On the stationary LCFS-PR single-server queue: a characterization via stochastic intensity (Q429007)

A stationary single-server queue with preemptive-resume last-come first-served discipline is studied. The input is stationary marked point process. The main tool is Palm-martingale calculus including connection between Palm probability and stochastic intensity. Some general formulas for the stationary joint distribution of queue size and remaining service time both at arbitrary instances and at arrival instances are derived. These relations generalize previous results to more general input process. The distribution of the stationary workload process is also derived. Connections between the obtained results and similar results for the classical system with the same discipline are discussed.