欢迎来到天天文库
浏览记录
ID:34552701
大小:2.02 MB
页数:98页
时间:2019-03-07
《基于蚁群算法的非结构化p2p搜索技术的研究》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、AbstractPeer-to-peer(P2P)networkhasbecomeanimportantapplicationinInternet.P2PnetworkcollectstheresourceonthenodesattheedgeofInternettoprovidestrongcomputingpowerandstoragecapability.Becauseofitsflexibilityonquerypatternandadaptabilitytodynamicenvironment,unstructuredP2Pnetworki
2、swidelydeployedandbecomethemainstreamofthecurrentP2Pnetworks.However,unstructuredP2Pnetworkssufferfrompoorperformance,thereforeimprovingsearchingalgorithmisoneimportantmajorresearchareaforunstructuredP2Pnetwork.Unfortunately,existingP2Psearchingmechanismsareusuallydissatisfied.
3、Forexample,structuredP2PsystemsareefficientbutlackofactualimplementsontheInternet,becauseoftheircomplicatedstructures.UnstructuredP2Psystemsareinefficientbutmorepopular.Fromvariousperspectives,ourworkfocusesonhowtoimproveretrievalefficiencyofunstructuredP2Pfile-sharingsystems.I
4、nthispaper,wepresentanewapproachtoP2Pinformationretrieval—ACODA,usingantcolonyalgorithmtosolvetheproblemsuchashugemessagepackets,lowsearchefficiency,redundancyroutesandothers.Thealgorithmsimulatestheprocessoftheants’searchingfoodandcandirectthequeryroutingefficientlyaccordingto
5、thepositivefeedbackprincipleoftheantpheromone.Itsendsthequerymessagestotheareawheretherecouldbetherequestedresourcestoimprovethesearchefficiency.Besides,informationrecommendationservicesbasedonACODAisintroducedinpaper.Thepeertransfersinformationperiodicallytootherpeersaccording
6、totheirinterest.Bythismethoditcanraisethefile-sharinglevelandreduceblindsearches.Averify-renascencemethodisprovidedinpaperto万方数据improvetheperformanceofindexcaching.Thismethodcanefficientlysolvetheinvalidindicesproblemandits’expansionproblem.Itcanimprovethesearchefficiencybyincr
7、easingthelifetimeofthevalidindex.ThesimulationbasedonPeerSimsoftwareshows,theinformationretrievestrategyofACODAalgorithmwithinformationrecommendationservicesandindexcachingcanefficientlyreducethetimeoffindingtheresources.Itcanachieveabettersearchperformanceandlowthenetworkloads
8、byincreasingtheresourcehitratioandreducingthequerymess
此文档下载收益归作者所有