Adaptive Congestion Management in Power Grids Using Grey Wolf Optimization and Multi-Agent Communication

Congestion management is a crucial aspect of maintaining stability and efficiency in modern power grids, particularly as networks expand and dynamic conditions evolve. This paper presents an adaptive congestion management approach using Grey Wolf Optimization (GWO) and Multi-Agent System (MAS) based communication for optimal power dispatch. The methodology is validated on the IEEE 57-bus system, where a manager agent coordinates with generator agents to iteratively adjust power outputs while ensuring economic load dispatch (ELD) and minimizing congestion. The system employs Smart Python Agent Development Environment (SPADE) based decentralized communication over an XMPP Openfire server, allowing seamless coordination among agents. A comparative analysis with Particle Swarm Optimization (PSO) highlights the limitations of PSO in resolving congestion in large-scale networks, demonstrating the superior effectiveness of GWO in handling transmission constraints and improving Available Transfer Capability (ATC). The results show that the proposed GWO-MAS approach achieves lower generation costs, better congestion mitigation, and enhanced grid flexibility. This research reveals the potential of bio-inspired optimization for adaptive and decentralized congestion management in smart grids and future power networks.