Kumaraswamy分布下负载共享并联系统的参数估计  

Estimation of Parameters for Load-Sharing Parallel Systems under Kumaraswamy Distribution

作　　者：李丹青 熊文洁 张正成 LI Dang-qing;XIONG Wen-jie;ZHANG Zheng-cheng(School of Mathematics and Physics,Lanzhou Jiaotong University,Lanzhou 730070,China;School of Mathematical Statistics,Hainan Normal University,Haikou 571158,China)

机构地区：[1]兰州交通大学数理学院,甘肃兰州730070 [2]海南师范大学数学与统计学院,海南海口571158

出　　处：《安徽师范大学学报（自然科学版）》2020年第4期307-314,共8页Journal of Anhui Normal University(Natural Science)

基　　金：海南省自然科学基金项目(SQ2019MSXM0967)。

摘　　要：在并联连接的多元件系统中,随着元件逐个失效,应用于系统的总负载在剩余幸存的元件之间重新分配,这通常称为负载共享。本文讨论了当元件的基础分布为Kumaraswamy分布时负载共享并联系统的参数估计。首先,我们列出了参数估计的具体方法并通过合适的迭代技术来获取参数估计的闭合值。其次获得了所提出的最大似然估计(MLE)的渐近分布,并对简化后的模型模拟出其参数的MLE和渐近最佳无偏估计(ABUE),再对两者进行渐近比较。最后,使用Bootstrap技术构造了参数的置信区间。In a multi-component system connected in parallel,as the components fail one by one,the total load applied to the system is redistributed among the remaining surviving components,which is commonly referred to as load sharing.This paper studies the parameter estimation of load sharing parallel system when the basic distribution of components is Kumaraswamy distribution.First,we list the specific methods of parameter estimation,and obtain the closed value of the parameter estimates through appropriate iterative techniques.Secondly,we obtain the asymptotic distribution of the proposed maximum likelihood estimation(MLE),and the simplified model is simulated with MLE and asymptotically best unbiased estimation(ABUE),then compare the asymptotics of MLE and ABUE.Finally,the confidence interval of the parameters is constructed using Bootstrap technique.

关 键 词：负载共享 Kumaraswamy分布 高斯-赛德尔迭代法 Bootstrap技术 参数估计 

分 类 号：O212.1[理学—概率论与数理统计]