系统可靠性评定的熵（法）近似限

APPROXIMATE LIMITS OF THE ENTROPY (METHOD) FOR RELIABILITY ASSESSMENT OF A SYSTEM

将近代Ｓｈａｎｎｏｎ信息论的信息概念广义化，利用信息量的“可加性”特性和离散变量熵的一般表达式，并与可靠性工程理论相结合，推导了由成败型或指数型单元所组成的串、并联系统可靠性评定用的熵（法）近似下限计算公式。应用本文公式所得出的评定结果总是介于经典和贝叶斯（法）近似下限之间，既不偏于保守，又不过分冒进。

Information concepts of modern Shannon′s information theory are extended in this paper. By the use of “addible” characteristics of information and general expression of the entropy for discrete variables in association with the reliability engineering theory, the calculation formulae of the approximate lower limits of the entropy (method) for reliability assessment of a series or parallel system consisting of success failure model or exponential model components are derived. The assessment results obtained by use of the formulae in this paper are always within the approximate lower limits of classical and Bayesian (methods). They are neither conservative nor radical, but very satisfactory.