A class of integration by parts formulae in stochastic analysis. I
File(s)1911.09733v1.pdf (178.03 KB)
Accepted version
Author(s)
Elworthy, KD
Li, Xue-Mei
Type
Chapter
Abstract
Consider a Stratonovich stochastic differential equation
dχt=X(χt)odBt+A(χt)dt
(1.1)
with C∞ coefficients on a compact Riemannian manifold M, with associated differential generator A=12ΔM+Z and solution flow {ξt : t ≥ 0} of random smooth diffeomorphisms of M. Let Tξt: TM → TM be the induced map on the tangent bundle of M obtained by differentiating ξt with respect to the initial point. Following an observation by A. Thalmaier we extend the basic formula of [EL94] to obtain
Edf(Tξ.(h.))=EF(ξ.(χ))∫T0⟨Tξs(h˙s),X(ξs(χ))dBs⟩
(1.2)
where F∈FC∞b(Cχ(M)), the space of smooth cylindrical functions on the space C x (M) of continuous paths γ : [0,T] → M with γ(0) = x, dF is its derivative, and h. is a suitable adapted process with sample paths in the Cameron-Martin space L
2,1
0
([0,T];T x M).Set F
x
t
= σ{ξs(x) : 0 ≤ s ≤ t} Taking conditional expectation with respect to.F
x
T
, formula (1.2) yields integration by parts formulae on C x (M) of the form
EdF(γ)(V¯¯¯¯h)=EF(γ)δV¯¯¯¯h(γ)
(1.3)
where V¯¯¯¯h is the vector field on C x(M)
V¯¯¯¯h(γ)t−E{Tξt(ht)|ξ.(χ)=γ}
and δV¯¯¯¯h:Cx(M)→ is given by
δV¯¯¯¯h(γ)=IE{∫T0<Tξs(h˙s),X(ξs(x))dBs>|ξ.(x)=γ}
dχt=X(χt)odBt+A(χt)dt
(1.1)
with C∞ coefficients on a compact Riemannian manifold M, with associated differential generator A=12ΔM+Z and solution flow {ξt : t ≥ 0} of random smooth diffeomorphisms of M. Let Tξt: TM → TM be the induced map on the tangent bundle of M obtained by differentiating ξt with respect to the initial point. Following an observation by A. Thalmaier we extend the basic formula of [EL94] to obtain
Edf(Tξ.(h.))=EF(ξ.(χ))∫T0⟨Tξs(h˙s),X(ξs(χ))dBs⟩
(1.2)
where F∈FC∞b(Cχ(M)), the space of smooth cylindrical functions on the space C x (M) of continuous paths γ : [0,T] → M with γ(0) = x, dF is its derivative, and h. is a suitable adapted process with sample paths in the Cameron-Martin space L
2,1
0
([0,T];T x M).Set F
x
t
= σ{ξs(x) : 0 ≤ s ≤ t} Taking conditional expectation with respect to.F
x
T
, formula (1.2) yields integration by parts formulae on C x (M) of the form
EdF(γ)(V¯¯¯¯h)=EF(γ)δV¯¯¯¯h(γ)
(1.3)
where V¯¯¯¯h is the vector field on C x(M)
V¯¯¯¯h(γ)t−E{Tξt(ht)|ξ.(χ)=γ}
and δV¯¯¯¯h:Cx(M)→ is given by
δV¯¯¯¯h(γ)=IE{∫T0<Tξs(h˙s),X(ξs(x))dBs>|ξ.(x)=γ}
Date Issued
1996
Citation
Itô’s stochastic calculus and probability theory, 1996, pp.15-30
ISBN
978-4-431-68532-6
Publisher
Springer, Tokyo
Start Page
15
End Page
30
Journal / Book Title
Itô’s stochastic calculus and probability theory
Copyright Statement
© 1996 Springer-Verlag Tokyo. The final publication is available at Springer via https://doi.org/10.1007/978-4-431-68532-6_2
Identifier
https://link.springer.com/chapter/10.1007/978-4-431-68532-6_2
Subjects
math.PR
math.PR
Notes
mrclass: 60H10 (60H05) mrnumber: 1439515 mrreviewer: Hélène Airault
Publication Status
Published