Arma-Based Confidence Intervals for Simulation Output Analysis

SYNOPTIC ABSTRACT

A method is presented for obtaining a confidence interval for the mean of a stationary stochastic process. The method fits an autoregressive moving average (ARMA) model to a sequence of sample outputs. The effectiveness of the confidence interval procedure is measured by applying the procedure to simulation output sequences generated by ARMA models and an M/M/1 queueing system model. Performance characteristics of the procedure are excellent for ARMA output sequences, and the coverage achieved with queueing data is promising.

Key Words and Phrases:

• autoregressive moving average
• confidence interval
• output analysis
• simulation