On Non-Point Invertible Transformations of Difference and Differential-Difference Equations

Non-point invertible transformations are completely described for difference equations on the quad-graph and for their differential-difference analogues. As an illustration, these transformations are used to construct new examples of integrable equations and autotransformations of the Hietarinta equ...

Повний опис

Збережено в:
Бібліографічні деталі
Опубліковано в: :Symmetry, Integrability and Geometry: Methods and Applications
Дата:2010
Автор: Startsev, S.Ya.
Формат: Стаття
Мова:English
Опубліковано: Інститут математики НАН України 2010
Онлайн доступ:https://nasplib.isofts.kiev.ua/handle/123456789/146524
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:On Non-Point Invertible Transformations of Difference and Differential-Difference Equations / S. Ya. Startsev // Symmetry, Integrability and Geometry: Methods and Applications. — 2010. — Т. 6. — Бібліогр.: 20 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-146524
record_format dspace
spelling Startsev, S.Ya.
2019-02-09T19:51:14Z
2019-02-09T19:51:14Z
2010
On Non-Point Invertible Transformations of Difference and Differential-Difference Equations / S. Ya. Startsev // Symmetry, Integrability and Geometry: Methods and Applications. — 2010. — Т. 6. — Бібліогр.: 20 назв. — англ.
1815-0659
2010 Mathematics Subject Classification: 39A14; 34K99; 37K35
DOI:10.3842/SIGMA.2010.092
https://nasplib.isofts.kiev.ua/handle/123456789/146524
Non-point invertible transformations are completely described for difference equations on the quad-graph and for their differential-difference analogues. As an illustration, these transformations are used to construct new examples of integrable equations and autotransformations of the Hietarinta equation.
This work is partially supported by the Russian Foundation for Basic Research (grant number 10-01-00088-a).
en
Інститут математики НАН України
Symmetry, Integrability and Geometry: Methods and Applications
On Non-Point Invertible Transformations of Difference and Differential-Difference Equations
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title On Non-Point Invertible Transformations of Difference and Differential-Difference Equations
spellingShingle On Non-Point Invertible Transformations of Difference and Differential-Difference Equations
Startsev, S.Ya.
title_short On Non-Point Invertible Transformations of Difference and Differential-Difference Equations
title_full On Non-Point Invertible Transformations of Difference and Differential-Difference Equations
title_fullStr On Non-Point Invertible Transformations of Difference and Differential-Difference Equations
title_full_unstemmed On Non-Point Invertible Transformations of Difference and Differential-Difference Equations
title_sort on non-point invertible transformations of difference and differential-difference equations
author Startsev, S.Ya.
author_facet Startsev, S.Ya.
publishDate 2010
language English
container_title Symmetry, Integrability and Geometry: Methods and Applications
publisher Інститут математики НАН України
format Article
description Non-point invertible transformations are completely described for difference equations on the quad-graph and for their differential-difference analogues. As an illustration, these transformations are used to construct new examples of integrable equations and autotransformations of the Hietarinta equation.
issn 1815-0659
url https://nasplib.isofts.kiev.ua/handle/123456789/146524
citation_txt On Non-Point Invertible Transformations of Difference and Differential-Difference Equations / S. Ya. Startsev // Symmetry, Integrability and Geometry: Methods and Applications. — 2010. — Т. 6. — Бібліогр.: 20 назв. — англ.
work_keys_str_mv AT startsevsya onnonpointinvertibletransformationsofdifferenceanddifferentialdifferenceequations
first_indexed 2025-11-26T00:08:35Z
last_indexed 2025-11-26T00:08:35Z
_version_ 1850589437566648320
fulltext Symmetry, Integrability and Geometry: Methods and Applications SIGMA 6 (2010), 092, 14 pages On Non-Point Invertible Transformations of Difference and Differential-Difference Equations Sergey Ya. STARTSEV Ufa Institute of Mathematics, Russian Academy of Sciences, 112 Chernyshevsky Str., Ufa, 450077, Russia E-mail: startsev@anrb.ru Received October 04, 2010, in final form December 03, 2010; Published online December 11, 2010 doi:10.3842/SIGMA.2010.092 Abstract. Non-point invertible transformations are completely described for difference equations on the quad-graph and for their differential-difference analogues. As an illustra- tion, these transformations are used to construct new examples of integrable equations and autotransformations of the Hietarinta equation. Key