ABSTRACT

In this paper, a new distribution-free double exponentially weighted moving average (DEWMA) control chart based on the Wilcoxon rank-sum (WRS) test without any distributional assumption of the underlying quality process is proposed. The proposed chart is the extension of the exponentially weighted moving average (EWMA) WRS control chart where the EWMA WRS statistic is applied twice. The ability of the EWMA and DEWMA WRS charts to detect small to large process mean shifts is further improved by combining each of them with the Shewhart control chart based on the WRS statistic. The in-control and out-of-control performances of the proposed charts are thoroughly investigated using extensive simulations. It is found that the proposed charts perform better than their competitors under both normal and non-normal distributions. A numerical illustration using real-life dataset is given in order to demonstrate the design and implementation of the new charts.

摘要

本文提出了一种基于Wilcoxon秩和（WRS）检验的新的无分布双指数加权移动平均值（DEWMA）控制图，该图不对基础质量过程进行任何分布假设。拟议的图表是对指数加权移动平均（EWMA）WRS控制图的扩展，其中两次应用了EWMA WRS统计信息。EWMA和DEWMA WRS图表检测小到大过程均值漂移的能力通过将它们与基于WRS统计信息的Shewhart控制图结合起来而得到了进一步提高。建议的图表的控制中和失控性能已通过广泛的模拟进行了全面研究。发现在正态分布和非正态分布下，拟议图表的性能均优于竞争对手。