A0230
Title: Increment degradation model: A Bayesian perspective
Authors: I-Tang Yu - Tunghai University (Taiwan) [presenting]
Abstract: One frequently employed approach for describing the degradation phenomenon involves using a degradation model that relies on stochastic processes. In a stochastic-process-based degradation model, the increments are assumed to follow a distribution with the additivity property. This property makes further inferences mathematically and statistically tractable. However, it limits the choices of the distributions. The aim is to use those distributions without the additivity property to model the increments. Under the frame of Bayesian analysis, Markov Chain Monte Carlo algorithms are developed to execute the necessary computations. Given that the proposed degradation models do not adhere to the additivity property, the challenges involved in predicting the lifetime of both online and offline products are tackled. The suitability of the proposed model is finally validated through a simulation study.
