Abstract: Deep reinforcement learning (DRL) methods have been applied to power system problems in active distribution networks, including the inverter-based volt/var control (VVC). However, existing ...