Large sample inference in retrial queues

We analyze retrial queueing systems from a statistical viewpoint. We consider a general G/G/1 retrial queue where the flow of repeated attempts can be non-Markovian. We observe the operation of the system over a random time interval (0, T] focusing on the case where T is a departure epoch. We obtain some asymptotic relationship among the random variables involved, even when the system in nonergodic. This allows us to get asymptotically Gaussian consistent estimators for the unknown parameters in a parametric context.