Skip to content

joaompereira/hipnex

Repository files navigation

HIPNEX

Matlab implementation of the algorithm proposed in

Alves, M. M., J. M. Pereira, and B. F. Svaiter, A search-free $O (1/k^{3/2})$ homotopy inexact proximal-Newton extragradient algorithm for monotone variational inequalities, arXiv:2308.05887 (2023).

Usage

This repository is self contained. After downloading, you may use the function hipnex right away to solve unconstrained variational inequality problems. Alternatively, we provide the function plain_npe for an implementation of the Newton Proximal-Extragradient method, introduced in:

See the documentation of hipnex and plain_npe regarding usage.

Reproducing results in the paper

To reproduce the experiments in the paper, run run_experiments to run the algorithms and then run plot_experiments and table_experiments to obtain the corresponding plot and table.

Other examples

Besides the experiments included in the paper, we also include in this repository a quadratic min-max experiment, including its setup, so that users can more easily understand how to set-up an optimization problem to use HIPNEX or NPE.

External software

Although the repository is self-contained, we include software from other authors, which we acknowledge here. Namely, we include:

  • ORN_ls_simple: code for Second Order Generalized Optimistic method, proposed in:
  • lsqrSOL: We used the Matlab implementation of LSQR available at https://web.stanford.edu/group/SOL/software/lsqr/, and modified it to include the stopping criteria in equation (26) of the HIPNEX paper.
  • minres_sol: We implement our version of the MINRES algorithm, using the Matlab implementation available in the Wikipedia page, as reference.

Feedback

Feel free to submit any feedback by submitting an issue in this repository, or alternatively by e-mailing to [email protected].

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages