Quasigraded Lie Algebras and Modified Toda Field Equations
We construct a family of quasigraded Lie algebras that coincide with the deformations of the loop algebras in "principal" gradation and admit Kostant-Adler-Symes scheme. Using them we obtain new Volterra coupled systems and modified Toda field equations for all series of classical matrix L...
Gespeichert in:
| Veröffentlicht in: | Symmetry, Integrability and Geometry: Methods and Applications |
|---|---|
| Datum: | 2006 |
| 1. Verfasser: | |
| Format: | Artikel |
| Sprache: | English |
| Veröffentlicht: |
Інститут математики НАН України
2006
|
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/146184 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Quasigraded Lie Algebras and Modified Toda Field Equations / T.V. Skrypnyk // Symmetry, Integrability and Geometry: Methods and Applications. — 2006. — Т. 2. — Бібліогр.: 29 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-146184 |
|---|---|
| record_format |
dspace |
| spelling |
Skrypnyk, T.V. 2019-02-07T20:35:44Z 2019-02-07T20:35:44Z 2006 Quasigraded Lie Algebras and Modified Toda Field Equations / T.V. Skrypnyk // Symmetry, Integrability and Geometry: Methods and Applications. — 2006. — Т. 2. — Бібліогр.: 29 назв. — англ. 1815-0659 2000 Mathematics Subject Classification: 37K05; 37K30 https://nasplib.isofts.kiev.ua/handle/123456789/146184 We construct a family of quasigraded Lie algebras that coincide with the deformations of the loop algebras in "principal" gradation and admit Kostant-Adler-Symes scheme. Using them we obtain new Volterra coupled systems and modified Toda field equations for all series of classical matrix Lie algebras g. Author is grateful to P. Holod, S. Zykov and M. Pavlov for discussions. The research was supported by INTAS YS Fellowship Nr 03-55-2233 and by the French-Ukrainian project “Dnipro”. en Інститут математики НАН України Symmetry, Integrability and Geometry: Methods and Applications Quasigraded Lie Algebras and Modified Toda Field Equations Article published earlier |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| title |
Quasigraded Lie Algebras and Modified Toda Field Equations |
| spellingShingle |
Quasigraded Lie Algebras and Modified Toda Field Equations Skrypnyk, T.V. |
| title_short |
Quasigraded Lie Algebras and Modified Toda Field Equations |
| title_full |
Quasigraded Lie Algebras and Modified Toda Field Equations |
| title_fullStr |
Quasigraded Lie Algebras and Modified Toda Field Equations |
| title_full_unstemmed |
Quasigraded Lie Algebras and Modified Toda Field Equations |
| title_sort |
quasigraded lie algebras and modified toda field equations |
| author |
Skrypnyk, T.V. |
| author_facet |
Skrypnyk, T.V. |
| publishDate |
2006 |
| language |
English |
| container_title |
Symmetry, Integrability and Geometry: Methods and Applications |
| publisher |
Інститут математики НАН України |
| format |
Article |
| description |
We construct a family of quasigraded Lie algebras that coincide with the deformations of the loop algebras in "principal" gradation and admit Kostant-Adler-Symes scheme. Using them we obtain new Volterra coupled systems and modified Toda field equations for all series of classical matrix Lie algebras g.
|
| issn |
1815-0659 |
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/146184 |
| citation_txt |
Quasigraded Lie Algebras and Modified Toda Field Equations / T.V. Skrypnyk // Symmetry, Integrability and Geometry: Methods and Applications. — 2006. — Т. 2. — Бібліогр.: 29 назв. — англ. |
| work_keys_str_mv |
AT skrypnyktv quasigradedliealgebrasandmodifiedtodafieldequations |
| first_indexed |
2025-11-26T05:07:29Z |
| last_indexed |
2025-11-26T05:07:29Z |
| _version_ |
1850612799055593472 |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications Vol. 2 (2006), Paper 043, 14 pages
Quasigraded Lie Algebras
and Modified Toda Field Equations
Taras V. SKRYPNYK †‡
† Bogolyubov Institute for Theoretical Physics, 14-b Metrologichna Str., Kyiv, 03143 Ukraine
‡ Institute of Mathematics, 3 Tereshchenkivs’ka Str., Kyiv-4, 01601 Ukraine
E-mail: tskrypnyk@imath.kiev.ua
Received October 31, 2005, in final form March 03, 2006; Published online April 16, 2006
Original article is available at http://www.emis.de/journals/SIGMA/2006/Paper043/
Abstract. We construct a family of quasigraded Lie algebras that coincide with the defor-
mations of the loop algebras in “principal” gradation and admit Kostant–Adler–Symes
scheme. Using them we obtain new Volterra coupled systems and modified Toda field
equations for all series of classical matrix Lie algebras g.
