A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆

For the -Painlevé equation with affine Weyl group symmetry of type ⁽¹⁾₆, a 2 × 2 matrix Lax form and a second-order scalar lax form were known. We give a new 3 × 3 matrix Lax form and a third order scalar equation related to it. Continuous limit is also discussed.

Saved in:
Bibliographic Details
Published in:Symmetry, Integrability and Geometry: Methods and Applications
Date:2023
Main Author: Park, Kanam
Format: Article
Language:English
Published: Інститут математики НАН України 2023
Online Access:https://nasplib.isofts.kiev.ua/handle/123456789/212037
Tags: Add Tag
No Tags, Be the first to tag this record!
Journal Title:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Cite this:A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆. Kanam Park. SIGMA 19 (2023), 094, 17 pages

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
_version_ 1859739395052011520
author Park, Kanam
author_facet Park, Kanam
citation_txt A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆. Kanam Park. SIGMA 19 (2023), 094, 17 pages
collection DSpace DC
container_title Symmetry, Integrability and Geometry: Methods and Applications
description For the -Painlevé equation with affine Weyl group symmetry of type ⁽¹⁾₆, a 2 × 2 matrix Lax form and a second-order scalar lax form were known. We give a new 3 × 3 matrix Lax form and a third order scalar equation related to it. Continuous limit is also discussed.
first_indexed 2026-03-15T14:41:36Z
format Article
fulltext Symmetry, Integrability and Geometry: Methods and Applications SIGMA 19 (2023), 094, 17 pages A 3 × 3 Lax Form for the q-Painlevé Equation of Type E6 Kanam PARK National Institute of Technology, Toba College, 1-1, Ikegami-cho, Toba-shi, Mie, Japan E-mail: paku-k@toba-cmt.ac.jp Received December 01, 2022, in final form November 05, 2023; Published online November 18, 2023 https://doi.org/10.3842/SIGMA.2023.094 Abstract. For the q-Painlevé equation with affine Weyl group symmetry of type E (1) 6 , a 2×2 matrix Lax form and a second order scalar lax form were known. We give a new 3×3 matrix Lax form and a third order scalar equation related to it. Continuous limit is also discussed. Key words: Lax formalism; q-Painlevé equation 2020 Mathematics Subject Classification: 14H70; 34M56; 39A13 1 Introduction The q-Painlevé equation with affine Weyl group symmetry of type E (1) 6 was first discovered in [9]. The well-known form of it is as follows T : (a1, a2, a3, a4, a5, a6, a7, a8; f, g) → ( a1/q, a2/q, a3/q, a4/q, a5, a6, a7, a8; f, g ) ,( fg − 1 ) (fg − 1) ff = (g − 1/a5)(g − 1/a6)(g − 1/a7)(g − 1/a8) (g − a3)(g − a4) ,( fg − 1 )( fg − 1 ) gg = ( f − a5 )( f − a6 )( f − a7 )( f − a8 )( f − a1/q )( f − a2/q ) , q = a1a2 a3a4a5a6a7a8 , where f and g are dependent variables, a1, a2, . . . , a8 are parameters, q is a constant and the overline symbol “¯” denotes the discrete time evolution. In previous works, the following Lax forms for the q-Painlevé equation for type E (1) 6 has been obtained. In [12], a 2 × 2 matrix Lax pair was first derived as a reduction of the q-Garnier system. The other approach gives a 2× 2 matrix Lax pair [2, 7]. In [16], a second order scalar Lax form was obtained as a reduction from the q-Painlevé equation of type E8. The relation between these Lax forms was given in [15]. In this article, we give a new Lax form with 3 × 3 matrix Lax pair. We derive such a Lax form as a special case of the system investigated in our previous work [8]. As a result, we derive an equation which is equivalent to the q-Painlevé equation of type E (1) 6 [4, 10]. In the previous work [8], we defined a nonlinear q-difference system as a connection preserving deformation of the following linear equation Ψ(qz) = Ψ(z)A(z), A(z) = DXε1 1 (z)Xε2 2 (z) · · ·XεM M (z), D = diag[d1, d2, . . . , dN ], mailto:paku-k@toba-cmt.ac.jp https://doi.org/10.3842/SIGMA.2023.094 2 K. Park Xi(z) = diag[u1,i, u2,i, . . . , uN,i] + Λ, Λ =  0 1 O . . . . . . 1 z 0  , (1.1) where the exponents are εi = ±1 (1 ≤ i ≤ M), uj,i (1 ≤ j ≤ N) are dependent variables and ci, dj are parameters which satisfy N∏ j=1 uj,i = ci. Since one can exchange the order of matrices X±1 i by suitable rational transformations of vari- ables uj,i, the equation (1.1) essentially depends on M+, M−, where M± = #{εi|εi = ±1}. The contents of this paper is as follows. In Section 2, we set up a linear q-difference equa- tion (2.1), which is a case of (M+,M−, N) = (3, 0, 3) for the equation (1.1). And we discuss about its two deformations. One deformation gives rise to a well known form of the q-Painlevé equation of type E (1) 6 , and the other deformation gives an equation for a non-standard direction. Namely, it does not give a q-shift deformation for parameters. In Section 3, we derive a scalar equation from the 3 × 3 matrix equation (2.1) and consider its characteristic properties. In Section 4, we study continuous limit of our constructions and its relation to the Boalch’s Lax pair [1]. In Appendix A, we give deformations considered in Section 2 on root variables. In Appendix B, we give explicit forms of coefficients of a single linear q-difference equation derived in Section 3. Remark 1.1. The equation (1.1) in case (M+,M−, N) = (2n + 2, 0, 2) is known that it is equivalent to the linear q-difference equation related to the 2n-dimensional q-Garnier system [11]. And in case of (M+,M−, N) = (2, 0, 2n + 2), the equation (1.1) is also equivalent to the linear q-difference equation related to the 2n-dimensional system q-P(n+1,n+1) [13]. In both cases, when n = 1, they give rise to the q-PVI equation [5]. We note that the equation (2.1) is coincide to the q-difference linear equation of the Lax form in [14] in case of (m, n)=(3, 1). 