Optimal burn-in procedure for mixed populations based on the device degradation process history

• This paper proposes a new degradation history-based burn-in procedure.
• The proposed procedure utilizes degradation process history of the item observed during burn-in.
• This paper provides a new ‘general insight/perspective’ on the burn-in models.
• This work opens new possibilities for further developments of similar approaches.

Burn-in is a method of ‘elimination’ of initial failures (infant mortality). In the conventional burn-in procedures, to burn-in an item means to subject it to a fixed time period of simulated use prior to actual operation. Then, the items which failed during burn-in are just scrapped and only those which survived the burn-in procedure are considered to be of satisfactory quality. Thus, when the items are subject to degradation phenomena, those whose degradation levels at the end of burn-in exceed a given failure threshold level are eliminated. In this paper, we consider a new burn-in procedure for items subject to degradation phenomena and belonging to mixed populations composed of a weak and a strong subpopulation. The new procedure is based on the ‘whole history’ of the degradation process of an item periodically observed during the burn-in and utilizes the information contained in the observed degradation process to assess whether the item belongs to the strong or weak subpopulation. The problem of determining the optimal burn-in parameters is considered and the properties of the optimal parameters are derived. A numerical example is also provided to illustrate the theoretical results obtained in this paper.