基于序贯蒙特卡洛法的自相关过程控制

Control of autocorrelation process based on sequential Monte Carlo method

针对参数未知的自相关过程在线质量控制问题,研究了基于序贯蒙特卡洛法(SMC)的过程控制策略。在给出过程状态空间方程模型的基础上,分析了由于参数未知使得运用Kalman滤波求解控制策略时存在的困难;通过设置未知参数的先验分布,运用序贯蒙特卡洛法得到各参数的后验估计,进而获得了使过程损失最小的控制策略。给出了仿真,以分析控制策略的有效性,结果表明所得到的控制策略具有较好的控制效果。

Aiming at the online quality control problem in an autocorrelation process with unknown parameters, the control strategy based on the sequential Monte Carlo （SMC） method is studied. Based on the state space process control model, the difficulty caused by the unknown parameters in solving the control strategy using the Kalman filter is analyzed. The posterior estimations for each parameter are achieved by building the prior distributions for the unknown parameters and using the SMC method, and then the control strategy for minimizing the loss of process is obtained. The simulations are given to analyze the effectiveness of the approach. The results show that the control strategy obtained by the approach has a good control performance.