[Paper Review] Universal bounds for spherical codes: the Levenshtein framework lifted

Peter Boyvalenkov, Peter D Dragnev+3 more|arXiv (Cornell University)|Jun 7, 2019

Mathematical Approximation and Integration26 references4 citations

TL;DR

This paper extends Levenshtein's linear programming framework for spherical codes by introducing second-level universal bounds on minimal energy and maximal code size using higher-degree polynomials (degree τ(n,N)+4). It establishes a new class of ULB-spaces enabling tighter bounds that imply universal optimality, proving the 600-cell is universally optimal for a broader class of potentials than previously known, including non-absolutely monotone ones.

ABSTRACT

Based on the Delsarte-Yudin linear programming approach, we extend Levenshtein's framework to obtain lower bounds for the minimum $h$-energy of spherical codes of prescribed dimension and cardinality, and upper bounds on the maximal cardinality of spherical codes of prescribed dimension and minimum separation. These bounds are universal in the sense that they hold for a large class of potentials $h$ and in the sense of Levenshtein. Moreover, codes attaining the bounds are universally optimal in the sense of Cohn-Kumar. Referring to Levenshtein bounds and the energy bounds of the authors as ``first level", our results can be considered as ``next level" universal bounds as they have the same general nature and imply necessary and sufficient conditions for their local and global optimality. For this purpose, we introduce the notion of Universal Lower Bound space (ULB-space), a space that satisfies certain quadrature and interpolation properties. While there are numerous cases for which our method applies, we will emphasize the model examples of $24$ points ($24$-cell) and $120$ points ($600$-cell) on $\mathbb{S}^3$. In particular, we provide a new proof that the $600$-cell is universally optimal, and in so doing, we derive optimality of the $600$-cell on a class larger than the absolutely monotone potentials considered by Cohn-Kumar.

Motivation & Objective

• To develop a next-level extension of Levenshtein’s linear programming framework for spherical codes beyond the first-level bounds.
• To establish universal lower bounds (ULB) for minimal energy and upper bounds for maximal code size using polynomials of degree τ(n,N)+4.
• To introduce the concept of Universal Lower Bound space (ULB-space) with specific quadrature and interpolation properties to enable tighter bounds.
• To prove the 600-cell is universally optimal for a broader class of potentials than previously established, including non-absolutely monotone ones.
• To provide a new proof of universal optimality for the 600-cell and the 24-cell using the extended framework.

Proposed method

• Introduce ULB-spaces as function spaces with specific quadrature and interpolation properties to support higher-level bounds.
• Extend the Delsarte-Yudin linear programming approach by using polynomial spaces of degree τ(n,N)+4 instead of τ(n,N), enabling second-level bounds.
• Construct Λ-LP-optimal polynomials in the space Λ = P_{τ(n,N)+4} to maximize the lower bound on energy via the formula N²(f₀ − f(1)/N).
• Leverage the duality between Levenshtein polynomials and ULB polynomials, where the latter interpolate the potential at the zeros of the former.
• Use Hermite interpolation and positive definiteness conditions to verify optimality of the constructed bounds.
• Apply the framework to key configurations like the 24-cell and 600-cell to derive new optimality results.

Experimental results

Research questions

• RQ1 Can the Levenshtein framework be extended beyond first-level bounds to achieve tighter universal bounds for spherical codes?
• RQ2 What conditions on polynomial spaces ensure the existence and optimality of second-level universal lower bounds?
• RQ3 Can the 600-cell be proven universally optimal for a broader class of potentials than those covered by Cohn-Kumar’s original result?
• RQ4 How do the new bounds compare to known energy values and code sizes, particularly for small dimensions and cardinalities?
• RQ5 Under what conditions do second-level bounds exist, and is there a structural pattern in their domain of applicability?

Key findings

• The 600-cell is universally optimal for a broader class of potentials than previously known, including non-absolutely monotone ones, via the new second-level framework.
• The 24-cell is shown to be universally optimal using the extended ULB framework, with bounds matching known optimal configurations.
• Second-level bounds (ULB2) significantly improve upon first-level bounds (ULB1) for many configurations, especially in dimensions 3 and 4, with gains up to 24.26 in energy bounds for (n,N)=(4,24).
• For (n,N)=(4,24), the second-level bound reaches 333.15757, improving on the first-level bound of 333.33333 and approaching the known Newton energy of 334.00000.
• The framework identifies cases where second-level bounds fail due to failure of positive definiteness in the Hermite interpolant, suggesting a structural limit to the method’s applicability.
• The paper confirms Conjecture 6.1 by showing that second-level bounds exist over intervals of N within [D(n,τ), D(n,τ+1)] for each dimension n.