logo资料库

StochasticNetworkOptimizationwithApplicationtoCommunicationandQueueingSystems.pdf

第1页 / 共212页
第2页 / 共212页
第3页 / 共212页
第4页 / 共212页
第5页 / 共212页
第6页 / 共212页
第7页 / 共212页
第8页 / 共212页
资料共212页,剩余部分请下载后查看
StochasticNetworkOptimizationwithApplicationtoCommunicationandQueueingSystems
Copyright©2010byMorgan&ClaypoolAllrightsreserved.Nopartofthispublicationmaybereproduced,storedinaretrievalsystem,ortransmittedinanyformorbyanymeans—electronic,mechanical,photocopy,recording,oranyotherexceptforbriefquotationsinprintedreviews,withoutthepriorpermissionofthepublisher.StochasticNetworkOptimizationwithApplicationtoCommunicationandQueueingSystemsMichaelJ.Neelywww.morganclaypool.comISBN:9781608454556paperbackISBN:9781608454563ebookDOI10.2200/S00271ED1V01Y201006CNT007APublicationintheMorgan&ClaypoolPublishersseriesSYNTHESISLECTURESONCOMMUNICATIONNETWORKSLecture#7SeriesEditor:JeanWalrand,UniversityofCalifornia,BerkeleySeriesISSNSynthesisLecturesonCommunicationNetworksPrint1935-4185Electronic1935-4193Thismaterialissupportedinpartbyoneormoreofthefollowing:theDARPAIT-MANETprogramgrantW911NF-07-0028,theNSFCareergrantCCF-0747525,andcontinuingthroughparticipationintheNetworkScienceCollaborativeTechnologyAlliancesponsoredbytheU.S.ArmyResearchLaboratory.
SynthesisLecturesonCommunicationNetworksEditorJeanWalrand,UniversityofCalifornia,BerkeleySynthesisLecturesonCommunicationNetworksisanongoingseriesof50-to100-pagepublicationsontopicsonthedesign,implementation,andmanagementofcommunicationnetworks.Eachlectureisaself-containedpresentationofonetopicbyaleadingexpert.Thetopicsrangefromalgorithmstohardwareimplementationsandcoverabroadspectrumofissuesfromsecuritytomultiple-accessprotocols.Theseriesaddressestechnologiesfromsensornetworkstorecon�gurableopticalnetworks.Theseriesisdesignedto:•Providethebestavailablepresentationsofimportantaspectsofcommunicationnetworks.•Helpengineersandadvancedstudentskeepupwithrecentdevelopmentsinarapidlyevolvingtechnology.•Facilitatethedevelopmentofcoursesinthis�eld.StochasticNetworkOptimizationwithApplicationtoCommunicationandQueueingSystemsMichaelJ.Neely2010SchedulingandCongestionControlforWirelessandProcessingNetworksLibinJiangandJeanWalrand2010PerformanceModelingofCommunicationNetworkswithMarkovChainsJeonghoonMo2010CommunicationNetworks:AConciseIntroductionJeanWalrandandShyamParekh2010PathProblemsinNetworksJohnS.BarasandGeorgeTheodorakopoulos2010
ivPerformanceModeling,LossNetworks,andStatisticalMultiplexingRaviR.Mazumdar2009NetworkSimulationRichardM.Fujimoto,KalyanS.Perumalla,andGeorgeF.Riley2006
StochasticNetworkOptimizationwithApplicationtoCommunicationandQueueingSystemsMichaelJ.NeelyUniversityofSouthernCaliforniaSYNTHESISLECTURESONCOMMUNICATIONNETWORKS#7CM&cLaypoolMorganpublishers&
ABSTRACTThistextpresentsamoderntheoryofanalysis,control,andoptimizationfordynamicnetworks.MathematicaltechniquesofLyapunovdriftandLyapunovoptimizationaredevelopedandshowntoenableconstrainedoptimizationoftimeaveragesingeneralstochasticsystems.Thefocusisoncommunicationandqueueingsystems,includingwirelessnetworkswithtime-varyingchannels,mobility,andrandomlyarrivingtraf�c.Asimpledrift-plus-penaltyframeworkisusedtooptimizetimeaveragessuchasthroughput,throughput-utility,power,anddistortion.Explicitperformance-delaytradeoffsareprovidedtoillustratethecostofapproachingoptimality.Thistheoryisalsoapplicabletoproblemsinoperationsresearchandeconomics,whereenergy-ef�cientandpro�t-maximizingdecisionsmustbemadewithoutknowingthefuture.Topicsinthetextincludethefollowing:•Queuestabilitytheory•Backpressure,max-weight,andvirtualqueuemethods•Primal-dualmethodsfornon-convexstochasticutilitymaximization•Universalschedulingtheoryforarbitrarysamplepaths•Approximateandrandomizedschedulingtheory•OptimizationofrenewalsystemsandMarkovdecisionsystemsDetailedexamplesandnumerousproblemsetquestionsareprovidedtoreinforcethemainconcepts.KEYWORDSdynamicscheduling,decisiontheory,wirelessnetworks,Lyapunovoptimization,con-gestioncontrol,fairness,networkutilitymaximization,multi-hop,mobilenetworks,routing,backpressure,max-weight,virtualqueues
viiContentsPreface..................................................................xi1Introduction..............................................................11.1ExampleOpportunisticSchedulingProblem................................11.1.1ExampleProblem1:MinimizingTimeAveragePowerSubjecttoStability.........................................................21.1.2ExampleProblem2:MaximizingThroughputSubjecttoTimeAveragePowerConstraints.........................................21.1.3ExampleProblem3:MaximizingThroughput-UtilitySubjecttoTimeAveragePowerConstraints.........................................31.2GeneralStochasticOptimizationProblems.................................41.3LyapunovDriftandLyapunovOptimization...............................51.4DifferencesfromourEarlierText.........................................71.5AlternativeApproaches..................................................71.6OnGeneralMarkovDecisionProblems....................................81.7OnNetworkDelay.....................................................91.7.1DelayandDynamicProgramming..................................91.7.2OptimalO(√V)andO(log(V))delaytradeoffs......................91.7.3Delay-optimalAlgorithmsforSymmetricNetworks..................101.7.4Order-optimalDelaySchedulingandQueueGrouping...............101.7.5HeavyTraf�candDecayExponents................................111.7.6CapacityandDelayTradeoffsforMobileNetworks..................111.8Preliminaries..........................................................122IntroductiontoQueues..................................................152.1RateStability.........................................................172.2StrongerFormsofStability..............................................182.3RandomizedSchedulingforRateStability................................192.3.1A3-Queue,2-ServerExample....................................202.3.2A2-QueueOpportunisticSchedulingExample......................222.4Exercises.............................................................25
分享到:
收藏