[Paper Review] Determinantal Correlations for Classical Projection Processes

Peter J. Forrester, Taro Nagao|arXiv (Cornell University)|Jan 1, 2008

Random Matrices and Applications18 references15 citations

TL;DR

This paper establishes a unified framework for computing determinantal correlation functions in classical projection processes derived from random matrix ensembles (GUE, LUE, JUE), using Rodrigues' formula for classical orthogonal polynomials. The key result is that scaling limits at the soft edge reproduce the correlations of Dyson Brownian motion, confirming universality in eigenvalue statistics for these processes.

ABSTRACT

Recent applications in queuing theory and statistical mechanics have isolated the process formed by the eigenvalues of successive minors of the GUE. Analogous eigenvalue processes, formed in general from the eigenvalues of nested sequences of matrices resulting from random corank 1 projections of classical random matrix ensembles, are identified for the LUE and JUE. The correlations for all these processes can be computed in a unified way. The resulting expressions can then be analyzed in various scaling limits. At the soft edge, with the rank of the minors differing by an amount proportional to $N^{2/3}$, the scaled correlations coincide with those known from the soft edge scaling of the Dyson Brownian motion model.

Motivation & Objective

• To identify and characterize eigenvalue processes arising from successive random corank-1 projections in classical random matrix ensembles (GUE, LUE, JUE).
• To develop a unified analytical method for computing correlation functions across these processes using classical orthogonal polynomials.
• To analyze scaling limits of the correlation functions, particularly at the soft edge and bulk regimes.
• To establish universality by showing that soft edge scaling limits match those of Dyson Brownian motion.
• To connect the results to broader statistical mechanical models via the RSK correspondence and last-passage percolation.

Proposed method

• Utilizes the Rodrigues formula for classical orthogonal polynomials as the core analytical tool to derive joint eigenvalue probability density functions.
• Applies the RSK correspondence to relate the eigenvalue processes to last-passage percolation models with geometric and exponential weights.
• Derives determinantal correlation functions using the kernel formalism, with kernels expressed in terms of orthogonal polynomials and their derivatives.
• Performs asymptotic analysis of orthogonal polynomials to compute scaling limits, including soft edge and bulk limits.
• Uses the scaled kernel expressions to compare with known models such as Dyson Brownian motion and the soft edge sine kernel.
• Validates results through consistency checks in limiting cases (e.g., α → 0 for bulk, α → −∞ for soft edge), confirming agreement with established kernels.

Experimental results

Research questions

• RQ1 Can determinantal correlation functions be computed in a unified way across classical projection processes derived from GUE, LUE, and JUE?
• RQ2 What are the scaling limits of the correlation functions in these processes, particularly at the soft edge and bulk?
• RQ3 Do the soft edge scaling limits of the classical projection processes match those of Dyson Brownian motion?
• RQ4 How do the eigenvalue statistics of nested minor processes relate to last-passage percolation and the RSK correspondence?
• RQ5 What is the role of classical orthogonal polynomials in enabling a unified treatment of correlation functions in these processes?

Key findings

• The correlation functions for classical projection processes (GUE, LUE, JUE) can be computed exactly and uniformly using the Rodrigues formula for classical orthogonal polynomials.
• At the soft edge, with minor rank differences proportional to $ N^{2/3} $, the scaled correlation functions coincide with those of the soft edge scaling of Dyson Brownian motion.
• In the bulk scaling limit, the correlation kernel reduces to the sine kernel $ K^{ m bulk}(X,Y) = rac{ an rac{ heta}{2}}{ heta} $, matching the sine process.
• The soft edge kernel $ K^{ m soft} $ is recovered in the limit $ eta o -rac{1}{2} $, confirming consistency with known universal edge behavior.
• The analysis confirms that the joint eigenvalue distribution of the projection process matches the last-passage percolation model with exponential weights, validating the RSK-based construction.
• In the limit $ ho_i o c $, the joint PDF of eigenvalues across the sequence of matrices $ A_{(n)} $ matches the form derived from directed percolation, verifying the model's consistency.