Analyzing multivariate output

ABSTRACT

This paper gives an overview of multivariate statistical techniques that can be useful for analyzing discrete-event simulation output, and describes some of the latest directions in research on multivariate output analysis. A general discussion is given of constructing joint confidence regions on the mean vector of multivariate output from independent replications of terminating models. The multivariate batch means method of simultaneous estimation of means from one long run of steady-state simulation models is described. References are also given for autoregressive, spectral analysis and regenerative methods of inference, as well as variance-reduction and sequential techniques.

REFERENCES

• 1: Anderson, T. W. 1978. Repeated measurements on autoregressive processes, Journal of The American Statistical Association, 73: 371-378.
• 2: Anderson, T.W. 1984. An Introduction to Multzvamate Statist,cal Analys,s, Second Edition. New York: John Wiley.
• 3: Paul Bratley , Bennett L. Fox , Linus E. Schrage, A guide to simulation (2nd ed.), Springer-Verlag New York, Inc., New York, NY, 1987
• 4: John M. Charnes, Power comparisons for the multivariate batch-means method, Proceedings of the 22nd conference on Winter simulation, p.281-287, December 09-12, 1990, New Orleans, Louisiana, United States
• 5: John M. Charnes, Multivariate simulation output analysis, Proceedings of the 23rd conference on Winter simulation, p.187-193, December 08-11, 1991, Phoenix, Arizona, United States
• 6: John M. Charnes , W. David Kelton, A comparison of confidence region estimators for multivariate simulation output, Proceedings of the 20th conference on Winter simulation, p.458-465, December 12-14, 1988, San Diego, California, United States [doi>10.1145/318123.318233]
• 7: John M. Charnes , W. David Kelton, Multivariate autoregressive techniques for constructing confidence regions on the mean vector, Management Science, v.39 n.9, p.1112-1129, Sept. 1993 [doi>10.1287/mnsc.39.9.1112]
• 8: D.-F. R. Chen, Multivariate inferences for regenerative simulations to a specified precision level, Proceedings of the 21st conference on Winter simulation, p.524-533, December 04-06, 1989, Washington, D.C., United States [doi>10.1145/76738.76806]
• 9: Chen, D.R. and Cheng, W.J. 1989. Multivariate statistical analysis for regenerative simulation, Journal of Management Science, 6: 43-56.
• 10: Chen, D.R. 1991. The relationship between multivari~te correlaticm and coverage rate, presentation at TIMS XXX~SOBRAPO XXIII Joint International Meeting.
• 11: Der-Fa R. Chen , Andrew F. Seila, Multivariate inference in stationary simulation using batch means, Proceedings of the 19th conference on Winter simulation, p.302-304, December 14-16, 1987, Atlanta, Georgia, United States [doi>10.1145/318371.318436]
• 12: Gallagher, M. A., Bauer, K. W., and Maybeck, P. S. 1994. Initial data truncation for multivariate output of discrete-event simulations using the Kalman filter. In Annals of Operatzons Research, ed. O. Balci, 53: 419.
• 13: R. A. Johnson , D. W. Wichern, Applied multivariate statistical analysis, Prentice-Hall, Inc., Upper Saddle River, NJ, 1988
• 14: Kabaila, P. and G. Nelson. 1985. On confidence regions for the mean of a multivariate time series, Communications in Statistics--Theory and Methods, 14: 735-753.
• 15: Averill M. Law , W. David Kelton, Simulation Modeling and Analysis, McGraw-Hill Higher Education, 1997
• 16: Miller, R.. G. Jr. 1981. Simultaneous StatzsticalInference, Second Edition. New York: Springer-Verlag.
• 17: Morrison, D.F. 1976. Multivariate Statistical Methods, New York: McGraw-Hill. Munoz, D. F. 1991. Cancellation methods in the analysis of simulation output (confidence intervals). Ph.D. thesis, Department of Operations Research, Stanford University.
• 18: Kimmo E. E. Raatikainen, A sequential procedure for simultaneous estimation of several means, ACM Transactions on Modeling and Computer Simulation (TOMACS), v.3 n.2, p.108-133, April 1993 [doi>10.1145/169702.169690]
• 19: Rao, C.R. 1951. An asymptotic expansion of the distribution of Wilk's criterion. Bulletin of the International Slatist~cal Institute, 33: 177-180.
• 20: Schmeiser, B. 1982. Batch size effects in the analysis of simulation output. Operations Research, 30: 556-568.
• 21: Lee W. Shruben, Control of initialization bias in multivariate simulation response, Communications of the ACM, v.24 n.4, p.246-252, April 1981 [doi>10.1145/358598.358634]
• 22: Seila, A. F. 1982. Multivariate estimation in regenerative simulation, Operatzons Research Letters, 1: 153-156.
• 23: Andrew F. Seila, Multivariate estimation in simulation, Proceedings of the 15th conference on Winter Simulation, p.693-698, December 12-14, 1983, Arlington, Virginia, United States
• 24: Seila, A. F. 1984. Multivariate simulation output analysis, A memcan Journal of Mathematical and Management Sciences, 4: 313-334.
• 25: Seila, A.F. 1990. Multivariate Estimation of Conditional Performance Measures in Regenerative Sireulation. American Journal of Mathematical and Management Sciences 10: 17-50.
• 26: Wei-Ning Yang , Barry L. Nelson, Multivariate estimation and variance reduction in terminating and steady-state simulation, Proceedings of the 20th conference on Winter simulation, p.466-472, December 12-14, 1988, San Diego, California, United States [doi>10.1145/318123.318234]
• 27: Haigh, J.C. Comfort, 466-472. Institute of Electrical and Electronics Engineers, San Diego, California.
• 28: Wei-Ning Yang , Barry L. Nelson, Multivariate batch means and control variates, Management Science, v.38 n.10, p.1415-1431, Oct. 1992 [doi>10.1287/mnsc.38.10.1415]

INDEX TERMS

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

Additional Classification:
  G. Mathematics of Computing
  G.1 NUMERICAL ANALYSIS
      G.1.4 Quadrature and Numerical Differentiation
          Subjects: Equal interval integration**
  G.3 PROBABILITY AND STATISTICS
      Subjects: Statistical computing

  I. Computing Methodologies
  I.6 SIMULATION AND MODELING
      I.6.1 Simulation Theory
          Subjects: Model classification
      I.6.8 Types of Simulation
          Subjects: Discrete event

General Terms:
Theory, Verification