Online Program Home
My Program

Abstract Details

Activity Number: 592 - New Developments in Experiment Design and Statistical Modeling
Type: Contributed
Date/Time: Wednesday, August 1, 2018 : 2:00 PM to 3:50 PM
Sponsor: International Chinese Statistical Association
Abstract #329687 Presentation
Title: Optimal Doubling Burn-In Policy Based on Tweedie Processes with Applications to Degradation Data
Author(s): Chien-Yu Peng*
Companies: Institute of Statistical Science, Academia Sinica
Keywords: compound Poisson; Tweedie process; degradation model
Abstract:

In the current competitive marketplace, manufacturers need to screen weak products in a short period of time. It is a challenge for manufacturers to implement a burn-in test that can screen out the weak products quickly and efficiently. When selecting an approach to determine the duration of the burn-in, one could build a criterion aiming to minimize the burn-in cost. In practice, when the optimal burn-in time is unreasonable (e.g., time 0) due to minimizing the cost, this means that the burn-in procedure is unnecessary to perform for manufacturers. In this study, we propose an optimal doubling burn-in policy to improve the predicament without additional experiments. The advantage of the proposed policy is to simultaneously determine the optimal burn-in time and the optimal cutoff point for classifying weak and strong components from the production. In addition, a new degradation model based on a Tweedie mixture process is used for a burn-in test. The proposed burn-in procedure is applied to a real data.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program