Epsilon greedy policy for DQN

Hello,

I have created DQN agent with epsilon greedy exploration which has 4 discrete actions and 10 observations.

Now, my policy is:

Epsilon = 0.9;

EpsilonDecay = 1e-3

EpsilonMin = 0.01

I want to plot the Epsilon value over the episodes during the training, or need to find the variable Epsilon over the training. But, i just can see the above described values even after the training has been done.

If you have idea how to plot/know the epsilon for particular episodes then please let me know?

Thanks

Answer 1

[Emmanouil Tzorakoleftherakis]

You can use the formula here to calculate the epsilon value