Simulating Stable Stochastic Systems, VI: Quantile Estimation

ABSTRACT

In this paper the author continues his study of the regenerative method for analyzing simulations of stable stochastic systems. The principal concern is to estimate the quantiles of the stationary distribution of a regenerative process. Markov chains in discrete or continuous time and multiple server queues in light traffic provide concrete examples of regenerative processes to which this technique applies. Approximate confidence intervals for these quantiles are derived from appropriate central limit theorems. The method has been applied to three stochastic simulations, and the numerical results are presented.

REFERENCES

• 1: Michael A. Crane , Donald L. Iglehart, Simulating Stable Stochastic Systems, I: General Multiserver Queues, Journal of the ACM (JACM), v.21 n.1, p.103-113, Jan. 1974 [doi>10.1145/321796.321805]
• 2: Michael A. Crane , Donald L. Iglehart, Simulating Stable Stochastic Systems, II: Markov Chains, Journal of the ACM (JACM), v.21 n.1, p.114-123, Jan. 1974 [doi>10.1145/321796.321806]
• 3: CRANE, M.A., AND IGLEHART, D.L. Simulating stable stochastzc systems, III: Regenerative processes and discrete-event simulations. Oper. Res 23 (1975), 33-45
• 4: CRANE, M.A, AND }GLEHART, D L. Simulating stable stochastic systems, IV: Approximation techniques. Manage. Sc~. 21 (1975), 1215-1224.
• 5: GOODMAN, A.S, LEwis, P.A.W., A~D ROBmNS, H.E. Simultaneous estimation of a large number of extreme quantiles in simulation experiments. (Submitted to a techmca} journal )
• 6: HORDIJK, A., IGLEHART, D.L., AND SCHASS~ERGER, R. Discrete time methods of simulating continuous time Markov chains. Tech. Rep. No. 35, Dep. of Oper. Res., Stanford U., Stanford, Calif., 1975.
• 7: IGLEHART, D.L. Simulating stable stochastic systems, VI: Quantile estimation. Tech. Rep. No. 86-15, Contro Analysis Corp., Palo Alto, Calif, 1974.
• 8: IGLEHART, D.L. Simulating stable stochastic systems, V: Comparison of ratio estimators. Naval Res. Log~st. Quart. ~ (1975), 553-565.
• 9: Donald E. Knuth, The Art of Computer Programming Volumes 1-3 Boxed Set, Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1998
• 10: LEwls, P.A.W. Large-scale computer-aided statistical mathematics Proc. of the Comp. Sci. and Star. Sixth Ann. Symposium on the interface, M.E. Tarter, Ed., U., of Cahfornia, Berkeley, Calif., 1973
• 11: MXLLBR, D.R. On the asymptotic behavior of regenerative processes and functionals of regenerative processes Tech Rep. No. 137, Dep. of Oper. Res., Cornell U., ithaca, N.Y., 1971.
• 12: MILLER, D.R Existence of limits in regenerative processes. Ann. Math. Statist. ~3 (1972), 1275-1282.
• 13: PROHOROV, Yu.V. Convergence of random processes and hmit theorems in probability theory. Theor Probability Appl. 1 (1956), 157-214 (Eng transl.)
• 14: WEISS, L. Asymptotic distributions of quantiles in some nonstandard cases. Nonparametric Techniques zn Statistical Inference, M. Purl, Ed., Cambridge U. Press, New York, 1970, pp. 343-348.

INDEX TERMS

Primary Classification:
  G. Mathematics of Computing
  G.3 PROBABILITY AND STATISTICS
      Subjects: Stochastic processes

Additional Classification:
  I. Computing Methodologies
  I.6 SIMULATION AND MODELING

General Terms:
Design, Measurement, Performance, Theory, Verification