2 A 3 × 3 matrix Lax form In this section, we consider two types of deformations for the linear q-difference equation (1.1) in a case (M+,M−, N) = (3, 0, 3). We consider the connection preserving deformation for the following q-difference equation for an unknown function Ψ(z) = [Ψ1(z),Ψ2(z),Ψ3(z)]: Ψ(qz) = Ψ(z)A(z) = Ψ(z)A(z, t), A(z) = DX1(z)X2(z)X3(z) = b1d1 ∗ ∗ 0 b2d2 ∗ 0 0 b3d3 + d1 0 0 ∗ d2 0 ∗ ∗ d3  z, (2.1) where the matrices D and Xi(z) (1 ≤ i ≤ 3) stand for D = diag[d1, d2, d3], Xi(z) = diag[u1,i, u2,i, u3,i] + Λ, Λ = 0 1 0 0 0 1 z 0 0  , and uj,i (1 ≤ i, j ≤ 3) are dependent variables and ci, dj are parameters which satisfy 3∏ i=1 uj,i = bj , 1 ≤ j ≤ 3, 3∏ j=1 uj,i = ci, 1 ≤ i ≤ 3, b1b2b3 = c1c2c3. (2.2) A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 3 The first equation in (2.2) is equivalent to that the characteristic exponents at z = 0 of the equation (2.1) are bjdj . The second equation in (2.2) is equivalent to the following condition: |A(z)| = d1d2d3(z + c1)(z + c2)(z + c3). (2.3) Through a gauge transformation by a 3×3 diagonal matrix we can take two components in (2.1) as 1. In the following, we use this kind of gauge fixings in case by case. By the condition (2.3), two of the remaining four components are determined by other components and parameters bj , ci, dj . In this article, we will consider two deformations T1 and T2 for the equation (2.1) which act on parameters bj , ci, dj as T1 : (b1, b2, b3, c1, c2, c3, d1, d2, d3) → ( b1, qb2, b3 q , c1, c2, c3, qd1, d2, qd3 ) , T2 : (b1, b2, b3, c1, c2, c3, d1, d2, d3) → ( b1d1 d2 , b2d2 d3 , b3, c1, c2, c3q, d2, d3, d1 q ) . 2.1 The deformation T1 We will show that the deformation T1 gives rise to the standard form of the q-E (1) 6 . We consider the following q-difference linear equation: Ψ(qz) = Ψ(z)A(z), A(z) = b1d1 1 v1 0 b2d2 1 0 0 b3d3 + d1 0 0 v2 d2 0 v3 v4 d3  z, (2.4) |A(z)| = d1d2d3(z + c1)(z + c2)(z + c3). (2.5) Although there are many ways of gauge fixings, fixing as the equation (2.4) makes relatively easier to find a new change of variables from indeterminate points of time evolution equations vi (∗ = T1(∗)). As a connection preserving deformation for the equation (2.4), (2.5) we take the following deformation for parameters T1 : (b1, b2, b3, c1, c2, c3, d1, d2, d3) → ( b1, qb2, b3 q , c1, c2, c3, qd1, d2, qd3 ) . (2.6) Then, there is a matrix B(z) which satisfies the following deformation equation: T1Ψ(z) = Ψ(z)B(z). We derive and show the matrix B(z). First, the matrix B(z) is a rational function for z [5]. In fact, by the argument [5] to determine a coefficient matrix of a deformation equation, the matrix B(z) is of degree one through the following: (i) The parameters ci which satisfy |A(ci)| = 0 are constant by the deformation T1 (2.6). Therefore, the matrix B(z) does not have poles at z = −qkci (k ∈ Z). Namely, the matrix B(z) has poles only at z = 0 or z = ∞. (ii) The deformation T1 (2.6) shifts characteristic exponents of the equation (2.4) at z = 0 as follows: T1 : (b1d1, b2d2, b3d3) → (qb1d1, qb2d2, b3d3). Therefore, the matrix B(z) behaves nearby z = 0 as B(z) = B0 +B1z +O ( z2 ) . 4 K. Park (iii) The deformation T1 (2.6) shifts characteristic exponents of the equation (2.4) at z = ∞ as follows: T1 : (d1, d2, d3) → (qd1, d2, qd3). Therefore, the matrix B(z) behaves nearby z = ∞ as B(z) = B1z +B0 +O ( 1 z ) . From the above (i)–(iii), the matrix B(z) is a polynomial in z of degree one. At last, we derive an explicit form of the matrix B(z). We express the matrix B(z) as follows: B(z) = B0 +B1z, Bi = [ βi j,k ] 1≤j,k≤3 , i = 1, 2. (2.7) Comparing coefficients of z for a compatibility condition equation B(z)A(z) = A(z)B(qz), ∗ = T1(∗), for the matrices A(z) (2.4) and B(z) (2.7), we have the following three equations: B0A0 = A0B0, B1A0 +B0A1 = qA0B1 +A1B0, B1A1 = qA1B1, (2.8) where matrices Ai denote coefficients of zi for the matrix A(z) (2.4). Solving the first and the third equations of (2.8), forms of the matrices B0 and B1 are as follows: B0 = 0 0 β0 2,3(b2d2v1−b3d3v1−1) b1d1−b3d3 0 0 β0 2,3 0 0 −β0 2,3 (b2d2 − b3d3)  , B1 =  (d1−d2)β1 2,1 v2 0 0 β1 2,1 0 0 d1qv3v4β1 2,1+d2(v2v3β1 3,2−qv3v4β1 2,1)+v2(qv4v4β1 2,1−β1 3,2(d3qv3+v2v4)) (d1−d3)qv2v4 β1 3,2 −β1 3,2(d2−d3q) v4  . From the second equation of (2.8), we obtain explicit forms of time evolutions vi (1 ≤ i ≤ 4), the components β0 2,3 and β1 2,1 of the matrix B(z) (2.7). Explicit forms of the remain components β0 2,3 and β1 2,1 are as follows: β0 2,3 = β1 3,2q(b1d1 − b3d3)(d1 − d2 − v1v2) ( −b22d 2 2qv1v2(d3q − d1 + v1v2)− b3d3d2qv2 − 2b23d 2 3d2qv1v2 + b3d3d2qv1v3 + b2d2 ( b3d3(d1q(−d3(q − 1)− 2v1v2) + qv1v2(d3q + 2v1v2) + d2(d3(q − 1)q + (2q − 1)v1v2)) + q ( d3qv2 + d1(v1v3 − v2) + v1 ( −d2v3 + v22 − v1v3v2 ))) + b3d3d2qv4 − b23d 2 3qv 2 1v 2 2 − b3d3qv1v 2 2 − b3d 2 3qv2 + b3d1d3qv2 + b23d1d 2 3qv1v2 − b3d1d3qv1v3 + b3d3qv 2 1v2v3 − b3d1d3qv4 + b3d3qv1v2v4 + b1d1 ( b3d3 ( d2qv1v2 − d3qv1v2 + d22(q − 1)− d1d2(q − 1) ) + b2d2q(d2(d3(−q) + d3 − v1v2) + d1d3(q − 1) + d3v1v2) + qv4(d1 − d2 − v1v2) ) − b23d 2 3d 2 2q + b23d1d 2 3d2q + b3d3d2v2 + b23d 2 3d2v1v2 + b23d 2 3d 2 2 − b23d1d 2 3d2 + d2qv3 − d1qv3 + qv1v2v3 )−1 , β1 2,1 = β1 3,2 qv2(b1d1v1 − b2d2v1 + 1) v2(1− b2d2v1) + b3d3(−d1 + d2 + v1v2) + b1d1(d1 − d2) . A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 5 We remark that time evolutions vi are independent for the component β1 3,2. Therefore, we set β1 3,2 = 1. From the above, a deformation equation T1Ψ(z) = Ψ(z)B(z) is expressed as follows: T1Ψ(z) = Ψ(z)B(z), B(z) = 0 0 w1 0 0 w2 0 0 w3 + w4 0 0 w5 0 0 w6 1 w7  z, (2.9) |B(z)| = (w1w5 − w2w4)z 2, T1 : (b1, b2, b3, c1, c2, c3, d1, d2, d3; v1, v4) → ( b1, qb2, b3 q , c1, c2, c3, qd1, d2, qd3; v1, v4 ) . (2.10) Theorem 2.1. Through a compatibility condition of the equations (2.4), (2.9), (2.10), B(z)A(z) = A(z)B(qz), (2.11) we obtain the following equations: (fg − 1) ( fg − 1 ) ff = (g − 1/c1)(g − 1/c2)(g − 1/c3)(g − d2/b3d3) (g − 1/b2)(g − d2/b1d1) ,( fg − 1 )( fg − 1 ) gg = ( f − c1 )( f − c2 )( f − c3 )( f − b3d3/d2 )( f − b3d3/d1 )( f − b3/q ) , (2.12) where f = b3d3 d3 − v1v4 , g = d2v1 b2d2v1 − 1 , (2.13) and ∗ stands for T1(∗). The equation (2.12) is the well known form of the q-Painlevé equation of type E6 [4, 10] (see also [6]). Proof. The result is obtained by a direct computation of the compatibility condition of (2.11). Since the computation is rather heavy, we will give a comment how to do it efficiently. Though the 2 variables v2, v3 can be represented by the rational functions of the remaining two variables v1, v4 by the relation (2.5), it is more efficient to do this elimination after the calculation of the compatibility condition (2.11) in 4 variables, and then reduce it to 2 variables. In this way, we get the following time evolutions for v1, v4 as rational functions of v1, v4: v1 = C1(v1, v4)C2(v1, v4) D1(v1, v4)D2(v1, v4) , v4 = C3(v1, v4) D3(v1, v4)D4(v1, v4) , (2.14) where C3(v1, v4), D2(v1, v4) : polynomials in v1, v4 of degree (4, 3), C2(v1, v4) : a polynomial in v1, v4 of degree (3, 1), C1(v1, v4), D3(v1, v4), D4(v1, v4) : polynomials in v1, v4 of degree (2, 1), D1(v1, v4) : a polynomial in v1, v4 of degree (1, 0). The remaining task is to rewrite the equation (2.14) to (2.12). A useful way to solve it is to look at the singularities of the equations [3]. Namely, we investigate the points at which the 6 K. Park right hand side of the equations (2.14) are indeterminate. We focus on the equation D3(v1, v4) of them D3(v1, v4) = d3(−b2d2v1 + b3d2v1 + 1) + v1v4(b2d2v1 − 1). (2.15) Investigating common zero points of the equation (2.15) and the other polynomials Ck(v1, v4), Dl(v1, v4), we find 4 indeterminate points as follows: (v1, v4) = ( − 1 (u−1 − b2)d2 ,− ( u−1 − b3 )( u−1 − b2 ) d3d2 u−1 ) , u = c−1 1 , c−1 2 , c−1 3 , d2/b3d3. (2.16) The other 4 points are the following: (v1, v1v4) = (∞,∞), ( 1 b2d2 − b1d1 ,∞ ) , ( 1 b2d2 , d3 − d1 ) , ( 1 b2d2 , 0 ) . In view of the form of the points in (2.16), we define the variables f , g as follows: (v1, v1v4) = ( − 1( g−1 − b2 ) d2 , (f − b3)d3 f ) . (2.17) By the transformation (2.17), the equation D3(v1, v4) = 0 is transformed to an equation fg = 1. Through the correspondence (2.17) and the time evolution equations for the variables v1, v4 (2.14), we have time evolution equations for f , g (2.12). ■ 2.2 The deformation T2 In this subsection, we take a deformation equation which is one of that considered in the previous work [8]. It corresponded to the permutations of the matricesXi(z) ±1. We consider the following Lax pair: Ψ(qz) = Ψ(z)A(z), A(z) = DX1(z)X2(z)X3(z) = b1d1 1 ∗ 0 b2d2 1 0 0 b3d3 + d1 0 0 ∗ d2 0 ∗ ∗ d3  z, (2.18) T2Ψ(z) = Ψ(z)B(z), B(z) = X3(z/q) −1, |B(z)| = q z + qc3 , T2 : (b1, b2, b3, c1, c2, c3, d1, d2, d3;x, y) → ( b1d1 d2 , b2d2 d3 , b3, c1, c2, c3q, d2, d3, d1 q ;x, y ) , (2.19) where we define variables x, y and gauge freedom w1, w2 with the variables uj,i in (2.18) as follows: x = u1,1u1,2u2,1u2,2 u1,1 + u2,2 , y = 1 u1,1u1,2 (u2,1 + u3,2) , w1 = u1,1, w2 = u1,3. (2.20) Theorem 2.2. Through a compatibility condition of the equations (2.18), (2.19), (2.20), B(z)A(z) = A(z)B(qz), (2.21) A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 7 we obtain the following equations: x+ c1 x+ c2 = E1E2 F1F2 , c1y + 1 c2y + 1 = E1E3 F1F3 , (2.22) where E1 = b1d1(b2d2(1− xy) + c1d3y(c2 + x)) + c1d2d3x(c2y + 1), E2 = b1d1(b2d2(1− xy) + c3d3qy(c1 + x)) + c3d2d3qx(c1y + 1), E3 = b1d1(b2d2(1− xy) + c2d3y(c1 + x)) + c3d2d3qx(c1y + 1), F1 = b1d1(b2d2(1− xy) + c2d3y(c1 + x)) + c2d2d3x(c1y + 1), F2 = b1d1(b2d2(1− xy) + c3d3qy(c2 + x)) + c3d2d3qx(c2y + 1), F3 = b1d1(b2d2(1− xy) + c1d3y(c2 + x)) + c3d2d3qx(c2y + 1), (2.23) and ∗ stands for T2(∗). Proof. Solving a compatibility condition (2.21) with (2.20) for the variables x, y, we obtain the following equations x = (xy − 1)G1(x, y) H1(x, y) , y = xG2(x, y) H2(x, y) , where Gk(x, y), Hk(x, y) (k = 1, 2) are polynomials in variables x, y. The polynomial G1(x, y) is of degree (1, 1), G2(x, y) is of degree (1, 2) and Hk(x, y) are of degree (2, 2). Using a method [6] for finding point configuration, a configuration of points for the equation (2.22), (2.23) is as follows: (x, y) = ( −c1,− 1 c1 ) , ( −c2,− 1 c2 ) , ( −b1d1 d2 ,− d2 b1d1 ) , ( −b1b2 c3 ,− c3 b1b2 ) , (−b2, 0), ( −b1b2d1 c3d3q , 0 ) , ( 0,− 1 b1 ) , ( 0,− b2d2 c1c2d3 ) . (2.24) Calculating x+c1 x+c2 and c1y+1 c2y+1 , respectively, we have equations (2.22), (2.23). ■ Remark 2.3. The time evolution equations uj,i are derived by solving the compatibility condi- tion (2.21). For another derivation using the transformations which correspond to permutations of the matrices Xεi i (z), see [8, Proposition 2.1]. Before ending this subsection, we show a relation between pairs of the variables (f, g) in (2.12) and (x, y) in (2.22). Proposition 2.4. The equations (2.4), (2.13) and (2.18), (2.20) are equivalent with each other if the variables f , g and x, y are related as f = c1c2c3(b2 + x)(xy − 1) b1b2y(c1 + x)(c2 + x) + c3x(b2(1− xy) + c2xy + c1y(c2 + x) + x) , g = − y(b1b2 + c3x) c3(b2(1− xy) + x) + b1b2xy . (2.25) Proof. Comparing the coefficient matrix A(z) of the equation (2.4), (2.13) to the coefficient matrix A(z) of the equation (2.18), (2.20), we can solve for the variables x, y and gauge freedom w1, w2 in terms of variables f , g and parameters bj , ci, dj (1 ≤ i, j ≤ 3). Then we have the desired relation between pairs of variables (f, g) and (x, y) (2.25). ■ In Appendix A, we give pictures of a point configuration of the equation (2.12) and the configuration (2.24), root basis associated with them, and three deformations T1, T2 and T 3 2 on root variables attached with root basis. 8 K. Park 3 A scalar equation related to the q-E (1) 6 In this section, we derive a scalar q-difference equation from the matrix q-difference equation (2.18), (2.20) for an unknown function Ψ(z) = [Ψ1(z),Ψ2(z),Ψ3(z)] and its properties. Before deriving, we state about a characterization of a linear q-difference equation. Linear differential equations are characterized by its singular points and characteristic exponents. Simi- larly, linear q-difference equations are also characterized by its singular points and characteristic exponents. We consider the following n-th order q-difference equation Pn(z)Φ(q nz) + Pn−1(z)Φ ( qn−1z ) + · · ·+ P0(z)Φ(z) = 0, Pk(z) = pk,0 + pk,1z + · · ·+ pk,n+l−kz n+l−k, 0 ≤ k ≤ n, l ∈ Z≥0, p0,0, p0,n+l, pn,0, pn,l ̸= 0. (3.1) In the q-difference equation (3.1), singular points are at z = 0 and z = ∞. And characteristic exponents of solutions at z = 0 and z = ∞ are given as solutions of the following characteristic equations respectively Pn(0)λ n + Pn−1(0)λ n−1 + · · ·+ P0(0) = 0, Pn(∞)qn−1µn + Pn−1(∞)qn−2µn−1 + · · ·+ P1(∞)µ+ P0(∞) = 0. (3.2) To characterize the q-difference equation (3.1) is namely to determine coefficients pk,m. The number of coefficients pk,m in the equation (3.1) is (n+1)(n+2l+2) 2 . Through the following, total 3n+ 2l − 1 coefficients pk,m are determined: (i) We express parameters ai and bj as zeroes of the coefficients Pn(z) and P0(z): Pn(z) = pn,l(z − a1)(z − a2) · · · (z − al), P0(z) = p0,n+l(z − b1)(z − b2) · · · (z − bn+l). (3.3) The parameters ai and bj indicate poles of solutions of the equation (3.1). By the expres- sions (3.3), (2l + n) coefficients pn,m, p0,m′ (m ̸= l, m′ ̸= n+ l) are determined. (ii) We put parameters ci and dj as characteristic exponents at z = 0 and z = ∞ respectively z = 0: c1, c2, . . . , cn, z = ∞ : d1, d2, . . . , dn. (3.4) There is the following relation between parameters ai, bj , ck and dl (q-Fuchs’ relation) (−1)nq n(n−1) 2 n∏ j=1 dj n+l∏ i=1 bi = l∏ k=1 ak n∏ j=1 cj . (3.5) By the conditions (3.4) and (3.5), solving relations between roots and coefficients of char- acteristic equations (3.2) at z = 0 and z = ∞ for the equation (3.1), (2n− 1) coefficients pk,n+l−k pk′,0 (k ̸= n, k′ ̸= 0) are determined. Example 3.1. In case (n, l) = (2, 0), the equation (3.1) has 6 coefficients pk,m. From the conditions (i) and (ii), the equation (3.1) is characterized uniquely up to normalization. This equation is equivalent to the q-hypergeometric equation via a gauge transformation. A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 9 From the above (i) and (ii), the number of remain coefficients pk,m is 1 2(n−1)(n+2l−2), which is the number of accessary parameters. If z = a1 is an apparent singularity for the equation (3.1), namely all solutions of the equation (3.1) are regular at z = a1, we have the following relations: P0(a1/q) = 0, and f := P0(a1) P1(a1/q) = P1(a1) P2(a1/q) = · · · = Pn−1(a1) Pn(a1/q) , (3.6) where f is a parameter. The above equations (3.6) correspond to a non-logarithmic condition via a Laplace transformation z ↔ Tz. The relations (3.6) determine n coefficients pk,m. From now on, we derive a scalar q-difference equation from the matrix q-difference equation (2.18), (2.20) for an unknown function Ψ(z) = [Ψ1(z),Ψ2(z),Ψ3(z)] and its properties. Elimi- nating functions Ψ2(z) and Ψ3(z) in the equation (2.18), (2.20), we obtain the following third linear q-difference equation for Φ(z) := Ψ1(z): L(z) := P3(z)Φ ( q3z ) + P2(z)Φ ( q2z ) + P1(z)Φ(qz) + P0(z)Φ(z) = 0, (3.7) where P3(z) = p31(z − u), P2(z) = p22z 2 + p21z + p20, P1(z) = p13z 3 + p12z 2 + p11z + p10, P0(z) = −P3(qz)d1d2d3(z + c1)(z + c2)(z + c3). (3.8) Here, the coefficients pk,l (1 ≤ k ≤ 3, 0 ≤ l ≤ 3) in the polynomials Pk(z) (3.8) depend on parameters bj , ci, dj , and a variable u defined as the zero of P3(z). The variable u is expressed in terms of x, y as follows u = I1(x, y)I2(x, y) J1(x, y)J2(x, y) , (3.9) where I1(x, y) = y(b1(c2xy + c1y(c2 + x) + x) + c1c2(1− xy)), I2(x, y) = b2d3I1(x, y)− b2d2J1(x, y) + c3d3x(c1y + 1)(c2y + 1), J1(x, y) = b2(b1y + 1)(xy − 1), J2(x, y) = d2(J1(x, y)− x(c1y + 1)(c2y + 1))− d3I1(x, y). Explicit forms of the polynomials Pj(z) (0 ≤ j ≤ 3) (3.8) are given in appendix. Then we have Lemma 3.2. The equation L(z) = 0 (3.7) has the following properties: (i) it is a linear four term equation between Φ ( qjz ) (0 ≤ j ≤ 3) and its coefficients Pj(z) are polynomials for z of degree 4− j, (ii) a polynomial P0(z) has four zero points at z = −ci (1 ≤ i ≤ 3), u/q, (iii) the exponents of solutions Φ(z) are b1d1, qb2d2, qb3d3 (at z = 0) and d1, d2, d3 (at z = ∞), (iv) a point z = u such that P3(z) = 0 is an apparent singularity, namely we have v := P0(u) P1(u/q) = P1(u) P2(u/q) = P2(u) P3(u/q) . (3.10) Conversely, the equation L(z) = 0 (3.7) is uniquely characterized by these properties (i)–(iv) up to normalization. 10 K. Park Proof. The properties (i)–(iv) follows by computation through eliminating Ψ2(z), Ψ3(z) in (2.18), (2.20). The converse can be confirmed that coefficients Pj(z) are defined uniquely by (i)–(iv) up to a normalization. To see this, we consider the following equation which satisfies the properties (i), (ii): L′(z) = P ′ 3(z)Φ ( q3z ) + P ′ 2(z)Φ ( q2z ) + P ′ 1(z)Φ(qz) + P ′ 0(z)Φ(z) = 0, (3.11) P ′ 3(z) = p′31(z − u), P ′ 2(z) = p′22z 2 + p′21z + p′20, P ′ 1(z) = p′13z 3 + p′12z 2 + p′11z + p′10, P ′ 0(z) = p′04(z − u/q)(z + c1)(z + c2)(z + c3). From the property (iii), the condition of the exponents of solutions Φ(z) at z = 0 determines the coefficients p′31, p ′ 20, p ′ 10 and the condition of the exponents of solutions Φ(z) at z = ∞ deter- mines the coefficients p′22, p ′ 13. The remaining coefficients except for p′04 are determined by the property (iv). If we put the normalization factor p′04 as q 5uvc1c2d1d2d3, the function L′(z) (3.11) equals to the function L(z) (3.7). ■ In the following, viewing Φ ( qiz ) (0 ≤ i ≤ 3) as parameters, we regard the scalar q-difference equation L(z) = 0 (3.7) as an algebraic curve in variables (u, v) ∈ P1 × P1. We represent as the curve as P (u, v) = 0. The features of the curve are the following. Lemma 3.3. The algebraic curve P (u, v) = 0 has the following properties: (i) The polynomial P (u, v) has the following form: P (u, v) = ∑ 0≤j≤3 0≤i≤4−j ci,ju ivj , c0,0 := c0z 2Φ(qz). The coefficients ci,j depend on bi, ci, di, q, z, Φ ( qiz ) . (ii) It passes the following 8 points: (u, v) = (0, qb1d1), ( 0, q2b2d2 ) , ( 0, q2b3d3 ) , (qz,∞), (z, 0), (−c1, 0), (−c2, 0), (−c3, 0), and 3 points in the coordinate (r, s) = (u, v/u) (r, s) = ( ∞, q2d1 ) , ( ∞, q2d2 ) , ( ∞, q2d3 ) . (iii) At u = z the equation P (u, v) = 0 has the following property:( ∑ 0≤i≤3−j ci,j+1z i )∣∣∣ z→qz,Φ(qkz)→Φ(qk+1z) = ∑ 0≤i≤4−j ci,j(qz) i, j = 0, 1, 2. (3.12) Conversely, the equation P (u, v) = 0 is uniquely characterized by these properties (i)–(iii) up to normalization factor c0. Proof. The properties (i)–(iii) follow for the polynomial P (u, v). Conversely, we consider a poly- nomial P ′(u, v) := ∑ 0≤j≤3 0≤i≤4−j c′i,ju ivj , c′0,0(z) = c′0Φ(qz). A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 11 The polynomial P ′(u, v) has 14 coefficients. From the property (ii), 10 coefficients are described in terms of parameters bj , cj , dj and the coefficient c′0 as follows: P ′(u, v) = c′0zΦ(qz) (u− qz) b1b2b3d1d2d3q6 v3 + ( c′1,2u− c′0 zΦ(qz) b1b2b3d1d2d3q4 ( (d1 + d2 + d3)u 2(b2d2q + b3d3q + b1d1)z )) v2 + ( c′2,1u 2 + c′1,1u+ c′0zΦ(qz) (( 1 d1 + 1 d2 + 1 d3 ) u3 q2b1b2b3 − z ( 1 b1d1q + 1 b2d2q2 + 1 b3d3q2 ))) v + c′0zΦ(qz) (c1 + u)(c2 + u)(c3 + u)(z − u) b1b2b3 . The remaining 3 parameters c′1,1, c ′ 1,2, c ′ 2,1 are determined in terms of parameters bi, ci, di, q, z, Φ ( qiz ) , c′0 by the property (iii). Namely, the property (iii) gives 3 linear inhomogeneous equa- tions among c′1,1, c ′ 1,2, c ′ 2,1, c ′ 1,1 ∣∣ z→qz , c′1,2 ∣∣ z→qz , c′2,1 ∣∣ z→qz . Though these relations are apparently q-difference equations, we can solve them algebraically. For example, in the equation (3.12), we solve c′12 when j = 2. Then when j = 0, we solve c′21|z→qz,Φ(qkz)→Φ(qk+1z). And finally solving c′11 when j = 1, they algebraically can be solved. ■ Explicit forms of the coefficients ci,j(z) of the polynomial P (u, v) are in Appendix B. 3.1 Relations among pairs of variables (f, g), (x, y) and (u, v) Proposition 3.4. Under the relations among variables (f, g), (x, y) and (u, v): f = c1c2c3(b2 + x)(xy − 1) b1b2y(c1 + x)(c2 + x) + c3x(b2(1− xy) + c2xy + c1y(c2 + x) + x) , g = − y(b1b2 + c3x) c3(b2(1− xy) + x) + b1b2xy , (3.13) x = K1(u, v)K2(u, v) L1(u, v) , y = K3(u, v) L2(u, v) , (3.14) K1(u, v) = b2v ( v − q2(b2d2 + d3u) ) , K2(u, v) = b1b2 ( q2 ( d3u(d2q 2(c3 + u)− v)− d2v(b2 + u) ) + v2 ) + c1c2d3q 2 ( b2d2q 2(c3 + u)− c3v ) , K3(u, v) = b2d2q 2u ( d3q 2(c3 + u)− v ) , L1(u, v) = q2 ( b22d2 ( d2q 2 ( d3q 2(c1 + u)(c3 + u)− uv )( d3q 2(c2 + u)(c3 + u)− uv ) + v ( d3q 2v(b1(2c3 + u) + u(c3 + 2u))− d23q 4u(b1 + u)(c3 + u)− uv2 )) − b32d 2 