基于马尔可夫链蒙特卡罗方法的统计过程调整技术研究

• 论文 •

基于马尔可夫链蒙特卡罗方法的统计过程调整技术研究

褚崴，于晓义，孙树栋

西北工业大学机电学院工业工程系，陕西西安710072

出版日期:2007-06-15 发布日期:2007-06-25

Statistical process adjustment technique based on Markov chain Monte Carlo approach

CHU Wei,YU Xiaoyi,SUN Shudong

Department of Industrial Eng., School of Mechatronics, Northwestern Polytechnical University ,Xi’an710072, China

Online:2007-06-15 Published:2007-06-25

摘要/Abstract

摘要： 建立了统计过程调整问题模型，研究了基于马尔可夫链蒙特卡罗方法的统计过程调整技术，并通过吉布斯抽样实现了偏移量的参数估计。在参数未知及参数已知的条件下，通过与其他方法的实例对比研究，验证了该方法的可行性及性能优势，并对参数未知条件下基于该方法的统计过程调整技术进行了改进。

关键词: 马尔可夫链蒙特卡罗方法, 统计过程调整, 单向分类随机效应模型, 吉布斯抽样

Abstract: Based on study of the Statistical Process Adjustment (SPA) problem in quality control area, the SPA problem model was set up. Then SPA techniques based on Markov Chain Monte Carlo (MCMC) approach were studied, and Gibbs sampling was used to estimate setup errors. Comparing to other methods which have been introduced to solve the problem on the premises of known parameter and unknown parameter, the feasibility and performance advantages of MCMC approach were proved. Furthermore, SPA techniques based on this approach with unknown parameters were also improved.

Key words: Markov chain Monte Carlo approach, statistical process adjustment, one-way random effects model, Gibbs sampling

中图分类号:

TP272

褚崴，于晓义，孙树栋. 基于马尔可夫链蒙特卡罗方法的统计过程调整技术研究[J]. .

CHU Wei,YU Xiaoyi,SUN Shudong. Statistical process adjustment technique based on Markov chain Monte Carlo approach[J]. .