Experimental performance evaluation of histogram approximation for simulation output analysis

We summarize the results of an experimental performance evaluation of using an empirical histogram to approximate the steady-state distribution of the underlying stochastic process. We use a runs test to determine the required sample size for simulation output analysis and construct a histogram by computing sample quantiles at certain grid points. The algorithm dynamically increases the sample size so that histogram estimates are asymptotically unbiased. Characteristics of the steady-state distribution, such as the mean and variance, can then be estimated through the empirical histogram. The preliminary experimental results indicate that the natural estimators obtained based on the empirical distribution are fairly accurate.