1
$\begingroup$

I'm looking for sources which elaborate a little bit on the fact that for Markov process $X_t$ with generator $L$, $\int_{}^{}\Gamma(f,f)(X_s)ds$ is a quadratic variation of $M_t := f(X_t) - f(x) - \int_{}^{t}Lf(X_s)ds$ (i.e. $M_t^2 - \int_{}^{}\Gamma(f,f)(X_s)ds$ is a martingale) where $\Gamma(f,f) = Lf^2 - 2fL(f)$. Thank you for all suggestions.

$\endgroup$
2
  • $\begingroup$ Did you find an answer to your question? $\endgroup$
    – user148364
    Commented Sep 14, 2020 at 12:12
  • $\begingroup$ no, I did not. I' ve only managed to prove it for diffusion processes (using Ito calculus) $\endgroup$
    – marcusy
    Commented Sep 15, 2020 at 13:03

2 Answers 2

0
$\begingroup$

Could these papers be of interest ?

http://djalil.chafai.net/blog/2017/04/09/carre-du-champ/

Show that the carré du champ operator is nonnegative

https://mathoverflow.net/questions/229226/carre-du-champ-subunit-paths-and-cc-metrics

http://www.numdam.org/article/SB_1976-1977__19__167_0.pdf

$\endgroup$
1
  • $\begingroup$ Unfortunately no, they do not contain the answer $\endgroup$
    – marcusy
    Commented Aug 21, 2020 at 7:56
0
$\begingroup$

Check out Proposition 3.26 in https://andre-schlichting.de/wp-content/uploads/2012/09/MP-script.pdf

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .