自相关过程的残差控制图  

The Residual Control Chart for Autocorrelated Processes

作　　者：张志雷[1] 

机构地区：[1]河南财经政法大学统计系,河南郑州450002

出　　处：《数理统计与管理》2011年第5期888-895,共8页Journal of Applied Statistics and Management

摘　　要：由于自相关过程违背了过程输出数据独立性的假定,使得传统休哈特图的有效性受到质疑。本文首先讨论控制图设计基本思想,然后分析了对自相关过程监控的残差控制图原理;进而以平均链长和各链点检出概率为准则,系统研究了AR(1)过程残差控制图的检测能力,并与休图进行了比较。最后,通过一个模拟验证了该方法的有效性。It is doubtful of the validity of the Shewhart chart that the observations from autocorrelated processes violate the assumption of independence. In the paper, we first discuss the design idea of control charts and analyse the theory of the residual control chart. Then we make a systematic study on the detection capability of the residual charts for AR（1） process and compare their performance to the performance of the Shewhart control chart by using average run length and the probability of detecting a shift. Finally a simulation is given to indicate the effectiveness of the proposed method.

关 键 词：自相关过程 控制图 平均链长(ARL) 模拟研究 

分 类 号：O213.1[理学—概率论与数理统计]