Abstract

In this paper, we propose a multi-objective optimal design method based on non-parametric, adaptive and dynamic EWMA control charts for the problem of a small number of samples and uncertain distribution faced by designing quality control charts for multi-variety and small batch manufacturing processes. Based on nonparametric statistical theory and adaptive control, a control chart statistic independent of sample data distribution is constructed, and a dynamic sampling method based on clustering distance is designed to realize sample sampling. On this basis, a control chart multi-objective optimal design model is established considering statistical and economic aspects, and the model is solved based on vector autoregressive algorithm, and then a non-parametric adaptive dynamic EWMA control chart for the multi-variety small batch system is constructed. The experimental results show that the root mean square error of processing quality prediction of the proposed vector autoregressive model is reduced by 4.67% and 4.14%, respectively, compared with the effects of SVM and PSO-SVM quality prediction models. The control chart based on the vector autoregressive model proposed in this paper can quickly monitor the quality abnormalities with high monitoring performance, which provides an effective way to monitor the quality of the multi-species and small-lot manufacturing process.