MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Assume we have a colored Gaussian process $z_t$, with an autocorrelation function $cov(z_t,z_s)$ given by an analytical function $\alpha(t,s)$ (if it helps, one can assume that $\alpha(t,s) = \kappa e^{-w(t-s)}$). Consider now a process defined by

$Z_t := \int_0^t z_s ds$

Now, my 3 questions:

  • is $Z_t$ well-defined?
  • if so, is $d Z_t = z_t dt$ true in any sense?
  • what is the quadratic variation of $Z_t$?
share|cite|improve this question
up vote 1 down vote accepted

Q1: Rather yes. If we assume that function $w$ is not crazy, e.g. $C^1$, then there exists a continuous version of $z$ and the integral can be computed path-by-path using the Lebesgue (or even Riemann) integral theory.

Q2:Under assumption above $Z$ has paths of the finite variation. Hence there exists the Stieltjes integral (for almost any path) and for any measurable function $f$ one have $\int_a^b f(t) dZ_t = \int_a^b f(t) z_t dt.$

Q3:I assume that you meant quadratic variation of Z, which is 0.

share|cite|improve this answer
BTW. I doubt if this question counts as "research level one" ;). – Piotr Miłoś Jan 3 '12 at 16:22
Thanks. I'm not that lazy, I did a calculation of quadratic variation and got zero, but wasn't sure if that's correct. I'm not a mathematician and I only have experience with "standard" Brownian motion, which behaves differently. This is a research question, but in physics ;-) – Katastrofa Jan 3 '12 at 19:41

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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