The paper solves a specific open problem in graph theory about how densely connected a certain type of graph can be while still avoiding a particular forbidden pattern. The setting involves "balanced multipartite graphs," which are networks where the nodes are divided into equal-sized groups, with edges only allowed between different groups. The question is: how high can the minimum number of connections per node get, if you also require that the graph contains no "complete subgraph on four nodes" (a K4, meaning four nodes all mutually connected)? The function f(n,r,t+1) measures this maximum possible minimum degree for various choices of the number of groups r and the forbidden clique size.
Prior work by Lo, Treglown, and Zhao had resolved nearly all cases of this problem for the K4-free setting across different numbers of parts, leaving exactly one case open: seven groups with no K4 allowed. This paper closes that gap. The author tackles it in two steps. First, they figure out the densest such graphs that can be properly colored with just three colors, reducing that question to a concrete optimization problem involving a small grid of integers, which they solve exactly. Three-colorability is a natural structural property here because a graph split into three color classes is automatically K4-free in a useful sense.
Second, the author proves that any balanced seven-part K4-free graph where every node has more than a certain threshold number of connections must itself be three-colorable. This structural insight bridges the two parts of the argument and pins down the answer precisely. The final result is that the maximum achievable minimum degree is the floor of 30n/7, with one small exception: when the group size n leaves a remainder of 4 when divided by 7 and is at least 11, the answer is exactly one less. This clean formula with a single exceptional family completes the classification of the K4-free multipartite problem.