Krall-type orthogonal polynomials and integrable isomonodromic deformations

Luc Haine · 2026

Read Krall-type orthogonal polynomials and integrable isomonodromic deformations by Luc Haine (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-2603-01132.txt.

Introduction

Haine shows Krall-type orthogonal polynomials, built by adding jumps at the range edge to Laguerre or Jacobi ones, are ruled by tidy systems. Their repeat rules and second-order laws follow plainly from chosen answers of those systems. One jump leads to Painlevé III or V cases, two jumps give a new Schlesinger-type partial system, and equal jumps from Gegenbauer ones reduce it to a Painlevé V case.

Catalog record

CREATORLuc Haine
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE101294
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.01132

Browse the full library · About the Tower of Babel

Full text

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

Krall-type orthogonal polynomials and integrable isomonodromic deformations Luc Haine Abstract . Krall-type polynomials are orthogonal polynomials for a Stieltjes’ measure obtained by adding jumps at the boundary of the interval of orthogonality of either the generalized Laguerre polynomials or the Jacobi polynomials. We show that both the recurrence relations and the second order linear differential equations defining these polynomials, are explicitly determined in terms of specific solutions of some integrable systems. When there is only one jump, we are led to integrable cases of the Painlevé III or the Painlevé V equation. In the case of two jumps, first studied by Koornwinder [ 22 ] , we obtain a new integrable system of partial differential equations of Schlesinger type. When the jumps are equal and the starting polynomials are the Gegenbauer polynomials, this system reduces to an integrable case of the Painlevé V equation. 2020 Mathematics Subject Classification Primary: 33C45; Secondary: 34M56 1. Introduction Let w(x) be the weight function of the generalized Laguerre or the Jacobi orthogonal polynomials. For the generalized Laguerre polynomials, the interval of orthogonality is [a,b[=[0,\infty[ , and for the Jacobi polynomials we shall pick it to be either [a,b]=[0,1] or [a,b[=[-1,1] , for convenience. Krall-type polynomials are orthogonal polynomials with respect to the Stieltjes’ measure d\sigma(x) with \sigma(x)=\begin{cases}0,\;\mbox{if}\;x\leq a,\\ \frac{1}{t_{1}}+\int_{a}^{x}w(t)\;dt,\;\mbox{if}\;a<x<b,\\ \frac{1}{t_{1}}+\frac{1}{t_{2}}+\int_{a}^{b}w(t)dt,\;\mbox{if}\;x\geq b,\end{cases} (1.1) with t_{1},t_{2}>0 , with the understanding that eventually t_{1} or t_{2}=\infty , in which case there is only one jump. Equivalently, Krall-type polynomials are orthogonal for the weight distribution \sigma^{\prime}(x)=w(x)\Big(H(x-a)-H(x-b)\Big)+\frac{1}{t_{1}}\delta(x-a)+\frac{1}{t_{2}}\delta(x-b), (1.2) where H(x) denotes the Heaviside function. In the sequel, we shall always normalize orthogonal polynomials p_{n}(x),n\geq 0 , to be monic, so that they satisfy a three-term recurrence relation xp_{n}=p_{n+1}+b_{n}p_{n}+a_{n}p_{n-1},\;n\geq 0, (1.3) with p_{-1}=0,p_{0}=1 . As a justification of our terminology, some special cases of Krall-type polynomials appeared first in H.L. Krall work [ 24 , 25 ] , devoted to the problem of finding all orthogonal polynomial sets which are eigenfunctions of a differential operator of arbitrary order, generalizing the classical orthogonal polynomials, for which the order