Heavy traffic limit for a processor sharing queue with soft deadlines (Q2456052)

The paper provides a framework for evaluating the heavy traffic performance of the GI/GI/1 processor sharing queue with respect to a fairly general structure of deadlines. The state of processor sharing queue with deadlines is tracked using a measure valued process in the right half-plane. The model consists of an infinite capacity buffer, to which jobs arrive according to a delayed renewal process, the service time requirements are determined by a sequence of independent, identically distributed positive random variables. The server works simultaneously on all jobs in the buffer providing an equal fraction of its capacity. In addition to its service time, each job arrives at the buffer with a deadline given by the jobs arrival time plus a real valued random variable called the initial lead time. The sequence of initial lead times is independent and identically distributed, but the job's initial lead time may be correlated with its service time. Job deadlines are soft, meaning that the job remains in the system until served to completion. The limit measure valued process is obtained under diffusion scaling and heavy traffic conditions and is characterized as a deterministic function of the limiting queue length process.