运行和方差控制图的改进

Improvement on Run Sum Control Charts

文章根据双边的运行和方差控制图,设计了加入负得分的单边运行和方差控制图,运用马尔科夫链方法研究它的平均链长表现。将新控制图与目前普遍使用的检测过程方差的累积和（CUSUM）控制图平均链长表现相比较,发现其表现不具有明显优越性。然后寻求改进方法,加入初值响应,数值结果证明了其检测效率明显地提高。

According to the two-sided run sum control charts, this paper designs one-sided run sum control charts with negative scores, and uses Markov chain method to study the average run length（ARL）. Comparing the new control charts with the widely used testing processes variability-CUSUM control charts ARL, the paper finds that its performance has no obvious superiority.And then the paper seeks an improved method by adding in initial response. The numerical simulation results prove that the test efficiency has improved significantly.