Burn-in is a widely used engineering method of elimination of defective items before they are shipped to customers or put into field operation. Under the assumption that a population is described by the decreasing or bathtub-shaped failure rate functions, various optimal burn-in problems have been intensively studied in the literature. In this paper, we consider a new model and assume that a population is composed of stochastically ordered subpopulations described by their own performance quality measures. It turns out that this setting can justify burn-in even in situations when it is not justified in the framework of conventional approaches. For instance, it is shown that it can be reasonable to perform burn-in even when the failure rate function that describes the heterogeneous population of items increases and this is one of the main and important findings of our study.
- Heterogeneous population
- Performance quality measures
- Stochastically ordered subpopulations