words: non-point transformation; Darboux integrability; discrete Liouville equation; higher symmetry 2010 Mathematics Subject Classification: 39A14; 34K99; 37K35 1 Introduction The present paper is devoted to invertible transformations for both discrete equations of the form ui+1,j+1 = F (ui,j , ui+1,j , ui,j+1), and “semi-discrete” chains of the differential equations (ui+1)x = F (x, ui, ui+1, (ui)x). Here i and j are integers, x is a continuous variable, u is a function of i, j and i, x for the first and the second equation, respectively. From now on, we shall omit i and j for brevity and, in particular, write the above equations in the form u1,1 = F (u, u1,0, u0,1) (1.1) and (u1)x = F (x, u, u1, ux). (1.2) We assume that FuFu1,0Fu1,0 6= 0 for equation (1.1) and Fux 6= 0 for equation (1.2). These conditions allows us to rewrite equation (1.1) in any of the following forms u−1,−1 = F (u, u−1,0, u0,−1), (1.3) u1,−1 = F̂ (u, u1,0, u0,−1), (1.4) u−1,1 = F̃ (u, u−1,0, u0,1), (1.5) and equation (1.2) – in the form (u−1)x = F̃ (x, u, u−1, ux). (1.6) mailto:startsev@anrb.ru http://dx.doi.org/10.3842/SIGMA.2010.092 2 S.Ya. Startsev Therefore, all “mixed shifts” um,n := ui+m,j+n (for both positive and negative non-zero n and m) can be expressed in terms of dynamical variables uk,0, u0,l by virtue of equations (1.1), (1.3)–(1.5). (A more detailed explanation of the dynamical variables, the notation um,n and the recursive procedure of the mixed shift elimination can be found, for example, in [11, 10].) Analogously, u (n) m := ∂nui+m/∂x n for any non-zero m ∈ Z and n ∈ N can be expressed in terms of x and dynamical variables ul := ui+l, u(k) := ∂kui/∂x k by virtue of equations (1.2), (1.6). The notation g[u] means that the function g depends on a finite number of the dynamical variables (and x if we consider equation (1.2)). The considerations in this paper are local (for example, we use the local implicit function theorem to obtain (1.3)–(1.6)) and, for simplicity, all functions are assumed to be locally analytical. In addition to the point transformations v = g(u), some of the equations (1.1) and (1.2) admit non-point transformations v = g[u] which are invertible in the sense of [16]. For example, the differential substitutions v = ux − sinu 2 (1.7) maps solutions of the differential-difference sine-Gordon equation [7, 14] (u1)x − sinu1 = ux + sinu (1.8) into solutions of the equation (v1 − v)x√ 1− (v1 − v)2 = ±(v1 + v), (1.9) which is a semi-discrete version of the complex sine-Gordon equation. Here the sign of the right- hand side of equation (1.9) coincides with the sign of the cosu value1. Indeed, v1 = (ux+sinu)/2 follows from equation (1.8) and, together with (1.7), gives us ux = v1 + v, sinu = v1 − v =⇒ (v1 − v)x = ux cosu = ±(v1 + v) √ 1− (v1 − v)2. The inverse transformation can be found in [12]: the formula u = π 2 ± (arcsin(v1− v)− π 2 ) maps any real solution of equation (1.9) into a solution of equation (1.8). This example belongs to the following class of non-point invertible transformations intro- duced in [17]. Let functions ϕ(x, y, z), ψ(x, y, z) satisfy the condition ϕyψz − ϕzψy 6= 0 and equation (1.2) can be written in the form ϕ(x, u1, (u1)x) = ψ(x, u, ux). (1.10) Then we rewrite (1.10) in the form of the system v = ϕ(x, u, ux), v1 = ψ(x, u, ux), (1.11) express u, ux in terms of v, v1 from (1.11) and obtain u = p(x, v, v1), ux = q(x, v, v1). (1.12) The system (1.12) is equivalent to the equation Dx(p(x, v, v1)) = q(x, v, v1), (1.13) 1A local transformation of an equation may, generally speaking, generate different equations for different domains of the “jet space”. This is true for both point and non-point local transformations. On Non-Point Invertible Transformations 3 where Dx denotes the total derivative with respect to x. The substitution v = ϕ(x, u, ux) maps solutions of (1.10) into solutions of (1.13) and the transformation u = p(x, v, v1) maps solutions of (1.13) back into solutions of (1.10). It is easy to see that the same scheme can be applied to the pure discrete equations of the form ϕ(u0,1, u1,1) = ψ(u, u1,0), (1.14) where ϕ(y, z) and ψ(y, z) are functionally independent. Indeed, expressing u and u1,0 from v = ϕ(u, u1,0), v0,1 = ψ(u, u1,0), (1.15) we obtain u = p(v, v0,1), u1,0 = q(v, v0,1) (1.16) and rewrite (1.16) in the form of the equivalent equation p(v1,0, v1,1) = q(v, v0,1). (1.17) Thus, the transformation v = ϕ(u, u1,0) maps solutions of (1.14) into solutions of (1.17) and the inverse transformation u = p(v, v0,1) maps solutions of (1.17) back into solutions of (1.14). The transformations (1.14)–(1.17) were, in fact, used in [18] without explicit formulation of the above scheme. The main result of this paper is the proof of the following