基于Bayes Monte Carlo方法的小样本模糊可靠性仿真研究

Research on Small-sample Fuzzy Reliability Simulation Based on Bayes Monte Carlo Method

针对基于Bayes解析方法的小样本模糊可靠性评估过程中存在的先验主观性和后验复杂性缺陷,通过引入贝叶斯蒙特卡罗(Bayes Monte Carlo,BMC)方法、融合信息熵原理与加权思想,建立了小样本模糊可靠性的BMC方法基本框架,给出了基于BMC方法的小样本模糊可靠性仿真流程,设计了相应算法。采用Matlab工具实现了该算法,并对指数寿命分布的小样本模糊可靠性进行了仿真实验。实验结果表明,运用BMC方法对小样本模糊可靠性进行评估能够有效地降低Bayes方法的先验主观性和后验复杂性。

There are two weaknesses—subjectivity and complexity in the process small-sample fuzzy reliability evaluation based on Bayes method.By introducing information entropy principle,weighted idea and Bayes Monte Carlo（BMC for short） method,the whole framework of BMC method was established for small-sample fuzzy reliability evaluation.The small-sample fuzzy reliability simulation flow was given,and the corresponded algorithm was proposed based on BMC method.Finally,the simulation algorithm was realized using the Matlab tool,and the experiment of the small-sample fuzzy reliability evaluation for exponent distributing was designed.The result shows prior subjectivity and posterior complexity of Bayes method are reduced effectively using BMC method for evaluating small-sample fuzzy reliability.