[Paper Review] Entropy Compression Method and Legitimate Colorings in Projective Planes

Rodrigo Bissacot, Luís Doin|arXiv (Cornell University)|Oct 19, 2017

Limits and Structures in Graph Theory12 references4 citations

TL;DR

This paper extends the entropy compression method to all problems formulated in the variable version of the Lovász Local Lemma, removing prior constraints on configuration extendability. It applies this to projective planes, proving that any finite projective plane admits a legitimate coloring with 42 colors, and establishes existence of 8-colorings for planes of order greater than $10^{54}$, improving prior bounds.

ABSTRACT

We prove that the entropy compression method systematized by L. Esperet and A. Parreau [11] can be applied to any problem formulated in the variable version of the Lovász Local Lemma. As an application, we prove the existence of legitimate colorings for projective planes with small orders, which extends results of N. Alon and Z. Füredi [2]. In fact, we allow different numbers of colors, proving that projective planes of any order can be legitimate colored with 42 colors.

Motivation & Objective

• To generalize the entropy compression method beyond its original constraints to apply to all problems in the variable version of the Lovász Local Lemma.
• To eliminate the need for fixed vertices to uniquely determine color extensions in forbidden configurations.
• To apply the refined method to the problem of legitimate colorings in finite projective planes.
• To improve existing bounds on the order of projective planes admitting legitimate colorings with a fixed number of colors.
• To demonstrate that 42 colors suffice for a legitimate coloring of any finite projective plane, regardless of order.

Proposed method

• Adapts the entropy compression algorithm by removing the requirement that a fixed set of vertices uniquely determines a forbidden coloring configuration.
• Reformulates the problem as a graph coloring task where certain configurations are forbidden, using the variable version of the Lovász Local Lemma.
• Introduces a modified dependency structure where the number of ways to extend partial colorings is bounded by $m_i = \underline{m}!$, with $\underline{m}$ chosen to minimize a key expression.
• Uses the generating function $\phi_E(\tau) = 1 + \tau^{\underline{m}}$ and solves $\phi_E(\tau) - \tau \phi_E'(\tau) = 0$ to derive convergence conditions for the algorithm.
• Applies the method to projective planes by modeling lines and points as hyperedges and vertices, defining 'dangerous pairs' of lines and their probabilities.
• Employs probabilistic bounds on the number of dangerous pairs and points involved in such pairs, using combinatorial estimates involving $\mathsf{d}(n,a,b)$, $K$, and binomial coefficients.

Experimental results

Research questions

• RQ1 Can the entropy compression method be generalized to all problems in the variable version of the Lovász Local Lemma, without requiring fixed vertices to determine color extensions?
• RQ2 What is the minimal number of colors required to guarantee a legitimate coloring of any finite projective plane?
• RQ3 For which orders of projective planes does an 8-color legitimate coloring exist, and how does this compare to previous bounds?
• RQ4 Can the entropy compression method be used to close the gap between small and large-order projective planes in terms of known existence results for legitimate colorings?
• RQ5 Is the improved bound on the order of projective planes admitting 8-color legitimate colorings an artifact of the method or an intrinsic feature of the problem?

Key findings

• The entropy compression method can be applied to any problem in the variable version of the Lovász Local Lemma, without requiring the existence of fixed vertices that uniquely determine a coloring extension.
• Any finite projective plane admits a legitimate coloring with 42 colors, regardless of its order.
• An 8-color legitimate coloring exists for all projective planes of order greater than $10^{54}$, improving upon the prior bound of $10^{250}$ from Alon and Füredi.
• The method provides a constructive existence proof via entropy compression, extending the applicability of the Moser-Tardos algorithm to broader classes of combinatorial problems.
• The bound on the order of projective planes admitting an 8-coloring is derived from an optimization involving $\min_m \frac{m}{m-1}(m! \cdot a \cdot b(m-1))^{1/m}$, with $a$ and $b$ controlling dangerous pair counts.
• The analysis shows that for large $n$, the minimal number of colors $\mathsf{d}$ required for a partial coloring to avoid too many dangerous pairs can be bounded using combinatorial estimates involving $K$, binomial coefficients, and logarithmic terms.