The Implementation of Deep Reinforcement Learning in E-Learning and Distance Learning: Remote Practical Work

Table 3

The options used to build our model.

	Layers	Sizes	Normalization
Image observation	Image input layer	(84, 84, 4)	None
State of the interrupter and the electronic corposant	Image input layer; Fully connected layer; ReLU layer; Fully connected layer	(1, 1); 500; —; 400	None
Action	Image input layer; Fully connected layer	(1, 1); 400	None
Output	Addition layer; ReLU layer; Fully connected layer	—; —; 1	—; —; —