Hankel Determinant for a Perturbed Laguerre Weight with Pole Singularities and Generalized Painlevé III ′ Equation

Shulin Lyu, Yuanfei Lyu · 2026

Read Hankel Determinant for a Perturbed Laguerre Weight with Pole Singularities and Generalized Painlevé III ′ Equation by Shulin Lyu, Yuanfei Lyu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-00391.txt.

Introduction

Lyu and Lyu study a Hankel determinant for a perturbed Laguerre weight with pole singularities, widening the range and adding a stronger zero at the origin. Ladder operators for the orthogonal polynomials plus three compatibility rules express recurrence coefficients through four helper values that repeat step by step. Two helpers also obey coupled differential rules that collapse to a Painlevé III prime formula in a limit. Under double scaling the team extracts the unitary ensemble density and extends the study to a cubic pole case.

Catalog record

CREATORShulin Lyu, Yuanfei Lyu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE98248
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.00391

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Hankel Determinant for a Perturbed Laguerre Weight with Pole Singularities and Generalized Painlevé III ′ Equation Shulin Lyu, Yuanfei Lyu Abstract We study the Hankel determinant for the weight x α exp ( − x − t 1 / x − t 2 / x 2 ) , x ∈ [ 0 , + ∞ ) x^{\alpha}{\rm exp}(-x-t_{1}/x-t_{2}/x^{2}),x\in[0,+\infty) , with α > − 1 , t 1 ∈ ℝ ∖ { 0 } , t 2 > 0 . \alpha>-1,~t_{1}\in\mathbb{R}\setminus\{0\},~t_{2}>0. Compared with the weight x α ​ e − x − t 1 / x x^{\alpha}{\rm e}^{-x-t_{1}/x} studied in prior work (where α , t 1 > 0 \alpha,t_{1}>0 ), the range of α \alpha in our work is extended and the parameter t 2 t_{2} introduces a “stronger” zero at the origin. This leads to more varied behavior of the Hankel determinant, and the interplay between t 1 t_{1} and t 2 t_{2} introduces uncertainty and complexity into the analysis. By using a pair of ladder operators satisfied by the associated monic orthogonal polynomials and three compatibility conditions, we show that the recurrence coefficients are expressed in terms of four auxiliary quantities which satisfy a system of difference equations that can be iterated. We also establish two coupled second order partial differential equations (PDEs) satisfied by two of the auxiliary quantities, which are reduced to a Painlevé III ′ equation when t 2 → 0 + t_{2}\rightarrow 0^{+} . Moreover, the logarithmic derivative of the Hankel determinant is shown to satisfy a second order six degree PDE which is reduced to the σ \sigma -form of the Painlevé III ′ equation when t 2 → 0 + t_{2}\rightarrow 0^{+} . Under suitable double scaling, we obtain the limiting forms of the above PDEs and deduce the equilibrium density of the eigenvalues for the unitary ensemble. We extend our analysis to the Hankel determinant for x α ​ exp ⁡ ( − x − ∑ k = 1 m t k / x k ) x^{\alpha}\exp(-x-\sum_{k=1}^{m}t_{k}/x^{k}) with m = 3 m=3 . For general m m , we outline a derivation that leads, at least in principle, to the PDE satisfied by the logarithmic derivative of the Hankel determinant. \mathbf{Keywords} : Laguerre unitary ensembles; Hankel determinant; Orthogonal polynomials; Painlevé equations \mathbf{Mathematics\>\>Subject\>\>Classification\>\>2020} : 15B52; 33E17; 42C05 1 Introduction We consider the n -dimensional singularly perturbed Laguerre unitary ensemble of n\times n Hermitian matrices whose eigenvalues have the following joint probability density function \displaystyle