Table 8.

Configuration hyperparameters for the IA trained by reinforcement through the PPO algorithm.

Trainer	PPO
Batch size	16
Beta	0.01
Buffer size	256
Epsilon	0.15
Gamma	0.9
Hidden units	64
Lambda	0.9
Learning rate	5 × 10−4
Max steps	10 × 104
Num epoch	10
Num layers	3
Time horizon	4