Transient and busy period analysis of the GIG/1 Queue as a Hilbert factorization problem

From MaRDI portal
Publication:3988129





The Laplace transform \(\sigma(s)\) of the busy period distribution and the double Laplace transform \(\Phi(s,\omega)=\iint e^{-sx-\omega y}f(x,y)dx dy\) of the waiting time density \(f(x,y)\) of a customer conditioned to arrive at time \(x\) are studied in a \(GI/G/1\) queue with Laplace transforms \(\alpha(s)\), \(\beta(s)\) of the interarrival time, resp. service time. For the waiting time results, it is assumed that the queue is initially empty and that the renewal arrival process is stationary. For the \(n\)th customer in a cycle, define \(f_ n(x,y)\) as the joint density of his interarrival time relative to the start of the cycle and his waiting time, and \(r_ n(x,y)\) as the joint density of the busy period and minus the idle period. It is shown that the double Laplace transforms \(\Phi^ +(s,\omega)\), \(\rho^ -(s,\omega)\) of \(\sum_ 1^ \infty f_ n(x,y)\), resp. \(\sum_ 1^ \infty r_ n(x,y)\) are solutions to the Hilbert factorization problem \[ \Phi^ +(s,\omega)(1- \alpha(s-\omega)\beta(\omega))=1-\rho^ -(s,\omega), \] and that \(\sigma\), \(\Phi\) are explicit functions of \(\Phi^ +\), \(\rho^ -\). Explicit expressions for \(\Phi^ +\), \(\rho^ -\) and thereby \(\sigma\), \(\Phi\) are derived in terms of the roots of the equation \(\alpha(s- z)\beta(z)=1\) for the case where either \(\alpha(s)\) or \(\beta(s)\) is a rational function of \(s\).











This page was built for publication: Transient and busy period analysis of the GIG/1 Queue as a Hilbert factorization problem

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3988129)