numerical_qkd

This repository contains the numerical method for calculating the secret key rate of QKD with the finite key analysis.

Installation

This package uses and is written in Python 3.x

It is optional to use/install the MATLAB tools, but we use MATLAB and CVX because certain SDPs take too long to be solved. Please install the following MATLAB tools:

• MATLAB Engine API for Python
• CVX for MATLAB
• cvxquad

Optional (but desirable): Install MOSEK solver for python.

Installation steps

1. If you'd like to use the MATLAB tools, please install the tools described above.

2. Install the required python packages using: pip install -r requirements.txt

3. Installation of Numerical QKD Python package can be done via setuptools >= 36.2.7: python setup.py install

Convex optimization tools

The convex optimization problems are solved using two disciplined convex programming (DCP) tools:

• cvxpy : www.cvxpy.org (Python)
• cvx : www.cvxr.com (MATLAB)

Usage

Please refer see the Jupyter notebooks in the examples folder to see how you can use this library. The simplest example is: simple_BB84_example.ipynb. To see how postselection is performed, see BB84_protocol.ipynb.