Short Rainbow Circuits in Regular Matroids
Sean McGuinness
TL;DR
This work resolves a conjecture of DeVos et al. by proving that every simple regular matroid $M$ with a $(r(M)+1)$-colouring, where each colour class has at least two elements, contains short rainbow circuits. The authors develop a unifying framework based on SRC$4$-tuples—collections of four rainbow circuits with total size at most $2r(M)+4$ and each element lying in at most two circuits—and show this structure persists under the standard Seymour decomposition of regular matroids. A key strategy is to analyze circuit-achromatic colourings on binary matroids (via stratifications) and to prove the main result first for graphic and cographic matroids, then extend to all regular matroids through sums and extensions. The paper also handles extremal edge-count cases and duals thegraphic/cographic arguments to secure the general theorem. The result thus yields a robust short rainbow-circuit guarantee, with implications for rainbow structures in matroids and their decompositions, and it provides a constructive approach that underpins potential algorithmic applications.
Abstract
DeVos et al conjectured that if $M$ is a simple, regular matroid and $c$ is a colouring of the elements of $M$ with $r(M)+1$ colours, where each colour class has at least two elements, then $M$ contains a rainbow circuit of size at most $\lceil \frac {r(M)+1}2 \rceil.$ We prove this conjecture by showing that for all such regular matroids there are four rainbow circuits $C_i,\ i = 1,2,3,4$ for which $\sum_i |C_i| \le 2r(M) +4$ and for which no element of $M$ belongs to more than two of the circuits.
