Tail asymptotics for the maximum of perturbed random walk (Q862212)

A perturbed random walk with a negative drift is considered. The main goal of the given approach is to compute the distribution of the all-time maximum \(M_{\infty} = \max\{S_{k} + \xi_{k}: k \geq 0\}\) of a perturbed random walk with a negative drift. The quantity \(M_{\infty}\) describes the steady-state of end-to-end delay time in a computer network modeling, as well as order-to-delivery time in a make-to-order production system, insurance risk evaluation, etc. As main results, the asymptotics for \(\mathbb{P}(M_{\infty} > x)\) as \(x \rightarrow \infty\) is obtained. It is derived that the tail asymptotics depend on whether the \(\xi_{n}\)'s are light- or heavy-tailed. Additionally, in the light-tailed setting, the tail asymptotic is closely related to the Cramér-Lundberg asymptotic for standard random walk.