基于T-S模糊控制理論的網絡擁塞控制算法研究.pdf

1、南京理工大學碩士學位論文基于TS模糊控制理論的網絡擁塞控制算法研究姓名：李彬申請學位級別：碩士專業(yè)：控制理論與控制工程指導教師：孫金生20100620碩士論文基于TS模糊控制理論的網絡擁寒控制算法研究AbstractWiththerapiddevelopmentofthecomputernetwork，soaringincreaseoftheIntemetUSerSandconstantappearanceofthenewnetwork

2、applications，havebroughtaseriesofproblemstothenetwork’Snormaloperation，oneofthemostseriousproblemisnetworkcongestionV～／hellthenetworkcongestionOccurs，itwillleadtothenetworkthroughput’Sdecrease，thedelay’Sincreaseandthepac

3、ketloss’Sincrease，seriouslyevencausenetworkcollapseTherefore，inordertopreventandcontrolthenetworkcongestion，weneedtodesigntheeffectivenetworkcongestioncontrolalgorithmCurrentlythenetworkcongestioncontrolalgorithmistheimp

4、ortantmeansforimprovingthenetwork’Sperformance，ensuringthenetwork’Sstableoperationandenhancingthenetwork’SQoS(Qualityofservice)Inrecentyears，expertsandscholarshaveachievedlotsofachievementinthefieldofnetworkcongestioncon

5、trolalgorithm’SresearchDuetotheend—toendTCPcongestioncontrolhascertainlimitation，ActiveQueueManagement(AQM)algorithmhasbecomethecurrentresearchhotspotinthenetworkcongestioncontrol’SfieldThisarticlemainlyresearchesthenetw

6、orkcongestionfromtheaspectofAQMalgorithmThemainresearchresultsareasfollows：(1)AnAQMalgorithmisproposedbasedonTS(TakagiSugeno)fuzzymodelForthenonlinearmodelofthenetworkcongestioncontrolsystem，theTSfuzzymodelofthecongestio

7、ncontrolsystemisestablishedforusingTSfuzzymodel’SfeaturewhichCanapproximatenonlinearsystemsverywell，andthestatefeedbackAQMcontrolalgorithmisdesigned，theconditionsofthesystem’Sstabilityaregiven，andthedesignmethodoftheAQMc

8、ontroller’SparametersisgivenbyusingLMI(LinearMatrixInequality)techologySimulationresultsshowthatthisalgorithmCallmaintainthequeuelengtharoundthegiventarget，hasgoodstabilityandrobustnessComparisonwithotherAQMschemes，sucha

9、sPI，REMandARED，hasdemostratedthesuperiorityofthisalgorithminconvergencerate，stabilityrobustnessandthefluctuationofqueuelength(2)AsingleneuronadaptivePIDcontrolalgorithmisdesignedbasedonTSfuzzymodelThroughintroducingtheTS

10、fuzzymodeltoadjusttheneuron’Sgain，thesingleneuronadaptivePIDcontrolalgorithmhasthefuctionthatthegaincanautotuneSimulationresultsshowthat，comparisonwiththeimprovingsingleneuronadaptivePIDcontrolalgorithm，thisalgorithmhasb