Table 3 GNN configurations.

From: Graph neural network-based cell switching for energy optimization in ultra-dense heterogeneous networks

Hyperparameters

Values

Number of hidden layers; Neuron size

3; 128 \(\times\) 128 \(\times\) 64

Hidden layer activation function

ReLU38

Output layer activation function

Sigmoid39

Optimizer

AdamW41

Learning rate (LR)

\(10^{-3}\)

LR scheduler

Reduce LR on Plateau40

Batch size

64

Maximum number of epochs

200

Back to article page