理学硕士学位论文马尔可夫链模型在某些经济预测中的应用姓名贺明明哈尔滨工业大学2008年6月国内图书分类号:O211.62国际图书分类号:519.217理学硕士学位论文马尔可夫链模型在某些经济预测中的应用硕士研究生:贺明明导师:王勇申请学位:理学硕士学科、专业:运筹学与控制论所在单位:理学院数学系答辩日期:2008年6月28日授予学位单位:哈尔滨工业大学ClassifiedIndex:O211.62U.D.C.:519.217DissertationfortheMasterDegreeinScienceTHEAPPLICATIONONSOMEECONOMICPREDICTIONWITHMARKOVCHAINMODELCandidate:HeMingmingSupervisor:Prof.WangYongAcademicDegreeAppliedfor:MasterofScienceSpeciality:OperationalResearchandCyberneticsUnit:DepartmentofMathematicsDateofOralExamination:June,28,2008University:HarbinInstituteofTechnology哈尔滨工业大学理学硕士学位论文摘要马尔可夫链是一种有着广泛应用的随机过程模型,它对一个系统由一种状态转移到另一种状态提出了定量分析,对定性问题的预测提供了一种思路,丰富了预测的内容。许多经济和社会现象中动态系统问题都可以采用马尔可夫链来描述。本文首先提出了两种划分状态空间的方法:样本均值均方差分级法和有序聚类法。建立了转移概率矩阵、提出了马氏性检验、介绍了两种马尔可夫链预测方法:叠加马尔可夫链预测方法以及加权马尔可夫链预测方法,并总结了这两种方法的计算步骤。然后根据全国居民消费价格指数历史资料,运用叠加马尔可夫链预测方法对全国居民消费价格指数进行了预测和分析,求得该序列的平稳分布(极限分布)与各个状态的重现期,为全国居民消费价格指数中短期预测提供了一个新途径;根据哈尔滨水稻年单位产量历史资料,运用加权马尔可夫链预测方法对哈尔滨水稻年单位产量进行预测和分析,结果表明该方法预测结果与实际情况吻合,证明了加权马尔可夫链模型的可行性和实用性。文章最后提出了灰色马尔可夫链模型,并根据全国社会消费品零售总额的历史资料,运用灰色马尔可夫链模型进行了预测。运用灰色模型预测曲线来反映全国社会消费品零售总额发展规律,马尔可夫链模型来反映波动规律。运用马尔可夫链对灰色模型预测结果进行优化,给出预测值的大体范围,结果表明两者相结合能很好地解决实际问题。关键词转移概率;叠加马尔可夫链;加权马尔可夫链;灰色马尔可夫模型-I-哈尔滨工业大学理学硕士学位论文AbstractMarkovchainisanextensivelyappliedstochasticprocessmodelwhichistoquantitativelyanalysisasystemtransferingfromonestatetoanother.Itoffersanewpathtopredictproblemandenrichesthecontentofprediction.ManydynamicsystematicproblemsintheeconomicandsocialphenomenoncanbedescribedbyMarkovchain.Thepaperfirstpresentstwokindsofmethodstothestateclassification:sampleaverage-meansquaredeviationandorderedclusteringmethod.Foundstransitionprobabilitymatrix,presentstheMarkovpropertytestofrandomvariableseriesandintroducestwokindsofMarkovchainpredictionmethods:foldadditionalMarkovchainandweightedMarkovchain,andsummarizesthecalculationstepsofthem.Theninthemainbodyofthepaper,basedonthehistorydataofcountrywideConsumerPriceIndex,usingthefoldadditionalMarkovchain,wepredictcountrywideConsumerPriceIndexandobtainit’ssmoothdistribution(terminaldistribution),recurringperiod.ThismethodprovidesanewpathtopredictConsumerPriceIndexinshort-term.Basedonthehistorydataoftherice’sunitoutputofHarbin,usingtheweightedMarkovchainmethod,wepredictthechangesofrice’sunitoutput.Itisconcludedthattheresultofpredictionbythismethodagreeswiththerealityandthismethodispracticalandfeasible.Finally,thepaperpresentsGreyMarkovmodel.Basedonthehistorydataofthetotalamountofthecountrywidecommunityconsumableretail,usingtheGreyMarkovchain,wepredictthetotalamountofconsumableretail.Greymodelforecastingcurveshowsthedevelopingofthetotalamountofconsumableretailarray,Markovmodelreflectthevolatilitylawandthenoptimizetheresult.Thescopeofpredictedvalueisgiven.Bythecombinationofthetwomentionedabove,thepracticalproblemissolved.KeywordsTransitionprobability;FoldadditionalMarkovchain;WeightedMarkovchain;GreyMarkovmodel-II-哈尔滨工业大学理学硕士学位论文目录摘要.............................................................................................................................IAbstract.......................................................................................................................II第1章绪论................................................................................................................11.1马尔可夫过程的历史背景...............................................................................11.2马尔可夫链及其应用综述...............................................................................21.2.1马尔可夫链在经济管理领域中的应用....................................................21.2.2马尔可夫链在教育领域中的应用............................................................41.2.3马尔可夫链理论在医学及自然灾害预测中的应用................................51.3主要研究内容...................................................................................................6第2章马尔可夫链的基本概念和性质....................................................................72.1马尔可夫链的定义和性质...............................................................................72.1.1马尔可夫链的定义....................................................................................72.1.2马尔可夫链的性质....................................................................................92.1.3切普曼—柯尔莫哥洛夫方程..................................................................102.2马尔可夫链中状态的分类.............................................................................112.2.1到达和互通..............................................................................................112.2.2状态空间的分解......................................................................................122.2.3常返态和非常返态..................................................................................132.3平稳分布.........................................................................................................162.3.1的渐进性质.....................................................................................16()ijpn2.3.2平稳分布..................................................................................................172.4本章小结........................................................................