摘要
复杂系统由若干个子系统构成,在复杂系统的全寿命过程中,从每个子系统到整个复杂系统,都与系统的可靠性密切相关.可靠性分析就显得极为重要,在复杂系统可靠性分析中,不可避免地遇到两类可靠性分析数据:失效或部分失效数据与很少或无失效数据,可概括为具有一定样本信息和少量或无样本信息两类数据.就两类可靠性分析数据信息,提出处理相应数据的可靠性统计分析方法:Bayes方法.Bayes方法是在小样本或无样本信息的前提下能较好得到参数估计的一种方法,它的基本出发点是:有效利用假定可得的待估参数的历史资料或先验知识,而不是基于样本数据的假定,这切合了复杂系统全过程中无失效数据的可靠性分析.
Complex system is made up of sub-systems. In the lifetime processes of the complex system, all systems including sub-systems and the complex system have close relationship with the system reliability. So the reliability analysis is of great importance. Two sorts of data are unavoidable in the reliability analysis of complex system： failure or partial failure data and seldom or un-failure data. E-qually they are said sample information and none sample information. This paper, based on these two kinds of reliability analysis information, develops one reliability statistic analysis method to process the corresponding data： Bayes statistic analysis. In the condition of small or none sample, Bayes method can achieve good parameter estimation. It starts from the point that we should effectively use obtainable historiced materials of uncertain parameter, or use prior knowl- edge instead of depending on the assumption of sample data.
出处
《贵州师范大学学报（自然科学版）》
CAS
2008年第4期61-63,68,共4页
Journal of Guizhou Normal University：Natural Sciences
基金
贵州省自然科学基金(黔科合J字[2008]2048号)
国家自然科学基金(10671045)
关键词
复杂系统
可靠性
BAYES推断
先验分布
后验密度
complex system
reliability
Bayes inference
prior distribution
posterior density
作者简介
胡尧（1971-），男，副教授，研究方向：应用统计学、可靠性工程、城市道路交通问题。
