[Paper Review] Minimal Gröbner bases and the predictable leading monomial property

Margreta Kuijper, Kristina Schindelar|arXiv (Cornell University)|Jun 25, 2009

Coding theory and cryptography29 references4 citations

TL;DR

This paper introduces the predictable leading monomial (PLM) property for minimal Gröbner bases in polynomial modules over rings, particularly ${\mathbb{Z}}_{p^r}$, where standard minimal Gröbner bases may lack this property. It proposes a construction of a minimal Gröbner $p$-basis from any minimal Gröbner basis that guarantees the $p$-PLM property, enabling efficient parametrization of solutions to interpolation and coding problems, such as list decoding of Reed-Solomon codes.

ABSTRACT

We focus on Gröbner bases for modules of univariate polynomial vectors over a ring. We identify a useful property, the "predictable leading monomial (PLM) property" that is shared by minimal Gröbner bases of modules in F[x]^q, no matter what positional term order is used. The PLM property is useful in a range of applications and can be seen as a strengthening of the wellknown predictable degree property (= row reducedness), a terminology introduced by Forney in the 70's. Because of the presence of zero divisors, minimal Gröbner bases over a finite ring of the type Z_p^r (where p is a prime integer and r is an integer >1) do not necessarily have the PLM property. In this paper we show how to derive, from an ordered minimal Gröbner basis, a so-called "minimal Gröbner p-basis" that does have a PLM property. We demonstrate that minimal Gröbner p-bases lend themselves particularly well to derive minimal realization parametrizations over Z_p^r. Applications are in coding and sequences over Z_p^r.

Motivation & Objective

• To identify and formalize the predictable leading monomial (PLM) property as a key structural feature of minimal Gröbner bases in polynomial modules over rings.
• To address the failure of minimal Gröbner bases over ${\mathbb{Z}}_{p^r}$ to satisfy the PLM property, which hinders their use in interpolation and coding applications.
• To construct a new type of basis, called the minimal Gröbner $p$-basis, from any minimal Gröbner basis that guarantees the $p$-PLM property, even when the original basis is not a module basis.
• To demonstrate the utility of the $p$-PLM property in deriving parametrized solutions for shortest linear recurrence relations over ${\mathbb{Z}}_{p^r}$, relevant to list decoding of Reed-Solomon codes.

Proposed method

• Define the PLM property as a strengthening of the predictable degree property, ensuring that the leading monomial of a polynomial combination is predictable from the leading monomials of the basis elements.
• Introduce the concept of a minimal Gröbner $p$-basis derived from a minimal Gröbner basis via a transformation that ensures the $p$-PLM property, even when the original basis is not a module basis.
• Use the TOP (Term Over Position) and POT (Position Over Term) monomial orders to define the structure of the polynomial modules and to ensure flexibility in basis construction.
• Apply the theory to construct parametrized families of solutions for linear recurrence relations by leveraging the $p$-PLM property to control the leading terms of combinations.
• Demonstrate the method on examples over ${\mathbb{Z}}_9$, showing how a minimal Gröbner $p$-basis yields a parametrization of all shortest linear recurrence relations.
• Leverage computational tools like Singular to compute minimal Gröbner bases and verify the $p$-PLM property in practice.

Experimental results

Research questions

• RQ1 Does the predictable leading monomial (PLM) property hold for minimal Gröbner bases over arbitrary rings, particularly ${\mathbb{Z}}_{p^r}$?
• RQ2 Can a minimal Gröbner basis over ${\mathbb{Z}}_{p^r}$ be transformed into a basis with the $p$-PLM property even when it is not a module basis?
• RQ3 How can the $p$-PLM property be used to parametrize solutions to interpolation problems, such as finding all shortest linear recurrence relations?
• RQ4 What advantages does the Gröbner approach offer over existing methods in list decoding of Reed-Solomon codes?
• RQ5 Can the $p$-PLM property be generalized to multivariate polynomial modules or used to derive canonical forms like Smith-McMillan for ${\mathbb{Z}}_{p^r}$?

Key findings

• The PLM property is a structural feature of minimal Gröbner bases over fields, but it does not necessarily hold for minimal Gröbner bases over rings like ${\mathbb{Z}}_{p^r}$, even for free modules.
• A minimal Gröbner $p$-basis can be constructed from any minimal Gröbner basis over ${\mathbb{Z}}_{p^r}$ such that it satisfies the $p$-PLM property, regardless of whether the original basis is a module basis.
• In Example 4.18 over ${\mathbb{Z}}_9$, the minimal TOP Gröbner $p$-basis $\{g_1, g_2\}$ with $g_1 = [x^3+4x^2+7x+4, x^2+3x]$ and $g_2 = [6x^2+8, x^3+5x^2+6x]$ yields a parametrization of all monic shortest linear recurrence relations as $x^3+4x^2+7x+4 + \Theta(6x^2+8)$ for $\Theta \in \mathbb{Z}_9$.
• For $\Theta = 3$, this parametrization recovers the shortest linear recurrence $x^3+4x^2+7x+1$ computed by the algorithm in [22], demonstrating the method's correctness and utility.
• The $p$-PLM property ensures that the leading monomial of any combination is predictable from the leading monomials of the basis elements, enabling systematic construction of solution families.
• The method provides a flexible framework for list decoding of Reed-Solomon codes by parametrizing all possible shortest linear recurrence relations from a single minimal Gröbner $p$-basis.