2q 2v(d3q 2(b1 + u)(c3 + u)− uv)− b2c3d3v ( b1v ( v − d3q 2u ) + d2q 2 ( d3q 2(c2u+ c1(2c2 + u))(c3 + u)− (c1 + c2)uv )) + c1c2c 2 3d 2 3q 2v2 ) , L2(u, v) = c1c2d3q 2 ( b2d2q 2(c3 + u)− c3v ) + b1b2v ( v − q2(b2d2 + d3u) ) , the corresponding Lax equations (2.4) with (2.17), (2.18) with (2.20) and (2.18) with (3.9), (3.10) are equivalent. Conversely, such relations among (f, g), (x, y) and (u, v) are uniquely determined as (3.13), (3.14). Proof. Using the relation (3.13), we can check that the equation (2.4) with (2.17) is equivalent to (2.18) with (2.20). Similarly, using the relation (3.14), we can check that the equation (2.18) 12 K. Park with (2.20) is equivalent to (2.18) with (3.9), (3.10). The converse is obvious from the form of the Lax matrix. ■ 4 Continuous limit In this section, we describe a relation between our result and the result of Boalch [1]. In [1], a Lax pair for the additional-difference Painlevé equation with affine Weyl symmetry group of type E6 was described. The linear differential equation of the Lax pair is as follows d dz Ψ(z) = Ψ(z) ( Ab 1 z + Ab 2 z − 1 ) , Ab 3 := − ( Ab 1 +Ab 2 ) , (4.1) where the matrices Ab i (1 ≤ i ≤ 3) are 3× 3 matrices with different eigenvalues. We show that the linear q-difference equation (2.1) reduces to the equation (4.1) via a continu- ous limit q → 1. The equation (2.1) takes the following form after a scale transformation z → −z and gauge transformations (1− z)Ψ(qz) = Ψ(z)A(z), A(z) = b1d1 k1 v1 0 b2d2 k2 0 0 b3d3 − d1 0 0 v2 d2 0 v3 v4 d3  z, |A(z)| = d1d2d3(z − c1)(z − c2)(z − c3), (4.2) where kj (j = 1, 2) are constants. We put q = eh and consider the limit h → 0. We set bi = qβi , ci = qγi , di = qδi , 1 ≤ i ≤ 3, kj = hlj , j = 1, 2, vm = hum, 1 ≤ m ≤ 4, (4.3) where lj are constants. By using Taylor’s expansion for (4.2), (4.3) (l.h.s.) = (1− z)Ψ(z) + h(1− z)z d dz Ψ(z) +O ( h2 ) , (r.h.s.) = (1− z)Ψ(z) + hΨ(z) β1 − δ1(z − 1) l1 u1 −u2z β2 − δ2(z − 1) l2 −u3z −u4z β3 − δ3(z − 1) +O ( h2 ) , we find the following limit as h → 0: d dz Ψ(z) = Ψ(z) ( A1 z + A2 z − 1 ) , A1 = β1 + δ1 l1 u1 0 β2 + δ2 l2 0 0 β3 + δ3  , A2 = −β1 −l1 −u1 u2 −β2 −l2 u3 u4 −β3  , A3 := −(A1 +A2) = δ1 0 0 u2 δ2 0 u3 u4 δ3  , (4.4) where eigenvalues of the matrix A2 are γi (1 ≤ i ≤ 3) by the condition of the determinant of the matrix A(z) (4.2). Therefore, the linear q-difference equation (2.1) reduces to the equation (4.1) via a continuous limit q → 1. In the following, we consider a continuous limit q → 1 of the result in Section 2.1. In Section 2.1, through a compatibility condition of the equations (2.11), we derived a standard A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 13 q-Painlevé equation of type E6. We take the following equation as a deformation equation for the differential equation (4.4) which is rewritten version of (2.10): TΨ(z) = B(z)Ψ(z), B(z) = w1 w2 w3 0 0 0 0 0 0 + w4 0 0 1 0 0 w5 w6 w7  z, |B(z)| = (w3w6 − w2w7)z 2, T : (β1, β2, β3, γ1, γ2, γ3, δ1, δ2, δ3) → (β1 − 1, β2 + 1, β3, γ1, γ2, γ3, δ1 + 1, δ2, δ3 + 1). (4.5) Solving a compatibility condition for the equation (4.4), (4.5), we obtain the following additional- difference Painlevé equation of type E6 [6, 10]: (f + g) ( f + g ) = (g + γ1)(g + γ2)(g + γ3)(g + β1 + δ1 − δ2) (g + β2 + 1)(g + β3 − δ2 + δ3 + 1) , ( f + g )( f + g ) = ( f − γ1 )( f − γ2 )( f − γ3 )( f − β1 − δ1 + δ2 )( f − β1 + β2 − 1 )( f − β1 − δ1 + δ3 + 1 ) , where f = β3 + u1u4 l1 , g = l1l2 u1 − β2, and ∗ stands for T (∗). From the above, we derive the additional-difference Painlevé equation of type E6 solving a compatibility condition of the Lax pair via a continuous limit q → 1. A Deformations T1, T2 and T 3 2 on root variables In this appendix, we show that how the deformations T1, T2 and T 3 2 act on root variables. We consider a pair of root basis of {αi} (i = 0, 1, . . . , 6) and {δj} (j = 0, 1, 2) as symmetry type E (1) 6 and surface type A (1) 2 , respectively. α3α2α1 α4 α5 α6 α0 δ1 δ2 δ0 Figure 1. The Dynkin diagram of E (1) 6 and A (1) 2 . Pictures of a point configuration of the equation (2.12) and the configuration (2.24) are presented in Figures 2 and 3, respectively. From these pictures, we take a pair of root basis {αi} and {δj} as symmetry type E (1) 6 and surface type A (1) 2 as follows: α0 = E7 − E8, α1 = E6 − E5, α2 = H2 − E1 − E6, α3 = E1 − E2, α4 = E2 − E3, α5 = E3 − E4, α6 = H1 − E1 − E7, δ0 = H1 +H2 − E1 − E2 − E3 − E4, δ1 = H1 − E5 − E6, δ2 = H2 − E7 − E8, δ = α0 + α1 + 2α2 + 3α3 + 2α4 + α5 + 2α6 = δ0 + δ1 + δ2, (A.1) where we stand for δ as a null root. The choice of the above root basis is the same as in [6]. 