Approximated optimal designs for a simple step-stress model with type-II censoring, and weibull distribution

This paper deals with theories for approximated optimal design for simple step-stress accelerated life testing (ALT) with type-II censoring, and Weibull distribution. Statistically approximated optimal ALT plans are developed to minimize the asymptotic variance of the maximum likelihood estimators (MLE) of the p-th percentile of lifetime at design stress. For the complex of calculating the expectation of order statistics, the Fisher information matrix for type-II censored data is approximated by that for type-I censored data. The approximated optimal plan doesn't depend on the values of accelerating model parameters. Simulation results also show that the optimal stress levels are highest possible stress and lowest possible stress.