Parameter Estimation Using EM Algorithm For Lifetimes From Step-Stress and Constant-Stress Accelerated Life Tests With Interval Monitoring

Stochastic information about the reliability parameters of a test unit can be rapidly obtained via accelerated life tests by running the tests at higher stress levels than normal operating conditions. Using a regression model, the reliability parameter at the normal design stress can be estimated via extrapolation. Recently, the design optimization of accelerated life tests has been investigated by many researchers but the associated inference for the regression parameters has not been. In this article, the Expectation-maximization algorithm is used to determine the maximum likelihood estimates of the regression parameters for time constrained exponential failure data from the step-stress and constant-stress accelerated life tests with interval monitoring. It is demonstrated that the method is feasible as well as easy to implement. Using the principle of missing information, the asymptotic variances and covariances of the maximum likelihood estimates are also calculated. The proposed method is illustrated using a real engineering case study.