which takes in a shared observation together with a local view, returns the action for each individual agent. As the bi-directional recurrent structure could serve not only as a communication channel but also as
Log In Sign Up NeurIPS 2022: CityLearn ChallengeCommNet and DDPGImplementing DDPG and CommNetmark_haoxiang 12 August 2022 9 Open in Colab A notebook containing a sample starter implementation of a centralised agent utilising the DDPG algorithm for training. This is not interfaced with order...