模糊神經(jīng)網(wǎng)絡(luò)技術(shù)在電梯群調(diào)度中的應(yīng)用.pdf

1、廣西師范大學碩士學位論文模糊神經(jīng)網(wǎng)絡(luò)技術(shù)在電梯群調(diào)度中的應(yīng)用姓名：鄧宗明申請學位級別：碩士專業(yè)：電路與系統(tǒng)指導教師：陳元琰20010101TheApplicationofFuzzyNeuralNetworkTechnologyinElevatorGroupDispatchAbstractwithmofeintelligentbuildingsHareconstructedandbeaer$elXicewasneedtheimprovem

2、entofElevatorGroupDispatchhasbecomeallemergentproblemManmethodssuchasFuzz3ControlApproachandNeuralNe“vorkApproacharealrea七usedinelevatorgroupdispatchBecauseofnon—linearrandomandambiguousfeaturesoftheelevatorgroupdispatch

3、therearevariousdrawbacks西thesemethodsCombiningFuzz7InferenceHj曲NeuralNetxvorkthefuzzyneuralnetworktechnologyhasgreatsuperiorit3inprocessingnonlinearrandomandambiguousproblemsSotheapplicationoffizZyneuralnemorkinelexatorg

4、roupdispatchbecomesoneoffocusinthefieldlnhomeButthereisfe’、thesestomtraducefuzzy“neuralnetlvorktechnologyaboutselectingelevatorproblemintheel口atorgroupdispatchSothethesismakesaprimitivestubindtisfieldTealizedispatchtheSe

5、rialCommunicationbetweenEpignComputerandPLCmustbeillstanhievedDtmtothepoorcalculaftOllcapabilit3andmutualit、ofthePLCwIlichcon口oIselevators，theEpignyComputermustrealizethecomplicateddispatchcalculationandthesimulationofth

6、eelevatormoving，collecttherunningdataoftheelevatorsandallocaterespondingelevatorThereforethefeatureofthesenajcommunicationmtheelevatorgroupdispatcharea_nalyzedhathethesis，suchascommunicationparametersettingonesidecommuni

7、cationprogrammingandrealtimeabilityandreliabilityofcommunicationAccordingtotheMEWTOCOLCOMcommunicationprotocolofFPlPLCweconslluctparallelmulticommunicationmodulenetworkstructureanduseVCtorealizemulti—threadinghigheffecuv

8、ecommumcafionneHallCallAssignmentStrate93thatisusually’usedintheelexatorgroupdispatchischosenandanal3’zedItisanEvaluationFunctionthattakesAverageWaitingTime(AWl)，LongW缸tiIIgProbabilifies(LWP)，PowerConsumptioils(I州C)asE、’

9、aluationStandards：S。=WlAwr。We’L、vPW3RNC。nethreevaluesarethekeytothef1_mctionandthecoreproblemoffuj霉rneuralnerworkapproach’stargetsAccordingtotheFuzzyRllles缸thefuzz3con訂olapproachthatcalculatethethreeevaluationstandardval

10、ues，andthelimitedlastingtimeofrespondinghallcalli11theelevatordispatchweconstructthreefoorlevelednew舭馮7neuralnetworkcalculatingtheevaluationcriteriavaluesThenetworkwithcharacterofsimplesWucturesmallcalculationandrapidari

11、thmetJcspeedrealizesthereflectionfromtherules(AWl坤andRNCltotheneuralnemarkItsfiLZZVnodeadoptsGaussMode】SubjccnonFunction，whosecentervalueand、、1dthvaltleexertsgreatinfluenceontheneBvork’soperationWansⅢissionandperformance

12、andneedsmodificafionb、，stodyhngThentheBPAlgorithmisinferredindetmlandaddingmomentumitemisusedtoa、aidlocalleastandspeeduptheconvergenceofA、erageErrorFunctionToguaranteetheph3sicalmeaningofthearea—valueandimpro、ethesignifi