The previous studies pertaining to the queues Geo/G/1 and Geo/G/1/N involve light-tailed distributions for service time. However, due to applications of heavy-tailed distributions in computer science and financial engineering, these distributions are used for service time. This article provides a simple and computationally efficient solution to the queues Geo/G/1 and Geo/G/1/N involving heavy-tailed distributions for service times.
TakagiH.Queuing analysis: A foundation of performance evaluation—Volume 3: Discrete time systems. New York, NY: North-Holland. 1993.
2.
HunterJJ.Mathematical techniques of applied probability, Volume II: Discrete time models—Techniques and applications. New York, NY: Academic Press. 1983.
3.
ChaudhryML.On the steady-state distribution of number in the queue Geom/G/1. Probab Eng Informat Sci. 1999; 13;121–125.
4.
ShortleJ.F.BrillP.H.FischerM.J.GrossD.MasiD.M.B.An algorithm to compute the waiting time distribution for the M/G/1 queue. INFORMS J Comput. 2004; 16;152–161.
5.
ChaudhryML.On numerical computations of some discrete-time queues. In: Grassmann WK, editor. Computational probability. Boston, MA: Kluwer Academic Publishers; 2000. pp. 365–407.
6.
BruneelHKimBG.Discrete-time models for communication systems including ATM. Volume 205. New York, NY: Springer Science & Business Media. 1993.
7.
GouweleeuwFN.A general approach to computing loss probabilities finite-buffer queues. PhD Thesis. Amsterdam: Vrije Universiteit. 1996.
8.
ChaudhryML. QROOT software package. Kingston ON: A& A Publications. 1991.
9.
ChaudhryMLHarisCMMarchalWG.Robustness of root finding in single-server queuing models. INFORMS J Comput1990; 2;273–286.
