时变参数自适应累积和图的经济统计优化设计  

Economic-statistical optimization design of time-varying parameters adaptive cumulative sum chart

作　　者：王海宇 Wang Haiyu(Business School,Zhengzhou University,Zhengzhou 450001,China)

机构地区：[1]郑州大学商学院,河南郑州450001

出　　处：《系统工程学报》2024年第2期233-243,共11页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(71672209,U1904211);国家社会科学基金资助项目(20BTJ059).

摘　　要：为了降低过程质量控制的成本,同时提高控制图对不同程度的异常偏移的监控效率,研究时变参数自适应累积和TVPACUSUM图的优化设计问题.首先提出了一种能够根据过程偏移的指数加权移动平均(EWMA)估计进行控制图参数调整的ACUSUM图方法;其次以平均产品长度和单位产品平均质量成本最小为目标,构建了该控制图的经济统计多目标优化设计模型;最后分别在可能出现的异常偏移程度固定或是在某一区间范围两种情形下,与已有文献中的控制图设计方法进行比较分析.结果表明时变参数设计能够有效提高控制图对过程状态变化的灵敏性,提出的优化设计方法能够在保持较低的整体质量成本的同时有效提升ACUSUM图的整体监控效率.In order to reduce the cost of process quality control and improve the monitoring efficiency of control chart for different degrees of shifts,the optimization design of time-varying parameter adaptive cumu-lative sum(TVPACUSUM)chart is discussed in this paper.Firstly,an ACUSUM chart which can adjust the control chart parameters according to exponential weighted moving average(EWMA)shift estimators is pro-posed.Secondly,aiming at minimizing the average product length and average quality cost per unit product,a economic-statistical multi-objective optimization design model of the TVPACUSUM chart is constructed.Finally,in the case that the possible process shift isfixed or in a certain range,it is compared with control chart design methods in the existing literature.The results show that the time-varying parameter design can effectively improve the sensitivity of the control chart to the change of process state,and the optimization de-sign proposed in this paper can effectively improve the overall monitoring efficiency of ACUSUM chart while maintaining a low overall quality cost.

关 键 词：时变参数自适应累积和控制图 多目标优化 平均产品长度 EWMA偏移估计 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]