Reformulation of q q -middle convolution and applications

Yumi Arai, Kouichi Takemura · 2025

Read Reformulation of q q -middle convolution and applications by Yumi Arai, Kouichi Takemura (2025) 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-2503-11214.txt.

Introduction

Yumi Arai and Kouichi Takemura recast a family of special functions in this 2025 math paper. Their tools are q-convolutions, deformations of classic convolution that depend on a parameter q. The middle version of the construction, due to Sakai and Yamaguchi, is the focus.

The new framing makes two middle convolutions compose cleanly, a property the old one lacked. The pair also define q-versions of addition tied to gauge changes. They give conditions for the key integrals to converge and obey the right equations. Several third-order q-difference equations are solved with the machinery. The work tidies a corner of special function theory.

Catalog record

CREATORYumi Arai, Kouichi Takemura
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE138 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-11214

Browse the full library · About the Tower of Babel

Full text

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

Reformulation of q q -middle convolution and applications Yumi Arai, Kouichi Takemura Abstract . We reformulate the q q -convolution and the q q -middle convolution introduced by Sakai and Yamaguchi, and we introduce q q -deformations of the addition which is related to the gauge-transformation. A merit of the reformulation is the additivity on composition of two q q -middle convolutions. We obtain sufficient conditions that the Jackson integrals associated with the q q -convolution converge and satisfy the q q -difference equation associated with the q q -convolution. We present several third-order linear q q -difference equations and solutions of them by using the q q -middle convolution and the q q -deformations of the addition. 2010 Mathematics Subject Classification 33D15,39A13 1. Introduction Differential equations and difference equations play important roles in investigating special functions, and the middle convolution has been applied for the study of special functions and differential equations. Katz introduced the middle convolution in his book [ 10 ] to study rigid local systems. Dettweiler and Reiter [ 3 , 4 ] reformulated the middle convolution for the Fuchsian equations in terms of the linear algebra. The Fuchsian equations is the system of linear differential equations written as (1.1) \frac{dY}{dx}=\left(\frac{A_{0}}{x-t_{0}}+\frac{A_{1}}{x-t_{1}}+\dots+\frac{A_{N}}{x-t_{N}}\right)Y, where Y is a column vector with m entries and A_{0},A_{1},\dots,A_{N} are constant square matrices of size m . We review briefly the definition of the convolution and the middle convolution for equation ( 1.1 ) (or the tuple of the matrices (A_{0},\dots,A_{N}) ). Let \lambda\in\mathbb{C} and F_{i} (i=0,\dots,N) be the square matrix of size (N+1)m of the form (1.2) F_{0}=\left(\begin{array}[]{@{\,}c@{\,}ccc@{\,}}A_{0}+\lambda I_{m}&A_{1}&\cdots&A_{N}\\ 0&0&\cdots&0\\ \vdots&\vdots&&\vdots\\ 0&0&\cdots&0\end{array}\right),\;F_{1}=\left(\begin{array}[]{@{\,}c@{\,}c@{\,}cc@{\,}}0&0&\cdots&0\\ A_{0}&A_{1}+\lambda I_{m}&\cdots&A_{N}\\ \vdots&\vdots&&\vdots\\ 0&0&\cdots&0\end{array}\right),\dots, where I_{m} is the identity matrix of size m . Then, the correspondence of the tuple of matrices (A_{0},\dots,A_{N})\mapsto(F_{0},\dots,F_{N}) (or the correspondence of the associated Fuchsian system) is called the convolution. It is shown that the following subspaces \mathcal{K},\mathcal{L} of \mathbb{C}^{(N+1)m} are preserved by the action of F_{i} (i=0,\dots,r) . (1.3) \mathcal{K}=\left(\begin{array}[]{@{\,}c@{\,}}\mathrm{ker}A_{0}\\ \vdots\\ \mathrm{ker}A_{N}\end{array}\right),\quad\mathcal{L}=\mathrm{ker}(F_{0}+F_{1}+\cdots+F_{N}). We denote the linear transformation induced from the action of F_{i} on the quotient space \mathbb{C}^{(N+1)m}/(\mathcal{K}+\mathcal{L}) by \overline{F}_{i} . The correspondence of the tuple of matrices (A_{0},\dots,A_{N})\mapsto(\overline{F}_{0},\dots,\overline{F}_{N}) (or the correspondence of the associated Fuchsian system) is called the middle convolution, and we denote it by mc_{\lambda} . See also Haraoka [ 7 , §7.5] for a detailed description. It is known that the middle convolution has additivity. Namely, mc_{0} is an identity map and we have (1.4) mc_{\lambda}\circ mc_{\mu}=mc_{\lambda+\mu}, under some conditions (see the conditions (*) and (**) in Definition 5.1 ). A theory of the q -deformation of the middle convolution was constructed by Yamaguchi and Sakai. A tentative theory was presented in master’s thesis of Yamaguchi [ 17 ] which was supervised by Sakai, and a full paper was published later by their joint paper [ 12 ] . They handled the system of q -difference equations described as (1.5) Y(qx)=\left(B_{\infty}+\frac{B_{1}}{1-x/b_{1}}+\dots+\frac{B_{N}}{1-x/b_{N}}\right)Y(x), where Y(x) is a column vector with m entries and B_{\infty},B_{1},\dots,B_{N} are square matrices of size (N+1)m . The construction of the q -middle convolution is similar to the case of the Fuchsian system of