An improved upper bound for the grid Ramsey problem

TL;DR

This paper improves the upper bounds for the grid Ramsey problem by employing a quasirandomness argument, advancing understanding of edge colourings in Cartesian products of complete graphs.

Contribution

It introduces a new upper bound for G(r) using quasirandomness, surpassing previous set system intersection bounds.

Findings

Established a new upper bound: G(r) ≤ (1 - 1/128 r^{-2}) r^{(r+1 choose 2)}

Demonstrated the effectiveness of quasirandomness methods in combinatorial bounds

Provided bounds for large r in the grid Ramsey problem.

Abstract

For a positive integer $r$, let $G(r)$ be the smallest $N$ such that, whenever the edges of the Cartesian product $K_N \times K_N$ are $r$-coloured, then there is a rectangle in which both pairs of opposite edges receive the same colour. In this paper, we improve the upper bounds on $G(r)$ by proving $G(r) \leq \Big(1 - \frac{1}{128}r^{-2}\Big) r^{\binom{r+1}{2}}$, for $r$ large enough. Unlike the previous improvements, which were based on bounds for the size of set systems with restricted intersection sizes, our proof is a form of a quasirandomness argument.