A polynomial kernel for vertex deletion into bipartite permutation graphs
Jan Derbisz
TL;DR
The paper studies the Bipartite Permutation Vertex Deletion problem, where one seeks to delete at most $k$ vertices to obtain a bipartite permutation graph (a NP-hard, FPT problem). It develops a polynomial kernel of size $O(k^{62})$ by combining a forbidden-subgraph characterization with a sunflower-based compression of small structures, followed by reduction rules that bound both isolated and non-isolated vertices after removing a modulator $T$. The analysis introduces explicit polynomial bounds $\\delta(k)$, $\\epsilon(k)$, and $\phi(k)$ to control different vertex groups and shows how they culminate in a kernel of size $\xi(k)=O(k^{62})$. This work advances kernelization techniques for vertex-deletion problems into graph classes and demonstrates a concrete, scalable pre-processing bound for this graph class.
Abstract
A permutation graph can be defined as an intersection graph of segments whose endpoints lie on two parallel lines $\ell_1$ and $\ell_2$, one on each. A bipartite permutation graph is a permutation graph which is bipartite. In the the bipartite permutation vertex deletion problem we ask for a given $n$-vertex graph, whether we can remove at most $k$ vertices to obtain a bipartite permutation graph. This problem is NP-complete but it does admit an FPT algorithm parameterized by $k$. In this paper we study the kernelization of this problem and show that it admits a polynomial kernel with $O(k^{62})$ vertices.
