Open Access
Research Article

Table 2.

Parameters used in DDQN

Parameter Value
The parameter ε of ε-greedy algorithm 0.8
Discount factor γ 0.9
Learning rate α 0.01
Maximum number of transitions M in replay memory 30,000
The size of mini-batch training set Ms 64
Descending degree tar 0.1
τ steps for updating the target network 200