Existence and uniqueness of a martingale problem in D(R_ +, L')

Let \(\Phi\) be a nuclear space and \(\Phi'\) be its dual. Let \((B_ t)_{t\in R_ +}\) be a family of continuous bilinear forms \(B_ t: \Phi\times\Phi\to R\) which are positive semidefinite and also fulfill: (i) \(t\to B_ t(\varphi,\varphi)\) is continuous and increasing for every \(\varphi\), (ii) \(B_ 0(\varphi,\varphi)=0\) for every \(\varphi\in\Phi\). Let \(A_ t(\varphi)\) be the quadratic form \(A_ t(\varphi)=B_ t(\varphi,\varphi)\). Now let \(\Omega=D(R_ +,\Phi')\) be the family of all right continuous functions \(w: R_ +\to\Phi'\) and \(X_ t(w)=w(t)\) be the canonical process adapted with respect to the usual filtration \({\mathcal D}_ t=\bigcap_{s>t}\sigma(X_ u;\;u\leq t)\). Finally, let \({\mathcal K}=\bigvee_{t\geq 0}{\mathcal D}_ t\). The author proves that on \((\Omega,{\mathcal K})\) there exists exactly one probability which solves the following martingale problem: (i) \(P(C(R_ +,\Phi'))=1\) (i.e. \(w_ t\) is continuous), (ii) \(w(0)=0\pmod P\), (iii) for every \(\varphi\in\Phi\) the process \(y_ t(w)=w_ t(\varphi)\) is a martingale with respect to \(P\) and its increasing process \(y_ t\) is exactly \(A(t,\varphi)\). The proof relies on a previous result of the author [REBRAPE 2, No. 2, 81-90 (1988; Zbl 0721.60025)] and is hardly intelligible.

Recommendations

Cites work

Cited in

(4)

This page was built for publication: Existence and uniqueness of a martingale problem in \(D({{\mathbb{R}{}}_ +},{\mathcal L}')\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1812902)