Mean buffer contents in discrete-time single-server queues with heterogeneous sources (Q1775809)

One considers a discrete-time work-conserving single-server queue with \(K\) independent, heterogeneous sources. Each source becomes idle and active alternately, the idle and active periods form a discrete-time alternating renewal process. The lengths of idle periods are geometrically distributed and the length of an active period is determined by the first passage time to the absorbing state in an underlying absorbing Markov chain. The number of cells generated by an active source can depend on the state of the Markov chain, the service times of cells are constant and equal to one slot. One generalizes the results of \textit{M. F. Neuts} [Commun. Stat., Stochastic Models 6, No.~1, 87--98 (1990; Zbl 0704.60093)] by allowing multiple cell arrivals during discrete phase-type periods. The mean buffer contents is given in terms of the traffic intensity, the first two moments of active period lengths and the number of cells generated during an active period of each source. The essential assumption in obtaining a general formula is that the lengths of idle periods of each source are geometrically distributed and the dynamics of each source during active periods are characterized by the underlying absorbing Markov chain with finite states. The finite-state assumption can be cancelled in two directions, by introducing a special structure into the absorbing chain or to provide a sufficient condition under which the general formula is applicable to queues with independent sources. As examples there are considered two specific queues: a queue with Markovian autoregressive sources and a queue with greedy sources.