Optimal burn-in for n-subpopulations with stochastic degradation

Often within any population of components there are strong and weak subpopulations, most notably in advanced electronic products. Burn-in test is often used to eliminate the early failures and reduce the warranty cost for those devices. This paper considers a burn-in process for a degradation-based heterogeneous population. An optimization model is formulated to achieve the optimal burn-in time. Numerical examples are provided to demonstrate the effectiveness of the burn-in process.