Advertisement

Boundary conditions for diffusion approximations to queueing problems

  • René Boel
Diffusion Approximations
Part of the Lecture Notes in Control and Information Sciences book series (LNCIS, volume 60)

Abstract

One of the main difficulties of the diffusion approximation to queues is that it underestimates the probability of an empty queue. This is due to the use of instantaneous reflection at the boundary. In this paper the use of diffusion processes with delayed reflections is suggested. A survey of the available mathematical literature on this topic is given. A method is suggested to give a rational choice of the stickyness of the boundary, based on how long the queue is empty in an interval up to the first buffer overflow. Some applications (ergodic theorems and recursive estimation) and some problems for further research are outlined.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

7. References

  1. 1.
    D. Gaver and G. Shedler: Multiprogramming system performance via diffusion approximations, IBM Res. Rep. RJ-938, 1971.Google Scholar
  2. 2.
    H. Kobayashi: Application of the diffusion approximation to queueing networks, parts I and II, J.A.C.M. 21 (1974), p. 316–328; p. 459–469.Google Scholar
  3. 3.
    W. Whitt: Weak convergence for queues in heavy traffic, Ph.D. Thesis, Dept. of Op. Res., Cornell University, 1968.Google Scholar
  4. 4.
    H. McKean: Stochastic Integrals, Academic Press, 1969.Google Scholar
  5. 5.
    K. Ito and H. McKean: Diffusion processes and their sample path, Springer, 2nd printing, 1974.Google Scholar
  6. 6.
    D. Williams: Diffusions, Markov Processes and Martingales, vol 1, Wiley, 1979.Google Scholar
  7. 7.
    M. Reiman: Queueing networks in heavy traffic, Ph.D. dissertation, Dept. of Op. Res., Stanford University, 1977.Google Scholar
  8. 8.
    R. Boel and M. Kohlmann: A control problem in a manifold with nonsmooth boundary, to appear in Proceedings of the 2nd Bad Honnef Workshop on Stochastic Differential Systems, 1982.Google Scholar
  9. 9.
    E. Gelenbe and I. Mitrani: Analysis and Synthesis of Computer Systems, Academic Press, 1980.Google Scholar
  10. 10.
    M. Harrison and A.J. Lemoine: Sticky Brownian motion as the limit of storage processes, J. Appl. Prob. 18 (1981), p. 886–905.Google Scholar
  11. 11.
    I. Gihman and A.V. Skorohod: Stochastic Differential equations, Springer, 1972.Google Scholar
  12. 12.
    D. Stroock and S. Varadhan: Diffusion, processes with boundary conditions, Comm. Pure and Appl. Math, 24 (1971), p. 147–225.Google Scholar
  13. 13.
    R. Boel: Martingale methods for the semi-Markov analysis of queues with blocking, Stochastics, 5 (1981), p. 115–133.Google Scholar
  14. 14.
    L. Breiman: Probability, Addison Wesley, 1968.Google Scholar
  15. 15.
    E. Wong: Stochastic Processes in Information and Dynamical Systems, McGraw Hill, 1971.Google Scholar
  16. 16.
    M. Davis and S. Marcus: An introduction to nonlinear filtering, in: Stochastic Systems: the mathematics of filtering and identification and applications, M. Hazewinkel and J. Willems, eds., p. 53–75, Reidel, 1981.Google Scholar
  17. 17.
    P. Brémaud: La méthode des semi-martingales en filtrage quand l'observation est un processus ponctuel marqué, Séminaire de Probabilités, X, Lecture Notes in Mathematics, vol. 511, Springer, 1976.Google Scholar
  18. 18.
    D. Kennedy: Some martingales related to cumulative sum tests and single server queues, Stochastic Processes and their Applications, 4 (1976), p. 261–269.Google Scholar
  19. 19.
    W. Rozenkrantz: Some martingales associated with queueing and storage processes, to appear in Zeit.für Wahrscheinlichkeitstheorie.Google Scholar
  20. 20.
    R. Boel: Some comments on control and estimation problems for diffusions in bounded regions, to appear in: Proceedings of the ENST-CNET Symposium on Estimation, February 1983.Google Scholar

Copyright information

© Springer-Verlag 1984

Authors and Affiliations

  • René Boel
    • 1
    • 2
  1. 1.Lab. for System DynamicsRijksuniversiteit GentZwijnaarde (Gent)Belgium
  2. 2.Department of Systems EngineeringA.N.U.CanberraAustralia

Personalised recommendations