is two. Except for the first non trivial case of order four, which was completely solved in [ 25 ] , the problem remains partially unsolved. We refer the reader to A.M. Krall [ 26 , 27 ] and Littlejohn [ 29 ] , who resurrected the subject in the 1980’s, for more references on the subject, and to Kwon an Lee [ 23 ] for characterizations of Bochner-Krall orthogonal polynomials of Jacobi type. Using an explicit formula for Krall-type polynomials discovered by Koornwinder [ 22 ] , J. Koekoek and R. Koekoek [ 20 , 21 ] have shown that Krall-type polynomials solve in general an ”infinite order version” of H.L. Krall problem, reducing to finite order only for special choices of the classical weights w(x) . In [ 12 ] , we showed that the polynomials discovered by H.L. Krall [ 25 ] can be obtained by the method of Darboux transformations, starting from some instances of the classical orthogonal polynomials. Further work in this direction [ 14 , 15 , 16 , 18 , 19 ] led to new solutions of H.L. Krall problem as posed in [ 24 ] . Krall-type polynomials, as defined in ( 1.1 ), were also constructed by Zhedanov [ 39 ] , using Darboux transformations. The present work doesn’t rely on Darboux transformations, instead it constructs explicitly Krall-type polynomials, using a method which originates in a remarkable work by Laguerre [ 28 ] . We refer the reader to [ 5 ] for an historical review of this work. As mentioned by Nuttall [ 36 ] (Section 5.1), G. Chudnovsky observed that, using Laguerre’s method, orthogonal polynomials corresponding to some generalized Jacobi weight with three factors, can be constructed in terms of a specific solution of the Painlevé VI equation, see also [ 2 , 33 ] . In a similar vein, in [ 17 ] we characterized some orthogonal polynomials, arising in the context of the Jacobi polynomial ensemble, in terms of the Painlevé VI equation, using at first sight unrelated tools from the theory of the Toda lattice hierarchy and its master symmetries [ 1 , 3 , 8 ] . As we shall see, these tools are closely tight up with isomonodromic deformations. The central result of this work is that Krall-type polynomials lead to integrable cases of the Painlevé equations and more generally to integrable isomonodromic deformations. The integrability of the equations leads to explicit expressions for Krall-type polynomials. In recent years, several intriguing connections between Painlevé equations and orthogonal polynomials have been unravelled, some in relation with rational solutions, but the integrable cases are ruled out in most of these studies, see the recent monograph [ 38 ] for an updated account. Introducing the Stieltjes’ transform f(x)=\int_{a}^{b}\frac{d\sigma(s)}{x-s},\;x\in\mathbb{C}\setminus[a,b], (1.4) our starting point is to observe that Krall-type polynomials satisfy Laguerre’s hypothesis in [ 28 ] , that is there exists polynomials U(x),V(x),W(x) such that Wf^{\prime}=2Vf+U. Under this condition, the functions p_{n}(x) and \frac{\varepsilon_{n}(x)}{v(x)} with \displaystyle\varepsilon_{n}(x)=\int_{a}^{b}\frac{p_{n}(s)}{x-s}\;d\sigma(s),x\in\mathbb{C}\setminus[a,b], (1.5) \displaystyle v(x)=e^{\int\frac{2V}{W}dx}, (1.6) form a basis of solutions of a linear second order differential equation W\Theta_{n}g_{n}^{\prime\prime}+[(2V+W^{\prime})\Theta_{n}-W\Theta_{n}^{\prime}]g_{n}^{\prime}+K_{n}g_{n}=0, (1.7) with \quad{}^{\prime}=\frac{d}{dx} , and \Theta_{n} and K_{n} some polynomials in x (depending on n and t_{1},t_{2} ) but of degrees