Cauchy approximation for sums of independent random variables英文文献资料.docVIP

CAUCHYAPPROXIMATIONFORSUMSOFINDEPENDENT RANDOMVARIABLES K.NEAMMANEE Received 6August 2002 We use Stein’s method to ?nd a bound for Cauchy approximation. The random variables which areconsidered need tobeindependent. 2000 Mathematics Subject Classi?cation: 60F05, 60G50. 1. Introduction. In Stein’s work [19], the aim was to show convergence in distributiontothenormal.Histechniquewasnovel.Stein’stechniquewasfree fromFourier methods andreliedinstead ontheelementary di?erential equa- tion f
(w)?wf(w)=h(x)?Nh (w∈R), (1.1) whereh:R→Rissuchthat
∞   h(x)e?(1/2)x2 dx∞ (1.2) ?∞ andNh=E(h(Z)),whereZ～N(0,1). Stein’s method was extended from normal distribution to the Poisson dis- tribution byChen[9].Stein’sequation forPoissonwithparameter λis   λf(w+1)?wf(w)=h(w)?Pλh wherePλh=E(h(Z)),Z～Poi(λ). w∈Z+ , (1.3) Sincethen,Stein’smethodhasfoundconsiderableapplicationsincombina- torics, probability, and statistics. Recent literature pertaining to this method includesArratiaetal.[1,2],BaldiandRinott[3],Barbour[4,5],Barbouretal.[6], Bolthausen andG?tze[7],Chen[10,11],GoldsteinandReinert[12],Goldstein andRinott[13],G?tze[14],andGreen[15];theworkofHolstandJanson[16] gives an excellent account of this method. In this paper, we further develop theSteintechnique tobounderrorsforaCauchyapproximation tothedistri- butionofW,thesumofindependentrandomvariables.Infact,therearesome literatures(e.g.,Boonyasombut andShapiro[8],Neammanee [17],andShapiro [18])giveaboundofCauchyapproximationinsomekindofrandomvariables. ButtheyusedFouriermethods.  1056 K. NEAMMANEE Thispaperisorganizedasfollows.MainresultsarestatedinSection2.Proof ofmainresultsisinSection3,whileanexampleisgiveninSection4. 2. Main results. At the heart of Stein’s method lies a Stein equation. For example, f
(w)?wf(w)=g(w), w∈R, λf(w+1)?wf(w)=g(w), w∈Z+ (2.1) areSteinequations fornormalandPoissondistribution, respectively.  Let?={h:R→R| ?∞∞ (|h(x)|/(1+x2))dx∞},andforeachh∈?,
1 ∞ h(x) Cau(