New Integer Linear Programming Models for the Vertex Coloring Problem

The vertex coloring problem asks for the minimum number of colors that can be assigned to the vertices of a given graph such that for all vertices v the color of v is different from the color of any of its neighbors. The problem is NP-hard. Here, we introduce new integer linear programming formulations based on partial orderings. They have the advantage that they are as simple to work with as the classical assignment formulation, since they can be fed directly into a standard integer linear programming solver. We evaluate our new models using Gurobi and show that our new simple approach is a good alternative to the best state-of-the-art approaches for the vertex coloring problem. In our computational experiments, we compare our formulations with the classical assignment formulation and the representatives formulation on a large set of benchmark graphs as well as randomly generated graphs of varying size and density. The evaluation shows that one of the new models dominates both formulations for sparse graphs, while the representatives formulation is the best for dense graphs. 1998 ACM Subject Classification G.1.6 Optimization: Integer Programming, G.2.2 Graph Theory: Graph algorithms, F.2.2 Nonnumerical Algorithms and Problems: Computations on discrete structures