Convergence Times of Decentralized Graph Coloring Algorithms

Ordinary graph coloring algorithms are nothing without their calculations, memorizations, and inter-vertex communications. We investigate a class of ultra simple algorithms which can find (Delta+1)-colorings despite drastic restrictions. For each procedure, conflicted vertices randomly recolor one at a time until the graph coloring is valid. We provide an array of run time bounds for these processes, including an O(n*log(Delta)) bound for a variant we propose, and an O(n*Delta) bound which applies to even the most adversarial scenarios