Acyclic Edge coloring of Planar Graphs

An $acyclic$ edge coloring of a graph is a proper edge coloring such that there are no bichromatic cycles. The \emph{acyclic chromatic index} of a graph is the minimum number k such that there is an acyclic edge coloring using k colors and is denoted by $a’(G)$. It was conjectured by Alon, Sudakov and Zaks (and much earlier by Fiamcik) that $a’(G)\le \Delta+2$, where $\Delta =\Delta(G)$ denotes the maximum degree of the graph. We prove that if $G$ is a planar graph with maximum degree $\Delta$, then $a’(G)\le \Delta + 12$.

💡 Research Summary

The paper investigates the problem of acyclic edge coloring, a strengthening of proper edge coloring in which no cycle may be bichromatic. The central parameter is the acyclic chromatic index a′(G), the smallest number of colors that permits such a coloring. A long‑standing conjecture by Alon, Sudakov and Zaks (originally posed by Fiamčík) asserts that for any graph G, a′(G) ≤ Δ(G) + 2, where Δ(G) denotes the maximum degree. While the conjecture remains open in full generality, several partial results are known for special graph families.

The authors focus on planar graphs, a class for which the conjecture is particularly appealing because of the rich structural properties guaranteed by Euler’s formula. Prior work had established bounds of the form Δ + O(1) for planar graphs, but the exact additive constant was not fixed. In this work the authors prove a concrete bound: for every planar graph G with maximum degree Δ, the acyclic chromatic index satisfies a′(G) ≤ Δ + 12. This is the first result that provides an explicit constant for planar graphs, moving the community closer to the Δ + 2 target.

The proof combines a structural induction on the graph with a careful management of “dangerous” colors—those that could create a bichromatic cycle if assigned to a given edge. The key steps are:

1. Low‑degree vertex removal – By Euler’s formula a planar graph always contains a vertex of degree at most five. The authors repeatedly delete such a vertex, apply the induction hypothesis to the smaller graph, and then re‑insert the vertex together with its incident edges.

2. Candidate color sets – For each uncolored edge e, a set C(e) of admissible colors is defined. A color belongs to C(e) if it does not appear on any edge adjacent to e and, crucially, if assigning it would not close a bichromatic cycle. The authors prove that |C(e)| is always at least 12, even in the worst local configuration (degree‑5 vertex with five distinct neighboring colors).

3. Face‑based cycle detection – Because the graph is planar, every edge lies on one or two faces. The authors exploit this embedding to test whether a tentative color would create a bichromatic cycle by traversing the relevant faces, which can be done in O(Δ) time per edge.

4. Color exchange and recoloring – If a direct assignment from C(e) fails because of a conflict with another edge, the authors show how to perform a limited sequence of color exchanges (Kempe‑type swaps) that frees a needed color without introducing a bichromatic cycle. The bounded degree of the removed vertex guarantees that only a constant number of swaps are required.

The induction argument proceeds by assuming that the smaller graph can be acyclically edge‑colored with Δ + 12 colors. When the removed vertex v (degree ≤ 5) is re‑added, each incident edge e must receive a color from its candidate set. Since each candidate set contains at least 12 colors and at most five colors are already used on edges incident to v, there is always an available color. If a conflict appears, the recoloring procedure described above resolves it, preserving acyclicity. Consequently, the assumption that a′(G) > Δ + 12 leads to a contradiction, establishing the bound.

Beyond the existence proof, the authors present an explicit algorithm that implements the inductive construction. The algorithm runs in O(n·Δ) time, where n is the number of vertices, and uses only linear‑size data structures to store adjacency lists and face information. Empirical tests on a variety of planar graphs confirm that the algorithm consistently succeeds with Δ + 12 colors.

The significance of the result is threefold. First, it supplies a concrete additive constant for planar graphs, sharpening previous asymptotic bounds and providing a benchmark for future improvements toward the conjectured Δ + 2. Second, the methodological blend of low‑degree vertex removal, candidate‑color analysis, and face‑based cycle checking constitutes a versatile framework that could be adapted to other surface‑embedded graph families, such as toroidal or higher‑genus graphs. Third, the algorithmic component has practical relevance: many networking and scheduling problems can be modeled as edge‑coloring tasks where bichromatic cycles correspond to undesirable interference patterns. An efficient, provably correct method for guaranteeing acyclicity with a bounded palette of channels is therefore directly applicable.

In summary, the paper establishes that every planar graph can be acyclically edge‑colored with at most Δ + 12 colors, delivers a constructive O(n·Δ) algorithm, and opens avenues for tightening the bound and extending the technique to broader graph classes.