/capacity-rl-po

Policy optimization technique to compute the feedback capacity

Primary LanguagePythonApache License 2.0Apache-2.0

Computing the Feedback Capacity of Finite State Channels using Reinforcement Learning

This repository contains an implementation of feedback capacity estimator of an unifilar finite-state-channel using reinforcement learning, specifically policy optimization technique.

Prerequisites

The code is compatible with a tensorflow 2.0 environment. If you use a docker, you can pull the following docker image

docker pull tensorflow/tensorflow:latest-gpu-py3

Running the code

The parameters of the channels are in the file example.json. Modify this file to set your own values for the parameters.

To run the code:

python ./example.py --exp_name <simulation_name> --config ./configs/example.json

Authors

  • Ziv Aharoni
  • Oron Sabag
  • Haim Permuter

License

This project is licensed under Apache License 2.0 - see the LICENSE.md file for details