[Paper Review] Noncolliding processes, matrix-valued processes and determinantal processes

Makoto Katori, Hideki Tanemura|arXiv (Cornell University)|May 4, 2010

Random Matrices and Applications57 references18 citations

TL;DR

This paper introduces generalized noncolliding diffusion processes, including temporally inhomogeneous and matrix-valued systems, and establishes their connection to determinantal and Pfaffian point processes. By extending Dyson's Brownian motion model and applying generalized Bru's theorem, the authors prove that multi-time correlation functions are determinantal or Pfaffian, and derive asymptotic laws in the N→∞ limit, linking them to Tracy-Widom distributions and random matrix theory.

ABSTRACT

A noncolliding diffusion process is a conditional process of $N$ independent one-dimensional diffusion processes such that the particles never collide with each other. This process realizes an interacting particle system with long-ranged strong repulsive forces acting between any pair of particles. When the individual diffusion process is a one-dimensional Brownian motion, the noncolliding process is equivalent in distribution with the eigenvalue process of an $N imes N$ Hermitian-matrix-valued process, which we call Dyson's model. For any deterministic initial configuration of $N$ particles, distribution of particle positions of the noncolliding Brownian motion on the real line at any fixed time $t >0$ is a determinantal point process. We can prove that the process is determinantal in the sense that the multi-time correlation function for any chosen series of times, which determines joint distributions at these times, is also represented by a determinant. We study the asymptotic behavior of the system, when the number of Brownian motions $N$ in the system tends to infinity. This problem is concerned with the random matrix theory on the asymptotics of eigenvalue distributions, when the matrix size becomes infinity. In the present paper, we introduce a variety of noncolliding diffusion processes by generalizing the noncolliding Brownian motion, some of which are temporally inhomogeneous. We report the results of our research project to construct and study finite and infinite particle systems with long-ranged strong interactions realized by noncolliding processes.

Motivation & Objective

• To generalize noncolliding Brownian motion to include temporally inhomogeneous and matrix-valued diffusion processes.
• To establish that the multi-time correlation functions of these processes are determinantal or Pfaffian, extending the framework of determinantal point processes.
• To analyze the asymptotic behavior of the system as the number of particles N approaches infinity, linking it to random matrix theory.
• To unify various known processes—such as Dyson’s model, Bessel processes, and generalized meanders—under a common mathematical framework using Fredholm determinants and Pfaffians.
• To demonstrate the existence of infinite-dimensional Pfaffian processes via scaling limits of finite Pfaffian systems.

Proposed method

• Use the Karlin-McGregor formula to express the transition density of noncolliding diffusion processes as a determinant of one-dimensional transition densities.
• Apply a generalized version of Bru’s theorem to derive stochastic differential equations for eigenvalue processes of matrix-valued diffusion processes.
• Characterize the correlation structure of noncolliding processes using Fredholm determinants and Pfaffians, generalizing determinantal point processes.
• Utilize Harish-Chandra/Itzykson-Zuber integral formulas and Riemann-Liouville differintegrals to describe correlation kernels in generalized meander processes.
• Analyze the N→∞ limit of Pfaffian and determinantal correlation functions to derive universal asymptotic laws, including Tracy-Widom distributions.
• Connect the resulting infinite particle systems to classical random matrix ensembles (GUE, GOE, GSE) and their β-ensembles via the Dyson model with β=1,2,4.

Experimental results

Research questions

• RQ1 How can noncolliding diffusion processes be generalized beyond standard Brownian motion to include inhomogeneous and matrix-valued dynamics?
• RQ2 Under what conditions are the multi-time correlation functions of noncolliding processes determinantal or Pfaffian?
• RQ3 What is the asymptotic behavior of these processes as the number of particles N tends to infinity?
• RQ4 How do the correlation kernels of generalized meanders and noncolliding bridges relate to special functions like Riemann-Liouville differintegrals?
• RQ5 What is the connection between the eigenvalue processes of random matrix ensembles and noncolliding diffusion processes in the limit N→∞?

Key findings

• The noncolliding Brownian motion is equivalent in law to the eigenvalue process of an N×N Hermitian matrix-valued diffusion, known as Dyson’s model with β=2.
• For any fixed time t>0, the particle position distribution of noncolliding Brownian motion is a determinantal point process with a correlation kernel derived from the Karlin-McGregor formula.
• The multi-time joint distributions of noncolliding processes are determinantal, with correlation functions represented by Fredholm determinants of N×N matrices.
• The temporally inhomogeneous noncolliding Brownian motion and noncolliding generalized meanders are Pfaffian processes, with correlation functions expressible via Pfaffians.
• In the N→∞ limit, the asymptotic analysis of Pfaffian and determinantal structures yields universal laws, including the Tracy-Widom distribution for the largest eigenvalue.
• The generalized meander process has a correlation kernel expressed using Riemann-Liouville differintegrals, extending the class of solvable noncolliding systems.