This code outputs a graphic that presents the fraction of successes in convex (LP) optimization recovering the combinatorial (NP) optimization outcomes. The equivalence phase transition is depicted as a function of the aspect ratio gamma and the sparsity through the phase transition rho. The user must specify the number of points generated, the length of the solution (n), the number of points for δ and ρ, as well as the number of trials performed to compute the fraction of equivalence. The larger the parameters specified (in particular the number of trials), the longer the code takes to run! For more information, please visit the SparseLab (Seeking Sparse Solutions to Linear Systems of Equations) website (http://sparselab.stanford.edu/).
Working Paper (2021)
Fernández-de-Marcos Alberto and García-Portugués Eduardo