Statistic Analysis For The Scaled Half Logistic Based On Censored Samples

In industrial experiments, products are often extremely reliable, with large mean time to failure under normal operating condition. Consequently, with conventional life testing experiment, it is almost impossobile to obtain adequate information about the failure time. To overcome these problems, the experimenter may retort to accelerated life testing. Accelerated life testing is the one of life test method to reduce test period by improving test stress. It was advanced for quickly evaluating the reliability of products and finding out the failure cause of them. From 1970s on,it has been extensively applied and researched in our country,and some new theory and method were advanced. Many scholars have studied scaled half logistic distribution based on the life test, in this paperm, I will do more in-depth study and discussion of it in the accelerated life test under the condition of the parameter estimation methods. In order to promote accelerated life testing in researching and application.This paper is divided into four chapters. Lifetime distribution, relliabitily characteristics, Some elementary theory and statistical analysis of accelerated life testing are the main constent in chapter one, such as the significance and classification of life test , some numeral characteristics which are distribution function, probability density, mean lifetime and so on, and the concept of failure rate and reliability. This chapter includes aslo some interrelated theory with scaled half logistic distribution.In chapte two, we introduce the estimation for parameters of life testing on scaled half logistic distribution. The discussion about the distribution is focus on the point estimation of parameter. In view of the current status of research and theoretical significance, This article first study the interval estimates and assumptions for further study, in the last, the example illustrate the methods and conclusions.In chapte three and four, we introduce the estimation for parameters of accelerated life testing on scaled half logistic distribution. First,relative concepts of accelerated life testing are briefly introduced. Its kinds and characterstics are given. The objectives, principles, ,accelerated methods and statistical analysis are discussed. Then this article illustrate the statistical analysis of scaled half logistic distribution both based on constant sress accelerated life testing with type-Ⅰcensering and type-Ⅱcensering and based on simple step stress accelerated life testing with and type-Ⅱcensering. We derive the maximum likelihood estimation of the parameters of a exposure model with scaled half logistic distribution. At last a comparison of deviation and the mean square error of the estimate was done and the approximate confidence lower bound, confidence interval inspected the true value of the coverage ratio by stochastic simulation, and have achieved relatively satisfied with the results.