Non-convex solvers compute improved lower bounds on nonnegative matrix rank, with a new algorithm for the self-scaled bound that establishes exact rank for some matrices.
Computing The Extension Complexities of All 4-Dimensional 0/1-Polytopes
1 Pith paper cite this work. Polarity classification is still indexing.
abstract
We present slight refinements of known general lower and upper bounds on sizes of extended formulations for polytopes. With these observations we are able to compute the extension complexities of all 0/1-polytopes up to dimension 4. We provide a complete list of our results including geometric constructions of minimum size extensions for all considered polytopes. Furthermore, we show that all of these extensions have strong properties. In particular, one of our computational results is that every 0/1-polytope up to dimension 4 has a minimum size extension that is also a 0/1-polytope.
fields
math.OC 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Computing Lower Bounds on the Nonnegative Rank via Non-Convex Optimization Solvers
Non-convex solvers compute improved lower bounds on nonnegative matrix rank, with a new algorithm for the self-scaled bound that establishes exact rank for some matrices.