Heavy‑Ball Projected Primal‑Dual Method (HBPPDM)

This repository provides a lightweight implementation of the Heavy‑Ball Projected Primal‑Dual Method (HBPPDM) for constrained optimization. The method accelerates a standard projected primal–dual algorithm by adding heavy‑ball momentum, which improves convergence speed when solving constrained linear and quadratic problems. The HBPPDM algorithm and its theoretical properties are based on peer‑reviewed research.

Citing This Work

If you use this repository in academic work, please cite the foundational publication:

• Y. J. J. Heuts, G. P. Padilla and M. C. F. Donkers, An Adaptive Restart Heavy-Ball Projected Primal-Dual Method for Solving Constrained Linear Quadratic Optimal Control Problems, 2021 60th IEEE Conference on Decision and Control (CDC), Austin, TX, USA, 2021, pp. 6722-6727, doi: 10.1109/CDC45484.2021.9683013.
• Y. J. J. Heuts, M. C. F. Donkers, A computationally efficient primal–dual solver for linear-quadratic optimal control problems, Automatica, Volume 177, 2025, 112341, ISSN 0005-1098, doi: 10.1016/j.automatica.2025.112341.

License

This project is released under the BSD‑3‑Clause License.