Evolving LLM-Derived Control Policies for Residential EV Charging and Vehicle-to-Grid Energy Optimization

Researchers have developed a new method to create rules for home electric vehicle charging. Instead of using opaque 'black box' AI, they use large language models to generate and refine simple, readable Python programs. These programs manage when to charge a car or sell energy back to the grid. The best-performing rules increased profits by 18% over a baseline while ensuring user comfort and safety.