Fortifying Multi-Agent Architectures in Smart Manufacturing: Leveraging Federated Learning and Blockchain for Security and Resilience
Petro Mushidi Tshakwanda, Henok Berhanu Tsegaye, Raddad Almaayn, Michael Devetsikiotis · 2025
This work presents a blockchain-based communication architecture for multi-agent smart manufacturing systems, enhanced with federated learning to improve security, data privacy, and scalability. Comparative evaluation with traditional centralized systems across various configurations reveals that our approach achieves up to 12 % higher model accuracy and 15 % greater security, particularly in large-scale deployments, with scalability scores surpassing those of centralized systems by 18 % as agent numbers increase. Although initial latency and energy consumption are higher, these factors improve significantly as the system scales, indicating suitability for complex manufacturing networks. This research underscores the potential of combining federated learning and blockchain to enhance resilience, security, and efficiency in Industry 4.0 smart manufacturing systems.