Table of Contents
Fetching ...

The complexity of matroid homomorphism reconfiguration

Cheolwon Heo, Mark Siggers

TL;DR

This work studies the reconfiguration of matroid homomorphisms ${\rm Recol}_\mathbb{M}(N)$ for binary matroids, defining adjacency via cocircuits and establishing a Recolouring framework analogous to graphs. The authors develop decision graphs $D(N)$ and their Tutte-inspired connections to bridge matroid recolouring with graph recolouring, and they introduce Kempe recolouring to transfer complexity results between settings. They prove a dichotomy: ${\rm Recol}_\mathbb{M}(N)$ is trivial when $N$ dismantles to $M(K_2)$ (or to a loop), and PSPACE-complete for $N=M(K_n)$ with $n\ge 3$, and for any graphic matroid containing $M(K_5)$, via gadget constructions and reductions. The work furthermore frames a conjecture that triviality corresponds precisely to dismantling to $M(K_2)$ and raises open questions about non-graphic matroids and girth-based tractability. This connects matroid theory, Tutte duality, and reconfiguration complexity, broadening understanding of colourings beyond graphs. $${\rm Recol}_\mathbb{M}(N)$$ thus inherits rich structure from graphic cases and Tutte-transformations, with clear avenues for future exploration.

Abstract

We consider a reconfiguration version of the homomorphism problem ${\rm Hom}_\mathbb{M}(N)$ for binary matroids $N$. This reconfiguration problem, ${\rm Recol}_\mathbb{M}(N)$, asks, for two homomorphisms $φ$ and $ψ$ of a matroid $M$ to $N$, if there is a path of homomorphism from $φ$ to $ψ$ such that consecutive homomorphism in the path differ on a single cocircuit of $N$. We show that this problem is trivial in the case that $N$ dismantles to the graphic matroid $M(K_2)$, and that the problem is ${\rm PSPACE}$-complete when $N$ is the graphic matroid $M(K_3)$, $M(K_4)$, or any graphic matroid containing $M(K_5)$.

The complexity of matroid homomorphism reconfiguration

TL;DR

This work studies the reconfiguration of matroid homomorphisms for binary matroids, defining adjacency via cocircuits and establishing a Recolouring framework analogous to graphs. The authors develop decision graphs and their Tutte-inspired connections to bridge matroid recolouring with graph recolouring, and they introduce Kempe recolouring to transfer complexity results between settings. They prove a dichotomy: is trivial when dismantles to (or to a loop), and PSPACE-complete for with , and for any graphic matroid containing , via gadget constructions and reductions. The work furthermore frames a conjecture that triviality corresponds precisely to dismantling to and raises open questions about non-graphic matroids and girth-based tractability. This connects matroid theory, Tutte duality, and reconfiguration complexity, broadening understanding of colourings beyond graphs. thus inherits rich structure from graphic cases and Tutte-transformations, with clear avenues for future exploration.

Abstract

We consider a reconfiguration version of the homomorphism problem for binary matroids . This reconfiguration problem, , asks, for two homomorphisms and of a matroid to , if there is a path of homomorphism from to such that consecutive homomorphism in the path differ on a single cocircuit of . We show that this problem is trivial in the case that dismantles to the graphic matroid , and that the problem is -complete when is the graphic matroid , , or any graphic matroid containing .

Paper Structure

This paper contains 9 sections, 14 theorems, 13 equations, 2 figures.

Key Result

Theorem 1.6

The problem $\mathop{\mathrm{\mathop{\mathrm{Recol}}\nolimits_\mathbb{M}}}\nolimits(M(K_n))$ is trivial if $n \leq 2$, and is otherwise $\mathop{\mathrm{PSPACE}}\nolimits$-complete.

Figures (2)

  • Figure 1: A $K_4$-colouring $\phi$ of a graph $G$, the $M(K_3)$-colouring $\tau_\phi$ of $M(G)$ that it defines, and the $K_4$-colouring $\phi_{\tau_\phi}$ of $G$ that $\tau_\phi$ defines.
  • Figure :

Theorems & Definitions (50)

  • Definition 1.1
  • Example 1.2
  • Example 1.3
  • Definition 1.4
  • Example 1.5
  • Theorem 1.6
  • proof
  • proof
  • proof
  • Example 2.4
  • ...and 40 more