Let $(X_n)_{n\in\mathbb{N}_0}$ be a time-homogeneus Markov chain, evoloving on a space $X$, with transition law $(x,A)\mapsto P(x,A)$. Further, let $\tau_n: \Omega\to \left[0,\infty\right)$, $n\in\mathbb{N}_0$, be a strictly increasing sequence of random variables with $\tau_0=0$ and $\lim_{n\to\infty} \tau_n=\infty$ (a.s.), whose increments $\Delta \tau_{n+1}:=\tau_{n+1} - \tau_n$ are mutually independent and have the conditional distribution of the form

$$\mathbb{P}(\Delta \tau_{n+1} \leq t\, | \, X_n=x)=1-\exp\left(-\int_0^t \lambda(x,s)\,ds\right).$$

Consider the process $(N_t)_{t\geq 0}$ defined by $$N_t=\sum_{n=0}^{\infty} \boldsymbol{1}_{\{\tau_n\leq t\}},$$ that is $\{N_t=n\}=\{\tau_n\leq t < \tau_{n+1}\}$. Clearly, if $\lambda$ is constant, then $(N_t)_{t\geq 0}$ is just a Poisson counting process.

Is anything known about $(N_t)_{t\geq 0}$ in the general case? Is there any book considering such processes? Specifically, I am interested in the following issues:

  1. Do the increments of $(N_t)_{t\geq 0}$ are conditionally independent given $X_0$?

  2. Is it possible to compute $\mathbb{P}(N_t=n\, | \, X_0=x)$ in terms of $\lambda$ and $P$?

  3. Can we determine the conditional distribution of $(\Delta \tau_1,\ldots,\Delta \tau_n)$ given $\{N_t=n, X_0=x\}$ (for a Poisson process it is the same as the joint distribution of the order statistics of $n$ i.i.d. random variables with the uniform distribution on $(0,t)$)?

I would be grateful for any information.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.