Rainbow forest packing at the boundary k = 2m

Determine whether, for every ε>0, there exists m₀ such that every properly edge-colored simple graph G with e(G)=km and maximum color-class size μ(G,c)≤m, where m≥m₀ and 1≤k≤2m, contains at least (1−ε)m pairwise edge-disjoint rainbow forests, each with exactly k edges.

Background

The paper proves that if 0<δ<1 is fixed, m is sufficiently large, and 1≤k≤(2−δ)m, then every properly edge-colored simple graph with km edges and color classes of size at most m contains at least (1−ε)m pairwise edge-disjoint rainbow forests with exactly k edges. The factor m is asymptotically optimal because the graph has only km edges.

The authors show that the endpoint k=2m+1 cannot be included: a copy of K_{2m+1} with every edge assigned a distinct color has m(2m+1)=km edges and satisfies the global color bound, but no forest in this graph has 2m+1 edges. The unresolved boundary is therefore whether the theorem remains valid all the way through k=2m, removing the fixed relative margin δ while stopping short of the known obstruction.

References

We ask whether the fixed relative margin in Theorem~\ref{thm:main} can be removed. Is it true that, for every $\varepsilon>0$, there exists $m_0$ such that, for all integers $m,k$ with $m\geq m_0$ and $1\leq k\leq2m$, every properly edge-colored simple graph $G$ with $e(G)=km$ and $\mu(G,c)\leq m$ contains at least $(1-\varepsilon)m$ pairwise edge-disjoint rainbow forests, each with exactly $k$ edges?

— Nearly optimal packings of equally sized rainbow forests  (2609.29351 - Xu, 24 Sep 2026) in Section 4, “A question at the boundary,” Question 4.1