Product quality, machine maintenance, and equipment upkeep are three closely interconnected factors that collectively influence the performance of a smart manufacturing system. While significant research has been conducted on production control, quality control, and maintenance scheduling as separate areas, there remains a gap in studies that integrate these aspects into a unified control framework.

This project addresses this gap by considering a mobile, multi-skilled robot-operated Flexible Manufacturing System (FMS). A comprehensive model is developed that integrates robot operations, individual workstation processes, and product quality management. A Multi-Agent Reinforcement Learning (MARL) approach is employed to design a control strategy that simultaneously manages robot assignments, preventive maintenance scheduling, and tool changes.

Furthermore, an extension to this framework is proposed using a Heterogeneous Graph Neural Network (HGNN) as an information aggregation module, enabling more coordinated and efficient decision-making within the system.