基于神经网络的维修物料需求预测系统的研究与开发

《基于神经网络的维修物料需求预测系统的研究与开发》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、上海交通大学硕士学位论文基于神经网络的维修物料需求预测系统的研究与开发姓名：薛锋申请学位级别：硕士专业：电子与通信工程指导教师：姜丽红;谢晨20060112上海交通大学工程硕士学位论文摘要物料需求预测模型综合考虑各种可供选用的神经网络模型的特点选定BP神经网络作为本文研究的重点在充分考虑维修物料的特点和影响维修物料需求的各种因素后对一系列输入参数进行有重点的筛选和量化以配合BP神经网络模型计算参数训练这样既能有效利用BP神经网络算法的优点又充分利用大量的历史数据积累从而大大提高了预测的准确性本文的意义在于基于神经网络预测技术的针对维修类业务的物料需求预测方法可以

2、为维修类业务企业的维修物料采购和维修物料库存管理提供辅助决策手段关键词维修业务物料需求预测神经网络训练优化第II页上海交通大学工程硕士学位论文ABSTRACTREARCHANDDEVELOPMENTOFREPAIRMATERIALREQUESTFORECASTSYSTEMBYNEURALNETWORKABSTRACTThesubjectofthispaperisthematerialrequestforecastsystembyneuralnetworkforrepairingbusiness.Firststageoftheresearchworkistobas

3、eonthecharacteristicofrepairingbusinesscompany,pointouttheimportanceofrepairmaterialintherepairingbusinesscompany.Thencarryoutthetheoreticalanalysisofcurrentmaterialrequestforecastsituation,describethedifficultlevelandthereasonsforsuchkindofphe-nomena.Becausetheinformationsystemsared

4、ividedinthisrepairingbusi-nesscompany,duringtheresearchprocessitisrequiredtocollectthedatawhichisstoreddifferentsystems,includingtheB/Sbasedrepairingrec-ords,theC/SbasedmanufactureengineeringinformationandalsotheTerminalbasedfinancialdata.Allthesedataneedtobecentralizedandstoredaccor

5、dingtorepairingmaterialforecastrequirements.Thispaperproposestheactivedatacollectingmethod,designsthespecificB/S,C/SandTerminalbaseddataintegrationprocedures.Afterdataintegrationisfinished,thispaperfiltersanddigitalizesthekeyinputsrelatedtotheforecastcalculationandthendesignsappropri

6、ateforecastcalculationprocess.Bycomparingtheoutputandhistoricaldata,repeattheneuralnetworkparameterstrainingprocedureuntiltheforecastoutputmeetstheexpectation.Thisresearchanddevelopmentoftherepairmaterialforecastsystemisfinishedandthissystemisimplementedforactualworking.Byusingthisre

7、pairmaterialforecastsystem,wecanhavetheaccuraterepairmate-rialforecastintime,sothiscanbeusedasthedecisionmakingaidformorereliablerepairmaterialpurchaseplanandinventorycontrol.Finallythissystemassiststhecompanytolowerthecapitaldepositforrepairmaterialinventory,reducetheoperationcostan

8、dcontroltheo