Improvement on the Natural Dictionary-ordering Method in Reliability Growth Tests
ZHENG Zhongguo, CAO Dachuang, ZHOU Zhengfa
(Department of Probability and Statistics, Peking University, Beijing, 100871)
Abstract:
In a m-stage reliability growth test, to estimate the lower confidence limit of the reliability of the system in the final stage, randomization is introduced to improve the exsiting natural chain method. Simulation study shows that when the number of stages is one, the randomization method is ideal.
Key words:
reliability; lower confidence limit; sample space ordering; ML ordering method; NATURAL dictionary ordering method