Firstly,throughtheresearchontheapplicabilityofthemul
5、it-agentsysteminthenetworkmanagement,Multi—agentbasedDistributedIntelligentNetworkManagementmodel(MADINM)forthelarge—scalecomputernetwork,whichutilizeskeycharacteristicsoftheagent,isputforward.Thismodeladoptingthepoliciesbasedonmanagementdomainnotonlyrealizesthedistributednetworkmanagementb
6、utalsoprovidesnetworkmanagementwithmoreintelligence.TheresultsofanalysisintheoryandexperimentindicatethatMADINMisbetterthantheexistingnetworkmanagementsolutionsonperformancefurthelarge·scalecomputernetwork.Secondly,throughtheresearchontheself-studymechanismofintelligentagentbasedondataminin
7、g,anovelsequentialpatternsminingalgorithm,namelyLocationIndex-ProjcotedSequentialPatternMining(LiSpan),isprovided.BasedonthethoughtonPattern·GrowthandDivide·and—Conquer,thisalgorithmavoidsscaningtheprojecteddatabasesandreducesthesystemoverloadbyusingthel
此文档下载收益归作者所有