This paper introduces a geometric method for proving ergodicity of degenerate noise driven stochastic processes. The driving noise is assumed to be an arbitrary Levy process with non-degenerate diffusion component (but that may be applied to a single degre
wouldliketodetermineifthisGibbsmeasureistheunique,invariantmeasureofthesystem.
AnewstrategybasedontheintroductionoftheasymptoticallystrongFellerpropertyhasbeenintroducedin[6].Thispaperproposesanalternativemethodbasedontwoconditions:weakirreducibilityandclosureundersecondrandomizationofthestochasticforcing(seetheorem3.1).OurstrategyisinsubstancesimilartotheoneproposedbyMeynandTweediefordiscreteMarkovChainsinchapter7of[9].
AlthoughtheH¨ormandercondition([11]38.16)canalsobeusedtoobtainlocalreg-ularitypropertiesofthesemi-group,hencealocalstrongFellerconditionandergodicproperties.Thealternativeapproachproposedheredoesn’trequiresmoothvector eldsormanifolds,itcandirectlybeappliedtoLevyprocessesand(thisisourmainmoti-vation)itallowsforanexplicitgeometricunderstandingofthemechanismssupportingergodicity.
Forrelatedpreviousworkwereferto[8],[7],[6],[3],[2],[5]and[4].
2Generalsetup.
Let(Xt)t∈R+beaMarkovstochasticprocessona(separable)manifoldMwithmodelspaceRn.
Let(ωt)0≤tbep-dimensionalLevyprocess,i.e.astochasticprocessonRpthathashasindependentincrements,isstationary,isstochasticallycontinuousandsuchthat(almostsurely)trajectoriesarecontinuousfromtheleftandwithlimitsfromtheright.
Weassumethatthereexistsafamilydeterministicmappings(indexedby0≤t)Ft:M×([0,t]→Rp)→Msuchthat
(2.1)Xt=Ft sXs,(ωs ωs)s≤s ≤t
Recallthatthe rstthreeconditionde ningaLevyprocessmeanthat(ωt ωs)t≥sisindependentof(ωs )0≤s ≤s),thelawofωt ωsdependsonlyont sandlims→0P[|ωs+t ωt|≥ ]=0.
Recallalso[11,12]thatsinceωisaLevyprocess,thereexistsaγ∈R,aconstantp×pmatrixσ,astandardp-dimensionalBrownianMotion(Bt)t≥0and( t)t≥0anindependentofjumpswithintensityofmeasuredt×ν(dx)ondt×Rp Poissonprocess
(suchthatRpmin(1,|zp|)ν(dz)<∞)suchthat
ωt=γt+σBt+Ct+Mt
(2.2)
WhereCt=s≤t s1| s|>1isacompoundPoissonpointprocess(ofjumpsofnormlargerthanone)and
Mt=lim s1 <| s|≤1 tzν(dz)(2.3) ↓0z∈Rp: <|z|≤1
isamartingale(ofsmalljumpscompensatedbyalineardrift).Recallalsothatanyprocessthatcanberepresentedas(2.2)isap-dimensionalLevyprocess,inparticulara
2
Subgeometric ergodicity of strong Markov processes We derive sufficient conditions...Markov processes, subgeometric f -ergodicity, drift criterion, Langevin di?...
Langevin diffusion dL t = dW t + 1 2 r log (L t)dt: We find ...processes, geometric ergodicity SHORT TITLE: CONVERGENCE OF LANGEVIN DIFFUSIONS ...
Ergodicity of Stochastic Differential Equations Driven...one attempts to derive the Langevin equation from...Self-similar stochastic processes like fractional ...
Ergodicity of Stochastic Differential Equations Driven...one attempts to derive the Langevin equation from...Self-similar stochastic processes like fractional ...
热门文档