English

Lower Bounds for Approximating the Matching Polytope

Computational Complexity 2017-11-29 v1

Abstract

We prove that any extended formulation that approximates the matching polytope on nn-vertex graphs up to a factor of (1+ε)(1+\varepsilon) for any 2nε1\frac2n \le \varepsilon \le 1 must have at least (nα/ε)\binom{n}{{\alpha}/{\varepsilon}} defining inequalities where 0<α<10<\alpha<1 is an absolute constant. This is tight as exhibited by the (1+ε)(1+\varepsilon) approximating linear program obtained by dropping the odd set constraints of size larger than (1+ε)/ε({1+\varepsilon})/{\varepsilon} from the description of the matching polytope. Previously, a tight lower bound of 2Ω(n)2^{\Omega(n)} was only known for ε=O(1n)\varepsilon = O\left(\frac{1}{n}\right) [Rothvoss, STOC '14; Braun and Pokutta, IEEE Trans. Information Theory '15] whereas for 2nε1\frac2n \le \varepsilon \le 1, the best lower bound was 2Ω(1/ε)2^{\Omega\left({1}/{\varepsilon}\right)} [Rothvoss, STOC '14]. The key new ingredient in our proof is a close connection to the non-negative rank of a lopsided version of the unique disjointness matrix.

Keywords

Cite

@article{arxiv.1711.10145,
  title  = {Lower Bounds for Approximating the Matching Polytope},
  author = {Makrand Sinha},
  journal= {arXiv preprint arXiv:1711.10145},
  year   = {2017}
}

Comments

To appear in proceedings of SODA '18

R2 v1 2026-06-22T22:59:02.730Z