Control variates for stochastic network simulation

Examines procedures for using path control variates to improve the accuracy of simulation-based point and confidence-interval estimators of the mean completion time of a stochastic activity networks (SAN). Because each path control variate is the duration of the corresponding directed path in the network from the source to the sink, the vector of selected path controls has a known mean and a known covariance matrix. This information is incorporated into control-variate estimation procedures that do not require normally distributed responses. The simulation-generated observations are split into three groups, and control-variate procedures, applied within each group, are combined in such a way that the overall point estimator and the associated variance estimator are always unbiased. The performance of these procedures is evaluated experimentally. In comparison to standard linear control-variate procedures, the proposed procedures can yield substantial improvements in point-estimator accuracy and confidence-interval coverage while achieving almost the same magnitude of variance reduction