Determining the stationary probabilities of states of the system GI|G|n|O with an almost-poisson input

From MaRDI portal
Publication:3682212





In the present article, which is a generalization of a previous paper of the author [in: Problems of stability of stochastic models, Proc. Sem., VNIISI, Moscow, 71-76 (Russian) (1981)], we propose a method for calculating the stationary probabilities of states of the system GI\(| G| n| 0\) with an incoming stream of calls that is almost- Poisson in a certain sense. We consider a queuing system in which a recurrent stream of calls, which is described by a distribution function F(x), enters n service channels. If there are already n calls present in the system at the arrival time of a given call, the new arrival is rejected and leaves the system unserved. The service times are random variables that are independent of the input process and among each other and have a distribution function G(x) with finite expected value \(\tau\). We assume that \(F(x)<1\) for any \(x\in [0,\infty)\) and that F(x) is an absolutely continuous function with probability density function \(f(x)<\infty\) for any \(x\in (0,\infty)\), i.e., \(F(x)=\int^{x}_{0}f(u)du\). Then the distribution function F(x) can be written in the form \[ F(x)=1-\exp \{-\int^{x}_{0}\lambda (u)du\},\quad x\geq 0, \] where \(\lambda (u)=f(u)/[1-F(u)]\), \(u\geq 0\).











This page was built for publication: Determining the stationary probabilities of states of the system GI|G|n|O with an almost-poisson input

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3682212)