Suppressing Uncommanded Roll-Yaw Motion by Jet Flow Control Based on Reinforcement Learning

Table 1

TD3 hyperparameters.

Parameter	Value
Optimizer	Adam [19]
Number of hidden layers (all networks)	2
Number of hidden units per layer	256
Critic learning rate
Actor learning rate
Discount factor ()	0.99
Exploration noise	0.1
Policy noise	0.2
Range to clip policy noise	0.5
Target smoothing coefficient ()	0.005
Number of samples per minibatch	256
Policy update frequency	2
Activation function	ReLU (rectified linear unit) [20]