Key words: infinite-dimensional Lie algebras; soliton equations
2000 Mathematics Subject Classification: 37K05; 37K30
1 Introduction
Integrability of equations of 1 + 1 field theory and condensed matter physics is based on the
possibility to represent them in the form of the zero-curvature equation [1, 2, 3]:
∂U(x, t, λ)
∂t
− ∂V (x, t, λ)
∂x
+ [U(x, t, λ), V (x, t, λ)] = 0.
The most productive interpretation of the zero-curvature equation is achieved (see [3, 4, 5, 6]) if
one treats it as a consistency condition for a set of commuting Hamiltonian flows on a dual space
to some infinite-dimensional Lie algebra g̃ of matrix-valued function of λ written in the Euler–
Arnold (generalized Lax) form. In this case the corresponding U–V pair coincides with the
matrix gradients of mutually commuting Hamiltonians with respect to the natural Lie–Poisson
bracket on g̃∗. The method that provides the needed set of the commuting Hamiltonian flows
is the famous Kostant–Adler scheme [7, 3]. The main ingredient of this scheme is an existence
of the decomposition of the algebra g̃ into the sum of two subalgebras: g̃ = g̃+ + g̃−.
Such approach was originally based on the graded loop algebras g̃ = L(g) = g ⊗ P (λ, λ−1)
[3, 4, 5, 6] that possess decompositions into sums of two subalgebras. In our previous papers
[8, 9, 10, 11, 12] it was shown that special quasigraded Lie algebras could be also used in
the framework of this approach. More precisely, we have constructed a family of quasigraded
Lie algebras g̃A parameterized by some numerical matrices A possessing the decomposition
g̃A = g̃+
A + g̃−A
1. The constructed quasigraded Lie algebras are multiparametric deformations of
the loop algebras: g̃A tends to L(g) in the limit A→ 0. They also generalize the special elliptic
so(3) algebra introduced in the papers [14, 15].
In our previous papers [12, 16, 17, 18], using constructed quasigraded Lie algebras g̃A we
have obtained new hierarchies of integrable equations that coincide with various generalizations
of Landau–Lifshitz and anisotropic chiral field hierarchies.
1Subalgebras isomorphic to g̃−A were independently constructed in [13] as a possible complementary subalgebras
to the Lie algebras of Taylor series in the Lie algebra of Laurent power series.
mailto:tskrypnyk@imath.kiev.ua
http://www.emis.de/journals/SIGMA/2006/Paper043/
2 T.V. Skrypnyk
In the present paper we combine our previous results from [11, 16, 17, 18] with ideas of [19]
and define a new type of the quasigraded Lie algebras g̃
pr
A admitting Kostant–Adler–Symes
scheme and coinciding with deformations of the loop algebras in the principal gradation. More
definitely, it turned out that for a special choice of the matrices A (that depends on the classical
matrix Lie algebras g) it is possible to define “principal” subalgebras g̃
pr
A ⊂ g̃A in the analogous
way as for the case of ordinary loop algebras [20, 21]. In the case A→ 0 the algebras g̃
pr
A coincide
with the ordinary loop algebras in the principal gradation.
