Red-QAOA Repository

This repository contains the source code and experimental scripts for the "Red-QAOA" (Quantum Approximate Optimization Algorithm) variant, as well as the relevant graph datasets used in our study.

Directory Structure

• src/: Contains the source files of the Red-QAOA implementation.
• experiments/: Houses the scripts for conducting various experiments as described in the paper.
• graph_sets/: Includes the graph datasets (Linux, AIDS, IMDb) used in our experiments.
• additional_experiments/: Houses the scripts for conducting additional experiments in the paper.

Installation

Prerequisites

• Python 3.11

Steps

1. Install Required Packages: Ensure Python 3.11 is installed on your system. Then, run the following command to install all necessary dependencies:

pip install -r requirements.txt

2. Optional GPU Support: For users with CUDA-enabled systems, the qiskit-aer-gpu package can be installed separately for GPU acceleration:

pip install qiskit-aer-gpu

Running Experiments

Each experiment script in the experiments/ folder corresponds to different evaluations of the Red-QAOA. Below are the key experiments:

1. MSE Analysis in Noisy and Ideal Conditions:

• Script: mse_noisy.py and mse_ideal.py
• Required and optional arguments are detailed within each script.

2. End-to-End Performance Evaluation:

• Script: end_to_end.py
• The script's arguments allow for customization and detailed performance analysis.

Refer to the individual script documentation for detailed usage instructions.

Additional Experiments

Additional experiments and their guides are available in the repository. These supplement the key experiments and provide further insights into Red-QAOA's capabilities.

Experiment Customization

Experiment parameters such as the number of QAOA layers are set as required arguments for consistency with the study. Optional arguments are available for more in-depth and varied testing.

Hardware and Software Dependencies

• Optional: CUDA-enabled NVidia GPUs for enhanced performance.
• Required software: Qiskit, Networkx, Scipy.
• Optional software: torch-geometric for comparing Red-QAOA with GNN-based methods.

Data Sets

The graph_sets/ folder includes the following datasets:

• Linux
• AIDS
• IMDb

These are used across various experiments to evaluate the Red-QAOA's performance.