Cannot find reference 'ActorMAPPO' in 'net.py'

[josyulakrishna]

Hi I want to use this library for Multiagent RL, in AgentMAPPO.py file there are two undefined references from elegantrl.agents.net import ActorMAPPO, CriticMAPPO ActorMAPPO and CriticMAPPO how can I fix can this?

[shixun404]

@josyulakrishna Hi, there is an available demo for QMIX in a past version of ElegantRL, https://github.com/AI4Finance-Foundation/ElegantRL/commit/02dd28aa25fce9db2a663dfa5d3052551bd83561.

• You can check out the version by running the following command

  git checkout 02dd28

• This demo tests the QMIX algorithm on SMAC tasks. You can install SMAC by following the instructions at https://github.com/AI4Finance-Foundation/ElegantRL/blob/02dd28aa25fce9db2a663dfa5d3052551bd83561/elegantrl/envs/starcraft/README.md
• Run the following command to test QMIX

  python3 elegantrl/Demo_SMAC.py --config=qmix --env-config=sc2 with env_args.map_name=corridor

Thanks for your participation :) For MAPPO, our team will deliver it as soon as possible.

[josyulakrishna]

@shixun404 thank you very much, i will use this right away, can i use QMIX for continuous action spaces?

[shixun404]

@josyulakrishna No problem. This QMIX demo is for discrete action space. For continuous action space, you can use MADDPG or MAPPO. I thought there was a MADDPG demo before. I will go to check it :)

[josyulakrishna]

Thank you @shixun404 ! My application is on continuous action spaces, if the maddpg demo is available too that'd be great, apologies for too many requests haha..

[shixun404]

@josyulakrishna Take it easy :) This is a MADDPG demo https://github.com/shixun404/MADDPG_Demo. Thanks for your interest in ElegantRL.

[josyulakrishna]

Haha thank you again @shixun404 ! :)

[shixun404]

@josyulakrishna You're welcome! Thanks for your support of ElegantRL.