Comparing ordered-entry queues with heterogeneous servers

From MaRDI portal





This paper discusses a queueing system with m heterogeneous exponential servers with service rates \(\mu_ 1\geq \mu_ 2\geq...\geq \mu_ m\) and a finite buffer with capacity N(\(\geq m)\). Arrivals follow an arbitrary point process. Let \(\pi\) be a permutation of m integers denoting an ``entry order, i.e. any arrival will receive the service from the first unoccupied server according to the order \(\pi\). If all servers are occupied but the buffer is not full, the arrival will join a single queue. The service pattern is FCFS. Once the service is completed, the customer leaves the system forever. If the buffer is full, the arrivals will be blocked and are deemed to be lost. This is precisely an ordered entry \(G/M(\mu_ i)/m/N\) model. This paper contains the best (worst) entry order in terms of minimizing (maximizing) congestion and maximizing (minimizing) the number of processed customers in the sense of stochastic ordering. The results complement previous results regarding the comparison of queues with parallel (homogeneous) servers.











This page was built for publication: Comparing ordered-entry queues with heterogeneous servers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1116560)