Convexity of Workload Autocorrelation in a Stationary Single Server Queue with Independent Increment Input

Abstract

We propose a method based on probabilistic arguments to study the convexity of the autocorrelation function of processes associated with a single server queue. To illustrate the power of the method we apply it in two cases: to the workload process seen by the customers of a GI/GI/1 queue, and to the workload process of a fluid queue fed by a general L\'evy process.