恒加速寿命试验广义半正态分布的统计推断-《集美大学学报（自然科学版）》

文章信息/Info

Title:: Statistical Inference of Generalized Half-Normal Distribution for Constant Accelerated Life Tests

Author(s):: HAN Xiurong1; CHEN Jianwei2; CAI Nanlian1; 1.School of Science,Jimei University,Xiamen 361021，China;2.Department of Mathematics and Statistics,San Diego State University,San Diego 92182,USA

Keywords:: accelerated life test; geometric process; maximum likelihood estimation; bootstrap confidence; asymptotic distribution

摘要:: 在恒加速寿命试验下，研究具有广义半正态分布寿命的参数估计问题。证明当恒加速寿命试验的应力和广义半正态分布参数满足一定条件时，寿命形成了一个几何过程；探讨模型参数的最大似然估计、渐近置信区间和bootstrap置信区间的求法。对得到的估计结果进行数值模拟，模拟结果表明，模型参数估计值及置信区间是有效的。

Abstract:: In this paper，under constant-stress accelerated life test，the parameter estimations on a population with generalized half-normal distribution were considered.Firstly，it was proved that the product life forms a geometric process when certain conditions of the stress and parameters of generalized half-normal distribution are satisfied.Secondly，the maximum likelihood estimations，asymptotic confidence interval and bootstrap confidence interval estimations of the parameters were proposed. Finally，the obtained estimation results were numerically simulated,which showed that the model parameter estimations and confidence interval estimations were valid.