Research Planning Grant: Inference Techniques for Estimatingthe Variance of Means of Simulation Output Data

本研究计划拨款的目的是从模拟输出数据中获得样本均值方差的精确、稳态估计。提出了一种新的方差估计方法，通过增强现有的估计器及其性能来估计方差。研究工作包括技术的理论发展和与其他估计器进行比较的实证研究，以基准性能。所需后续研究的长期目标是设计用于稳态仿真研究的自动运行长度控制系统。本研究可望为计算机仿真实验分析知识库的建立做出理论和实践上的贡献。

The objective of this Research Planning Grant is to obtain precise, steady-state estimates of the variance of the sample mean from simulation output data. A new inference technique is proposed for estimating the variance by augmenting existing estimators and their performance. The research effort includes theoretical development of the techniques and empirical studies in comparison with other estimators to benchmark performance. The long-term goal of the required follow-up study is the design of an automatic run length control system for steady state simulation studies. This research is anticipated to make theoretical and practical contributions to the knowledge base of computer simulation experiment analysis.