Altitute for Reinforcement Learning based Routing Optimization

Jing Chen, Zili Meng, Mingwei Xu. Always Heading for the Peak: Learning to Route with Domain Knowledge. In Proceedings of IEEE INFOCOM 2020 (Poster).

Prerequisites

• Python
• Tensorflow
• CPU or NVIDIA GPU + CUDA CuDNN

Getting started

train altitude-based RL model

Train with the command line:

python agent_cap.py --topology ...

See more training options with python agent_cap.py --help

train Learning-To-Route model

Train with the command line:

python agent.py --topology ...