-1
$\begingroup$

Let $\left(\Omega\ ,\mathcal{F}\ ,\mathbb{P} \right)$ be a probability space with filtration $F=\left\{ \mathcal{F_t} \right\}_{t\geqslant0}$ , and $X=\left\{ X_t \right\}_{t\geqslant0}$ is an adapted stochastic process w.r.t. $F$.

Definition 1: A process $X$ is called stochastically continuous , if for all $t\geqslant0$ , $X_s\overset{\mathbb{P}}{\rightarrow} X_t$ as $s\rightarrow t$.
Definition 2: An adapted process $X$ is called a process with independent increments w.r.t. $F=\left\{ \mathcal{F_t} \right\}_{t\geqslant0}$ , if for all $t>s\geqslant0$ , $X_t-X_s$ is independent of $\mathcal{F_s}$.

Proposition: If a stochastically continuous process $X$ has independent increments w.r.t. $F=\left\{ \mathcal{F_t} \right\}_{t\geqslant0}$ , then $X$ has independent increments w.r.t. $F_+:=\left\{ \mathcal{F_{t+}} \right\}_{t\geqslant0}$ , where $\mathcal{F_{t+}}:=\bigcap_{r>t}{\mathcal{F_r}} $ .

Now we want to prove that if $t>s\geqslant0$ , then $X_t-X_s$ is independent of $\mathcal{F_{s+}}=\bigcap_{n\in\mathbb{N}}{\mathcal{ F_{s+1/n} }} $ . The condition tell us that $X_t-X_{s+1/n}\overset{\mathbb{P}}{\rightarrow}X_t-X_s$ as $n\rightarrow\infty$. In fact, $\overset{\mathbb{P}}{\rightarrow}$ can be replaced by $\overset{a.s.}{\rightarrow}$ because of independent increments.
Note that $X_t-X_{s+1/n}$ is independent of ${\mathcal{ F_{s+1/n} }}$ , it seems that $$X_t-X_s=\lim_{n\rightarrow\infty}\left(X_t-X_{s+1/n}\right) \text{ is independent of }\ \mathcal{F_{s+}}=\bigcap_{n\in\mathbb{N}}{\mathcal{ F_{s+1/n} }}\ ,$$ but I don't know how to get the result.

$\endgroup$
  • $\begingroup$ Personal input? An approach seems natural, did you try to make it work? $\endgroup$ – Did Jun 29 '16 at 9:17
  • $\begingroup$ @Did I have tried but failed. I don't konw how to keep the independence with the limit in probability. $\endgroup$ – Wang Ke Jun 29 '16 at 9:26
  • $\begingroup$ Please include whichever tries you made, in the question. $\endgroup$ – Did Jun 29 '16 at 9:28
  • $\begingroup$ @Did I added my thinking. $\endgroup$ – Wang Ke Jun 29 '16 at 9:53
  • $\begingroup$ And now we see that $X_t-X_{s+1/n}$ is independent of $\mathcal F_{s+}$ and that $X_t-X_{s+1/n}\to X_t-X_s$ almost surely hence the question is: "Assume that $(Y_n)$ is independent of $\mathcal G$ and that $Y_n\to Y$ almost surely then $Y$ is independent of $\mathcal G$." Any idea for this? $\endgroup$ – Did Jun 29 '16 at 10:56
1
$\begingroup$

Thanks for @Did's help. Now I have got the answer.

Because $X_t-X_{s+1/n}$ is independent of ${\mathcal{ F_{s+1/n} }}$ and $\mathcal{ F_{s+} }\subset\mathcal{F_{s+1/n}}\ $ , $X_t-X_{s+1/n}\ $ is independent of ${\mathcal{ F_{s+} }}$ .
Now the question is to prove "Assume that $\left\{Y_n\right\}$ is independent of $\mathcal{G}$ and that $Y_n\overset{a.s.}{\rightarrow}Y$ , then $Y$ is independent of $\mathcal{G}$."

Proof:

  • " $Y$ is independent of $\mathcal{G}$ " is equivalent to " $\forall A\in\mathcal{G}$ , $Y$ is independent of $I_A$ ".
  • $Y_n\overset{a.s.}{\rightarrow}Y$ , hence $\exp\left(it_1Y_n\right) \overset{a.s.}{\rightarrow} \exp\left(it_1Y\right)$ and $\exp\left(it_1Y_n+it_2I_A\right) \overset{a.s.}{\rightarrow} \exp\left(it_1Y+it_2I_A\right)$ .
  • By dominated convergence theorem, we have that $\mathbb{E}\left[\exp\left(it_1Y_n\right)\right] \rightarrow \mathbb{E}\left[\exp\left(it_1Y\right)\right]$ and that $\mathbb{E}\left[\exp\left(it_1Y_n+it_2I_A\right)\right] \rightarrow \mathbb{E}\left[\exp\left(it_1Y+it_2I_A\right)\right]$ .
  • For each $n$, $Y_n$ is independent of $I_A$, hence $$\mathbb{E}\left[\exp\left(it_1Y_n+it_2I_A\right)\right]=\mathbb{E}\left[\exp\left(it_1Y_n\right)\right]\mathbb{E}\left[\exp\left(it_2I_A\right)\right], $$
    then we have that
    $$\begin{align*} &\mathbb{E}\left[\exp\left(it_1Y+it_2I_A\right)\right] = \lim_{n\rightarrow\infty}{\mathbb{E}\left[\exp\left(it_1Y_n+it_2I_A\right)\right]} \\ &\qquad\qquad\qquad\qquad\quad =\lim_{n\rightarrow\infty}{\mathbb{E}\left[\exp\left(it_1Y_n\right)\right]\mathbb{E}\left[\exp\left(it_2I_A\right)\right]} \\ &\qquad\qquad\qquad\qquad\quad =\mathbb{E}\left[\exp\left(it_1Y\right)\right]\mathbb{E}\left[\exp\left(it_2I_A\right)\right]. \end{align*}$$
    It is concluded that $Y$ is independent of $I_A$.
$\endgroup$
  • $\begingroup$ The fact that, for every $n$, $\sigma\left(Y_n\right)$ is independent of $\mathcal{G}$, does not imply that $\sigma\left(Y_n,n\in\mathbb{N}\right)$ is independent of $\mathcal{G}$. $\endgroup$ – Did Jun 29 '16 at 11:48
  • $\begingroup$ @Did I have edited. $\endgroup$ – Wang Ke Jun 29 '16 at 12:36

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.