English

Sparse Approximation Over the Cube

Optimization and Control 2022-10-07 v1 Discrete Mathematics

Abstract

This paper presents an anlysis of the NP-hard minimization problem min{bAx2: x[0,1]n,supp(x)σ}\min \{\|b - Ax\|_2: \ x \in [0,1]^n, | \text{supp}(x) | \leq \sigma\}, where supp(x)={i[n]:xi0}\text{supp}(x) = \{i \in [n]: x_i \neq 0\} and σ\sigma is a positive integer. The object of investigation is a natural relaxation where we replace supp(x)σ| \text{supp}(x) | \leq \sigma by ixiσ\sum_i x_i \leq \sigma. Our analysis includes a probabilistic view on when the relaxation is exact. We also consider the problem from a deterministic point of view and provide a bound on the distance between the images of optimal solutions of the original problem and its relaxation under AA. This leads to an algorithm for generic matrices AZm×nA \in \mathbb{Z}^{m \times n} and achieves a polynomial running time provided that mm and A\|A\|_{\infty} are fixed.

Keywords

Cite

@article{arxiv.2210.02738,
  title  = {Sparse Approximation Over the Cube},
  author = {Sabrina Bruckmeier and Christoph Hunkenschröder and Robert Weismantel},
  journal= {arXiv preprint arXiv:2210.02738},
  year   = {2022}
}