Point and interval estimation of decomposition error in discrete-time open tandem queues

We analyze the approximation quality of the discrete-time decomposition approach, compared to simulation, and with respect to the expected value and the 95th-percentile of waiting time. For both performance measures, we use OLS regression models to compute point estimates, and quantile regression models to compute interval estimates of decomposition error. The ANOVA reveal major influencing factors on decomposition error while the regression models are demonstrated to provide accurate forecasts and precise confidence intervals for decomposition error.