search

HumanCompatibleAI / overcooked_ai

A benchmark environment for fully cooperative human-AI performance.
https://arxiv.org/abs/1910.05789
MIT License
683 stars 144 forks source link

Is It Possible to Configure the Environment for Single Agent? #108

Closed mahaozhe closed 1 year ago

mahaozhe commented 1 year ago

Hi, thanks for all of the great work and the nice environment! May I ask if it is possible to configure the environment to have only one single agent to work? Maybe in this case, the task can be easily developed as one single-agent RL task. Thank you in advance!

jyan1999 commented 1 year ago

Hi,

Have you checked out this issue? I think currently the best way to accomplish this is still to substitute one agent with a StayAgent and put it in an insignificant spot on the layout. If you want to train with one using the existing framework in this repo you might need to define another policy class. It can be a static class so it won't be too complicated to implement.

Let me know if this helps!

mahaozhe commented 1 year ago

Thank you for your reply, to set one stay-agent should work very well, thank you very much!