Integrating production, maintenance, and quality concepts has yielded positive results for imperfect processes that degrade over time due to various specific causes. In practice, identified correlations within monitored data over time challenges the traditional assumption of independence. To address this, the Autoregressive Moving Average (ARMA) control chart has been utilized in integrated models. This study proposes a novel control chart based on the support vector regression (SVR) method to apply within integrated models. In an integrated model, the performance of both monitoring techniques are assessed. A solution procedure based on particle swarm optimization (PSO) algorithm is employed. An industrial case study and comparative analyses are presented for further examination. Under smaller shifts, SVR performs better, whereas monitoring with the ARMA chart yields infeasible solutions in six scenarios. However, due to the time-consuming procedure of the SVR, the ARMA chart becomes a more desirable option for larger shifts.

Keywords