Computing Capacity-Cost Function for Continuous Channels in Wasserstein Space

Introduction

This repo implements algorithms to computing the capacity-cost function of discrete and continuous memoryless channel.

See the paper Computing Capacity-Cost Function for Continuous Channels in Wasserstein Space for more details.

Getting Started

Install this package into your conda environment, say cdopt.

conda create -n cdopt python=3.11 pip
conda activate cdopt
pip install -r requirements.txt
pip install -e .

Reproduce the Results

To reproduce the results in the paper, you can run the following commands.

# MIMO-AWGN
bash scripts/mimoawgn.sh

# Rayleigh Fading without CSIR
bash scripts/fading.sh

# Rayleigh Fading with CSIR
bash scripts/fading_csir.sh

The plots can be generated by the notebook notebooks/plot.ipynb.

For more details about the simulation configurations, please refer to the configs folder and the comments in the source code.

Examples

You can find examples in the folder notebook.

Acknowledgement

This repo is generated from the python template https://github.com/xinyanglii/python-template.