Statistical Analysis And Optimal Design Of Progressive Censored Life Test

The difference between life test researched by this paper and general life test is some unfailed products be moved out during the test, we name it progressive censoring scheme which is more and more popular in recent years. First we introduce the basic process of Type-I, Type-II and Hybrid progressive censoring scheme under exponential distribution, obtain the maximum likelihood estimator of the unknown parameter in an exact form. Confidence interval based on likelihood ratio test is also proposed, bayes estimate and credible interval of the unknown parameter are obtained under the assumption of gamma prior of the unknown parameter. Then we research the optimal designs of progressive censoring accelerated life test under the assumption that there are k stress and k unknown parameters in the accelerated equation. There are two optimal rules which minimize the asymptotic variance of the eatimated mean or maximize the det of Fisher matrix.