Large finite population queueing systems: The single-server model

From MaRDI portal





The single-server model is investigated where a finite number of \(n\) customers arrive at some facility. This model can be used to model a computing center, where programs are run once a day, for instance. For large \(n\) there are obtained Gaussian limiting processes and diffusion approximations for a wide variety of stochastic variables. In particular, the approximations for asymptotic in heavy traffic (as \(n\) goes to infinity) for distributions of queue size process, workload process and total storage occupied process are obtained (it is assumed that each customer asks for some storage). Several regions of the time axis are considered, leading to oversaturated, undersaturated and transition behaviours. The corresponding limiting processes (Brownian motion and reflected Brownian motion, as a rule) for normalized and centered basic ones are investigated in detail. Some other variables of interest (idle time, busy period, cost distribution) are considered, too.



Cites work









This page was built for publication: Large finite population queueing systems: The single-server model

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1338765)