/incorporating-history-and-deviations

Supplementary code for a paper.

Primary LanguageJupyter Notebook

Incorporating history and deviations in forward–backward splitting

This repository contains the code associated with our paper

Hamed Sadeghi, Sebastian Banert, and Pontus Giselsson. Incorporating history and deviations in forward–backward splitting. 2022. arXiv: 2208.05498v2 (math.OC)

Requirements

Running the notebook for symbolic verification requires an installation of Python, Jupyter, and SymPy. Running the numerical experiments requires Python, NumPy, and matplotlib.

I tested both with a system-wide installation on Arch Linux, using

  • Python 3.11.3
  • SymPy 1.12
  • Jupyterlab 4.0.2
  • NumPy 1.25.0
  • Matplotlib 3.7.1

Versions

The notebook uses notation and references from arXiv version v2 of the above-mentioned article. The calculation in Section 6 of arXiv version v1 can be verified in the same way when using $β$ instead of $\bar β$ and $\ell_n^2$ instead of $\ell_n$ everywhere.

The numerical experiments were added in arXiv version v2.

Authors

Sebastian Banert, Pontus Giselsson, Hamed Sadeghi,
Department of Automatic Control, Lund University.

In case of any problems or questions, feel free to open an issue for this repository.