I have a question regarding Kallenberg's "Foundations of Modern Probability" 2nd ed, a statement right after Lemma 11.2: "...$\eta\{0\}=1$ a.s.".

Let me state the setup:

$\xi$: a random measure on $\mathbb{R}^d$

$X$: an $\mathbb{R}^d$-indexed process taking values in a measurable space $S$.

$\mathcal{M}(\mathbb{R}^d)$: the space of locally finite measures on $\mathbb{R}^d$.

($X$,$\xi$) are jointly stationary w.r.t. a shift $\theta_t$, $t\in \mathbb{R}^d$ (i.e. $\theta_t x=x+t$, $x\in \mathbb{R}^d$), namely, $\theta_t(X,\xi):= (\theta_tX,\theta_t\xi)$ has the same distribution as $(X,\xi)$.

Assume $\mathrm{E}\xi=c\lambda^d$, where $\lambda^d$ is the Lebesgue measure and $c\in \mathbb{R}_+$ (this is almost a consequence of stationarity).

For a function $f\ge 0$ on $S^{R^d}\times \mathcal{M}(\mathbb{R}^d)$, define $$ Q_{X,\xi} f=\mathrm{E} \int_B f(\theta_s(X,\xi))\xi(ds) / \mathrm{E}\xi B, $$ for some Borel set $B\subset\mathbb{R}^d$ with $\lambda^d B\in \mathbb{R}_+$. By Lemma 11.2 and Theorem 2.6 of Kallenberg, the preceding expression actually does not depend on $B$.

$Q_{X,\xi}$ can be viewed a probability measure on $S^{R^d}\times \mathcal{M}(\mathbb{R}^d)$. Let a random pair $(Y,\eta)$ have distribution $Q_{X,\xi}$.

Then the book claims that "When $\xi$ is a simple point process, then ...and we note that $\eta\{0\}=1$ a.s.". I just do not know why "$\eta\{0\}=1$ a.s.".


Just figured out:

Take a measurable $A\in \mathcal{M}(\mathbb{R}^d)$. Then $$ Q_{X,\xi}(\eta\in A)= \mathrm{E} \int_{[0,1]^d}1_{\{\theta_s \xi\in A\}}\xi(ds)=\\ \mathrm{E} \int_{[0,1]^d}1_{\{\theta_s \xi\in A, \theta_s\xi\{0\}=1\}}\xi(ds)+\mathrm{E} \int_{[0,1]^d}1_{\{\theta_s \xi\in A, \theta_s\xi\{0\}=0\}}\xi(ds). $$ Since the simple point process $\xi$ satisfies $\xi\{s\}=0$ or $1$, the second term in the preceding line must be $0$ because $\theta_s\xi\{0\}=\xi\{s\}$. Therefore, $$ Q_{X,\xi}(\eta\in A)=\mathrm{E} \int_{[0,1]^d}1_{\{\theta_s \xi\in A, \theta_s\xi\{0\}=1\}}\xi(ds) = Q_{X,\xi}(\eta\in A, \eta\{0\}=1). $$

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.