考虑调整随机误差的设置调整问题研究  

Research on Setup Adjustment Problem Considering Random Adjustment Error

作　　者：刘利平[1] 马义中[1] 汪建均[1] 周晓剑[1] 

机构地区：[1]南京理工大学经济管理学院,江苏南京210094

出　　处：《数理统计与管理》2011年第5期896-903,共8页Journal of Applied Statistics and Management

基　　金：国家自然科学基金(70672088)

摘　　要：针对存在设置偏差的小批量离散制造过程,研究了每次调整成本固定且调整存在随机误差情形下的质量控制问题。在建立过程状态空间方程模型的基础上,通过贝叶斯方法估计过程的未知参数,利用动态规划的方法得到了边界形式的过程最优调整策略。通过算例验证了所提出调整策略的有效性,并利用仿真对本文提出的调整策略与其他调整策略进行了比较分析,结果表明,本文提出的方法能够更好地减少过程总体质量损失。Aiming at the short-run discrete manufacturing process with unknown initial offset, this paper studies the quality control problem of this process under the circumstances of adjustment with random error and fixed cost. First, based on the state-space process-control model, the unknown parameter of process is estimated by Bayesian method, and then the deadband form optimal process adjustment solution is presented by using a dynamic programming. Second, the effectiveness of the solution is verified by an example, and it is also compared with another adjustment solution by simulations, which shows the result that the adjustment solution presented by this paper is more effective than the other one to reduce the total quality loss of the process.

关 键 词：统计过程控制 边界调整 贝叶斯方法 状态空间模型 

分 类 号：O212.8[理学—概率论与数理统计] O212[理学—数学]