Want more on DQN, basically on multi-agent

I have taken the whole specialization and it helps a lot.
The last part, namely DQN, is the most intriguing for me.
In the class the moon landing theme is a single-agent model, but in my research area, there are many cases that there are many agents on the stage and interacting with each other, which may lead to a multi-agent model.
Thus, I am wondering is there any introduction-level paper or material of any kind on that? I don’t major in CS and a very cutting-edge paper will kill me instantly. :skull:
Any suggestions? Thanks in advance! :smiley:

Oh, also it will be better if the action space is discrete (to be or not to be), which sounds a little bit easier for me to understand!

Hey Tianqi,

Sorry I am not an expert on that, but perhaps you may want to search some literature review on Google Scholar?


Hi @Hu_Tianqi,

There are a few things to look at. First, there is a nice “Medium” article which goes into these ideas. The title is:

Deep Reinforcement Learning DQN for Multi-Agent Environment

Next, to go in to much greater detail, there is the paper:

Deep Q-Network Based Multi-agent Reinforcement Learning with Binary Action Agents

which is on arxiv.org. This is written at quite a high level so probably best to just stick with the Medium article unless you are feeling particularly ambitious.

These should provide a good starting point.


1 Like

Dear Professor chandler,
Thank you very much. It helps a lot. I think I will start from the Medium one. :smiley:
Best regards,

1 Like