Note on batch arrival LCFS and related symmetric queues (Q1317021)

It is well-known that the stationary distribution of the number of customers in a single server queue with Poisson input flow and Last Come First Served-Preemptive/Resume service discipline depends on the service time distribution only through its mean. This result is generalized to the case of batch arrivals when acceptance policy depends only on the number of customers in the system just before arrivals and related symmetric queues. For the system with general interarrival time distribution stochastic bounds for the queue length distribution are obtained. The analysis is based on general relationships between time- and embedded-stationary distributions of a stationary jump process, some modification of the notion of generalized semi-Markov process and properties of NBUE and NWUE distributions.