A statistical method of accelerated life testing based on fuzzy theory

Failure data in accelerated life test are imprecise due to kind of subjectivity. This makes the assessment of testing lack of accuracy. A statistical model of accelerated life testing based on fuzzy theory is proposed in this paper. The model aims to describe the subjectivity of failure data and give a fuzzy interval assessment of lifetime and reliability comparing to the traditional point estimation. Firstly, triangular membership function is chosen to describe Type II censored data and their a-level cut sets are settled under acceptable level. Then, the statistical modelling for ALT data is established based on fuzzy theory by integrating maximum likelihood estimation to determine the membership functions of parameters. And using Particle Swarm Optimization algorithm calculates their fuzzy evaluation values. Finally, the proposed method is verified by the simulation study.