14 K. Park g = 0 f = 0 fg = 1 p4 (c3, 1 c3 ) p3 (c2, 1 c2 ) p2 (c1, 1 c1 ) p1 ( b3d3d2 , d2 b3d3 )(0, 1 b2 ) p6 (0, d2 b1d1 ) p5 p7 (b3, 0) p8 ( qb3d3d1 , 0) Figure 2. A point configuration of the equation (2.12). y = 0 x = 0 xy = 1 q4 (−c2,− 1 c2 ) q3 (−c1,− 1 c1 ) q2 (− b1d1 d2 ,− d2 b1d1 ) q1 (− b1b2 c3 ,− c3 b1b2 ) (0,− 1 b1 ) q6 (0,− b2d2 c1c2d3 ) q5 q7 (−b2, 0) q8 (− b1b2d1 qc3d3 , 0) Figure 3. A point configuration (2.24). A.1 A deformation T1 on root variables We take variables ai (i = 0, 1, . . . , 6) as root variables attached to the root αi (A.1) associated with a point configuration in coordinate (f, g) (see Figure 2) a0 = qd3 d1 , a1 = b1d1 b2d2 , a2 = b2d2 b3d3 , a3 = b3d3 c1d2 , a4 = c1 c2 , a5 = c2 c3 , a6 = d2 d3 , (A.2) which satisfy q = a0a1a 2 2a 3 3a 2 4a5a 2 6. Then we have the following statement. Proposition A.1. The action T1 (2.10) on the root variables ai in (A.2) is given by the trans- lation T1(a0, a1, a2, a3, a4, a5, a6) = (a0, a1, qa2, a3, a4, a5, a6/q). (A.3) Proof. Applying (2.10), we obtain the desired result (A.3). ■ A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 15 A.2 Deformations T2 and T 3 2 on root variables The deformation T2 (2.19) is not a translation on parameters bi, ci, dj but a deformation T 3 2 gives a translation on them T 3 2 : (b1, b2, b3, c1, c2, c3, d1, d2, d3) → ( qb1, qb2, b3, c1, c2, q 3c3, d1 q , d2 q , d3 q ) . (A.4) We show that how the deformations T2 (2.19) and T 3 2 (A.4) on root variables. We take variables a′i (i = 0, 1, . . . , 6) as root variables attached to the root αi (A.1) associated with a point configuration in coordinate (x, y) (see Figure 3) a′0 = qc3d3 b1d1 , a′1 = b1b2d2 c1c2d3 , a′2 = b2 c3 , a′3 = c3d1 b2d2 , a′4 = c1d2 b1d1 , a′5 = c2 c1 , a′6 = b1 c3 , (A.5) which satisfy q = a′0a ′ 1a ′ 2 2a′3 3a′4 2a′5a ′ 6 2. Then we have the following statement. Proposition A.2. The actions T2 (2.19) and T 3 2 on the root variables a′i in (A.5) are given as follows: T2(a ′ 0, a ′ 1, a ′ 2, a ′ 3, a ′ 4, a ′ 5, a ′ 6) = ( q a′6 , a′0 2 a′1a ′ 2a ′ 3a ′ 6 2 , 1 a′0a ′ 3a ′ 6 , q a′2 , a′0a ′ 2a ′ 3a ′ 4a ′ 6 q , a′5, a′2a ′ 3a ′ 6 q ) , T 3 2 (a ′ 0, a ′ 1, a ′ 2, a ′ 3, a ′ 4, a ′ 5, a ′ 6) = ( a′0q 2, a′1q 3, a′2 q2 , a′3q 2, a′4 q , a′5, a′6 q2 ) . (A.6) Proof. Applying (2.19) and (A.4), we obtain the desired results (A.6). ■ B Explicit forms of coefficients in Section 3 In this appendix, we give explicit forms of Pj(z) (3.8) and the coefficients cij (0 ≤ j ≤ 3, 0 ≤ i ≤ 4− j) of the polynomial P (u, v) in variables u and v. Explicit forms of Pj(z) (3.8) are as follows: P3(z) = p31(z − u), P2(z) = p22z 2 + p21z + p20, P1(z) = p13z 3 + p12z 2 + p11z + p10, P0(z) = −P3(qz)d1d2d3(z + c1)(z + c2)(z + c3), where the coefficients pj,k (1 ≤ j ≤ 3, 0 ≤ k ≤ 3) are p31 = −q2uvc1c2, p22 = q4uvc1c2(d1 + d2 + d3), p21 = −quvc1c2 ( ud3q 3 + (qu− b2)d2q 2 − b3d3q 2 − vq + v + ( q3u− qb1 ) d1 ) , p20 = −q2u2vc1c2(b1d1 + qb2d2 + qb3d3), p13 = −q5uvc1c2(d1d2 + d2d3 + d3d1), p12 = −q ( ub1b2b3(u+ c2)d1d2d3q 5 + uc21c2(u+ c2)d1d2d3q 5 + c1 ( u2c22d1d2d3q 5 + ub1b2b3d1d2d3q 5 − c2 ( qd1 ( qu ( qud2 ( − ud3q 2 + vq + v ) + v(q(q + 1)ud3 − v) ) 16 K. Park − b1 ( v − q2b2d2 )( v − q2b3d3 )) + v ( d2 ( u(q(q + 1)ud3 − v) + b2 ( q2b3d3 − v )) q2 + v ( v − q2(u+ b3)d3 ))))) , p11 = u ( ub1b2b3(u+ c2)d1d2d3q 6 + uc21c2(u+ c2)d1d2d3q 6 + c1 ( u2c22d1d2d3q 6 + ub1b2b3d1d2d3q 6 − c2 ( qd1 ( b1 ( b2d2 ( − b3d3q 3 + vq + v ) q2 + v ( q2(q + 1)b3d3 − v )) − qu ( v − q2ud2 )( v − q2ud3 )) + v ( d2 ( u ( q2ud3 − v ) + b2 ( q2(q + 1)b3d3 − v )) q2 + v ( v − q2(u+ b3)d3 ))))) , p10 = q3u2vc1c2(qb2b3d2d3 + b1b2d1d2 + b1b3d1d3). We give also explicit forms of the coefficients cij (0 ≤ j ≤ 3, 0 ≤ i + j ≤ 4) of the polynomial P (u, v) in variables u and v: c01 = −c0 z2 q2 ( q b1d1 + 1 b2d2 + 1 b3d3 ) Φ(qz), c02 = c0 z2(b1d1 + qb2d2 + qb3d3)Φ(qz) q4b1b2b3d1d2d3 , c03 = −c0 z2Φ(qz) q5b1b2b3d1d2d3 , c10 = c0 ( z2 c1 + z2 c2 + z2 c3 − z ) Φ(qz), c11 = −c0z ( (c1 + z)(c2 + z)(c3 + z)Φ(z) b1b2b3q − ( 1 b1d1q + 1 b3d3q2 + 1 b2d2q2 + 1 b1d1q2 + 1 b3d3q3 + 1 b2d2q3 + z2 b1b2b3d3q + z2 b1b2b3d2q + z2 b1b2b3d1q ) Φ(qz) + ( d1 + d2 + d3 q2b1b2b3d1d2d3 z − b1d1 + qb2d2 + b3d3 q4b1b2b3d1d2d3 ) Φ ( q2z ) − 1 b1b2b3d1d2d3q4 Φ ( q3z )) , c12 = c0z (( d1 + d2 + d3 q3b1b2b3d1d2d3 z − b1d1 + qb2d2 + qb3d3 q5b1b2b3d1d2d3 ) Φ(qz) + (−1 + q) q5b1b2b3d1d2d3 Φ ( q2z )) , c13 = c0z Φ(qz) q6b1b2b3d1d2d3 , c20 = c0z (( 1 c1c2 + 1 c2c3 + 1 c3c1 ) z − ( 1 c1 + 1 c2 + 1 c3 )) Φ(qz), c21 = c0 ( (c1 + z)(c2 + z)(c3 + z) b1b2b3q2 Φ(z) − ( (1 + q)(d1d2 + d2d3 + d3d1) q2b1b2b3d1d2d3 + 1 q3b3d3 + 1 q3b2d2 + 1 q2b1d1 ) Φ(qz) + ( d1 + d2 + d3 q2b1b2b3d1d2d3 z + b1d1 + qb2d2 + qb3d3 q4b1b2b3d1d2d3 ) Φ ( q2z )) , c22 = −c0 (d1 + d2 + d3)Φ(qz) q4b1b2b3d1d2d3 zΦ(qz), c30 = c0 ( z2 c1c2c3 − ( 1 c1c2 + 1 c2c3 + 1 c3c1 ) z ) Φ(qz), c31 = (d1d2 + d2d3 + d3d1)z b1b2b3d1d2d3q2 Φ(qz), c40 = −c0 zΦ(qz) b1b2b3 . A 3× 3 Lax Form for the q-Painlevé Equation of Type E6 17 Acknowledgements The author would like to express her gratitude to Professor Yasuhiko Yamada for valuable sug- gestions and encouragement. And the author is grateful to referees for giving helpful comments to improve the manuscript. She also thanks supports from JSPS KAKENHI Grant Numbers 17H06127 and 26287018 for the travel expenses in accomplishing this study. References [1] Boalch P., Quivers and difference Painlevé equations, in Groups and Symmetries,CRM Proc. Lect. Notes, Vol. 47, American Mathematical Society, Providence, RI, 2009, 25–51, arXiv:0706.2634. [2] Dzhamay A., Knizel A., q-Racah ensemble and q-P(E (1) 7 /A (1) 1 ) discrete Painlevé equation, Int. Math. Res. Not. 2020 (2020), 9797–9843, arXiv:1903.06159. [3] Dzhamay A., Takenawa T., On some applications of Sakai’s geometric theory of discrete Painlevé equations, SIGMA 14 (2018), 075, 20 pages, arXiv:1804.10341. [4] Grammaticos B., Ramani A., The hunting for the discrete Painlevé equations, Regul. Chaotic Dyn. 5 (2000), 53–66. [5] Jimbo M., Sakai H., A q-analog of the sixth Painlevé equation, Lett. Math. Phys. 38 (1996), 145–154, arXiv:chao-dyn/9507010. [6] Kajiwara K., Noumi M., Yamada Y., Geometric aspects of Painlevé equations, J. Phys. A 50 (2017), 073001, 164 pages, arXiv:1509.08186. [7] Knizel A., Moduli spaces of q-connections and gap probabilities, Int. Math. Res. Not. 2016 (2016), 6921– 6954, arXiv:1506.06718. [8] Park K., A certain generalization of q-hypergeometric functions and their related connection preserving deformation II, Funkcial. Ekvac. 