第25卷第1期2023年3月测绘技术装备GeomaticsTechnologyandEquipmentVoe.25No.1Mae.2023马尔可夫链理论与灰色GM(1"1)组合模型预测地表变形张西步，栾洲，王义昌(山东科技大学测绘与空间信息学院，山东青岛266000)摘要：为了提高矿产开采过程中地表沉降变形的预测精度，本文结合灰色理论和马尔可夫链理论,建立了一种组合预测模型。首先，以GM(1,1)预测模型为基础，采用马尔可夫链理论和残差修正进行优化，结合工程验证，证明了组合模型精度可显著提高地表沉降变形预测精度；其次,将组合模型应用于实际矿产生产的变形监测中，通过工程实例分析，组合模型预测值的平均相对误差仅为3.26%，且模型精度达到1级。试验结果表明，组合模型的预测值与实际值的误差较传统的灰色预测模型有明显减小，结果更接近于实际情况，证明了本文模型在矿区沉降分析与预测中的合理性#关键词：GM(1，1)模型；马尔可夫链理论;组合模型；残差修正；沉降预测中图分类号:TD325.2DOI：10.20006/j.cnki.61-1363/P.2023.01.008PrePictionofsurfacedeformationbyMarkovchaintheoryandgreyGM(1,1)combinermodelZHANGXibu,LUANZhou,WANGYichang(ShandongUniversityofScienceandTechnology,ColleyeofGeodesyandGeomatics,Qingdao,Shandong266000,China)Abstract：InoCertoimprovethepredictionaccuccyofsurfacesubsdencedeformationintheprocessofmining,acombinedpredictionmodelisestablishedbycombininggcythereandMarkovchaintheore.FDstly,theMarkovchaintheoreandresidualcofectionwereadoptedOooptimizationandcombinedwithengineeringverificationbasedonGM(1,1)predictionmodel,itwasprovedthattheaccuccyolcombinedmodelcansienificantlyimprovethepredictionaccuracyofsurfacesubsidencedeformation.Secondly,thecombinedmodelwasusedinthedeformationmonitoringofactualminingproduction.Byanalyzingtheengineeringexamples,theaveragerelativeerrocofthepredictedveluvofthecombinedmodelisonly3.26%,andthemodelaccuracyreacheslevel1.ThetestresultsshowthattheefOTbetweenthepredictedveluvandtheactualveluvofthecombinedmodelissigniDcantlyreducedcomparedwiththatofthetraditionalgreypredictionmodel,andtheresultsaremoreconsistenttotheactualsituation,whichprovestherationalityoftheproposedmodelinsubsidenceanalysisandpcdictioninmonongaeea.Keywords:GM...
