基于自回归移动平均过程的贝叶斯质量控制方法研究

Bayesian Quality Control for Autoregressive Moving-average Processes

针对自回归移动平均过程中控制变量的观测值并不具有相互独立性,引入贝叶斯分析方法研究过程质量控制问题.通过模型结构的贝叶斯分析,利用残差序列建立了基于自回归移动平均过程的贝叶斯质量控制模型,解决了观察数据相关条件下的过程质量监控问题.仿真分析结果表明:贝叶斯ARMA质量控制方法能够有效地避免了在受控状态下使用常规控制图造成的漏发或虚发报警现象,解决了自回归移动平均过程情况下的质量控制问题.

To explore the quality control under the condition that the sample data in the autoregressive moving-average processes are not independent,time series models were introduced to fit these data.The Bayesian ARMA control charts were constructed with independent residual series data,and used to monitor the quality in autocorrelative processes.The results from simulation show that Bayesian ARMA quality control charts can effectively carry out quality control autoregressive moving-average processes,and avoid alarming incorrectly when the processes are under statistical control and not alarming when processes are out of control.