Clever Geek Handbook

Openai gym for multi-agent games

Can openai game environments be used for multi-agent games? In particular, I would like to simulate a card game with four players (agents). The player who has gained momentum starts the next turn. How would I model the necessary coordination between players (eg who will it be next)? Ultimately, I would like to use amplification training for four agents playing against each other.

+3


source to share


2 answers


The OpenAI team has implemented a multi-agent deep deterministic policy gradient. The MADDPG method was implemented by the OpenAI team.



This is a repo to get started. https://github.com/openai/multiagent-particle-envs

0


source


There is a specific multi-agent reinforcement learning environment here . It supports any number of agents written in any programming language. An example game has already been implemented, which is a card game.



0


source






More articles:


All Articles