statement: any invertible trans- formation of equations (1.1), (1.2) is a composition of shifts, point transformations and trans- formations (1.10)–(1.13), (1.14)–(1.17). Roughly speaking, equations (1.1) and (1.2) have no non-point invertible transformations other than (1.14)–(1.17) and (1.10)–(1.13), respectively. The proof is similar to that was used in [16] for continuous equations (hyperbolic PDEs). The invertible transformations allow us to obtain objects associated with integrability of equations (1.13), (1.17) (such as conservation laws and higher symmetries) from the correspon- ding objects of equations (1.10), (1.14) because we can express shifts and derivatives of u in terms of shifts and derivatives of v. Therefore, the invertible transformations may be useful for constructing new examples of integrable equations of the form (1.1), (1.2). To illustrate this, in Section 4 we construct Darboux integrable equations related via invertible transformations to difference and differential-difference analogues of the Liouville equation. In addition, an example of constructing an equation possessing the higher symmetries is contained at the end of Section 2. In this section we also demonstrate that the scheme (1.14)–(1.17) generates autotransformations of the Hietarinta equation. 2 Invertible transformations of discrete equations We let Ti and Tj denote the operators of the forward shifts in i and j by virtue of equation (1.1). These operators are defined by the following rules: Ti(f(a, b, c, . . . )) = f(Ti(a), Ti(b), Ti(c), . . . ) and Tj(f(a, b, c, . . . )) = f(Tj(a), Tj(b), Tj(c), . . . ) for any function f ; Ti(um,0) = um+1,0 and Tj(u0,n) = u0,n+1; Ti(u0,n) = Tn−1 j (F ) for positive n and Ti(u0,n) = Tn+1 j (F̂ ) for negative n, Tj(um,0) = Tm−1 i (F ) for positivem and Tj(um,0) = Tm+1 i (F̃ ) for negativem (i.e. mixed variables u1,n and um,1 are expressed in terms of the dynamical variables by virtue of equations (1.1), (1.4), (1.5)). The inverse (backward) shift operators T−1 i and T−1 j are defined in the similar way. 4 S.Ya. Startsev Definition 1. We say that a transformation v = f [u] maps the equation (1.1) into an equation v1,1 = G(v, v1,0, v0,1) if TiTj(f) = G(f, Ti(f), Tj(f)). (2.1) Definition 2. A transformation v = f [u] of equation (1.1) is called invertible if any of the dynamical variables u, uk,0, u0,l, k, l ∈ Z, can be expressed as a function of a finite subset of the variables v := f, vr,0 := T r i (f), v0,s := T s j (f), r, s ∈ Z. (2.2) We exclude all mixed variables vr,s, rs 6= 0, from (2.2) because we consider only the cases when the transformation maps (1.1) into an equation of the form v1,1 = G(v, v1,0, v0,1), GvGv1,0Gv0,1 6= 0 (2.3) and the mixed variables can be expressed in terms of (2.2) by virtue of this equation. It is easy to see that any shift w = vr,s maps equation (2.3) into equation (2.3) again and the composition of the shift and an invertible transformation v = f [u] is invertible too. This leads to the following Definition 3. Transformations v = f [u] and w = g[u] are called equivalent if there exist r, s ∈ Z such that g = T s i T r j (f). Proposition 1. Let v = g[u] be an invertible transformation that maps equation (1.1) into equation (2.3). Then this transformation is equivalent to either a transformation of the form w = f(u, u1,0, u2,0, . . . , um,0), (2.4) or a transformation of the form w = f(u, u0,1, u0,2, . . . , u0,n). (2.5) Proof. The transformation is equivalent to that of the form v = h(u, u1,0, . . . , uk,0, u0,1, . . . , u0,l) (2.6) because we can eliminate “negative” variables ur,0, u0,s, r, s < 0 from the transformation by shifts of g. We can express u as u = P (va,0, va+1,0, . . . , vb,0, v0,c, v0,c+1, . . . , v0,d) := P ( T a i (h), T a+1 i (h), . . . , T b i (h), T c j (h), T c+1 j (h), . . . , T d j (h) ) (2.7) if the transformation is invertible. Differentiating equation (2.7) with respect to uk+b,0, we obtain Pvb,0 T b i (huk,0 ) = 0 ⇒ Pvb,0 = 0 if b, k > 0. The analogous reasoning gives Pv0,d = 0 if d, l > 0. Thus, b, d ≤ 0 if kl 6= 0. Let ( T−1 i (h) ) u−1,0 ( T−1 j (h) ) u0,−1 6= 0. Then (T a i (h))ua,0 ( T c j (h) ) u0,c 6= 0 for any negative a and c, and we obtain Pva,0 = Pv0,c = 0 by differentiating equation (2.7) with respect to ua,0 and u0,c. Therefore, either a = c = b = d = 0 (i.e. u = P (h) that is possible only if k = l = 0) or ( T−1 i (h) ) u−1,0 ( T−1 j (h) ) u0,−1 = 0. The latter equality means that either T−1 i (h) = h̃(u, u1,0, . . . , uk−1,0, u0,1, . . . , u0,l) or T−1 j (h) = h̃(u, u1,0, . . . , uk,0, u0,1, . . . , u0,l−1), i.e. any invertible transformation of the form (2.6) with kl 6= 0 is equivalent to a transformation ṽ = h̃(u, u1,0, . . . , uk̃,0, u0,1, . . . , u0,l̃) such that k̃l̃ < kl. Applying this conclusion several times, we obtain that (2.6) is equivalent to a transformation w = f(u, u1,0, . . . , um,0, u0,1, . . . , u0,n) with mn = 0. � On Non-Point Invertible Transformations 5 Definition 4. A transformation is called non-point if this transformation is not equivalent to any point transformation of the form w = g(u). Because the transformations v = f(um,0) and v = f(u0,n) are equivalent to the point trans- formation w = f(u), a transformation of the form (2.4) or (2.5) is non-point only if f depends on more than one variable. We use only this property of the non-point transformations in the proof of the following Theorem 1. Let a non-point invertible transformation of the form (2.4) map equation (1.1) into equation (2.3). Then equation (1.1) can be written in the form ϕ(u0,1, u1,1) = ψ(u, u1,0), where ϕ(y, z) and ψ(y, z) are functionally independent, and the transformation is equivalent to the composition of the invertible transformation w = ϕ(u, u1,0) and an invertible transformation of the form v = h(w,w1,0, w2,0, . . . , wm−1,0). In particular, any non-point invertible transformation of the form v = f(u, u1,0) is equivalent to the composition of the transformation w = ϕ(u, u1,0) and a point transformation v = h(w). Proof. If fu = 0 and s is the smallest integer for which fus,0 6= 0, then the equivalent transfor- mation v = T−s i (f [u]) depends on u. Therefore, we can, without loss of generality, assume that fu 6= 0. We also can write ul,0 = Pl(va,0, va+1,0, . . . , vb,0, v0,c, v0,c+1, . . . , v0,d) = Pl ( T a i (f), T a+1 i (f), . . . , T b i (f), T c j (f), T c+1 j (f), . . . , T d j (f) ) , l = 0,m because the transformation is invertible. Here the notation l = 0,m means that l runs over all integers from 0 to m. Differentiating these equalities with respect to ua,0, we obtain (Pl)va,0T a i (fu) = 0 ⇒ (Pl)va,0 = 0 if a < 0. The similar reasoning gives (Pl)vb,0 = 0 if b > 0. Thus, ul,0 = Pl ( T c j (f), T c+1 j (f), . . . , T d j (f) ) , l = 0,m. Let c < 0 and s be the biggest negative integer such that (T s j (f))u0,−1 6= 0. If s ≥ c, then T c j (f) depends on u0,c−s−1 and (Pl)u0,c−s−1 = (Pl)v0,c(T c j (f))u0,c−s−1 = 0 ⇒ (Pl)v0,c = 0. Hence s < c, i.e. (T r j (f))u0,−1 = 0 for all r ≥ c. This implies T c j (f) = g(u, u1,0, . . . , um,0) and ul,0 = Pl ( g, Tj(g), . . . , T d̃ j (g) ) , l = 0,m. (2.8) If c ≥ 0, then equations (2.8) holds too, with g = f and d̃ = d. Repeating the above argumentation, we prove that (T r j (g))u0,1 = 0 for all r ≤ d̃. Let us consider the operators X = T−1 j ∂u0,1Tj (cf. [4]) and Y = [∂u0,−1 , X], where ∂z := ∂ ∂z . These operators have the form X = ∂u + m∑ l=1 ξl∂ul,0 , Y = m∑ l=1 νl∂ul,0 for functions of u, u1,0, . . . , um,0. According to equation (2.8), the set {g, Tj(g), . . . , T d̃ j (g)} must contain m + 1 functionally independent functions because ul,0, l = 0,m, are functionally in- dependent. Hence the system X(z) = 0, Y (z) = 0 has m functionally independent solutions depending on u, u1,0, . . . , um,0 and the vectors (1, ξ1, . . . , ξm), (0, ν1, . . . , νm) must be collinear. The latter is possible only if νl = 0 for all l = 0,m. In particular, ν1 = [T−1 j (Fu0,1)]u0,−1 = 0 ⇒ T−1 j (Fu0,1) = α(u, u1,0) ⇒ Fu0,1 = Tj(α) = α(u0,1, F ) 6 S.Ya. Startsev ⇒ Fuu0,1 = αu1,0(u0,1, F )Fu, Fu1,0u0,1 = αu1,0(u0,1, F )Fu1,0 ⇒ (ln(Fu1,0)− ln(Fu))u0,1 = 0 ⇒ Fu1,0 − β(u, u1,0)Fu = 0 ⇒ F = E(ψ(u, u1,0), u0,1), where ψ is a solution of the equation ψu1,0 − β(u, u1,0)ψu = 0. Thus, equation (1.1) can be written in the form (1.14). We can express g in terms of u, ϕ(u, u1,0), ϕ(u1,0, u2,0), . . . , ϕ(um−1,0, um,0): g = h(u, ϕ(u, u1,0), ϕ(u1,0, u2,0), . . . , ϕ(um−1,0, um,0)). It is proved above that X(g) = 0. Taking this fact into account, we obtain hu = 0 because X(g) = X(h) = T−1 j [hu(u0,1, ψ(u, u1,0), . . . , ψ(um−1,0, um,0))] = hu. This means that the trans- formation (2.4) is equivalent to the composition of the transformation w = ϕ(u, u1,0) and the transformation v = h(w,w1,0, w2,0, . . . , wm−1,0). The latter transformation is invertible because w = ϕ(P0, P1) = P̃0 ( g, Tj(g), . . . , T d̃ j (g) ) = P̃0 ( h, Tj(h), . . . , T d̃ j (h) ) by virtue of equation (2.8). The expressions for other dynamical variables can be obtained by the formulas w0,r = T r j (P̃0) and ws,0 = T s i (P̃0). Let ϕ and ψ be functionally dependent. Under this assumption equation (1.14) has the form ϕ(u0,1, u1,1) = E(ϕ(u,u1,0)) and all functions T r j (g) can be expressed in terms of ϕ(u, u1,0), ϕ(u1,0, u2,0), . . . , ϕ(um−1,0, um,0) (Tj(g) = h(E(ϕ(u, u1,0)), . . . , E(ϕ(um−1,0, um,0)) and so on). Hence the set {g, Tj(g), . . . , T d̃ j (g)} contains no more than m functionally independent functions. But we prove above that this set must