Litcius/Paper detail

Planar orthogonal polynomials and boundary universality in the random normal matrix model

Håakan Hedenmalm, Aron Wennman

2021Acta Mathematica51 citationsDOIOpen Access PDF

Abstract

We show that the planar normalized orthogonal polynomials $P_{m,n}(z)$ of degree $n$ with respect to an exponentially varying planar measure $\mathrm{e}^{-2mQ}\mathrm{dA}$ enjoy an asymptotic expansion \[ P_{m,n}(z)\sim m^{\frac{1}{4}}\sqrt{\phi_\tau'(z)}[\phi_\tau(z)]^n \mathrm{e}^{m\mathcal{Q}_\tau(z)}\left(\mathcal{B}_{\tau, 0}(z) +m^{-1}\mathcal{B}_{\tau, 1}(z)+m^{-2} \mathcal{B}_{\tau,2}(z)+\ldots\right), \] as $n,m\to\infty$ while the ratio $\tau=\frac{n}{m}$ is fixed. Here $\mathcal{S}_\tau$ denotes the droplet, the boundary of which is assumed to be a smooth simple closed curve, and $\phi_\tau$ is a conformal mapping from the complement $\mathcal{S}_\tau^c$ to the exterior disk $\Bbb{D}_\mathrm{e}$. The functions $\mathcal{Q}_\tau$ and $\mathcal{B}_{\tau, j}$ are bounded holomorphic functions which may be expressed in terms of $Q$ and $\mathcal{S}_\tau$. We apply these results to obtain boundary universality in the random normal matrix model for smooth droplets, i.e., that the limiting rescaled process is the random process with correlation kernel \[ \mathrm{k}(\xi,\eta)= \mathrm{e}^{\xi\bar\eta\,-\frac12(\lvert\xi\rvert^2+\lvert \eta\rvert^2)} \,\mathrm{erf}\,(\xi+\bar{\eta}). \] A key ingredient in the proof of the asymptotic expansion of the orthogonal polynomials is the construction of an orthogonal foliation -- a smooth flow of closed curves near $\partial\mathcal{S}_\tau$, on each of which $P_{m,n}$ is appropriately orthogonal to lower order polynomials. To compute the coefficient functions, we develop an algorithm which determines the coefficients $\mathcal{B}_{\tau, j}$ successively in terms of inhomogeneous Toeplitz kernel conditions. These inhomogeneous Toeplitz kernel conditions may be understood in terms of scalar Riemann-Hilbert problems.

Topics & Concepts

MathematicsOrthogonal polynomialsUniversality (dynamical systems)Pure mathematicsClassical orthogonal polynomialsRandom matrixPlanarCombinatoricsEigenvalues and eigenvectorsComputer graphics (images)Computer scienceQuantum mechanicsPhysicsGeometry and complex manifoldsRandom Matrices and ApplicationsStochastic processes and statistical mechanics