Fast algorithms for Vizing's theorem on bounded degree graphs

Vizing's theorem states that every graph of maximum degree can be properly edge-colored using colors. The fastest currently known -edge-coloring algorithm for general graphs is due to Sinnamon and runs in time , where and . We investigate the case when is constant, i.e., . In this regime, the runtime of Sinnamon's algorithm is , which can be improved to , as shown by Gabow, Nishizeki, Kariv, Leven, and Terada. Here we give an algorithm whose running time is only , which is obviously best possible. Prior to this work, no linear-time -edge-coloring algorithm was known for any . Using some of the same ideas, we also develop new algorithms for -edge-coloring in the model of distributed computation. Namely, when is constant, we design a deterministic algorithm with running time and a randomized algorithm with running time . Although our focus is on the constant regime, our results remain interesting for up to , since the dependence of their running time on is polynomial. The key new ingredient in our algorithms is a novel application of the entropy compression method.
View on arXiv