5th International ICST Conference on Performance Evaluation Methodologies and Tools

Research Article

A Taylor series expansions approach to queues with train arrivals

Download587 downloads
  • @INPROCEEDINGS{10.4108/icst.valuetools.2011.245768,
        author={Koen De Turck and Dieter Fiems and Sabine Wittevrongel and Herwig Bruneel},
        title={A Taylor series expansions approach to queues with train arrivals},
        proceedings={5th International ICST Conference on Performance Evaluation Methodologies and Tools},
        publisher={ICST},
        proceedings_a={VALUETOOLS},
        year={2012},
        month={6},
        keywords={queueing theory; Taylor series; M/G/infinity},
        doi={10.4108/icst.valuetools.2011.245768}
    }
    
  • Koen De Turck
    Dieter Fiems
    Sabine Wittevrongel
    Herwig Bruneel
    Year: 2012
    A Taylor series expansions approach to queues with train arrivals
    VALUETOOLS
    ICST
    DOI: 10.4108/icst.valuetools.2011.245768
Koen De Turck1,*, Dieter Fiems1, Sabine Wittevrongel1, Herwig Bruneel1
  • 1: Ghent University
*Contact email: koen.deturck@gmail.com

Abstract

We assess the performance of a discrete-time queueing system with train arrivals. Arrivals at the queue stem from a number of active sessions, each generating a packet in a slot with fixed probability $q$. Since an exact analysis is not feasible for $q \ne 1$, we rely on Taylor-series expansions around $q=0$ of the joint probability generating functions of the number of active sessions and the queue content. These expansions are then either combined with the known generating function for $q=1$ if the system is stable for $q=1$, or with heavy-traffic results if this is not the case. In both cases, we obtain expressions for the moments of queue content and packet delay and assess the accuracy of our approximations by means of simulation.