Asymptotic property of the occupation measures in a multi-dimensional skip-free Markov modulated random walk

Abstract

We consider a discrete-time $d$-dimensional process $\{\boldsymbol{X}_n\}=\{(X_{1,n},X_{2,n},...,X_{d,n})\}$ on $\mathbb{Z}^d$ with a background process $\{J_n\}$ on a countable set $S_0$, where individual processes $\{X_{i,n}\},i\in\{1,2,...,d\},$ are skip free. We assume that the joint process $\{\boldsymbol{Y}_n\}=\{(\boldsymbol{X}_n,J_n)\}$ is Markovian and that the transition probabilities of the $d$-dimensional process $\{\boldsymbol{X}_n\}$ vary according to the state of the background process $\{J_n\}$. This modulation is assumed to be space homogeneous. We refer to this process as a $d$-dimensional skip-free Markov modulate random walk. For $\boldsymbol{y}, \boldsymbol{y}'\in \mathbb{Z}_+^d\times S_0$, consider the process $\{\boldsymbol{Y}_n\}_{n\ge 0}$ starting from the state $\boldsymbol{y}$ and let $\tilde{q}_{\boldsymbol{y},\boldsymbol{y}'}$ be the expected number of visits to the state $\boldsymbol{y}'$ before the process leaves the nonnegative area $\mathbb{Z}_+^d\times S_0$ for the first time. For $\boldsymbol{y}=(\boldsymbol{x},j)\in \mathbb{Z}_+^d\times S_0$, the measure $(\tilde{q}_{\boldsymbol{y},\boldsymbol{y}'}; \boldsymbol{y}'=(\boldsymbol{x}',j')\in \mathbb{Z}_+^d\times S_0)$ is called an occupation measure. Our primary aim is to obtain the asymptotic decay rate of the occupation measure as $\boldsymbol{x}'$ go to infinity in a given direction. We also obtain the convergence domain of the matrix moment generating function of the occupation measures.