Degradation modeling considering unit-to-unit heterogeneity-A general model and comparative study

The performance of units in the same batch can exhibit considerable heterogeneity due to variation raw materials and fluctuation manufacturing process. For products suffering degradation their use, such often results an increase dispersion paths a population. rate be unit-specific treated as random effects. This paper develops novel random-effects Wiener process model account for unit-to-unit degradation, where generalized inverse Gaussian (GIG) distribution is used rate. GIG very general with broad applications, which includes (IG) Gamma special cases. We investigate properties develop expectation maximization (EM) algorithm parameter estimation. By comparing proposed existing models on two real datasets infrared LEDs GaAs lasers, we show that quite effective modeling heterogeneous rates.