2
$\begingroup$

Suppose $u(t,\omega), t \geq 0$ is a continuous time stochastic process with smooth paths so $\frac{d}{dt} u(t,\omega)$ exists for all $\omega$. Suppose you know the distribution of $\frac{d}{dt} u(t,\omega)$ and the initial distribution of $u$. Can you use this information to determine the distribution of $u(t,\omega)$ for all $t \geq 0$? That is, you know

  1. $P(\frac{d}{dt}u(t,\omega) \leq A)$ for all $A \in R, t\geq 0$
  2. $P(u(0,\omega) \leq A)$ for all $A \in R$.

Can you use 1 and 2 to determine $P(u(t,\omega) \leq B)$ for all $B \in R$?

Can you go the other direction? That is, if you know $P(u(t,\omega) \leq A$ for all $A \in R$, what is $P(\frac{d}{dt} u(t,\omega) \leq B)$?

$\endgroup$

1 Answer 1

0
$\begingroup$

I would think the answer is "no" to both questions.

Let me abbreviate $v(t)=du/dt$, so $u(t)=u(0)+\int_0^t v(t')dt'$. Now ask for the expectation of $u^2(t)$. You'll need to know how $v(t')$ and $v(t'')$ are correlated for any $t',t''$ in the interval $(0,t)$. That information is not given.

Similarly, in the other direction, to find the distribution of the velocity $v(t)$ you need to know how the position $u(t)$ at nearby times is correlated, and that information is not given.

$\endgroup$
1
  • $\begingroup$ Makes sense. A simple counterexample to the first question would be $v_1(t) = X$, and $v_2(t) = Y$ for $t < 1, v_2(t) = Z$ for $t >= 1$ with $X, Y, Z$ all being independent standard normals. $v_1$ and $v_2$ are equal in distribution but their integrals are not. Thank you very much for your help! $\endgroup$
    – bryan
    May 2, 2013 at 2:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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