[Paper Review] Inequivalent Representations of Matroids over Prime Fields

Jim Geelen, Geoff Whittle|arXiv (Cornell University)|Jan 24, 2011

Cooperative Communication and Network Coding18 references4 citations

TL;DR

This paper establishes that 4-connected matroids over any prime field GF(p) have at most a finite, bounded number of inequivalent representations—specifically, an integer f(p) depending only on p. The key result is a structural theorem on k-coherent matroids, which generalizes 3- and 4-connectivity and enables control over representation diversity. The proof relies on deep structural decompositions involving flowers, k-skeletons, and paths of separations, culminating in bounds on inequivalent representations and applications to certifying non-representability with O(n²) rank evaluations.

ABSTRACT

It is proved that for each prime field $GF(p)$, there is an integer $f(p)$ such that a 4-connected matroid has at most $f(p)$ inequivalent representations over $GF(p)$. We also prove a stronger theorem that obtains the same conclusion for matroids satisfying a connectivity condition, intermediate between 3-connectivity and 4-connectivity that we term "$k$-coherence". We obtain a variety of other results on inequivalent representations including the following curious one. For a prime power $q$, let ${\mathcal R}(q)$ denote the set of matroids representable over all fields with at least $q$ elements. Then there are infinitely many Mersenne primes if and only if, for each prime power $q$, there is an integer $m_q$ such that a 3-connected member of ${\mathcal R}(q)$ has at most $m_q$ inequivalent GF(7)-representations. The theorems on inequivalent representations of matroids are consequences of structural results that do not rely on representability. The bulk of this paper is devoted to proving such results.

Motivation & Objective

• To establish a uniform bound on the number of inequivalent representations of 4-connected matroids over any prime field GF(p).
• To extend this bound to a weaker connectivity condition—k-coherence—intermediate between 3- and 4-connectivity.
• To develop structural tools, including flowers, k-skeletons, and paths of separations, that control representation diversity independently of field representability.
• To apply these structural results to certify non-representability over GF(p) using only O(n²) rank evaluations.
• To explore a deep number-theoretic connection: the infinitude of Mersenne primes is equivalent to a uniform bound on GF(7)-representations of 3-connected matroids in R(q), the class representable over all fields of size ≥ q.

Proposed method

• Introduces the concept of k-coherence as a connectivity condition that ensures structural control and bounds on representation diversity.
• Develops the theory of flowers and their structural properties, particularly tight and equivalent flowers, to analyze 3-separations and modular structure.
• Defines k-skeletons as a framework to track fixed and free elements, using notions of freedom, cofreedom, and gangs of three to analyze element behavior under deletion/contraction.
• Applies a chain theorem for k-skeletons to decompose matroids into simpler components and control the growth of inequivalent representations.
• Uses paths of 3- and 2-separations, especially sequentially and simply-bridged paths, to analyze how representations can be extended or blocked.
• Applies the structural results to prove that near k-coherent matroids have bounded representation diversity over GF(p), enabling efficient non-representability certification.

Experimental results

Research questions

• RQ1 Does every 4-connected matroid over a prime field GF(p) have only finitely many inequivalent representations, and can this number be uniformly bounded by a function f(p)?
• RQ2 Can this bound be extended to a weaker connectivity condition than 4-connectivity, such as k-coherence, to capture more matroids with bounded representation diversity?
• RQ3 What is the connection between the infinitude of Mersenne primes and the boundedness of inequivalent GF(7)-representations in the class R(q) of matroids representable over all fields of size at least q?
• RQ4 Can structural decompositions of matroids—such as flowers, k-skeletons, and paths of separations—be used to control representation diversity without relying on field-specific properties?
• RQ5 Can the structural results be leveraged to certify non-representability over GF(p) using only O(n²) rank evaluations, as required in algorithmic matroid theory?

Key findings

• For each prime p, there exists an integer f(p) such that any 4-connected matroid has at most f(p) inequivalent representations over GF(p), resolving a long-standing question in matroid representability.
• The bound extends to k-coherent matroids, a connectivity condition weaker than 4-connectivity but stronger than 3-connectivity, ensuring bounded representation diversity under this framework.
• A matroid M has at most µp inequivalent representations over GF(p) if it is near k-coherent for k ≥ 5, where µp depends only on p, enabling efficient non-representability certification.
• The infinitude of Mersenne primes is equivalent to the existence of a uniform bound mq on the number of inequivalent GF(7)-representations for 3-connected members of R(q), the class of matroids representable over all fields of size at least q.
• The structural theory of k-skeletons, including gangs of three and bogan couples, provides a mechanism to control element behavior and representation growth in decompositions.
• The results imply that non-representability over GF(p) can be certified using only O(n²) rank evaluations, a significant improvement over exponential methods.