Multi-Agent Simulation of Manufacturing Networks: Autonomous and Adaptive Decision-Making for Industry 4.0

This work addresses the modeling and simulation of Flexible Manufacturing Systems (FMS) to optimize industrial production in Industry 4.0. The research integrates intelligent agents and reinforcement learning to improve efficiency and flexibility in decision-making in autonomous units. It proposes a multi-criteria framework that balances profit, sustainability, and variability, allowing real-time adaptation. The multi-agent environment facilitates unit communication, redistributing tasks to reduce rejects and increase productivity. The validation was carried out by computational simulations of a small manufacturing network comprising six manufacturing units. Each unit has a dedicated agent to decide the demand acceptance according to its policy, resource, and performance measurements. The simulation results have shown that the decision model can pursue the target and equilibrate the economic, sustainability, and customization criteria. The study highlights the potential of autonomous and connected manufacturing networks to optimize industrial processes, promoting sustainability and competitiveness in Industry 4.0.