Performance evaluation of a wavelet-based spectral method for steady-state simulation analysis

We summarize the results of an experimental performance evaluation of WASSP, an automated wavelet-based spectral method for constructing an approximate confidence interval on the steady-state mean of a simulation output process so that the delivered confidence interval satisfies user-specified requirements on absolute or relative precision as well as coverage probability. We applied WASSP to test problems designed specifically to explore its efficiency and robustness in comparison with ASAP3 and the Heidelberger-Welch algorithm, two sequential procedures based respectively on the methods of nonoverlapping batch means and spectral analysis. Concerning efficiency, WASSP compared favorably with its competitors, often requiring smaller sample sizes to deliver confidence intervals with the same nominal levels of precision and coverage probability. Concerning robustness against the statistical anomalies commonly encountered in simulation studies, WASSP outperformed its competitors, delivering confidence intervals whose actual half-lengths and coverage probabilities were frequently closer to the corresponding user-specified nominal levels.