Some new results on the initial transient problem

ABSTRACT

This paper contains two new results pertaining to the initial transient problem for steady-state simulations. Our first result rigorously establishes the asymptotic superiority of a few long replications relative to a large number of shorter replications, assuming that no initial transient deletion is attempted. Our second result concerns an initial transient detection test proposed by Schruben; we develop asymptotics that are suggestive of the types of initial transients that the test is capable of detecting. As one might expect, the ability to detect a non-stationarity in the simulation output depends both on the magnitude of the non-stationarity of the initial condition, and the degree of autocorrelation in the process.

REFERENCES

• 1: Paul Bratley , Bennett L. Fox , Linus E. Schrage, A guide to simulation (2nd ed.), Springer-Verlag New York, Inc., New York, NY, 1987
• 2: CsSrgS, M. and P. R@v@sz. 1981. Strong approximations in probability and statistics. Academic Press, New York.
• 3: Glynn, Peter W. 1984. Some asymptotic formulas for Markov chains with applications to simulation. Journal of Statistical Computation and Simulation 19:97-112.
• 4: Glynn, Peter W. 1987. Limit theorems for the method of replication. Stochastic Models 4:343-350.
• 5: P. W. Glynn , D. L. Iglehart, Simulation output analysis using standardized time series, Mathematics of Operations Research, v.15 n.1, p.1-16, Feb. 1990 [doi>10.1287/moor.15.1.1]
• 6: Lindvall, T. 1992. Lectures on the Coupling Method. John Wiley, New York.
• 7: Meyn, S. P. and R. L. Tweedie. 1993. Markov chains and stochastic stability. Springer-Verlag, New York.
• 8: Philipp, W. and W. Stout. 1975. Almost sure invariance principles for partial sums of weakly dependent random variables. Mem. Amer. Math. Soc. No. 161.
• 9: Schruben, L. W. 1982. Detecting initialization bias in simulation output. Operations Research 30:569- 590.
• 10: Ward Whitt, The efficiency of one long run versus independent replication in steady-state simulation, Management Science, v.37 n.6, p.645-666, June 1991 [doi>10.1287/mnsc.37.6.645]

INDEX TERMS

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

Additional Classification:
  G. Mathematics of Computing
  G.1 NUMERICAL ANALYSIS

  I. Computing Methodologies
  I.6 SIMULATION AND MODELING

General Terms:
Theory, Verification