上海师范大学毕业论文EDA样式.doc

EDA技术及应用 题目（中文）指数分布冷贮备单元的串联系统产品的统计分析 （英文）Statistical Analysis of Cold Reserve Series System for Exponential Distribution 学 院 信息与机电工程学院 年级专业 2008级计算机科学与技术 学生姓名 XXXXXX 学 号 080143504 指导教师 XXXXXX 完 成 日 期 2012 年 4 月 摘要 考虑寿命服从指数分布的冷贮备单元串联而成的系统，先介绍了其工作原理并做出一些基本假设。利用概率论中指数分布无记忆性，推出整个系统寿命的分布函数。给出了参数的矩估计、极大似然估计。给出了适用于不同样本容量的含参数的逆矩估计和修正逆矩估计的方程，并证明了方程的解的唯一性。通过大量Monte-Carlo模拟考察估计的精度，得到矩估计和极大似然估计优于逆矩估计。同时，利用枢轴量法和中心极限定理，还给出了参数的精确区间估计和近似区间估计，通过大量Monte-Carlo模拟考察了区间估计的精度，得到参数的精确区间估计优于近似区间估计。最后，随机产生一组数据应用以上方法举例说明。 关键词：冷贮备串联系统；矩估计；极大似然估计；逆矩估计；区间估计 Abstract Basing on the series system in which the life of each cold reserve unit conforms to the exponential distribution, the paper firstly introduces the working principle of this system and proposes some basic assumptions. Using the memoryless property of exponential distribution functions, the paper gives the whole system life’s distribution function. the moment estimate, maximum likelihood estimate are given. the Equations including the inverse moment estimate and revised inverse moment estimate of the parameter which fit different sample size are given. Then the uniqueness of solution is also proved. In order to compare the precision of the estimate, a large amount of Monte-Carlo simulations are performed and it can be concluded that the moment estimate and maximum likelihood estimate are better than the inverse moment estimate. In addition, using the pivot method and central limit theorems，the paper gives the precise interval estimate and approximate interval estimate of the parameter. The precisions of both interval estimates are compared by Monte-Carlo simulations. It can be concluded that the precise interval estimate is better than the approximate interval estimate. At last, a group of data is produced as an example to apply this method. Keywords: cold reserve series system；moment estimate；maximum