Bootstrap generated confidence interval for time averaged measure
作者机构:School of Management and AdministrationYong In University YonginRepublic of Korea School of EngineeringSungkyunkwan University SuwonRepublic of Korea
出 版 物:《International Journal of Modeling, Simulation, and Scientific Computing》 (建模、仿真和科学计算国际期刊(英文))
年 卷 期:2015年第6卷第3期
页 面:107-115页
核心收录:
学科分类:07[理学] 0701[理学-数学] 070101[理学-基础数学]
主 题:Simulation output analysis confidence interval time averaged measure bootstrap
摘 要:In the simulation output analysis,there are some measures that should be calculated by time average concept such as the mean queue ***,the confidence interval of those measures might be required for statistical *** this situation,the traditional method that utilizes the central limit theorem(CLT)is inapplicable if the output data set has autocorrelation *** bootstrap is one of the most suitable methods which can reflect the autocorrelated phenomena in statistical ***,the confidence interval for a time averaged measure having autocorrelation structure can also be calculated by the bootstrap *** study introduces the method that constructs these confidence intervals applying the *** bootstraps proposed are the threshold bootstrap(TB),the moving block bootstrap(MBB)and stationary bootstrap(SB).Finally,some numerical examples will be provided for verification.