A note on hyperplane generation
This short and elegant note describes an algorithm that generates all the hyperplanes spanned by a finite point set \(X\) ``at a polynomial rate. That is, although the set \(\mathcal H\) of hyperplanes may have exponential size, the first \(i\) hyperplanes in \(\mathcal H\) are generated in a number of steps that is polynomial in \(i\) and the size of the input. In particular, one can check whether a given list of hyperplanes is complete, in a time that is polynomial in \(| X|+ |{\mathcal H}|\), the number of points plus the number of hyperplanes. The proof is based on a matroid formulation of the problem: ``The introduction of matroid theory here is just for clarity and maximum generality, but later, in the proof, the use of matroid theory is essential. A problem of Lovász that motivated this note, how to check whether a list \(\mathcal H\) of hyperplanes completely describes the convex hull of \(X\), remains open.
- Computing the spark: mixed-integer programming for the (vector) matroid girth problem
- Which nonnegative matrices are slack matrices?
- a-tint: a polymake extension for algorithmic tropical intersection theory
- On the (co)girth of a connected matroid
- From the zonotope construction to the Minkowski addition of convex polytopes
- Unique key Horn functions
- scientific article; zbMATH DE number 4125543 (Why is no real title available?)
- Algorithmic aspects of partial convexity
This page was built for publication: A note on hyperplane generation
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1328388)