p(x_{1},\dots,x_{n})=\frac{1}{n!D_{n}}\prod_{1\leq i<j\leq n}(x_{i}-x_{j})^{2}\prod_{k=1}^{n}w(x_{k};t_{1},t_{2}), (1.1) where the weight function w(x;t_{1},t_{2}) is obtained through a deformation of the Laguerre weight, i.e. \displaystyle w(x;t_{1},t_{2})=x^{\alpha}{\rm exp}\left(-x-\frac{t_{1}}{x}-\frac{t_{2}}{x^{2}}\right),\quad x\in[0,+\infty),~\alpha>-1,~t_{1}\in\mathbb{R}\setminus\{0\},~t_{2}>0. (1.2) The reason why we assume t_{1}\neq 0 is that the approach presented in this paper is inapplicable for t_{1}=0 . The term n!D_{n} in ( 1.1 ) is known as the partition function and D_{n} is given by \displaystyle D_{n}(t_{1},t_{2})= \displaystyle\frac{1}{n!}\int_{[0,+\infty)^{n}}\prod\limits_{1\leq i<j\leq n}(x_{i}-x_{j})^{2}\prod_{k=1}^{n}w(x_{k};t_{1},t_{2})dx_{1}\cdots dx_{n} \displaystyle= \displaystyle\det\left(\int_{0}^{+\infty}x^{i+j}w(x;t_{1},t_{2})dx\right)_{i,j=0}^{n-1}, (1.3) where the second equality is due to Heine’s formula [ 34 ] . This paper is concerned with D_{n}(t_{1},t_{2}) . The weight w(x;t_{1},0) , i.e. x^{\alpha}{\rm e}^{-x-t_{1}/x} with t_{1}>0 , arises from integrable quantum field theories at finite temperature and was studied in [ 9 ] under the assumption that \alpha>0 . By using the ladder operators and the Lax pair of the Riemann-Hilbert problem satisfied by the associated monic orthogonal polynomials, D_{n}(t_{1},0) was expressed as an integral in terms of the solution of a Painlevé III ′ equation. Compared with the above case, the range of \alpha is extended in w(x;t_{1},t_{2}) and the factor {\rm e}^{-t_{2}/x^{2}} introduces a “stronger” zero at the origin, which leads to more varied behavior of D_{n} . In addition, the interplay between t_{1} and t_{2} introduces uncertainty and complexity into the analysis. The Hankel determinant generated by the weight function y^{\alpha}{\rm e}^{-n(y+t^{2}/y^{2})} was studied in [ 1 ] (see (1.51) therein with k=2 ). It multiplied by n! denotes the partition function of the associated unitary ensemble and also the moment generating function of the charge relaxation resistance [ 1 ] . For other motivations to study this Hankel determinant, refer to Section 1.1 of [ 1 ] . Under a suitable double scaling, the eigenvalue correlation kernel of this ensemble was characterized by a hierarchy of higher order Painlevé III equation (see Theorem 1.11 of [ 1 ] ). By a change of variable y=x/n , this Hankel determinant is transformed into the one for the weight w(x;0,n^{3}t^{2}) , up to a multiplicative constant. It is well-known that the Hankel determinant admits the following representation [ 17 ] : \displaystyle D_{n}(t_{1},t_{2})=\prod_{j=0}^{n-1}h_{j}(t_{1},t_{2}), (1.4) where h_{j} is the square of the L^{2} -norm of the j th-degree monic polynomial P_{j}(x;t_{1},t_{2}) orthogonal with respect to w(x;t_{1},t_{2}) , namely, \displaystyle\int_{0}^{+\infty}P_{j}(x;t_{1},t_{2})P_{k}(x;t_{1},t_{2})w(x;t_{1},t_{2})dx=h_{j}(t_{1},t_{2})\delta_{jk},\qquad j,k=0,1,\ldots. (1.5) Here \delta_{jk}=1 for j=k and 0 otherwise, and \displaystyle P_{j}(x;t_{1},t_{2})=x^{j}+p(j,t_{1},t_{2})x^{j-1}+\cdots+P_{j}(0;t_{1},t_{2}) (1.6) for j\geq 1 and P_{0}(x;t_{1},t_{2}):=1 . According to ( 1.5 )-( 1.6 ), one shows that \{P_{j}(x;t_{1},t_{2})\} satisfy a three-term recurrence relation \displaystyle xP_{j}(x;t_{1},t_{2})=P_{j+1}(x;t_{1},t_{2})+\alpha_{j}(t_{1},t_{2})P_{j}(x;t_{1},t_{2})+\beta_{j}(t_{1},t_{2})P_{j-1}(x;t_{1},t_{2}),\quad j=0,1,\ldots, (1.7) subject to the initial condition \beta_{0}(t_{1},t_{2})P_{-1}(x;t_{1},t_{2}):=0. Inserting ( 1.6 ) into ( 1.7 ) and equating the coefficients of x^{j} , we find \displaystyle\alpha_{j}(t_{1},t_{2})=p(j,t_{1},t_{2})-p(j+1,t_{1},t_{2}),\qquad\qquad j\geq 0, (1.8) with the initial value p(0,t_{1},t_{2}):=0 , which immediately gives us \displaystyle\sum_{j=0}^{n-1}\alpha_{j}(t_{1},t_{2})=-p(n,t_{1},t_{2}). (1.9) Multiplying both sides of ( 1.7 ) by P_{j-1}(x;t_{1},t_{2})w(x;t_{1},t_{2}) and integrating it from 0 to +\infty , in view of ( 1.5 ), we get \displaystyle\beta_{j}(t_{1},t_{2})h_{j-1}(t_{1},t_{2})=\int_{0}^{+\infty}xP_{j-1}(x;t_{1},t_{2})P_{j}(x;t_{1},t_{2})w(x;t_{1},t_{2})dx=h_{j}(t_{1},t_{2}), (1.10) where the second equality is obtained via a replacement of xP_{j-1} by using ( 1.7 ) with j-1 in place of j . We readily see from ( 1.10 ) that \displaystyle\beta_{j}(t_{1},t_{2})=\frac{h_{j}(t_{1},t_{2})}{h_{j-1}(t_{1},t_{2})}, (1.11) According to the recurrence relation ( 1.7 ), one derives the Christoffel-Darboux formula \displaystyle\sum_{j=0}^{n-1}\frac{P_{j}(x;t_{1},t_{2})P_{j}(y;t_{1},t_{2})}{h_{j}(t_{1},t_{2})}=\frac{P_{n}(x;t_{1},t_{2})P_{n-1}(y;t_{1},t_{2})-P_{n}(y;t_{1},t_{2})P_{n-1}(x;t_{1},t_{2})}{h_{n-1}(t_{1},t_{2})(x-y)}. By using this formula, the recurrence relation ( 1.7 ) and the definition of P_{n}(z):=P_{n}(z;t_{1},t_{2}) given by ( 1.5 )-( 1.6 ), the following pair of lowering and raising operators were deduced in [ 28 ] : \displaystyle\left(\frac{d}{dz}+B_{n}(z)\right)P_{n}(z)= \displaystyle\beta_{n}A_{n}(z)P_{n-1}(z), (1.12) \displaystyle\left(\frac{d}{dz}-B_{n}(z)-v^{\prime}(z)\right)P_{n-1}(z)= \displaystyle-A_{n-1}(z)P_{n}(z), (1.13) for n\geq 0 , where v(z)=v(z;t_{1},t_{2}):=-\ln\>w(z;t_{1},t_{2}) and \displaystyle A_{n}(z):=\frac{1}{z}\cdot\frac{1}{h_{n}}\int_{0}^{+\infty}\frac{zv^{\prime}(z)-xv^{\prime}(x)}{z-x}P_{n}^{2}(x)w(x;t_{1},t_{2})dx,\qquad n\geq 0, (1.14) \displaystyle B_{n}(z):=\frac{1}{z}\left(\frac{1}{h_{n-1}}\int_{0}^{+\infty}\frac{zv^{\prime}(z)-xv^{\prime}(x)}{z-x}P_{n}(x)P_{n-1}(x)w(x;t_{1},t_{2})dx-n\right),\qquad n\geq 1, (1.15) with the initial conditions A_{-1}(z)=B_{0}(z):=0 . They can also be derived by employing the Riemann-Hilbert problem for P_{n}(z) [ 21 ] . Moreover, it was shown in [ 28 ] and [ 21 ] that A_{n} and B_{n} satisfy the following identities \displaystyle B_{n+1}(z)+B_{n}(z)=\left(z-\alpha_{n}\right)A_{n}(z)-v^{\prime}(z), ( S_{1} ) \displaystyle 1+\left(z-\alpha_{n}\right)\left(B_{n+1}(z)-B_{n}(z)\right)=\beta_{n+1}A_{n+1}(z)-\beta_{n}A_{n-1}(z), ( S_{2} ) \displaystyle\left(B_{n}(z)+v^{\prime}(z)\right)B_{n}(z)+\sum_{j=0}^{n-1}A_{j}(z)=\beta_{n}A_{n}(z)A_{n-1}(z). ( S_{2}^{\prime} ) Actually, it was proved in [ 28 ] and [ 21 ] that these three compatibility conditions and the above ladder operators are valid for a more general Laguerre-type weight function of the form x^{\alpha}w_{0}(x),x\in[0,+\infty) with \alpha>-1 , where w_{0}(x)>0 was assumed to be continuously differentiable and \lim\limits_{x\rightarrow+\infty}\pi_{n}(x)x^{\alpha}w_{0}(x)=0 for an arbitrary polynomial \pi_{n}(x) of order n\geq 0 . For more discussions on ladder operators, we refer the readers to [ 7 , 11 , 22 , 31 , 35 ] . For w(x;t_{1},t_{2}) given by ( 1.2 ), we get the expressions for A_{n} and B_{n} from ( 1.14 )-( 1.15 ), with four auxiliary quantities introduced. Substituting them into (S_{1}) and (S_{2}^{\prime}) , we express \alpha_{n},\beta_{n} and p(n,t_{1},t_{2}) in terms of the auxiliary quantities which are shown to satisfy a system of difference equations that can be iterated in n . In addition, by differentiating

Read the complete text