This research creates digital twin models of EV users powered by large language models to intelligently manage charging stations and optimize grid resources.

• Develops an LLM-driven multi-agent framework that accurately simulates user decision-making for charging/discharging EVs
• Implements dynamic incentive mechanisms to influence user behavior and balance grid needs
• Enables more efficient vehicle-to-grid (V2G) integration by predicting and adapting to user preferences
• Demonstrates how AI can enhance energy management through personalized user modeling

This approach matters for engineering because it provides a scalable solution for managing distributed energy resources while respecting individual user preferences and constraints, helping utilities balance grid stability with user satisfaction.

Dynamic Incentive Strategies for Smart EV Charging Stations: An LLM-Driven User Digital Twin Approach