A Faster Algorithm for the 4-Coloring Problem

We explore the 4-coloring problem, a fundamental combinatorial NP-hard problem. Given a graph G, the 4-coloring problem asks whether there exists a function f from the vertex set of G to {1,2,3,4} such that f(u)≠ f(v) for each edge uv of G. Such function f is referred to as a 4-coloring of G. The fastest known algorithm for the 4-coloring problem, introduced by Fomin, Gaspers, and Saurabh (COCOON 2007), exhibits a time complexity of O(1.7272ⁿ) and exponential space. In this paper, we propose an enhanced algorithm for the 4-coloring problem with a time complexity of O(1.7159ⁿ) and polynomial space. Our algorithm is deterministic and built upon a novel method. Specifically, inspired by previous algorithmic approaches for the 4-coloring problem, such as the aforementioned O(1.7272ⁿ) time algorithm, we consider the instance (G,I,S), where G is a graph and I,S are subsets of its vertex set representing vertices colored with 1 and vertices unable to be colored with 1, respectively. For a given instance (G,I,S), we aim to determine the existence of a 4-coloring f of G such that f(v) = 1 for v ∈ I and f(v)≠ 1 for v ∈ S. Our key innovation lies in recognizing that, leveraging certain combinatorial properties, the instance (G,I,S) can be efficiently solved when G-I-S is a union of K₃’s and K₄’s (where K₃ and K₄ denote complete graphs with 3 and 4 vertices, respectively). The ability to efficiently solve instances (G,I,S), where G-I-S is comprised solely of K₃’s and K₄’s, enables us to devise a branching algorithm capable of efficiently addressing instances (G,I,S), where G-I-S is not a union of K₃’s and K₄’s (the other case). Based on this innovative method, we derive our final enhanced algorithm.