This file contains the MATLAB implementation of the active-set method proposed in the following article
"An active-set method for sparse approximations. Part II: General piecewise-linear terms", S. Pougkakiotis, J. Gondzio, D. S. Kalogerias (10.13140/RG.2.2.16392.08962),
for the solution of convex quadratic programming problems with piecewise-linear terms in the objective.
The repository constists of the following files:
- An outer proximal method of multipliers (PMM),
- an inner semismooth Newton method (SSN),
- two ADMM-based warmstarting procedures,
- as well as three problem examples: i) from portfolio optimization ii) from quantile regression and iii) from binary classification with linear SVMs.
Links to the datasets are provided.