contain m+ 1 functionally independent functions if the transformation is invertible. Therefore, ϕ and ψ must be functionally independent if equa- tion (1.14) admits an invertible transformation of the form v = f(u, u1,0, u2,0, . . . , um,0). � It is not always easy to see whether equation (1.1) can be represented in the form (1.14). For example, at first glance it seems that the equation v1,1 = v(v1,0 + 1) v(v0,1 − v1,0) + v0,1 + 1 (2.9) does not admit an invertible transformation of the form u = ϕ(v, v1,0). But in reality we can rewrite this equation as v1,1 + 1 v0,1v1,1 − 1 = v + 1 v1,0v − 1 and relate it to the equation (u1,1 − 1)(u0,1 + 1) = (u1,0 + 1)(u− 1) (2.10) via the invertible transformation u = −2 v1,0 + 1 vv1,0 − 1 − 1, v = u0,1 − 1 u+ 1 . Therefore, it is useful to reformulate our result in the following form. Corollary 1. The equation (1.1) admits a non-point invertible transformation of the form (2.4) into an equation of the form (2.3) if and only if both the conditions( Fu1,0 Fu ) u0,1 = 0, Fu + Fu1,0T −1 j (Fu0,1) 6= 0 are satisfied. On Non-Point Invertible Transformations 7 Proof. If equation (1.1) is represented in the form (1.14), then the right-hand side F of (1.1) is determined as an implicit function from the identity ϕ(u0,1, F ) = ψ(u, u1,0). (2.11) Differentiating this identity with respect to u and u1,0, we obtain Tj(ϕu1,0(u, u1,0))Fu = ψu(u, u1,0), Tj(ϕu1,0(u, u1,0))Fu1,0 = ψu1,0(u, u1,0). (2.12) Therefore, Fu1,0/Fu does not depend on u0,1. Conversely, if Fu1,0/Fu = β(u, u1,0), then F = E(ψ(u, u1,0), u0,1) and (1.1) can be rewritten in the form (1.14). Differentiating (2.11) with respect to u0,1, we obtain Fu0,1 = −Tj ( ϕu/ϕu1,0 ) . This expression and equation (2.12) allow us to rewrite the functional independence condition for ϕ, ψ in the following way ϕu1,0(u, u1,0)ψu(u, u1,0)− ϕu(u, u1,0)ψu1,0(u, u1,0) = ϕu1,0(ψu + T−1 j (Fu0,1)ψu1,0) = Tj(ϕu1,0)ϕu1,0(Fu + T−1 j (Fu0,1)Fu1,0) 6= 0. � Naturally, the propositions analogous to Theorem 1 and Corollary 1 are valid for invertible transformations of the form v = f(u, u0,1, u0,2, . . . , u0,n) too. Returning to equations (2.9), (2.10), we note that equation (2.10) was introduced in [13] in a slightly different form. This equation has also been used in [10] as an example of an equation which is inconsistent around the cube (in the sense of [1]) but possesses the higher symmetries. Therefore, we can obtain symmetries of equation (2.9) from symmetries of equation (2.10). Indeed, if a transformation v = f(u, u0,1) maps equation (1.1) into equation (2.3), then differentiation of (2.1) with respect to τ by virtue of a symmetry uτ = ξ[u] of equation (1.1) gives us LG ( (fu0,1Tj + fu)(ξ[u]) ) = (λ[u]Tj + µ[u]) ( LF (ξ[u]) ) , where LG = TiTj +Gv1,0Ti +Gv0,1Tj +Gv, LF = TiTj + Fu1,0Ti + Fu0,1Tj + Fu. Because LF (ξ[u]) = 0 by definition of symmetry, we see that vτ = fu0,1Tj(ξ[u]) + fuξ[u] (after rewriting in terms of v and its shifts) is a symmetry of equation (2.3). Applying this, for example, to the three-point symmetries uτ = (u2 − 1)(u1,0 − u−1,0), uτ = (u2 − 1) ( 1 u0,1 + u − 1 u+ u0,−1 ) of equation (2.10), we obtain the symmetries vτ = (v + 1)2 ( 1 vv1,0 − 1 − 1 vv−1,0 − 1 ) , vτ = v ( 1 v0,1 + 1 − 1 v0,−1 + 1 ) of equation (2.9). The Hietarinta [6] equation2 u1,1(u+ β)(u0,1 + α) = u0,1(u+ α)(u1,0 + β) (2.13) 2We write this equation in the form used in [15]. 8 S.Ya. Startsev is another interesting example. The invertible transformations v = u1,0(u+ α) u − α, w = βu0,1 β + u− u0,1 map this equation into equation (2.13) again. In addition, the Hietarinta equation is lineari- zable [15]. We note that the above properties of equation (2.13) are similar to those of the continuous equation uxy = (α(x, y)eu)x + (β(x, y)e−u)y + γ(x, y) which was considered in [16]. 3 Invertible transformations of differential-difference equations We let T denote the operator of the forward shift in i by virtue of equation (1.2). This operator is defined by the following rules: T (f(a, b, c, . . . )) = f(T (a), T (b), T (c), . . . ) for any function f ; T (um) = um+1; T (u(n)) = Dn−1 x (F ) (mixed variables u(n) 1 are expressed in terms of the dynamical variables by virtue of equation (1.2)). Here Dx = ∂ ∂x + u(1) ∂ ∂u + ∞∑ k=1 ( u(k+1) ∂ ∂u(k) + T (k−1)(F ) ∂ ∂uk + T (1−k)(F̃ ) ∂ ∂u−k ) , i.e. Dx is the total derivative with respect to x by virtue of equations (1.2), (1.6). The inverse (backward) shift operator T−1 is defined in the similar way. Definition 5. We say that a transformation v = f [u] maps equation (1.2) into an equation (v1)x = G(x, v, v1, vx), Gvx 6= 0 (3.1) if DxT (f) = G(x, f, T (f), Dx(f)). Definition 6. A transformation v = f [u] of equation (1.2) is called invertible if any of the dynamical variables