[Paper Review] Global Asymptotics for the Christoffel-Darboux Kernel of Random Matrix Theory

Thomas Kriecherbauer, Schubert, Kristina+2 more|arXiv (Cornell University)|Jan 27, 2014

Random Matrices and Applications50 references8 citations

TL;DR

This paper establishes global asymptotics for the Christoffel-Darboux kernel in unitary random matrix ensembles using the nonlinear steepest descent method, extending prior local results to provide uniform, global asymptotics under perturbations of the potential. The key contribution is a uniform leading-order asymptotic expansion valid across the entire spectrum, enabling analysis of transitions from universal to non-universal eigenvalue statistics, including moderate deviations of the largest eigenvalues.

ABSTRACT

The investigation of universality questions for local eigenvalue statistics continues to be a driving force in the theory of Random Matrices. For Matrix Models [53] the method of orthogonal polynomials can be used and the asymptotics of the Christoffel-Darboux kernel [59] become the key for studying universality. In this paper the existing results on the CD-kernel will be extended in two directions. Firstly, in order to analyze the transition from the universal to the non-universal regime, we provide leading order asymptotics that are global rather than local. This allows e.g. to describe the moderate deviations for the largest eigenvalues of unitary ensembles ($β$ = 2), where such a transition occurs. Secondly, our asymptotics will be uniform under perturbations of the probability measure that defines the matrix ensemble. Such information is useful for the analysis of a different type of ensembles [25], which is not known to be determinantal and for which the method of orthogonal polynomials cannot be used directly. The just described applications of our results are formulated in this paper but will be proved elsewhere. As a byproduct of our analysis we derive first order corrections for the 1-point correlation functions of unitary ensembles in the bulk. Our proofs are based on the nonlinear steepest descent method [20]. They follow closely [17] and incorporate improvements introduced in [36, 64]. The presentation is self-contained except for a number of general facts from Random Matrix theory and from the theory of singular integral operators.

Motivation & Objective

• To extend existing local asymptotics of the Christoffel-Darboux kernel to global asymptotics valid across the entire spectrum of unitary random matrix ensembles.
• To derive uniform asymptotics under perturbations of the underlying probability measure, enabling applications to non-determinantal ensembles where orthogonal polynomial methods fail.
• To provide a rigorous framework for analyzing the transition regime between universal and non-universal local eigenvalue statistics, particularly for moderate deviations of the largest eigenvalues.
• To derive first-order corrections to the 1-point correlation functions in the bulk of the spectrum for unitary ensembles.

Proposed method

• Application of the nonlinear steepest descent method to the Riemann-Hilbert problem formulation of orthogonal polynomials associated with the Christoffel-Darboux kernel.
• Use of a Deift-Zhou-type asymptotic analysis with explicit parametrix constructions near spectral edges and bulk regions, incorporating improvements from [36, 64].
• Construction of a global parametrix via a model Riemann-Hilbert problem with a complex phase function and a singular integral operator framework.
• Introduction of a local parametrix model near the soft edge using the Airy function and its derivatives, with uniform error bounds in the parameter space.
• Derivation of error bounds for the jump matrices in the Riemann-Hilbert problem via uniform estimates on the transition regions, ensuring $ \mathcal{O}(N^{-1}) $ convergence.
• Use of a partition of unity and local coordinate transformations to glue parametrix solutions across different spectral regions, maintaining analyticity and uniformity.

Experimental results

Research questions

• RQ1 How can global asymptotics for the Christoffel-Darboux kernel be derived beyond the local bulk and edge scaling limits?
• RQ2 What is the uniform behavior of the kernel under small perturbations of the potential measure in unitary ensembles?
• RQ3 Can the transition from universal to non-universal eigenvalue statistics be described using global asymptotics?
• RQ4 What are the first-order corrections to the 1-point correlation function in the bulk of the spectrum for unitary ensembles?
• RQ5 How can the nonlinear steepest descent method be adapted to yield uniform error bounds across the entire spectrum and under measure perturbations?

Key findings

• The Christoffel-Darboux kernel admits a leading-order global asymptotic expansion valid uniformly across the entire spectrum of the unitary ensemble, extending beyond local scaling limits.
• The asymptotics are uniformly valid under perturbations of the potential measure, providing robustness for applications to non-determinantal ensembles.
• The method yields $ \mathcal{O}(N^{-1}) $ error bounds for the jump matrices in the Riemann-Hilbert problem, uniform in the spectral parameter and in a neighborhood of the reference potential.
• First-order corrections to the 1-point correlation function in the bulk are derived explicitly, improving upon leading-order Wigner semicircle approximations.
• The analysis enables the description of moderate deviations for the largest eigenvalues in unitary ensembles, capturing the transition from universal to non-universal behavior.
• The global parametrix construction ensures analytic continuation of the solution across the entire complex plane, with uniform error control in compact subsets of the parameter space.