对数正态分布下恒定应力加速寿命试验的Bayes方法

Bayesian Approach to an Accelerated Life Test under Log-normal Distribution

恒定应力加速寿命试验是产品寿命试验中一种有效的、经济的试验方法。其理论日趋成熟,并在实践中得到了广泛的应用。本文在产品寿命服从对数正态分布的情况下,讨论了恒定应力加速寿命试验数据的Bayes统计分析方法,并用Gibbs抽样的迭代算法成功地解决了Bayes分析中极为复杂的后验边际分布的计算问题。实例表明这种方法是有效且可行的。

Constant-stress accelerated life test is a kind of effective and economic trial method in the product life span experiment. The theory of test is gradually maturing, and applied extensively in practice. The Bayesian statistical data analysis method for a constant stress, accelerated life test to non-censoring data under log-normal distribution was discussed, and by using the iterative algorithm in Gibbs sampling the extremely complex calculating problem of the posterior marginal distribution is solved successfully. The example illustrates that the approach is effective and practical.