LLM-Enhanced MARL for Unified EV Charging and Grid Optimization

This framework integrates LLMs into Multi-Agent Reinforcement Learning to optimize electric vehicle charging, station profitability, and grid stability simultaneously. The LLM performs interpretable feature selection from high-dimensional IoT data and dynamically adjusts weights for conflicting multi-objective functions.