Qoncord Experiment Reproduction

This repository contains the scripts necessary to reproduce the key results presented in our paper: Qoncord: A Multi-Device Job Scheduling Framework for Variational Quantum Algorithms. Each experiment has its dedicated Python script for execution, and a separate script for generating the corresponding plots.

Table of Contents

• Prerequisites
• Setup
• Experiments
  • Queue Simulation
  • Multi-Restart QAOA Optimization
  • Single-Restart VQE Optimization
  • Comparison to Asynchronous Gradient Descent
• License

Prerequisites

• Conda (for environment management)
• Python 3.11+

Setup

1. Clone the repository:

   git clone https://github.com/meng-ubc/Qoncord.git
   cd Qoncord

2. Create and activate the Conda environment:

   conda create -n qoncord_env python=3.11
   conda activate qoncord_env

3. Install the required packages:

   pip install -r requirements.txt

Experiments

Each experiment has its own script. Running these scripts will execute the experiment and save the results for plotting.

Queue Simulation

• Script: main_queue_sim.py
• Expected Runtime: Approximately 2 minutes
• Output: queue_sim_results.csv

To run the experiment:

cd 1_queue_simulation
python queue_simulation.py

To plot the result fiugre:

python plot_queue_sim_result.py

Multi-Restart QAOA Optimization

• Script: 2_qaoa_optimization.py
• Expected Runtime: Approximately 10 minutes
• Output: qaoa_results.json

To run the experiment:

python 2_qaoa_optimization.py

To plot the result fiugre:

python plot_results.py 2

Single-Restart VQE Optimization

• Script: 3_vqe_optimization.py
• Expected Runtime: Approximately 4 minutes
• Output: vqe_results.json

To run the experiment:

python 3_vqe_optimization.py

To plot the result fiugre:

python plot_results.py 3

Comparison to Asynchronous Gradient Descent

• Script: 4_ae_async.py
• Expected Runtime: Approximately 6 minutes
• Output: async_results.json

To run the experiment:

python 4_ae_async.py

To plot the result fiugre:

python plot_results.py 4

License

This project is licensed under the MIT License. See the LICENSE file for details.