A novel approach for solving convex problems with cardinality constraints

G. Banjac and P. Goulart

in IFAC World Congress, Toulouse, France, July 2017.
BibTeX  URL 

  author = {G. Banjac and P. Goulart},
  title = {A novel approach for solving convex problems with cardinality constraints},
  booktitle = {IFAC World Congress},
  year = {2017},
  url = {https://doi.org/10.1016/j.ifacol.2017.08.2174},
  doi = {10.1016/j.ifacol.2017.08.2174}

In this paper we consider the problem of minimizing a convex differentiable function subject to sparsity constraints. Such constraints are non-convex and the resulting optimization problem is known to be hard to solve. We propose a novel generalization of this problem and demonstrate that it is equivalent to the original sparsity-constrained problem if a certain weighting term is sufficiently large. We use the proximal gradient method to solve our generalized problem, and show that under certain regularity assumptions on the objective function the algorithm converges to a local minimum. We further propose an updating heuristic for the weighting parameter, ensuring that the solution produced is locally optimal for the original sparsity constrained problem. Numerical results show that our algorithm outperforms other algorithms proposed in the literature.