• DocumentCode
    3094983
  • Title

    Study on Buffer Optimal Control Algorithm Based on Infinitesimal Perturbation Analysis for Broadband Satellite Networks

  • Author

    Lei, Wang ; Ning, He

  • Author_Institution
    Chinese People´´s Liberation Army, Beijing, China
  • fYear
    2009
  • fDate
    12-14 Dec. 2009
  • Firstpage
    523
  • Lastpage
    526
  • Abstract
    To solve the buffer management problem in Broadband satellite Networks, based on the IPA (Infinitesimal Perturbation Analysis) method, a new algorithm was proposed. The algorithm regarded the buffer management problem as one minimization of the performance function with the decision variable (specially, the buffer threshold) describing the nod with SFM(Stochastic Fluid Model). IPA technique was used to derive sensitivity estimators for the performance function, thus provided SA(Stochastic Approximation) algorithms capable to optimally minimize the performance cost based on IPA gradient. The algorithm can be evaluated based on data observed from the sample path of the real system without any pre-knowledge of the inflow.
  • Keywords
    broadband networks; optimal control; perturbation techniques; satellite communication; stochastic processes; IPA gradient; SA algorithm; SFM; broadband satellite networks; buffer management; buffer optimal control algorithm; infinitesimal perturbation analysis; sensitivity estimators; stochastic approximation; stochastic fluid model; Algorithm design and analysis; Approximation algorithms; Artificial satellites; Communication system control; Computer networks; Cost function; Optimal control; Optimization methods; Stochastic processes; Stochastic systems; IPA; buffer management; gradient; stochastic approximation.;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Dependable, Autonomic and Secure Computing, 2009. DASC '09. Eighth IEEE International Conference on
  • Conference_Location
    Chengdu
  • Print_ISBN
    978-0-7695-3929-4
  • Electronic_ISBN
    978-1-4244-5421-1
  • Type

    conf

  • DOI
    10.1109/DASC.2009.165
  • Filename
    5380408