Reinforcement Learning: competitive or collaborative options in MARL Matlab

Question

ali farid 2024-6-12

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2127896-reinforcement-learning-competitive-or-collaborative-options-in-marl-matlab

编辑： Shivani 2024-6-13

Hello,

I am trying to set up three explorer agents to explore the unknown area in collobrative or competive manners. I am wondering how I can set up this in Matlab, is there any options?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Shivani 2024-6-13

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2127896-reinforcement-learning-competitive-or-collaborative-options-in-marl-matlab#answer_1471166

编辑：Shivani 2024-6-13

Hello,

You can set up three explorer agents to collaboratively perform a particular task in Simulink by using the Reinforcement Learning Toolbox.

Kindly refer to this example, which elaborates on training multiple agents to perform a collaborative task, for more details on implementation: https://www.mathworks.com/help/reinforcement-learning/ug/train-2-agents-to-collaborate.html

It is important to mention here that multi-training is only supported in Simulink. If you have an environment created in MATLAB you could copy and paste the core parts like the reward and step function into a MATLAB Fcn block in Simulink as a workaround. You can look at the following MATLAB answer thred for more details: https://www.mathworks.com/matlabcentral/answers/654148-how-to-implement-multi-agent-rl-training-with-custom-matlab-environment-function

Hope this helps!