Hybrid Multi-Agent Framework for Steel Production Planning

This document is a technical report that presents HyMAS, a Hybrid Multi-Agent System designed for autonomous steel production planning. The report outlines the complexities of steel manufacturing, including high energy consumption and strict quality demands, which traditional planning methods struggle to address. HyMAS aims to enhance decision-making in steel production by employing a hierarchical network of autonomous agents that manage different aspects of the production process. The architecture includes resource agents, process agents, and process chain agents, allowing for localized decision-making and rapid adjustments to production plans. The integration of Industry 4.0 principles, such as digital twins and machine learning, supports real-time data-driven scheduling. A case study demonstrates HyMAS's effectiveness, achieving an 18% efficiency gain and an 8% improvement in on-time delivery. The report concludes by discussing the contributions of HyMAS to resilient autonomous scheduling and future research directions.