16.08.2013 Views

Projection markovienne de processus stochastiques

Projection markovienne de processus stochastiques

Projection markovienne de processus stochastiques

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

tel-00766235, version 1 - 17 Dec 2012<br />

Hence, the limit when ǫ → 0 of<br />

1<br />

ǫ [Uǫ −I]Uth<br />

exists, implying that Uth belongs to the domain of A for any h ∈ D 0 .<br />

Thus,<br />

T<br />

belongs to the domain of A and<br />

T<br />

ǫ<br />

ǫ<br />

due −u Uuh<br />

due −u T<br />

UuAh = A due<br />

ǫ<br />

−u Uuh.<br />

Since U is a contraction semigroup and given the continuity property of<br />

Ut on the space D0 , one may take ǫ → 0 and T → ∞ in (2.15), leading<br />

to ∞<br />

e −t ∞<br />

Uthdt = U0 +A due −u Uuh.<br />

Thus<br />

0<br />

<br />

I −A ∞<br />

due −u Uuh(s,x) = U0h(s,x) = h(s,x),<br />

0<br />

yielding h ∈ Im(I − A). We have shown that (Ut,t ≥ 0) generates<br />

a strongly continuous contraction on C0([0,∞[×R d ) with infinitesimal<br />

generator A (see [40, Theorem 2.2, Chapter 4]). The Hille-Yosida theorem<br />

[40, Proposition 2.6, Chapter 1] then implies that for all λ > 0<br />

Im(λ−A) = C0([0,∞[×R d ).<br />

5. Now let pt(x0,dy) be another solution of (2.7). First, consi<strong>de</strong>ring equation<br />

(3.13) for the particular function g(y) = 1, yields<br />

<br />

∀t ≥ 0 pt(x0,dy) = 1,<br />

and pt(x0,dy) has mass 1.<br />

R d<br />

34<br />

0

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!