Pilot Study on the Applicability of Variance Reduction Techniques to the Simulation of a Stochastic Combat Model

摘要

This thesis investigations the applicability of VRTs to simulation of stochastic combat models. Ways of measuring the efficiency of a VRT are explored. Antithetic variates and stratified sampling are applied to the simulation of a trivariate Markovian combat model means of programming the antithetic variates and stratified sampling to reduce the inherent variability of uncertainty in the output data of the model are illustrated. Response surface regression models are used to characterize the performance of the antithetic variates and stratified sampling in the Markovian combat model.