Hostname: page-component-6766d58669-76mfw Total loading time: 0 Render date: 2026-05-16T02:03:12.236Z Has data issue: false hasContentIssue false
  • English
  • Français

Simulation-Based Computation of the Workload Correlation Function in a Lévy-Driven Queue

Part of: Stochastic processes Special processes

Published online by Cambridge University Press:  14 July 2016

Peter W. Glynn  and
Michel Mandjes
Peter W. Glynn*
Affiliation:
Stanford University
Michel Mandjes*
Affiliation:
University of Amsterdam, EURANDOM and CWI
*
∗ Postal address: Department of Management Science & Engineering, Stanford University, Stanford, CA 94305, USA. Email address: glynn@stanford.edu
∗∗ Postal address: Korteweg-de Vries Institute for Mathematics, University of Amsterdam, Science Park 904, 1098 XH Amsterdam, The Netherlands. Email address: m.r.h.mandjes@uva.nl
Rights & Permissions [Opens in a new window]

Abstract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the 'Save PDF' action button.

In this paper we consider a single-server queue with Lévy input, and, in particular, its workload process (Q t )t≥0, focusing on its correlation structure. With the correlation function defined as r(t):= cov(Q 0, Q t ) / varQ 0(assuming that the workload process is in stationarity at time 0), we first study its transform ∫0 ∞ r(t)e-ϑt dt, both for when the Lévy process has positive jumps and when it has negative jumps. These expressions allow us to prove that r(·) is positive, decreasing, and convex, relying on the machinery of completely monotone functions. For the light-tailed case, we estimate the behavior of r(t) for large t. We then focus on techniques to estimate r(t) by simulation. Naive simulation techniques require roughly (r(t))-2 runs to obtain an estimate of a given precision, but we develop a coupling technique that leads to substantial variance reduction (the required number of runs being roughly (r(t))-1). If this is augmented with importance sampling, it even leads to a logarithmically efficient algorithm.

Keywords

Lévy processreflectionworkload processcorrelation functionsimulationcouplingimportance sampling

MSC classification

Secondary: 60G51: Processes with independent increments; L'evy processes 60K25: Queueing theory

Information

Type
Research Article
Information
Journal of Applied Probability , Volume 48 , Issue 1 , March 2011 , pp. 114 - 130
Copyright
Copyright © Applied Probability Trust 2011 

References

[1] Abate, J. and Whitt, W. (1994). Transient behavior of the M/G/1 workload process. Operat. Res. 42, 750–764.CrossRefGoogle Scholar
[2] Abate, J. and Whitt, W. (1997). Asymptotics for M/G/1 low-priority waiting-time tail probabilities. Queueing Systems 25, 173–233.CrossRefGoogle Scholar
[3] Asmussen, S. (2003). Applied Probability and Queues, 2nd edn. Springer, New York.Google Scholar
[4] Asmussen, S. and Glynn, P. W. (2007). Stochastic Simulation: Algorithms and Analysis. Springer, New York.CrossRefGoogle Scholar
[5] Beneš, V. E. (1957). On queues with Poisson arrivals. Ann. Math. Statist. 28, 670–677.CrossRefGoogle Scholar
[6] Bernstein, S. N. (1929). Sur les fonctions absolument monotones. Acta Math. 52, 1–66.CrossRefGoogle Scholar
[7] Debicki, K., Es-Saghouani, A. and Mandjes, M. (2010). Transient asymptotics of Lévy-driven queues. J. Appl. Prob. 47, 109–129.CrossRefGoogle Scholar
[8] Dembo, A. and Zeitouni, O. (1998). Large Deviations Techniques and Applications, 2nd edn. Springer, New York.CrossRefGoogle Scholar
[9] Devroye, L. (1986). Nonuniform Random Variate Generation. Springer, Berlin.CrossRefGoogle Scholar
[10] Doney, R. A. (2005). Some excursion calculations for spectrally one-sided Lévy processes. In Séminaire de Probabilités XXXVIII (Lecture Notes Math. 1857), Springer, Berlin, pp. 5–15.CrossRefGoogle Scholar
[11] Ensor, K. and Glynn, P. (2000). Simulating the maximum of a random walk. J. Statist. Planning Infer. 85, 127–135.CrossRefGoogle Scholar
[12] Es-Saghouani, A. and Mandjes, M. (2008). On the correlation structure of a Lévy-driven queue. J. Appl. Prob. 45, 940–952.CrossRefGoogle Scholar
[13] Feller, W. (1971). An Introduction to Probability Theory and Its Applications, Vol. II, 2nd edn. John Wiley, New York.Google Scholar
[14] Kella, O., Boxma, O. J. and Mandjes, M. (2006). A Lévy process reflected at a Poisson age process. J. Appl. Prob. 43, 221–230.CrossRefGoogle Scholar
[15] Kyprianou, A. E. (2006). Introductory Lectures on Fluctuations of Lévy Processes with Applications. Springer, Berlin.Google Scholar
[16] Mandjes, M. (2007). Large Deviations for Gaussian Queues. John Wiley, Chichester.CrossRefGoogle Scholar
[17] Morse, P. M. (1955). Stochastic properties of waiting lines. Operat. Res. 3, 255–261.Google Scholar
[18] Ott, T. J. (1977). The covariance function of the virtual waiting-time process in an M/G/1 queue. Adv. Appl. Prob. 9, 158–168.CrossRefGoogle Scholar
[19] Pistorius, M. R. (2004). On exit and ergodicity of the spectrally one-sided Lévy process reflected at its infimum. J. Theoret. Prob. 17, 183–220.CrossRefGoogle Scholar
[20] Reynolds, J. F. (1975). The covariance structure of queues and related processes—a survey of recent work. Adv. Appl. Prob. 7, 383–415.CrossRefGoogle Scholar
[21] Williams, D. (1991). Probability with Martingales. Cambridge University Press.CrossRefGoogle Scholar
[22] Zolotarev, V. M. (1964). The first passage time of a level and the behaviour at infinity for a class of processes with independent increments. Theoret. Prob. Appl. 9, 653–661.CrossRefGoogle Scholar