Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

For a Markov process $X$ on the Polish space $\mathscr X$ its transition probability is given by $$ P(x,A) :=\mathsf P_x (X_1\in A) $$ and $X$ is time-reversible if there is a probability measure $\pi$ such that for any $A,B$ in a Borel $\sigma$-algebra of $\mathscr X$ denoted by $\mathscr B(\mathscr X)$ it holds that $$ \int\limits_A P(x,B)\pi(dx) = \int\limits_B P(x,A)\pi(dx). $$ The discrete-time Laplacian $\Delta$ and the gradient $\nabla_{xy}$ are defined as $$ \nabla_{xy}f = f(y)-f(x) $$ and $$ \Delta f(x) = \int\limits_\mathscr Xf(y)P(x,dy)-f(x) = \int\limits_\mathscr X(\nabla_{xy}f)P(x,dy). $$

In the case when $\mathscr X$ is countable, for any measurable $A$ the following Green's formula holds: $$ \int\limits_A\Delta f(x)g(x)\pi(dx) = -\frac12\int\limits_{A}\nabla_{xy}f\cdot\nabla_{xy}g\;\pi(dx)+\int\limits_{A}\int\limits_{A^c}(\nabla_{xy}f)g(x)P(x,dy)\pi(dx). $$ This formula can be generalized for an uncountable state space as well. I wonder though if there is an analogue for the case when $X$ is not time-reverisble.

share|improve this question
    
I don't know if this is what you are looking for, but the left-hand side in your last identity is called the Dirichlet form $\mathcal{E}(f,g)$. For a Markov chain on a countable space and with invariant measure $\pi$, (not necessarily reversible), it is always true that $$\mathcal{E}(f,g) = \sum_{x,y} \pi(x) P(x,y) g(x) \nabla_{x,y} f.$$ But this is an easy calculation, so presumably you are aware of it. –  Nathanael Berestycki Dec 20 '11 at 22:05
    
@Nathanael: but this is only for the case $A=\mathscr X$, isn't it? In that case you obtain an answer just by writing $\Delta f(x)$ through the integral (or the sum in you example) –  Ilya Dec 21 '11 at 7:06
    
@Nathanael: could you refer me to any book (or maybe, lecture notes) on Dirichlet forms for discrete time Markov processes with general state spaces (I guess you call them simply Markov chains)? –  Ilya Jan 12 '12 at 13:03

Your Answer

 
discard

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

Browse other questions tagged or ask your own question.