简介:摘要细颗粒物浓度是决定大气能见度的关键因子,利用2013年11月1日~2014年3月31日期间成都市人民南路四段逐时PM2.5质量浓度和大气能见度监测数据以及温江站同期地面气象观测资料,探讨了在不同相对湿度条件下大气能见度与PM2.5质量浓度的关系。结果表明(1)成都市能见度水平整体较低,2000m以下和2000m-3000m的样本数分别为1335、542,共占有效样本数的71%;(2)PM2.5质量浓度以及相对湿度是影响能见度的主要因子,大气能见度随平均相对湿度和平均PM2.5质量浓度的增大而减小;(3)利用Mie散射理论,构建了大气能见度与PM2.5的函数关系,并进一步基于幂函数分别在RH<70%、70%≤RH<80%以及80%≤RH<90%三种相对湿度条件下对二者进行拟合研究,对应的相关系数依次为-0.41、-0.47、-0.52。
简介:Nowstudyingfordoctor'sdegreeatinstituteofSystemsScience,AcademiaSinica.TheoremDI4]Letk21beaninteger.IfGisaconnectedclaw-freegmphwithhiV(G)levenandwithminimumdegreee(G)atleastZk,thenGhasak-factor.Inthispaper,wegeneralizedtheresultofTheoremC,andobtainthefollowingTheoremifGisanN'-locallyconnectedclawtheegraphwithb(G)22,thenGhasa2-factor.2.LemmasLemma1IfGisanN'-locallyconnectedclaw-acegashwith6(G)22,thenforeachxo6V(G),Ghasashonestcyclecontainingxoandhavingatmost5venices.Lemma2IfGisanN'--locallyconnectedclaw-fr?
简介:TheemergenceofB2Belectronicmarketshasgreatlychangedtherelativebargainingpowerofbuyersandsellers.Westudytheequilibriummarketstructureinabuyer’smarket.Wefindthatbuyer-controlledB2BmarketsandneutralB2Bmarketshavedifferentequilibriumstructures,andtheemergenceofB2Bmarketswillincreasesocialwelfare,butitseffectonbuyersandsellerswillbedifferent:B2Bmarketsincreasetheconsumersurplusoftheendmarket,buttheireffectsonbuyerandsellerprofitsaremoderatedbytherelativebargainingpowerofbuyersandsellers.TheprofitsofthesidewithmuchweakerbargainingpowerwilldecreaseduetotheintroductionofB2Bmarkets.
简介:ThemixedH2/H∞guaranteedcostcontrolproblemviastatefeedbackcontrollawsisconsideredinthispaperforlinearsystemswithnorm-boundedparameteruncertainty.Basedonthelinearmatrixinequality(LMI)approach,sufficientconditionsarederivedfortheexistenceofguaranteedcostcontrollerswhihcguaranteenotonlyaprespecifiedH∞disturbanceattenuationlevelononecontrolledoutputforalladmissibleparameteruncertainties,butalsotheworst-caseH2performanceindexontheothercontrolledoutputtobenomorethanaspecifiedbound.Furthermore,aconvexoptimizationproblemisformulatedtodesignanoptimalH2/H∞guaranteedcostcontroller.