Deep Reinforcement Learning-Based Trading Strategy for Load Aggregators on Price-Responsive Demand

Table 2

Description of the parameters of the DDPG.

Parameters	Meaning	Value
TAU	Smoothing coefficient of target network in actor and critic network	0.001
α	Actor network and critic network learning rate	0.0005
Batch_size	Number drawn from the experience pool per training	64
Capacity	Size of the experience pool	100000
γ	Discount factor	0.99