串联系统的可靠性评估方法

A Method of Evaluating the Reliability of Series Systems

若要对某个对象的可靠性进行评估,首先必须获得大量与对象的可靠性有关的初始信息。这些信息可能同时蕴含多种不确定性,包括随机性、模糊性、不完全可信性、不确知性等。初始信息中的不确定性必然导致评估结果也蕴含不确定性。但目前常用的评估方法既不能同时处理、也不能在评估结果中综合体现这些本质各异的不确定性。为此,提出了一种基于D-S证据理论的评估方法,以解决上述问题。该方法适用于任一具有独立功能的对象,其主要特点为:针对对象的每个评估指标建立一项证据,并通过"证据的组合"来得到评估结果。本文围绕"证据组合"做了充分的预处理和后处理工作,从而避免了信息丢失,保障了评估结果的合理性。在此基础上,提出了一种适用于串联系统的可靠性评估方法。

In order to evaluate the reliability of some object,firstly,a great deal of initial information relating with the reliability of the object must be obtained.In the information,perhaps there are many kinds of uncertainty,including randomness,fuzziness,imperfect credibility,ignorance and so on.However,the evaluation methods used at present not only can not deal with all kinds of uncertainties synchronously but also can not reflect all kinds of uncertainties synthetically in the evaluation result.Therefore,an evaluation method based on dempster-shafer evidential theory（called D-S evidential theory） is proposed in this paper to solve the above problems.This method is suitable for any object with independent function,and its main feature is that an evidence is established according to each evaluation index,and the evaluation result is obtained by combining the evidences.Adequate pre-processing and post-processing are done for the combination of evidences so as to avoid loss of information,and ensure the reasonableness of the evaluation result.Furthermore,a method used to evaluate the reliability of a series system is proposed on the base of the above work.