A study on Bayesian design of degradation tests with the inverse Gaussian processes

摘要

Besides the Wiener and Gamma processes, the inverse Gaussian (IG) process is recently proposed as an attractive yet flexible family for degradation modeling. Since degradation test depends heavily on the degradation model chosen for a product's degradation process, we discuss the optimal design for degradation tests specifically based on the IG process. Other than an optimal design with pre-estimated planning values of model parameters, we handle the situation with uncertainty in the planning values using the Bayesian method. The inspection frequency and measurement numbers are included as design variables. The average pre-posterior variance of reliability is defined as the optimization criterion. An application to the degradation test planning of a GaAs Laser device is used to demonstrate the proposed method.