نمذجة وتقدير دالة معولية النظام المتسلسل في حالة عدم تماثل مركباته

Abstract

    This research includes a function modeling the reliability of each compound from the series system compounds assuming anon symmetry of these compounds both by the nature of operational data have it as a mathematical model represented. There are some operational data which are binary value represents the success and the failure of the compound and is modeled by the Logistic Regression Model to represent Reliability, there some compounds that operate between each consecutive failures and modeled by exponential failure After modeling process is completed, reliability series system function is estimated in ways that classical appreciation and Bayes Method using the following methods:

1. 
   The method of maximum likelihood (Maximum Likelihood Method)
   2. Uniformly Minimum Variance Unbiased Estimator Method (UMVUE Method)
   3. Standard Bayesian Method

- Standard Bayesian Estimator By Using Non-Informative Prior p.d.f .
- Standard Bayesian Est. by Using Natural Conjugate Prior p.d.f .

    For the purpose of comparison between the preference of these estimators were employed style simulation way (Monte-Carlo) using a statistical measure (MSE),we reach an estimated preference is UMVU Estimator Method than others estimators .