Exercise about strong Markov property and measurability

34 Views Asked by At

I'm trying to solve the first question of the exercise 3.19 from the Daniel Revuz, Continuous Martingale and Brownian motion book (strong Markov property section). The text is the following.

Let $X$ be a Feller process, $T$ a finite stopping time. Prove that any $\mathcal{F}_{\infty}$-measurable and positive r.v. $Z$ may be written as $\phi(\omega, \theta_T(\omega))$ where $\phi$ is $\mathcal{F}_T \otimes \mathcal{F}_\infty$-measurable. Then $$ $$ \begin{equation} \tag{1}\label{eq:id} E_\nu [Z | \mathcal{F}_T](\omega) = \int \phi(\omega,\omega') P_{X_T(\omega)}(\text{d}\omega') \quad P_\nu-\text{a.s.} \end{equation} where

  • $X$ is the canonical process $\Omega := E^{\mathbb{R}_+} \to E$ where $E$ is the state space.
  • $\mathcal{F}_\infty$ is the $\sigma$-algebra over $\Omega$.
  • $(\mathcal{F}_t)_{t \geq 0} := \sigma(X_u : u \leq t)$
  • $T$ is a stopping time wrt to $(\mathcal{F}_t)$
  • $\theta_T$ is the shift operator over $\Omega$

I'm not really sure about how to proceed.

  • About writing $Z$ as $\phi$ I suppose is something similar to the Doob–Dynkin lemma.
  • Instead, for the identity \eqref{eq:id} I don't know what to do. To be honest, it confuses me a bit. And I know that I should use the strong Markov property but I cannot see how.

Any hint, just to start from somewhere, would be really appreciated.