The limiting distribution of the time to attain the level for a random walk with reflection onto the superposition of two regeneration processes (Q1085522)

Suppose that a random walk has a reflection in zero whose generating function of jumps, as a superposition of Poisson and recurrent regeneration processes, \(\chi =\sum^{\nu_ x}_{1}\theta_ m-1\) is \(g_ 0(z)=Mz^{\chi}=g(\alpha (1-p(z)))/z\), \(p(z)=MZ^{\theta_ m}\), \(g(\lambda)=Me^{-\lambda x}\), where \(\alpha\) is the parameter of the Poisson process \(\nu_ t\). Under various conditions on the existence of a stationary distribution of the walk, this article gives the limiting relation \[ (1)\quad \lim_{N\to \infty} M \exp (-\Lambda z_ 0^{-N- 1}\tau^ N_{\gamma})=1-W_{\gamma}+\frac{c}{c+\Lambda}W_{\gamma}, \] here \(z_ 0>1\) satisfying \(g(\alpha (1-p(z_ 0)))=z_ 0\), \(W_{\gamma}=P({\bar \chi}<\gamma)\), \({\bar \chi}=\max_{n\geq 0}\sum^{n}_{k=1}\chi_ k\), and c is the parameter of the limiting exponential law. Using (1) the trajectories of the process with reflection can be splitted into two types.