Do reinforcement learning problems on networks problems require mathematic modeling in ns3?

[Simo Abdo]

I am a PhD student in computer science, and I want to apply reinforcement learning methods to computer networks problems in ns3 and get results in order to publish a research paper in international journals, but I am not very good in mathematics, so my question is: can I model a problem and apply reinforcement learning to networks problems(like energy efficiency , congestion control ,...) without needing mathematics?

[Charles Pandian]

There are some ns-3 extensions such as ns3-Gym and ns3-AI to implement your RL-based protocol under ns-3.

You may search for ns3-Gym and ns3-AI and learn to use them from their official documentation and installation procedure.

You may have a rough idea of about these AI/RL frameworks by looking into the following posts. Especially, you may see the References section at the bottom to follow the original installation procedures (since the following versions may not be complete ones). Anyway, you may get some general idea if you see the following links:

https://www.projectguideline.com/installing-ns3-gym-ai-model-under-chroot-jail/
https://www.projectguideline.com/installing-drl-based-cw-optimization-framework-under-ns3-gym/
https://www.projectguideline.com/rl-fec-a-self-contained-reinforcement-learning-code-of-fec-for-ns-3/
https://www.projectguideline.com/installing-ns3-ai-model-under-ns-3-35/

Charles Pandian,

On Sat, Jun 11, 2022 at 6:46 AM Simo Abdo <simoab...@gmail.com> wrote:

I am a PhD student in computer science, and I want to apply reinforcement learning methods to computer networks problems in ns3 and get results in order to publish a research paper in international journals, but I am not very good in mathematics, so my question is: can I model a problem and apply reinforcement learning to networks problems(like energy efficiency , congestion control ,...) without needing mathematics?

--
Posting to this group should follow these guidelines https://www.nsnam.org/wiki/Ns-3-users-guidelines-for-posting
---
You received this message because you are subscribed to the Google Groups "ns-3-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ns-3-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ns-3-users/edbc8cf6-9d40-4781-8c1e-93874caefa3an%40googlegroups.com.

[Simo Abdo]

Thank you for your response, I have already installed these modules on NS3 and tested some examples , but my question is , can i find some good results to publish in international journals , without being very good on mathematics ?

To view this discussion on the web visit https://groups.google.com/d/msgid/ns-3-users/CAHXeiM9rsSPKKwHOvAC5rk-o%2BXR1q%2BkVB-zbN0nt5jDWag9cRQ%40mail.gmail.com.

[Mubanga Nsofu]

My friend, 

RL and Math are not mutually exclusive. I suggest you rather start learning the foundations from Sutton's book or begin with basic examples from the entry point which would be bandits.

[Ali H. Wheeb]

i s there any lectures or tutorials explaining how to use AI or Rl in ns-3 networks?

[Mubanga Nsofu]

Hi Ali,

You can check out this lecture tutorial that describes how to use RL in ns-3 using ns3-ai

https://vimeo.com/groups/779512/videos/566296651

I have not found any video lecture that shows how to use ns3-gym in ns-3 only a github link
and their introductory paper.

Mubanga