A single X chart outperforming the joint (X)over-bar &R and (X)over-bar &S charts for monitoring mean and variance

被引:8
|
作者
Haridy, Salah [1 ]
Ou, Yanjing [2 ]
Wu, Zhang [3 ]
Khoo, Michael B. C. [4 ]
机构
[1] Benha Univ, Fac Engn, Banha, Egypt
[2] Singapore Inst Mfg Technol SIMTech, Singapore, Singapore
[3] Nanyang Technol Univ, Sch Mech & Aerosp Engn, Singapore, Singapore
[4] Univ Sains Malaysia, Sch Math Sci, George Town, Malaysia
来源
关键词
Quality control; statistical process control; control chart; Shewhart chart; Average extra quadratic loss; STATISTICAL PROCESS-CONTROL; VARIABLE SAMPLE-SIZES; SHORT PRODUCTION-RUNS; CUSUM CHART; (X)OVER-BAR; INTERVALS; LOCATION; SHIFTS; OPTIMIZATION; DESIGN;
D O I
10.1080/16843703.2016.1189181
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
The Shewhart (X) over bar & R and (X) over bar & S control charts have traditionally been used for detecting mean shift and standard deviation shift (sigma). This article studies and compares the overall performance of the (X) over bar chart with that of the (X) over bar & R and (X) over bar & S charts, as well as the X&MR chart. The comparative study led to surprising results that contradict the conventional wisdom in Statistical Process Control (SPC) niche. It is found that the simplest single X chart (i.e., the (X) over bar chart with a sample size n=1) is always the optimal version of the (X) over bar chart for detecting and (sigma). Moreover, the single X chart even outperforms the joint (X) over bar & R and (X) over bar & S charts in overall detection effectiveness. On average, the X chart is more effective than the (X) over bar & R and (X) over bar & S charts by around 5% under different circumstances. Most importantly, the X chart is very simple to understand, implement and design. As a result, it may be highly preferred for many SPC applications, in which both the mean and variance of a variable need to be monitored.
引用
收藏
页码:289 / 308
页数:20
相关论文
共 50 条