Research
152d agoAutoB2G: LLM‑Driven Building‑Grid Co‑Simulation
AutoB2G harnesses large language models to automate building‑grid co‑simulation, turning natural‑language task descriptions into end‑to‑end workflows. By integrating reinforcement learning with grid‑level impact assessment, the framework eliminates manual configuration and expert coding, accelerating research across global smart‑city initiatives.
The Signal
The system builds on CityLearn V2 and demonstrates scalable, data‑driven control for power networks worldwide.