Integrated production, quality and maintenance of multivariate processes with multiple assignable causes and timely restocking

Many integration studies of production, quality and maintenance suffer from these limitations: (1) Concentrating on monitoring a single quality characteristic, despite the need for simultaneous multivariate monitoring in many cases. (2) Assuming equipment degradation stems from a single assignable cause, whereas real-world systems often experience multiple assignable causes. (3) Relying on static parameter control charts, which are insensitive to small and moderate process shifts. (4) Establishing the buffer stock at the initiation of a production cycle may lead to excessive inventory costs. To overcome these limitations, an integrated model addressing production, quality control (based on an adaptive T2 control chart), and maintenance for imperfect systems is proposed. The model accounts for multiple assignable causes and allows timely buffer replenishment. A genetic algorithm is utilized to minimize the expected total cost per cycle under statistical quality constraints. Comparative studies demonstrate the benefits of adopting adaptive quality control and considering multiple assignable causes in terms of cost efficiency. Finally, a sensitivity analysis using Taguchi's Design of Experiments method helps provide deeper insight into key parameters affecting the total cost.