A Spectral Bundle Method for Sparse Semidefinite Programs

This repository contains an implementation of the algorithm described in the paper titled A Spectral Bundle Method for Sparse Semidefinite Programs.

Table of Contents

• Prerequisites
• Getting Started
• Run
• Usage
• References
• License
• Contact

Prerequisites

Before you begin, ensure that you have the following software installed on your system:

• MATLAB

Getting Started

To get started with this project, follow these steps:

1. Install Solvers:

   • Sedumi
   • Mosek

2. Install CDCS:

   • Download CDCS.
   • Install CDCS and ensure that it's correctly set up.
   • Add "+cdcs_utils" in the CDCS package to your MATLAB path.

Run

You can run the project by executing the "Experiment1.m" script.

Usage

This project is designed to solve standard Semidefinite Programming (SDP) problems. If your problem data is structured with a sparsity pattern, this algorithm decomposes the large semidefinite cone constraint into several smaller ones, improving efficiency.

References

• A Spectral Bundle Method for Sparse Semidefinite Programs.

License

This project is licensed under the Your License Name or Link. See the LICENSE file for details.

Contact

If you have any questions, feedback, or need assistance, please don't hesitate to contact the project maintainer:

• Hesam Mojtahedi
• Email: [email protected]