u, uk, k ∈ Z, u(l), l ∈ N can be expressed as a function of a finite subset of the variables x, v := f, vr := T r(f), v(s) := Ds x(f), r ∈ Z, s ∈ N. Definition 7. Transformations v = f [u] and w = g[u] are called equivalent if there exists r ∈ Z such that g = T r(f). Proposition 2. Let a transformation of the form v = g[u] be invertible and map equation (1.2) into equation (3.1). Then this transformation is equivalent to either a transformation of the form w = f(x, u, u1, u2, . . . , um), (3.2) or a transformations of the form w = f ( x, u, u(1), u(2), . . . , u(n) ) . (3.3) Definition 8. A transformation is called non-point if this transformation is not equivalent to any point transformation of the form w = g(x, u). On Non-Point Invertible Transformations 9 It is easy to see that a transformation of the form (3.2) or (3.3) is non-point only if f depends on more than one of the variables u, u1, . . . , um or on at least one of the variables u(1), . . . , u(n), respectively. Theorem 2. Let a non-point invertible transformation of the form (3.3) map equation (1.2) into equation (3.1). Then equation (1.2) can be written in the form ϕ(x, u1, (u1)x) = ψ(x, u, ux), where ϕ(x, y, z) and ψ(x, y, z) satisfy the condition ϕyψz −ϕzψy 6= 0, and the transformation is equivalent to the composition of the invertible transformation w = ϕ(x, u, ux) and an invertible transformation of the form v = h(x,w,w(1), w(2), . . . , w(n−1)). In particular, any non-point invertible transformation of the form v = f(x, u, ux) is equivalent to the composition of the transformation w = ϕ(x, u, ux) and a point transformation v = h(x,w). Corollary 2. The equation (1.2) admits a non-point invertible transformation of the form (3.3) into an equation of the form (3.1) if and only if both the conditions FuFuxu1 − Fuu1Fux = 0, Fu + FuxT −1(Fu1) 6= 0 are satisfied. Theorem 3. Let a non-point invertible transformation of the form (3.2) map equation (1.2) into equation (3.1). Then equation (1.2) can be written in the form Dx(p(x, u, u1)) = q(x, u, u1), where p(x, y, z) and q(x, y, z) satisfy the condition pyqz − pzqy 6= 0, and the transformation is the composition of the transformation w = p(x, u, u1) and an invertible transformation of the form v = h(x,w,w1, w2, . . . , wm−1). In particular, any non-point invertible transformation of the form v = f(x, u, u1) is the composition of the transformation w = p(x, u, u1) and a point transformation v = h(x,w). Corollary 3. The equation (1.2) admits a non-point invertible transformation of the form (3.2) into an equation of the form (3.1) if and only if equation (1.2) has the form (u1)x = a(x, u, u1)ux + b(x, u, u1), where a and b satisfy the condition ax + au1b− abu1 − bu 6= 0. For brevity, we omit the proofs of the above propositions because they are very similar to the proofs for discrete equations. 4 Examples: the transformations of Liouville equation analogues A special class of integrable equations of the form uxy = F (x, y, u, ux, uy) (4.1) consists of equations for which there exist both a differential substitution of the form v = X(x, y, ux, uxx, . . . ) and a substitution of the form w = Y (x, y, uy, uyy, . . . ) that map (4.1) into the equations vy = 0 and wx = 0, respectively. Such equations are called Darboux integrable or equations of the Liouville type. They not only are C-integrable (in accordance with the term of [3]) but also possess infinitely many symmetries of arbitrary high order [19, 20]. The complete classification of the Darboux integrable equations (4.1) has been performed in [20]. Equations with the analogous properties exist among equations of the form (1.1) and (1.2) too, but the classification of such equations is completed for a special case of equation (1.2) only [5]. Therefore, deriving new examples of discrete and semi-discrete Darboux integrable equations from already known equations may be useful (for example, to check the completeness of a future classification). 10 S.Ya. Startsev 4.1 Discrete equations The first example is the discrete Liouville equation u1,1 = (u1,0 − 1)(u0,1 − 1) u (4.2) from [9]. According to [2], this equation has the integrals I[u] = ( u2,0 u1,0 − 1 + 1 ) ( u− 1 u1,0 + 1 ) , J [u] = ( u0,2 u0,1 − 1 + 1 ) ( u− 1 u0,1 + 1 ) , (4.3) i.e. functions I[u], J [u] such that Tj(I[u]) = I[u], Ti(J [u]) = J [u]. In addition, equation (4.2) is linearizable: the substitution u = z0,1z1,0 (z1,0 − z)(z0,1 − z) (4.4) maps solutions of the equation z1,1 = z1,0 + z0,1 − z (4.5) into solutions of (4.2). It is easy to see that (4.2) can be written in the form Tj(ϕ(u, u1,0)) = ψ(u, u1,0) and the scheme (1.14)–(1.17) is applicable to this equation: v = ϕ = u1,0 u− 1 , v0,1 = ψ = u1,0 − 1 u , u = p = v + 1 v − v0,1 , u1,0 = q = v v0,1 + 