Extending the Affirmative Action Problem: mixing numbers and integrated colorings of graphs

Abstract

Consider a graph whose vertices are colored in one of two colors, say black or white. A white vertex is called integrated if it has at least as many black neighbors as white neighbors, and similarly for a black vertex. The coloring as a whole is integrated if every vertex is integrated. A classic exercise in graph theory, known as the Affirmative Action Problem, is to prove that every finite simple graph admits an integrated coloring. The solution can be neatly summarized with the one-liner: "maximize the number of balanced edges," that is, the edges that connect neighbors of different colors. However, not all integrated colorings advertise the maximum possible number of balanced edges. In this paper, we characterize and enumerate the integrated colorings for complete graphs, bicliques, paths, and cycles. We also derive the distributions and extremal values for the mixing numbers (the number of balanced edges) across all integrated colorings over these families of graphs. For paths and cycles in particular, we use the quasi-powers framework for probability generating functions to prove a central limit theorem for the mixing number of random integrated colorings. Lastly, we obtain an upper bound for the number of integrated colorings over any fixed graph via the second-moment method. A specialized bound for the number of integrated colorings over a regular graph is also obtained.