Extension Complexity Lower Bounds for Mixed-Integer Extended Formulations
Abstract
We prove that any mixed-integer linear extended formulation for the matching polytope of the complete graph on vertices, with a polynomial number of constraints, requires many integer variables. By known reductions, this result extends to the traveling salesman polytope. This lower bound has various implications regarding the existence of small mixed-integer mathematical formulations of common problems in operations research. In particular, it shows that for many classic vehicle routing problems and problems involving matchings, any compact mixed-integer linear description of such a problem requires a large number of integer variables. This provides a first non-trivial lower bound on the number of integer variables needed in such settings.
Cite
@article{arxiv.1611.00707,
title = {Extension Complexity Lower Bounds for Mixed-Integer Extended Formulations},
author = {Robert Hildebrand and Robert Weismantel and Rico Zenklusen},
journal= {arXiv preprint arXiv:1611.00707},
year = {2022}
}
Comments
Unfortunately, the proof technique seems to have a flaw in Lemma 8. Specifically, there was an error in rearranging formulas involving projections of mixed integer sets. The overall main results, it turns out hold true. Please see the improved techniques (quite different) by Cevallos, Weltge, and Zenklusen. https://epubs.siam.org/doi/pdf/10.1137/1.9781611975031.51