Small eigenvalues of large Hankel matrices: The indeterminate case

Christian Berg, Yang Chen, Mourad E. H. Ismail

TL;DR

This paper investigates the behavior of small eigenvalues of large Hankel matrices, specifically in the "indeterminate" case of moment problems. The authors establish a criterion for determinacy based on these eigenvalues and provide explicit lower bounds, offering a new perspective that relies solely on the original moment sequence.

What Problem Does This Paper Address?

The paper delves into the theory of moment problems, which are fundamental in various areas of mathematics and physics. A moment problem involves reconstructing a measure from its moments (integrals of powers of the variable with respect to the measure). A key distinction is between "determinate" and "indeterminate" moment problems. In a determinate problem, the moments uniquely define the measure. In an indeterminate problem, multiple measures can share the same moments.

This paper focuses on the "indeterminate case" and its connection to the eigenvalues of large Hankel matrices. Hankel matrices are constructed from the moments of the measure. For a positive definite Hankel matrix, its eigenvalues are positive. The behavior of the smallest eigenvalue of these matrices as their size grows is a crucial indicator of whether the moment problem is determinate or indeterminate. Previous work primarily focused on determinate cases, showing that the smallest eigenvalue tends to zero. This paper aims to characterize the indeterminate case by finding a condition related to these smallest eigenvalues and to derive explicit bounds for them.

Methods

The core of the paper's contribution lies in establishing a new criterion for the determinacy of a moment problem. The authors introduce Theorem 1.1, which states that a moment problem is determinate if and only if the smallest eigenvalue of the $N \times N$ Hankel matrix, denoted $\lambda_N$, tends to zero as $N$ approaches infinity. This is a significant result as it directly links the spectral properties of Hankel matrices to the determinacy of the moment problem, using only the original moment sequence.

They also derive an explicit lower bound for these smallest eigenvalues in the indeterminate case (Theorem 1.2). This bound is expressed in terms of the harmonic mean of a specific function related to the orthonormal polynomials associated with the moment problem, evaluated along the unit circle.

The paper then proceeds to explore several examples of indeterminate moment problems, calculating or estimating this lower bound for specific classes of orthogonal polynomials, such as Stieltjes-Wigert, Al-Salam-Carlitz, and q-Hermite polynomials. These calculations involve advanced techniques from the theory of q-special functions and hypergeometric series.

Datasets

Not discussed in the paper. The paper deals with theoretical mathematical concepts and calculations based on defined mathematical objects (moment sequences, orthogonal polynomials) rather than empirical data.

Key Insights and Ideas

  • Eigenvalue Criterion for Determinacy: The most significant insight is the direct link established between the asymptotic behavior of the smallest eigenvalue of Hankel matrices and the determinacy of the moment problem. Specifically, $\lambda_N \to 0$ if and only if the problem is determinate. This provides a powerful spectral tool for analyzing moment problems.
  • Explicit Lower Bound for Indeterminate Cases: The paper provides a concrete, calculable lower bound for the smallest eigenvalue in the indeterminate case. This bound is given by the harmonic mean of a function related to orthonormal polynomials on the unit circle, offering a way to quantify the "degree" of indeterminacy.
  • Connection to Hamburger's Theorem: The authors show how their result relates to and potentially simplifies existing criteria like Hamburger's theorem, which involves shifted moment sequences. Their criterion uses only the original moments, making it more direct.
  • Richness of Indeterminate Problems: The detailed examples with q-special functions highlight the complexity and variety of indeterminate moment problems and the sophisticated mathematical machinery required to analyze them.

Key Results

  • Theorem 1.1: A moment problem is determinate if and only if $\lim_{N \to \infty} \lambda_N = 0$, where $\lambda_N$ is the smallest eigenvalue of the $N \times N$ Hankel matrix.
  • Theorem 1.2: In the indeterminate case, the smallest eigenvalue $\lambda_N$ is bounded below by the harmonic mean of a function related to orthonormal polynomials on the unit circle.
  • The paper provides explicit calculations and estimations of these lower bounds for several families of orthogonal polynomials, demonstrating the practical application of the theoretical results. For instance, in Example 1 (Stieltjes-Wigert polynomials), they compare a numerically computed lower bound with an extrapolated value of the smallest eigenvalue, showing a percentage error.

Limitations and Future Work

  • Computational Complexity: While explicit formulas are derived, the calculations for the examples, especially those involving q-special functions, are highly technical and computationally intensive.
  • Generalization: The paper focuses on specific types of moment problems and orthogonal polynomials. Further work could explore the applicability of these results to a broader range of problems.
  • Nature of the Lower Bound: The paper provides a lower bound, but further investigation into the precise asymptotic behavior of $\lambda_N$ in the indeterminate case could be a direction for future research.
  • Numerical vs. Analytical: The paper combines analytical derivations with numerical computations and extrapolations for specific examples. Expanding the analytical treatment for more cases would be valuable.

Summary updated

Abstract

In this paper we characterize the indeterminate case by the eigenvalues of the Hankel matrices being bounded below by a strictly positive constant. An explicit lower bound is given in terms of the orthonormal polynomials and we find expressions for this lower bound in a number of indeterminate moment problems.
BibTeX
@inproceedings{berg2002small,
  author = {Christian Berg and Yang Chen and Mourad E. H. Ismail},
  title = {{Small eigenvalues of large Hankel matrices: The indeterminate case}},
  booktitle = {MATHEMATICA},
  year = {2002},
  doi = {10.7146/math.scand.a-14379},
}

Loading related content...

Figure