The use of relative priorities in optimizing the performance of a queueing system

Control of queueing systems to maximize profits or welfare has been the subject of numerous papers. The common methods are by setting adequate price and priority regimes (see [10] for a survey of such models). In other cases, the service provider also sets and advertises waiting time standards [1]. The common priority regime is that of (preemptive or non-preemptive) absolute priorities, where the customer classes are ranked and customers are called to be served according to this order. There is a voluminous literature analyzing and comparing different priority disciplines, see for instance the survey texts by Gelenbe and Mitrani [8] and Kleinrock [13]. A notable generalization of this concept was offered by Federgruen and Groenevelt [7] who considered work conservU 48 49 50 51 52 53 54 55