Analysis of the Number of Orbiting Customers in M/g/1 Retrial Queue with General Retrial Times

The research in the area of retrial queues focused mainly on the fact that the retrials operate under classical retrial policy, which suppose that the intervals between successive repeated attempts are exponentially distributed with total rate jθ (j is the number of customers in the orbit). There is another discipline, called constant retrial policy, where the total retrial rate does not depend on the number of customers in the retrial group: the customers form a FCFS queue, and then only the customer at the head of this queue can request a service. It was introduced by Fayolle [7] for an M/M/1 queue with exponentially distributed retrial times, and studied by Choi et al. [4] for general retrial times. Farahmand [6] named the queueing systems in question as retrial queues with FCFS orbit. In recent years, several retrial models with FCFS orbit and general retrial times have been analyzed, details of which may be found in [2], [3], [11], [12], [14], [15], [16]. The stability of single server queues under general distribution for retrial times was discussed in [9].