Title: Planning accelerated life tests via GLMs
Authors: Rong Pan - Arizona State University (United States) [presenting]
Abstract: A generalized linear model (GLM) approach to reliability data analysis is presented. Specifically, the failure time or censoring time data that are generated from an accelerated life test can be formulated by the GLM that defines the relationship between these test response data and the stress factor that was applied in the test. Through this modeling approach, it is easy to derive the test plan that can maximize its statistical efficiency in terms of the D-optimality or U-optimality criterion, etc. We also developed an R package, ALTopt, which can help practitioners to obtain optimal test plans even with multiple stress factors.