Abstract
The study of an alternative robustness definition for batching estimators, related with the variance of sample mean, was investigated. The steady-state simulation experiments were considered for studying the estimators of sample mean variance. The alternative definition proved that the overlapping-batch-means (OBM) estimators possessed both the optimal mean squared error (MSE) and optimal robustness. The comparison results of different estimators based on MSE optimal performance and minimal asymptotic MSE were also discussed.
Original language | English |
---|---|
Pages (from-to) | 293-298 |
Number of pages | 6 |
Journal | Operations Research Letters |
Volume | 32 |
Issue number | 3 |
DOIs | |
State | Published - May 2004 |
Keywords
- Monte Carlo
- Simulation
- Standard error
- Statistical analysis