The Impact of Transients on Simulation Variance Estimators

Given a stationary simulation process with unknown mean µ , interest frequently lies in, and various methods exist for, developing estimates and confidence intervals for µ . Typically, the sample mean is used as the point estimate for µ . It is also useful to estimate the variance parameter, σ 2 , a measure of sample mean’s precision. While there are many methods for estimating the variance parameter for such processes, they usually assume that the process has reached steady state before data collection begins. If this is not the case, then transient behavior can have a significant impact on the estimates of µ and σ 2 . We present empirical evidence which suggests that transient behavior distorts some variance estimators much more than others. Specifically we consider batchmeans estimators and standardized time series based Lpnorm estimators; and we show that the batch-means estimators appear to be significantly less robust to bias.

• Type: Conference
• Year: 1997
• Where: WSC
• Authors: Daniel H. Ockerman, David Goldsman
• Keywords: Modeling And Simulation, Sample Mean, Transient Behavior, Variance Parameter, WSC 1997