search

PKU-RL / I2C

41 stars 12 forks source link

Computing Metrics #1

Open varun-intel opened 3 years ago

varun-intel commented 3 years ago

I'm trying to understand how the metrics in the paper were generated from this code, particularly in the Cooperative Navigation environment. Could you please explain how you aggregate the metrics for each episode? Thanks for your help.

leonkding commented 3 years ago

In the paper, the report reward is the reward of the last step in one episode. The occupied landmark is the same case.