Expert-guided deep reinforcement learning for HVAC system optimization control

doi:10.1016/j.applthermaleng.2026.130640

Expert-guided deep reinforcement learning for HVAC system optimization control

2026-03-19

0

PRE

AI

DOI：10.1016/j.applthermaleng.2026.130640

Original

Original request for help

Save

Abstract

En 中文

• Two-stage offline-to-online expert-guided RL for stable and efficient HVAC control. • Dual signals for RL-based HVAC control with both imitation and environment guidance. • Superior performance in energy, indoor temperatures and cost over advanced baselines. • Strong robustness across weather datasets and additional thermal test environments.

Expert-guided deep reinforcement learning for HVAC system optimization control

Abstract

Journal

Organization