In statistical process control, some variables do not follow normal distribution. To resolve the problem of non-normal distribution multivariable process control, based on data depth theory, a change-point control chart is proposed. In order to draw such a chart, methodology and control process are presented for the collection of statistical data. To test the performance of the proposed method, samples that follow binary Gamma-distribution are collected. It is done under different conditions with the location parameter shift ranging from 0.2 to 1.0, and change point being 14, 24, and 34, respectively. The simulation results show that, the larger the shift is, the better the performance is. When the shift is less than 0. 7, the larger the change point is, the better the performance is. However, when the shift ranges from 0.1 to 0.4, the marginal effect is reduced.%为了解决多元非正态分布情况下的过程控制问题,提出基于数据深度的变点控制图,并对构建该控制图检验统计量的具体方法及控制流程进行了详细描述.为了检验该控制图的控制效果,采用服从二元伽马分布的样本数据对其进行了验证,并设置位置参数偏移范围为0.2至1.0,变点为14、24、34,几种情况分别检验其控制效果.数据仿真的结果表明:偏移越大,检测效果越好；偏移量小于0.7时,变点越大,检测效率越高；而当变点大于0.7时变点对检测效果的影响不明显.偏移量在0.1至0.4的范围内,变点越大,检测效果越好,但是这种边际效果在减小.
展开▼
