• DocumentCode
    2816511
  • Title

    Sensitivity analysis of loss volume in stochastic fluid systems with flow control

  • Author

    Adams, R. ; Wardi, Y. ; Riley, G.

  • Author_Institution
    Georgia Inst. of Technol., Atlanta
  • fYear
    2007
  • fDate
    12-14 Dec. 2007
  • Firstpage
    1510
  • Lastpage
    1515
  • Abstract
    This paper concerns infinitesimal perturbation analysis (IPA) for fluid queues with finite buffers, where input fluid in excess of buffer capacity is turned away and considered lost. In particular, it considers the loss-volume (and hence average loss rate) performance measure with respect to the buffer sizes. Most previous results pertained to open- loop systems, but recently there has been a mounting interest in feedback control, and especially in workload-based flow control systems. This paper considers an alternative flow- control mechanism whereby the inflow rate is restricted by an amount that is proportional to the loss rate. It first discusses a single-server queue, and then it analyzes a two-queue tandem network, while pointing out possible extensions to more general- topology networks and other performance functions. The main results are comprised of recursive algorithms for computing the IPA derivatives in nonparametric and model-free forms.
  • Keywords
    feedback; queueing theory; sensitivity analysis; telecommunication control; telecommunication network topology; buffer capacity; feedback control; flow control; fluid queues; general-topology networks; infinitesimal perturbation analysis; loss volume; open-loop systems; sensitivity analysis; single-server queue; stochastic fluid systems; two-queue tandem network; Control systems; Fluid flow control; Loss measurement; Open loop systems; Particle measurements; Performance loss; Queueing analysis; Sensitivity analysis; Size measurement; Stochastic systems;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Decision and Control, 2007 46th IEEE Conference on
  • Conference_Location
    New Orleans, LA
  • ISSN
    0191-2216
  • Print_ISBN
    978-1-4244-1497-0
  • Electronic_ISBN
    0191-2216
  • Type

    conf

  • DOI
    10.1109/CDC.2007.4434139
  • Filename
    4434139