Exact Tail Asymptotics --- Revisit of a Retrial Queue with Two Input Streams and Two Orbits

TL;DR

This paper analyzes the tail asymptotics of a complex retrial queue with two customer types and orbits, using a kernel method and a novel approach involving a censored random walk, providing insights without solving the full generating function.

Contribution

It introduces a new kernel method approach for tail asymptotics in a modulated two-dimensional QBD process, avoiding full generating function solutions.

Findings

Derived tail asymptotics for the joint distribution of two orbits.

Identified a censored random walk technique applicable to similar models.

Provided a framework for analyzing modulated random walks in queueing systems.

Abstract

We revisit a single-server retrial queue with two independent Poisson streams (corresponding to two types of customers) and two orbits. The size of each orbit is infinite. The exponential server (with a rate independent of the type of customers) can hold at most one customer at a time and there is no waiting room. Upon arrival, if a type $i$ customer $(i=1,2)$ finds a busy server, it will join the type $i$ orbit. After an exponential time with a constant (retrial) rate $\mu_i$, an type $i$ customer attempts to get service. This model has been recently studied by Avrachenkov, Nain and Yechiali~\cite{ANY2014} by solving a Riemann-Hilbert boundary value problem. One may notice that, this model is not a random walk in the quarter plane. Instead, it can be viewed as a random walk in the quarter plane modulated by a two-state Markov chain, or a two-dimensional quasi-birth-and-death (QBD) process. The special structure of this chain allows us to deal with the fundamental form corresponding to one state of the chain at a time, and therefore it can be studied through a boundary value problem. Inspired by this fact, in this paper, we focus on the tail asymptotic behaviour of the stationary joint probability distribution of the two orbits with either an idle or busy server by using the kernel method, a different one that does not require a full determination of the unknown generating function. To take advantage of existing literature results on the kernel method, we identify a censored random walk, which is an usual walk in the quarter plane. This technique can also be used for other random walks modulated by a finite-state Markov chain with a similar structure property.