independent of n . When t_{1},t_{2}\to\infty , Laguerre’s equation reduces to the second order differential equation satisfied by the generalized Laguerre or the Jacobi polynomials, in which case the polynomials \Theta_{n} is independent of n . It follows from ( 1.2 ) and ( 1.5 ) that \varepsilon_{n}(x)=\int_{a}^{b}\frac{p_{n}(s)}{x-s}w(s)ds+\frac{1}{t_{1}}\frac{p_{n}(a)}{x-a}+\frac{1}{t_{2}}\frac{p_{n}(b)}{x-b}, (1.8) hence the monodromy in the complex plane of the solutions of ( 1.7 ) is independent of t_{1} and t_{2} . When there is only one free parameter involved i.e. t_{1}=t and t_{2}=\infty or t_{1}=t_{2}=t , this indicates that some Painlevé equations should be lurking around. Among the six Painlevé equations P_{I}-P_{VI} , the only ones which admit a first integral which is an elementary function of q,\dot{q} and t , with \dot{\quad}=\frac{d}{dt} , are the P_{III} equation (P_{III})\;\ddot{q}=\frac{\dot{q}^{2}}{q}-\frac{\dot{q}}{t}+\frac{1}{t}(aq^{2}+b)+cq^{3}+\frac{d}{q}, (1.9) when a=c=0 , with first integral \Big(\frac{t\dot{q}}{q}\Big)^{2}-\frac{2t\dot{q}}{q}+\frac{dt^{2}}{q^{2}}+\frac{2bt}{q}=C, or when b=d=0 , with first integral \Big(\frac{t\dot{q}}{q}\Big)^{2}+\frac{2t\dot{q}}{q}-2atq-ct^{2}q^{2}=C, (1.10) and the P_{V} equation (P_{V})\;\ddot{q}=\Big(\frac{1}{2q}+\frac{1}{q-1}\Big)\dot{q}^{2}-\frac{\dot{q}}{t}+\frac{(q-1)^{2}}{t^{2}}\Big(aq+\frac{b}{q}\Big)+\frac{cq}{t}+\frac{dq(q+1)}{q-1}, (1.11) when c=d=0 , with first integral \frac{1}{2q}\Big(\frac{t\dot{q}}{q-1}\Big)^{2}-aq+\frac{b}{q}=C, (1.12) with C an arbitrary constant. This result was first established by Gromak [ 10 , 11 ] , following work by Lukashevich [ 32 ] where some integrable cases where discovered. A new and different proof was provided by Zoladek and Filipuk [ 40 ] . Precisely, we shall establish the following Theorem. Theorem 1.1 . The recurrence relation ( 1.3 ) and the Laguerre differential equation ( 1.7 ) for (a) the Krall-Laguerre type polynomials, with weight distribution \frac{1}{\Gamma(\alpha+1)}x^{\alpha}e^{-x}H(x)+\frac{1}{t}\delta(x),\;\alpha>-1, (1.13) (b) the Krall-Jacobi type polynomials, with weight distribution \frac{\Gamma(\alpha+\beta+2)}{\Gamma(\alpha+1)\Gamma(\beta+1)}x^{\beta}(1-x)^{\alpha}\big(H(x)-H(x-1)\big)+\frac{1}{t}\delta(x),\;\alpha,\beta>-1, (1.14) (c) the Krall-Gegenbauer type polynomials with weight distribution \frac{\Gamma(\alpha+3/2)}{\sqrt{\pi}\;\Gamma(\alpha+1)}(1-x^{2})^{\alpha}\big(H(x+1)-H(x-1)\big)+\frac{1}{t}\big(\delta(x+1)+\delta(x-1)\big),\;\alpha>-1, (1.15) are completely determined by a sequence of solutions q_{n}(t),n\geq 1 , of integrable cases of the P_{III} or the P_{V} equation, namely for cases (a), (b) and (c) it is \begin{cases}(a)\;P_{III}\;\mbox{with}\;a=-\frac{2n+1+\alpha}{(\alpha+1)^{2}},\;b=0,\;c=\frac{1}{(\alpha+1)^{2}},\;d=0,\\ (b)\;P_{V}\;\mbox{with}\;a=\frac{(\alpha+\beta+2n+1)^{2}}{2(\beta+1)^{2}},\;b=-\frac{\alpha^{2}}{2(\beta+1)^{2}},\;c=0,\;d=0,\\ (c)\;P_{V}\;\mbox{with}\;a=\frac{(2\alpha+2n+1)^{2}}{8(\alpha+1)^{2}},\;b=-\frac{1}{8(\alpha+1)^{2}},\;c=0,\;d=0.\end{cases} For Koornwinder’s polynomials sudied in [ 22 ] , with weight distribution \frac{\Gamma(\alpha+\beta+2)}{2^{\alpha+\beta+1}\Gamma(\alpha+1)\Gamma(\beta+1)}(1-x)^{\alpha}(1+x)^{\beta}\big(H(x+1)-H(x-1)\big)\\ +\frac{1}{t_{1}}\delta(x+1)+\frac{1}{t_{2}}\delta(x-1),\;\alpha,\beta>-1, (1.16) one has a similar result, except that now the polynomials \Theta_{n},K_{n} in Laguerre’s equations ( 1.7 ) involve two

Read the complete text