1 v − v0,1 , v1,0 + 1 v1,0 − v1,1 = v0,1 + 1 v − v0,1 v. Thus, we obtain the equation v v1,1 − v1,0 v0,1 − v = v1,0 + 1 v0,1 + 1 (4.6) that is related to the discrete Liouville equation via the invertible transformation v = u1,0/(u−1). Substituting the expressions of u, u0,1, u1,0, . . . in terms of v, v0,1, v1,0, . . . into (4.3), we obtain the integrals of equation (4.6): I[v] = v1,0 + v1,0 + 1 v , J [v] = (v0,3 − v0,1)(v0,2 − v) (v0,3 − v0,2)(v0,1 − v) . The composition v = z0,2(z1,0 − z) z(z2,0 − z1,0) of the transformation v = u1,0/(u− 1) and (4.4) allows us to construct the solution v = (αi+2 + βj)(αi+1 − αi) (αi + βj)(αj+2 − αi+1) of equation (4.6) from the general solution z = αi + βj of (4.5), where αi and βj are arbitrary. On Non-Point Invertible Transformations 11 Equation (4.6) can be written in the form Tj(ϕ) = ψ but ϕ and ψ are functionally dependent (ϕ = ψ = I[v]). According to Theorem 1, this fact implies that (4.6) has no non-point invertible transformation of the form ṽ = f(v, v1,0, . . . , vn,0) and hence equation (4.2) admits, up to equiva- lence, only the first order invertible transformations (v = f(u1,0/(u−1)) and w = g(u0,1/(u−1)) only). Applying Corollary 1, we see that the other discrete version [8] of the Liouville equation v1,1 = v1,0v0,1 − 1 v does not admit a non-point invertible transformation. This equation is mapped into (4.2) via the non-invertible transformation u = v1,0v0,1 and has the integrals I[v] = ( v3,0 v1,0 + 1 ) ( v v2,0 + 1 ) , J [v] = ( v0,3 v0,1 + 1 ) ( v v0,2 + 1 ) . 4.2 Differential-difference equations Let us consider the following analogue of the Liouville equation: (u1)x = u1 ( u1 + ux u + u ) . (4.7) This equation has the integrals X[u] = 2 uxx u − 3 u2 x u2 − u2, I[u] = ( 1 + u1 u2 ) ( 1 + u1 u ) , i.e. functions X[u], I[u] such that T (X) = X, Dx(I) = 0. Like the discrete and continuous Liouville equations, equation (4.7) is linearizable: the substitution u = (z1 − z)zx z1z (4.8) maps solutions of the equation (z1)x = zx (4.9) into solutions of (4.7). The above information and some other details about equation (4.7) can be found in [2]. Equation (4.7) can be written as (u1)x u1 − u1 = ux u + u. Applying the scheme (1.10)–(1.13), we obtain v = 1 2 (ux u − u ) , v1 = 1 2 (ux u + u ) , u = v1 − v, ux = v2 1 − v2, (v1 − v)x = v2 1 − v2. (4.10) Thus, the invertible transformation v = (ux/u− u)/2 maps equation (4.7) into the sequence of the coupled Riccati equations (4.10). Expressing X[u] and I[u] in terms of v, v1, vx, . . . , we obtain the integrals X[v] = vx − v2, I[v] = (v3 − v1)(v2 − v) (v3 − v2)(v1 − v) . 12 S.Ya. Startsev of equation (4.10). The composition v = zxx 2zx − zx z of the invertible transformation and (4.8) generates the solution v = βxx 2βx − βx αi + β (4.11) of equation (4.10) from the general solution z = αi + β(x) of (4.9), where αi and β(x) are arbitrary. Equation (4.10) was used in [2] as an example of an equation admitting the integrals and the solution (4.11) was constructed in this article by another method (directly form the equation X[v] = ξ(x)). Moreover, equation (4.7) can be represented in the form (1.13) too. Applying the scheme (1.10)–(1.13) in the reverse order, we get w = p = u1 u , wx = q = u2 1 u + u1, u = wx (w + 1)w , u1 = wx w + 1 , (w1)x (w1 + 1)w1 = wx w + 1 and see that the invertible transformation w = u1/u maps (4.7) into the equation (w1)x = wxw1 w1 + 1 w + 1 . (4.12) As above, we construct the integrals X[w] = 2 wxxx wx − 3 w2 xx w2 x , I[w] = (w1 + 1)(w + 1) w1 of equation (4.12) by expressing X[u] and I[u] in terms of w,w1, wx, . . . , and obtain its solution w = (αi+2 − αi+1)(αi + β(x)) (αi+1 − αi)(αi+2 + β(x)) with arbitrary αi and β(x) by applying the composition w = (z2 − z1)z (z1 − z)z2 of the transformations w = u1/u and (4.8) to the general solution z = αi+β(x) of equation (4.9). The semi-discrete Liouville equation (4.7) is a special case of the Darboux integrable equation (u1)x = ux + √ Ce2u1 +Be(u1+u) + Ce2u. (4.13) that was introduced in [5]. Indeed, replacing u in (4.7) by exp(u), we obtain equation (4.13) with C = 1, B = 2. Without loss of generality, we can assume that the constant C in equation (4.13) equals 1 or 0 because C can be scaled via the point transformation u = ũ + γ. Applying Corollary 2, we see that equation (4.13) admits an invertible transformation of the form v = f(x, u, ux) only if B = 2C. The invertible transformations w = eu1−u, eu = wx w √ Cw2 +Bw + C On Non-Point Invertible Transformations 13 relate (4.13) to the equation (w1)x = w1wx √ Cw2 1 +Bw1 + C Cw2 +Bw + C . (4.14) The later equation has the integrals X[w] = 2 wxxx wx − 3 ( wxx wx )2 + 3w2 x(B2 − 4C2) 4(Cw2 +Bw + C) , I[w] = ∫ w1 ds s √ Cs2 +Bs+ C − ∫ w ds√ Cs2 +Bs+ C and can not