65 (2022), 311–328, arXiv:2005.04992. [9] Ramani A., Grammaticos B., Hietarinta J., Discrete versions of the Painlevé equations, Phys. Rev. Lett. 67 (1991), 1829–1832. [10] Ramani A., Grammaticos B., Tamizhmani T., Tamizhmani K.M., Special function solutions of the discrete Painlevé equations, Comput. Math. Appl. 42 (2001), 603–614. [11] Sakai H., A q-analog of the Garnier system, Funkcial. Ekvac. 48 (2005), 273–297. [12] Sakai H., Lax form of the q-Painlevé equation associated with the A (1) 2 surface, J. Phys. A 39 (2006), 12203–12210. [13] Suzuki T., A q-analogue of the Drinfeld–Sokolov hierarchy of type A and q-Painlevé system, in Algebraic and Analytic Aspects of Integrable Systems and Painlevé Equations, Contemp. Math., Vol. 651, American Mathematical Society, Providence, RI, 2015, 25–38, arXiv:1105.4240. [14] Suzuki T., A Lax formulation of a generalized q-Garnier system, Math. Phys. Anal. Geom. 24 (2021), 38, 12 pages, arXiv:2103.15336. [15] Witte N.S., Ormerod C.M., Construction of a Lax pair for the E (1) 6 q-Painlevé system, SIGMA 8 (2012), 097, 27 pages, arXiv:1207.0041. [16] Yamada Y., Lax formalism for q-Painlevé equations with affine Weyl group symmetry of type E (1) n , Int. Math. Res. Not. 2011 (2011), 3823–3838, arXiv:1004.1687. https://doi.org/10.1090/crmp/047/04 https://arxiv.org/abs/0706.2634 https://doi.org/10.1093/imrn/rnz211 https://doi.org/10.1093/imrn/rnz211 https://arxiv.org/abs/1903.06159 https://doi.org/10.3842/SIGMA.2018.075 https://arxiv.org/abs/1804.10341 https://doi.org/10.1007/BF00398316 https://arxiv.org/abs/chao-dyn/9507010 https://doi.org/10.1088/1751-8121/50/7/073001 https://arxiv.org/abs/1509.08186 https://doi.org/10.1093/imrn/rnv366 https://arxiv.org/abs/1506.06718 https://doi.org/10.1619/fesi.65.311 https://arxiv.org/abs/2005.04992 https://doi.org/10.1103/PhysRevLett.67.1829 https://doi.org/10.1016/S0898-1221(01)00180-8 https://doi.org/10.1619/fesi.48.273 https://doi.org/10.1088/0305-4470/39/39/S13 https://doi.org/10.1090/conm/651/13037 https://doi.org/10.1090/conm/651/13037 https://arxiv.org/abs/1105.4240 https://doi.org/10.1007/s11040-021-09412-3 https://arxiv.org/abs/2103.15336 https://doi.org/10.3842/SIGMA.2012.097 https://arxiv.org/abs/1207.0041 https://doi.org/10.1093/imrn/rnq232 https://doi.org/10.1093/imrn/rnq232 https://arxiv.org/abs/1004.1687 1 Introduction 2 A 3 times 3 matrix Lax for 2.1 The deformation T_1 2.2 The deformation T_2 3 A scalar equation related to the q-E_6^(1) 3.1 Relations among pairs of variables (f, g), (x, y) and (u, v) 4 Continuous limit A Deformations T_1, T_2 and T_2^3 on root variables A.1 A deformation T_1 on root variables A.2 Deformations T_2 and T_2^3 on root variables B Explicit forms of coefficients in Section 3 References
id nasplib_isofts_kiev_ua-123456789-212037
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
issn 1815-0659
language English
last_indexed 2026-03-15T14:41:36Z
publishDate 2023
publisher Інститут математики НАН України
record_format dspace
spelling Park, Kanam
2026-01-23T10:10:27Z
2023
A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆. Kanam Park. SIGMA 19 (2023), 094, 17 pages
1815-0659
2020 Mathematics Subject Classification: 14H70; 34M56; 39A13
arXiv:2211.16706
https://nasplib.isofts.kiev.ua/handle/123456789/212037
https://doi.org/10.3842/SIGMA.2023.094
For the -Painlevé equation with affine Weyl group symmetry of type ⁽¹⁾₆, a 2 × 2 matrix Lax form and a second-order scalar lax form were known. We give a new 3 × 3 matrix Lax form and a third order scalar equation related to it. Continuous limit is also discussed.
The author would like to express her gratitude to Professor Yasuhiko Yamada for valuable suggestions and encouragement. And the author is grateful to referees for giving helpful comments to improve the manuscript. She also thanks supports from JSPS KAKENHI Grant Numbers 17H06127 and 26287018 for the travel expenses in accomplishing this study.
en
Інститут математики НАН України
Symmetry, Integrability and Geometry: Methods and Applications
A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆
Article
published earlier
spellingShingle A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆
Park, Kanam
title A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆
title_full A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆
title_fullStr A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆
title_full_unstemmed A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆
title_short A 3 × 3 Lax Form for the -Painlevé Equation of Type ₆
title_sort 3 × 3 lax form for the -painlevé equation of type ₆
url https://nasplib.isofts.kiev.ua/handle/123456789/212037
work_keys_str_mv AT parkkanam a33laxformforthepainleveequationoftype6
AT parkkanam 33laxformforthepainleveequationoftype6