Abstract: To choose the multiple dynamic quality control charts with more sensitive control performance,a performance analysis model for multiple dynamic control charts was established and the corresponding selection strategy was proposed.In Matlab environment,Markov chain method and Monte Carlo simulation method were used to solve the model,and the computational efficiency and precision of two methods were compared and analyzed when the process was in control and out of control.To solve the problem of too many conversions of optimal selection strategy,Variable Sampling Inlerval(VSI) and Variable Sampling Size and Interval(VSSI) multiple quality control charts should be given priority in the practical selection.If the size of off-target   was small,VSI Multi-variate Cumulative Sum Control Chart(MCUSUM) control chart was optimal in practice;if 0.9＜δ＜1.9,VSSI Hotelling T2 control chart should be selected;if δ＞2,any kind of multivariate adaptive control chart would be suitable.

Key words: multiple dynamic control chart, average time to signal, average number of samples, Markov chain method, Monte Carlo simulation

摘要： 为选择控制性能更加灵敏的多元动态质量控制图,建立常见的多元动态控制图性能对比分析模型并提出相应的选择策略。在MATLAB环境下求解模型时,以Markov链法与蒙特卡洛仿真方法为求解工具,对比分析两种方法在过程受控及失控时的计算效率及精度;为解决最优选择策略转换次数过多的问题,在实际应用中应优先选择变抽样区间策略和变抽样容量与区间策略多元质量控制图;在均值偏移量δ较小的情况下选择变抽样区间多元累积和控制图,09<δ<19时选择变抽样策略霍特林T2控制图,δ>2时任选一种多元动态质量控制图。

关键词: 多元动态控制图, 平均报警时间, 平均样本观测数, Markov链法, 蒙特卡洛仿真