Entropy maximisation is applied to characterise the distributional form of the steady-state probabilities of a G/G/c/PR queue with c (⩾ 2) parallel servers and R (⩾ 2) priority classes under preemptive resume rule. For R = 2, closed-form expressions are derived as explicit functions of appropriate but unknown constraints which are approximately determined using analytical results for GE-type queues. The applicability of the maximum entropy (ME) solutions for R = 2 is also extended to the case of R > 2 via the method of class aggregation. Typical numerical results are presented to illustrate the credibility of ME approximations against simulations. Comments on the role of the GE/GE/c/PR queue as a building block for the approximate analysis of arbitrary queueing networks are included.
