A new estimation approach based on phi-divergence measures for one-shot device accelerated life testing

Abstract

One-shot device testing data are used only once and they get destroyed when tested. As these products usually have large mean times to failure under normal operating conditions, accelerated life tests are commonly used to infer their lifetime distribution. While much work has been done to determine the maximum likelihood estimates (MLEs) of model parameters for one-shot device accelerated life testing, the efficiency of these methods may not be guaranteed for small to moderate sample sizes. In this paper, we develop new estimators and confidence intervals based on phi-divergences, we show that they outperform the conventional MLE under different lifetime distributions and present an example to illustrate all the inferential methods developed here