仿真输出中位置参数的区间估计方法比较

概述和比较了仿真输出分析中对位置参数进行置信区间估计的方法 ,分析了这些方法的适用条件及传统方法的局限性 .探讨了当经典方法所依赖的有关总体分布的假定不成立时进行区间估计的选择方法 ,并通过仿真实验进行了验证 .仿真实验结果表明 ,该方法在大多数情形中能给出有效的置信区间 ,而且仿真精度较传统方法有明显提高 .

Comparison for Interval Estimators of the Locational Parameters in Simulation Output

The methods for making confidence interval estimator in simulation output analysis are reviewed and compared. Classical estimation procedures that are often used in confidence interval estimation rely on the assumption about the population distribution. Empirical evidence suggests however, that this may not be a valid assumption. This study illustrates the use of the different estimating methods, the suited conditions and their limits. The results of Monte Carlo simulations indicate that the method (method 2) proposed in this study is more effective and efficient than the classical procedure.