Combined Variance Reduction and Output Analysis in Stochastic Simulation

提出的研究旨在开发新的方法，同时应用方差减少技术，精确的点估计，输出分析方法，有效的区间估计，在随机模拟实验。将执行三项任务：研究稳态模拟实验的组合方差约简技术和输出分析方法，将现有的方差约简技术扩展到终止模拟的多变量参数估计，并开发了一个框架，用于在依次进行的实验中选择和参数化方差约简技术。比如通过模拟进行优化。这些任务的成功完成将导致更广泛地应用有效的方差缩减和适当的输出分析。

The proposed research seeks to develop new methods for simultaneously applying variance reduction techniques, for precise point estimation, and output analysis methods, for valid interval estimation, in stochastic simulation experiments. Three tasks will be performed: examination of combined variance reduction techniques and output analysis methods for steady-state simulation experiments, the extension of existing variance reduction techniques to multivariate parameter estimation in terminating simulations, and the development of a framework for selecting and parameterizing variance reduction techniques in experiments performed sequentially, such as optimization via simulation. Successful completion of these tasks will lead to wider application of effective variance reduction and appropriate output analysis.