Three-Tier Computing Platform Optimization: A Deep Reinforcement Learning Approach

Table 3

Actor-critic DRL parameters.

Entropy weight	0.005
Clip value of the gradient clipping	40.0
Buffer size	10,000
Minibatch size of DNN	64
Maximum episode	1000
Maximum number of steps in each episode	1000
Actor learning rate	0.0001
Critic learning rate	0.001
Number of iterations X	300
Activation function of DNN	ReLu
Number of hidden layers of DNN	2
Number of neurons in the hidden layers	300