Model-Free Attitude Control of Spacecraft Based on PID-Guide TD3 Algorithm

Table 2

Hyperparameter settings.

Hyperparameters	Symbol	Value
Random seed	—	2
Max episodes		400
Max steps per episode		200
Sample time		1
Replay buffer size		10^6
Batch size		250
Policy network learning rate		0.0003
Critic network learning rate		0.001
Exploration noise scale		0.1
Delay update		3
Discount factor		0.99
Soft update rate		0.01