Diffusion approximation for a processor sharing queue in heavy traffic.

From MaRDI portal



Abstract: Consider a single server queue with renewal arrivals and i.i.d. service times in which the server operates under a processor sharing service discipline. To describe the evolution of this system, we use a measure valued process that keeps track of the residual service times of all jobs in the system at any given time. From this measure valued process, one can recover the traditional performance processes, including queue length and workload. We show that under mild assumptions, including standard heavy traffic assumptions, the (suitably rescaled) measure valued processes corresponding to a sequence of processor sharing queues converge in distribution to a measure valued diffusion process. The limiting process is characterized as the image under an appropriate lifting map, of a one-dimensional reflected Brownian motion. As an immediate consequence, one obtains a diffusion approximation for the queue length process of a processor sharing queue.


The paper presents a heavy traffic diffusion approximation for a measure valued process \(\{\mu(t): t\geq 0\}\) that keeps track of the residual service times of jobs in the buffer of a \(GI/G/1\) processor sharing queue. Each job in the buffer is being processed at a rate that is the reciprocal of the number of jobs in the buffer. When the server has fulfilled a given job's service time requirement, the job exits the buffer. The residual service time at time \(t\geq 0\) of a job which has entered the buffer by time \(t\), is given by the amount of processing time originally requested by the job minus the total amount of processing time it has received by time \(t\). For each \(t\geq 0\), \(\mu(t)\) is the random element of the space of finite, nonnegative Borel measures on \([0,\infty)\), that has a unit of mass at the residual service time of each job currently in the buffer at time \(t\).




Cited in
(29)








This page was built for publication: Diffusion approximation for a processor sharing queue in heavy traffic.

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1879882)