TCP Congestion Control - A Machine Learning Approach

An attempt to apply machine learning for tcp congestion control. This is our CS359 computer networks project. We have made a simple network simulator to test our algorithms. First we implemented a simplified TCP Tahoe. Then we collected data from it and trained a polynomial regression model. Finally we used the model and collected the results below.

Install dependencies using:

python3 -m pip install -U -r requirements.txt

To run the model:

python3 ./src/main.py

Results:

1 host

2 hosts

3 hosts

4 hosts

As can be seen from above, the algorithm divides the available bandwidth fairly equally. ## Contributors: - [aryadas98](https://github.com/aryadas98) - [ankurD98](https://github.com/ankurD98) - [atul337](https://github.com/atul337)