Construction of a random variable

64 Views Asked by At

I'm reading Dirichlet Forms and Symmetric Markov Processes by M. Fukushima, Y. Oshima, and M. Takeda. In Appendix A.2, where they discuss the construction of a random variable, there is the statement:

"..., we need a non-negative random variable $Z(\omega)$ on $(\Omega,\mathcal{M},P_{x})$ which is of exponential distribution with mean $1$; independent of $(X_{t})_{t \ge 0}$ under $P_{x}$ for every $x \in S$ satisfying $Z(\theta_{s}(\omega))=(Z(\omega)-s)\vee 0$."

In the statement, $(\Omega, \mathcal{M}, (X_{t})_{t \ge 0}, (P_{x})_{x \in S})$ is a Hunt process with state space $S$ and $(\theta_{t})_{t \ge 0}$ is a shift operator of this Hunt process.

This book says,

"This requirement is fulfilled by replacing $\Omega$ by its direct with $[0,\infty)$ if necessary. To see this, let $Z(\xi)$ be a non-negative random variable on $([0,\infty), \mathcal{B}([0,\infty)), \lambda)$ of exponent distribution with mean $1$; $\lambda(\left\{ \xi \ge 0; Z(\xi) >a \right\})=e^{-a}$, $\forall a \ge 0$. We can consider the functions $X_{t},\theta_{t}$ and $Z$ as function on $\Omega \otimes [0,\infty)$ by putting $X_{t}(\omega,\xi)=X_{t}(\omega), \theta_{t}(\omega ,\xi)=(\theta_{t}(\omega),(\xi-t)^{+})$ and $Z(\omega,\xi)=Z(\xi)$, $\forall (\omega, \xi) \in \Omega \otimes [0,\infty)$. It is clear, under this identification, that $(\Omega \otimes [0,\infty), \mathcal{M} \otimes \mathcal{B}([0,\infty)), X_{t}, P_{x}\otimes \lambda)$ is a Hunt process with admissible filtration $\mathcal{M}_{t}\otimes \mathcal{B}([0,\infty))$ and that the random variable $Z$ has the desired properties."

Question

I don't understand $Z$ satisfies $Z(\theta_{s}(\omega ,\xi))=(Z(\omega,\xi)-s)\vee 0$. By definition, \begin{align} Z(\theta_{s}(\omega,\xi))&=Z(\theta_{s}(\omega), (\xi-s)^{+}) \\ &=Z((\xi-s)^{+}) \\ &=Z(\omega, (\xi-s)^{+} ). \end{align}

How do I get the equation $Z(\theta_{s}(\omega ,\xi))=(Z(\omega,\xi)-s)\vee 0$ ? Doing somethig wrong?

Thank you in advance.