[Paper Review] Polynomial method in coding and information theory

Alexei Ashikhmin, Alexander Barg+1 more|ArXiv.org|Oct 31, 1999

Coding theory and cryptography25 references3 citations

TL;DR

This paper generalizes the polynomial (Delsarte) method in coding and information theory to derive new asymptotic bounds on code performance over binary-input memoryless and Gaussian channels, improving upon classical results from Shannon (1959–1967). The framework unifies prior structural results and yields tighter bounds on code size and error rates using orthogonal polynomials and duality in metric spaces.

ABSTRACT

Polynomial, or Delsarte's, method in coding theory accounts for a variety of structural results on, and bounds on the size of, extremal configurations (codes and designs) in various metric spaces. In recent works of the authors the applicability of the method was extended to cover a wider range of problems in coding and information theory. In this paper we present a general framework for the method which includes previous results as particular cases. We explain how this generalization leads to new asymptotic bounds on the performance of codes in binary-input memoryless channels and the Gaussian channel, which improve the results of Shannon et al. of 1959-67, and to a number of other results in combinatorial coding theory.

Motivation & Objective

• To extend the applicability of the polynomial (Delsarte) method beyond classical coding theory to broader problems in information theory.
• To derive tighter asymptotic bounds on the performance of codes in binary-input memoryless and Gaussian channels.
• To unify and generalize prior structural results on extremal codes and designs in metric spaces.
• To establish new quantitative limits on code size and error rates using polynomial inequalities and duality principles.
• To provide a systematic framework that subsumes earlier results and enables new combinatorial bounds in coding theory.

Proposed method

• Applies the Delsarte polynomial method to metric spaces, using orthogonal polynomials to bound code sizes via duality.
• Employs linear programming bounds based on the Krawtchouk and Hermite polynomials for binary and Gaussian channels, respectively.
• Introduces a generalized framework that incorporates symmetry and duality in metric spaces to analyze extremal configurations.
• Uses the duality between the weight distribution of a code and the distance distribution of its dual to derive bounds.
• Applies the method to derive bounds on the maximal size of codes with given minimum distance and rate in various channels.
• Leverages properties of orthogonal polynomials and integral transforms to analyze extremal configurations in combinatorial metric spaces.

Experimental results

Research questions

• RQ1 Can the polynomial method be extended to yield improved asymptotic bounds on code performance in memoryless and Gaussian channels?
• RQ2 How can the Delsarte method be generalized to unify existing results on extremal codes and designs in metric spaces?
• RQ3 What new structural constraints or bounds emerge when applying the polynomial method to binary-input memoryless channels?
• RQ4 To what extent do the new bounds surpass the classical results of Shannon from the late 1950s and 1960s?
• RQ5 What role do orthogonal polynomials and duality play in deriving tighter bounds on code size and error probability?

Key findings

• The generalized polynomial method yields new asymptotic bounds on code performance in binary-input memoryless and Gaussian channels that improve upon the classical bounds of Shannon (1959–1967).
• The framework unifies and extends previous results on extremal codes and designs in metric spaces through a common duality principle.
• New bounds on the maximal size of codes with given minimum distance are derived using Krawtchouk and Hermite polynomials in the respective channel models.
• The method provides tighter upper bounds on the rate of codes for a given minimum distance, particularly in the high-rate and low-rate regimes.
• The approach demonstrates that the polynomial method can be systematically applied to derive non-trivial bounds in combinatorial coding theory beyond traditional linear programming.
• The results show that the use of orthogonal polynomials and duality leads to improved estimates of code capacity and error exponents.