Data analysis of reliability growth test for electronic equipments based on Fiducial inference

Reliability growth is an important part of reliability engineering, but for the problem that small samples are often occurred during reliability growth test, there is almost no effective method so far. To solve this problem, fiducial inference with restricted parameter space was studied, for non-location-scale families, it was generalized by standardized likelihood, and it was pointed out that evaluation results obtained by fiducial distribution are equal to the corresponding results obtained by frequency statistics. Under the reliability growth model with monotone constraints, reliability evaluation of electronic equipments for reliability growth test was analyzed by fiducial inference with restricted parameter space, and for cases with complete failure time data or interval-censored failure time data during each stage of the reliability growth test, the upper confidence limit of failure rate for electronic equipments was given in detail.