Discussion of 'Stochastic modelling and analysis of degradation for highly reliable products'

Abstract

The analysis of degradation data in fields of reliability research has become increasingly important over the past 20 years. Lu and Meeker 1 applied general regression techniques to repeated measurements data to help translate a common problem in both manufacturing and medical life testing into a convenient statistical framework, providing the genesis of statistical methodologies that followed. While this general path model was convenient to implement (for instance, Meeker and Escobar 2 provided an add-on S+ code (SPLIDA) for their textbook problems), by the turn of the last century, more research would focus on characterizing the degradation process of various materials using basic Wiener and other suitable processes. As the author implied in Section 3.4, the more sophisticated stochastic process models allowed for a more natural and realistic explanation for a degrading mechanism, but at the cost of using a more complex class of failure time models.