Analysis of the asymmetrical shortest two-server queueing model (Q1270299)

This lengthy paper deals with the analytic solution of a two server queueing model with Poisson arrivals in which an arriving customer joins the shorter queue if the queue lengths are unequal; if they are equal he joins the queue in front of server \(i\) with probability \(\pi_i\) \((i=1,2)\). The service time at counter \(i\) has negative exponential distribution with rate \(1/\beta_i\) \((i=1,2)\). The study of this model known as ``two server shortest queue'' or ``two queues in parallel'' is mostly concerned with the symmetrical case in which \(\pi_1= \pi_2=1/2\) and \(\beta_1= \beta_2\). In the present analysis of the asymmetrical model, the bivariate generating function of the stationary joint distribution of the queue length, which provides the necessary performance measures, is determined in terms of four generating functions, each of which is a sum of a polynomial and a meromorphic function. It is shown that the poles and the residues of the meromorphic functions can be calculated recursively and the coefficients of the polynomials can be easily found, in particular, if the asymmetry in the model is not very large. The approach developed in this article seems to be applicable to a larger class of random walks modelling asymmetrical two-dimensional queues.