be reduced to equation (4.12) via a point transformation because equation (4.14), in contrast to equation (4.12), does not admit an invertible transformation of the form v = f(x,w,wx, wxx) if B 6= 2C. Acknowledgments This work is partially supported by the Russian Foundation for Basic Research (grant number 10-01-00088-a). References [1] Adler V.E., Bobenko A.I., Suris Yu.B., Classification of integrable equation on quad-graphs. The consistency approach, Comm. Math. Phys. 233 (2003), 513–543, nlin.SI/0202024. [2] Adler V.E., Startsev S.Ya., Discrete analogues of the Liouville equation, Teoret. Mat. Fiz. 121 (1999), 271–284 (English transl.: Theoret. and Math. Phys. 121 (1999), 1484–1495), solv-int/9902016. [3] Calogero F., Why are certain nonlinear PDEs both widely applicable and integrable?, in What is integra- bility?, Editor V.E. Zakharov, Springer Ser. Nonlinear Dynam., Springer, Berlin, 1991, 1–62. [4] Habibullin I.T., Characteristic algebras of fully discrete hyperbolic type equations, SIGMA 1 (2005), 023, 9 pages, nlin.SI/0506027. [5] Habibullin I.T., Zheltukhina N., Pekcan A., Complete list of Darboux integrable chains of the form t1x = tx + d(t, t1), J. Math. Phys. 50 (2009), 102710, 23 pages, arXiv:0907.3785. [6] Hietarinta J., A new two-dimensional lattice model that is ‘consistent around a cube’, J. Phys. A: Math. Gen. 37 (2004), L67–L73, nlin.SI/0311034. [7] Hirota R., Nonlinear partial difference equations. III. Discrete sine-Gordon equation, J. Phys. Soc. Japan 43 (1977), 2079–2086. [8] Hirota R., Nonlinear partial difference equations. V. Nonlinear equations reducible to linear equations, J. Phys. Soc. Japan 46 (1979), 312–319. [9] Hirota R., Discrete two-dimensional Toda molecule equation, J. Phys. Soc. Japan 56 (1987), 4285–4288. [10] Levi R., Yamilov R.I., The generalized symmetry method for discrete equation, J. Phys. A: Math. Theor. 42 (2009), 454012, 18 pages, arXiv:0902.4421. [11] Mikhailov A.V., Wang J.P., Xenitidis P.D., Recursion operators, conservation laws and integrability condi- tions for difference equations, arXiv:1004.5346. [12] Nijhoff W.F., Quispel G.R.W., Capel H.W., Linearization of nonlinear differential-difference equations, Phys. Lett. A 95 (1983), 273–276. [13] Nijhoff F.W., Capel H.W., The discrete Korteweg–de Vries equation, Acta Appl. Math. 39 (1995), 133–158. [14] Orfanidis S.J., Discrete sine-Gordon equations, Phys. Rev. D 18 (1978), 3822–3827. [15] Ramani A., Joshi N., Grammaticos B., Tamizhmani N., Deconstructing an integrable lattice equation, J. Phys. A: Math. Gen. 39 (2006), L145–L149. [16] Sokolov V.V., Svinolupov S.I., On nonclassical invertible transformation of hyperbolic equations, European J. Appl. Math. 6 (1995), 145–156. http://dx.doi.org/10.1007/s00220-002-0762-8 http://arxiv.org/abs/nlin.SI/0202024 http://dx.doi.org/10.1007/BF02557219 http://arxiv.org/abs/solv-int/9902016 http://dx.doi.org/10.3842/SIGMA.2005.023 http://arxiv.org/abs/nlin.SI/0506027 http://dx.doi.org/10.1063/1.3251334 http://arxiv.org/abs/0907.3785 http://dx.doi.org/10.1088/0305-4470/37/6/L01 http://dx.doi.org/10.1088/0305-4470/37/6/L01 http://arxiv.org/abs/nlin.SI/0311034 http://dx.doi.org/10.1143/JPSJ.43.2079 http://dx.doi.org/10.1143/JPSJ.46.312 http://dx.doi.org/10.1143/JPSJ.56.4285 http://dx.doi.org/10.1088/1751-8113/42/45/454012 http://arxiv.org/abs/0902.4421 http://arxiv.org/abs/1004.5346 http://dx.doi.org/10.1016/0375-9601(83)90016-6 http://dx.doi.org/10.1007/BF00994631 http://dx.doi.org/10.1103/PhysRevD.18.3822 http://dx.doi.org/10.1088/0305-4470/39/8/L01 http://dx.doi.org/10.1017/S0956792500001741 http://dx.doi.org/10.1017/S0956792500001741 14 S.Ya. Startsev [17] Yamilov R.I., Invertible changes of variables generated by Bäcklund transformations, Teoret. Mat. Fiz. 85 (1990), 368–375 (English transl.: Theoret. and Math. Phys. 85 (1991), 1269–1275). [18] Yamilov R.I., Construction scheme for discrete Miura transformation, J. Phys. A: Math. Gen. 27 (1994), 6839–6851. [19] Zhiber A.V., Sokolov V.V., Startsev S.Ya., On nonlinear Darboux-integrable hyperbolic equations, Dokl. Ross. Akad. Nauk 343 (1995), 746–748 (English transl.: Doklady Math. 52 (1996), 128–130). [20] Zhiber A.V., Sokolov V.V., Exactly integrable hyperbolic equations of Liouville type, Usp. Mat. Nauk 56 (2001), no. 1, 63–106 (English transl.: Russ. Math. Surv. 56 (2001), no. 1, 61–101). http://dx.doi.org/10.1007/BF01018403 http://stacks.iop.org/0305-4470/27/6839 http://dx.doi.org/10.1070/RM2001v056n01ABEH000357 1 Introduction 2 Invertible transformations of discrete equations 3 Invertible transformations of differential-difference equations 4 Examples: the transformations of Liouville equation analogues 4.1 Discrete equations 4.2 Differential-difference equations References