The Algorithm Of Expert Information Process And Mechanical Fuzzy Reliability Design

Great progress in mechanical fuzzy reliability theory and its application have been made since the eighties of the twentieth century, but the meanings of some important concepts in fuzzy reliability theory are not accepted by all of people. Although some correlative theory have been proposed, most of these theory are obscure, and the expressions are complex, so it is difficult to put them into practice.Firstly, in this paper some basic math knowledge which are needed for the fuzzy reliability theory are described, including the probability theory, mathematics statistics and the fuzzy mathematics. Secondly, an important model on fuzzy decision in fuzzy reliability theory is discussed, which starts with some supposes and three decision-making models are given with expressions. Based on these models, the solution of two types of problems often found in practice is given. A method based on numerical estimation is used to calculate the reliability because of complexity of the expressions and the difficulty of analytic expression calculation, and a soft-ware named expert information process is developed. Finally, digital simulation of fuzzy reliability is discussed. The Monte Carlo method is utilized to compute the reliability after the fuzzy variable is transformed to the equivalent random variable, but is substituted by the importance sampling method for its low-effectiveness sometimes. The genetic algorithm is introduced to calculate the design point, which is a very important parameter of the importance sampling. The fuzzy reliability theory is put into practice better through these improved methods, and will have a good perspective.