Performance Evaluation of Onboard Processing Capability Reduction in Cooperative Vehicles Using 5G and Artificial Intelligence

Table 3

Summary of the key parameters of our algorithms proposed in previous works [57, 71].

Parameters	Architecture
	First	Second
Hidden layers	2
Hidden layer neurons	526
Dropout	0.2
Replay memories	100000
Mini batch size	96
Update target	Soft mode
Optimizer	RMSProp
Loss	MSE
End of episode	500 steps
	0.0025
Output layer neurons	6	5
Input layer neurons	28	28–30
Policy-based RL	DDQL	DDPG