differential equations. They investigated several properties of the q -middle convolution and related topics. On the composition of the middle q -middle convolution, they obtained (1.6) \displaystyle\overline{\Psi}_{\mu}\circ\overline{\Psi}_{\lambda}\simeq\overline{\Psi}_{\log(q^{\lambda}+q^{\mu}-1)/\log q}. See the appendix for the notation. The theory by Sakai and Yamaguchi was applied to studies of some q -difference equations in [ 13 , 2 ] . Recall that the convolution in equation ( 1.2 ) is related with Euler’s integral transformation. Let Y(x) be a solution of equation ( 1.1 ). Set (1.7) W_{j}(x)=\frac{Y(x)}{x-t_{j}},\quad W(x)=\left(\begin{array}[]{@{\,}c@{\,}}W_{0}(x)\\ \vdots\\ W_{N}(x)\end{array}\right). We apply Euler’s integral transformation for each entry of W(x) , i.e. we set (1.8) \widetilde{Y}(x)=\int_{\Delta}W(s)(x-s)^{\lambda}ds, where \Delta is an appropriate cycle in \mathbb{C} with the variable s . Dettweiler and Reiter [ 3 ] established that the function \widetilde{Y}(x) satisfies the following Fuchsian system of differential equation (1.9) \frac{d\widetilde{Y}}{dx}=\left(\frac{F_{0}}{x-t_{0}}+\frac{F_{1}}{x-t_{1}}+\dots+\frac{F_{N}}{x-t_{N}}\right)\widetilde{Y}, where F_{0},\dots,F_{N} were defined in equation ( 1.2 ). Sakai and Yamaguchi [ 12 ] introduced the q -deformation of the convolution by considering a q -deformation of the integral transformation of Dettweiler and Reiter. See the appendix for a review of them. In [ 13 ] , the theory of Sakai and Yamaguchi was applied to the q -Heun equation. Consequently, a q -integral transformation of the q -Heun equation was suggested. Note that it was discussed in [ 16 ] from another approach based on the kernel function identity, and the q -integral transformation was extended to the variants of the q -Heun equations. In our previous paper [ 2 ] , the q -convolution of Sakai and Yamaguchi was reconsidered. Namely, one parameter which corresponds to the integral cycle was added to the q -integral transformation of Sakai and Yamaguchi, and convergence of the q -integral transformation was considered. As an application, q -integral representations of solutions to the q -hypergeometric equation and its variants were obtained. In this paper, we reformulate the q -convolution of Sakai and Yamaguchi. Our definition of the q -convolution is simpler than that of Sakai and Yamaguchi, and our discussion takes care of convergence. We rewrite equation ( 1.5 ) to (1.10) \frac{Y(qx)-Y(x)}{-x}=\Big[\sum^{N}_{i=0}\frac{B_{i}}{x-b_{i}}\Big]Y(x), where b_{0}=0 , and define the q -convolution and the q -middle convolution for the tuple (B_{0},B_{1},\dots,B_{N}) . A merit of our definition is the additivity of two q -middle convolutions. Namely, we have (1.11) mc^{q}_{\lambda}\circ mc^{q}_{\mu}=mc^{q}_{\lambda+\mu}, instead of equation ( 1.6 ) (see section 3 for the definition). The q -convolution in this paper is also related to a q -integral transformation, and the q -middle convolution induces the q -integral transformation for solutions to q -difference equations. We pay attention to a q -analogue of the addition for the Fuchsian equations, which is induced from a gauge-transformation of solutions. Let Y be a solution to the system of differential equation given in equation ( 1.1 ), and let (a_{0},a_{1},\dots,a_{N}) be a tuple of complex numbers. Then, the function Z=(x-t_{0})^{a_{0}}(x-t_{1})^{a_{1}}\dots(x-t_{N})^{a_{N}}Y satisfies the following system of differential equations; (1.12) \frac{dZ}{dx}=\left(\frac{A_{0}+a_{0}I_{m}}{x-t_{0}}+\frac{A_{1}+a_{1}I_{m}}{x-t_{1}}+\dots+\frac{A_{N}+a_{N}I_{m}}{x-t_{N}}\right)Z. The correspondence of the tuple of the matrices (A_{0},A_{1},\dots,A_{N})\mapsto(A_{0}+a_{0}I_{m},A_{1}+a_{1}I_{m},\dots,A_{N}+a_{N}I_{m}) is called the addition. We introduce an analogue of the addition for systems of q -difference equations. In particular, we investigate the transformations of the tuples (B_{0},B_{1},\dots,B_{N}) induced from the transformations of the solutions by (1.13) Y(x)\mapsto x^{\mu}Y(x)\mbox{ and }Y(x)\mapsto\frac{(x/\gamma;q)_{\infty}}{(x/\delta;q)_{\infty}}Y(x), where (1.14) (\alpha;q)_{\infty}=(1-\alpha)(1-q\alpha)(1-q^{2}\alpha)\cdots. See section 2 for details of the q -analogue of addition. In our previous paper [ 2 ] , we obtained q -integral representations of solutions to the q -hypergeometric equation and the variants of the q -hypergeometric equation by applying the q -middle convolution. A standard form of the q -hypergeometric equation is given as (1.15) (x-q)g(x/q)+(abx-c)g(qx)-\{(a+b)x-q-c\}g(x)=0. In this paper, we obtain several q -difference equations by applying q -deformations of the addition and the reformulated q -middle convolution repeatedly. Consequently, we have formal q -integral representations of solutions. The reformulated q -middle convolution would be better for discussing a sequence of q -difference equations. We discuss sufficient conditions that the q -integral representations converge and they satisfy the attached q -difference equation, which is used to generate actual solutions. The generalized q -hypergeometric equation of order 3 is written in the

Read the complete text