Using the technique of [4, 5, 3, 6] we develop a general scheme of obtaining of non-linear
partial differential equations admitting the zero-curvature representation with the values in
the Lie algebras g̃
pr
A . For this purpose we explicitly construct dual space, coadjoint action,
its invariants and Lie–Poisson brackets for the case of the Lie algebra g̃
pr
A . In the result we
obtain new integrable hierarchies satisfying zero-curvature conditions associated with the Lie
algebras g̃
pr
A . We show that among the corresponding equations there are remarkably simple
partial differential equations which we call “generalized Volterra coupled systems”. They have
the following form:
∂tui = uiβ
A
i (v), ∂xvi = viαi(u), (1)
where βAi are linear forms (“pseudo-roots”) on the subspace g−1 depending on the “deformation”
matrix A and Lie algebra g, αi ∈ Π
⋃
{−Θ} are the linear forms (roots) on the Cartan subalgebra
g0 = h, Hi is the basis in h, E−αi , EΘ is the basis in the subspace g−1 and
U = u ≡
dim h∑
i=1
uiHi, V = λ−1v ≡ λ−1
∑
αi∈Π
⋃
Θ
viE−αi ,
is the U–V -pair for the corresponding “deformed” zero-curvature conditions.
It is necessary to notice that in the case g = gl(n) our equations (1) coincide with the
periodic closure of the infinite Volterra coupled system considered in [22]. In the non-periodic
gl(n) case it was also considered in [23]. For the case of the other series of classical Lie algebras
equations (1) seem to be new.
Using parametrization generalizing one from [22] equations (1)could be re-written in the form
of the “modified Toda-field equations”:
∂2
xtψi = ∂xψi
∑
αj∈Π
⋃
{−Θ}
βAi,je
αj(ψ)
.
The structure of the present article is as follows: in the second section we define the quasi-
graded Lie algebras g̃A and their “principal” subalgebras. In the third section we define dual
spaces, Lie–Poisson brackets and Casimir functions on g̃
pr
A . In the fourth section we obtain the
zero-curvature equations with the values in the twisted subalgebras g̃
pr
A and consider the case of
modified Toda field hierarchies as the main examples they yield.
2 Principal grading of simple Lie algebras
In this subsection we will introduce necessary notations and remind some important facts from
the theory of semisimple Lie algebras [20, 21]. Let algebra g with the bracket [·, ·] be a simple
(reductive) classical Lie algebra of the rank n. Let h ⊂ g be its Cartan subalgebra, ∆± be its set
of positive(negative) roots, Π – the set of simple roots, Hi ∈ h the basis of Cartan subalgebra Eα,
α ∈ ∆ the corresponding root vectors.
Quasigraded Lie Algebras and Modified Toda Field Equations 3
Let us define the so-called “principal” grading of g [20], putting:
degHi = 0, degEαi = 1, degE−αi = −1.
It is evident that in such a way we obtain the grading of g: g =
h−1∑
k=0
gk with the graded subspaces
gk defined as follows: gk = SpanC{Eα} , where α is the root of the height k, i.e. α =
r∑
i=1
kiEαi
if α ∈ ∆+, α =
r∑
i=1
kiE−αi if α ∈ ∆− and k =
r∑
i=1
ki, h is the Coxeter number of g. In particular
g0 = h, g1 = SpanC{Eαi , E−θ|αi ∈ Π}, g−1 = SpanC{E−αi , Eθ|αi ∈ Π} and θ is the highest root
of the height h− 1.
Let us consider the following examples of classical matrix Lie algebras:
Example 1. Let g = gl(n) with the basis (Xij)ab = δiaδjb, i, j ∈ 1, n and the standard commu-
tation relations:
[Xij , Xkl] = δkjXil − δilXkj .
In this case g1 = SpanC{Eαi ≡ Xii+1, E−θ ≡ Xn1|i ∈ 1, n− 1}, g0 = SpanC{Hi ≡ Xii|i ∈ 1, n},
g−1 = SpanC{E−αi ≡ Xi+1,1, Eθ ≡ X1n|i ∈ 1, n− 1} and the Coxeter number is h = n.
Example 2. Let g = so(2n + 1), where so(2n + 1) = {X ∈ gl(2n + 1)|X = −sX>s} and
s = diag(1, s2n), s2n =
(
0 1n
1n 0
)
. In such realization the Cartan subalgebra has a basis
Hi = Xi+1,i+1 − Xi+n+1,i+n+1 where i = 1, n, generators of algebra that correspond to the
simple roots are Eαi = Xi+1,i+2 − Xn+i+2,n+i+1, i = 1, n − 1, Eαn = Xn+1,1 − X1,2n+1, their
negative counterparts are E−αi = Xi+2,i+1−Xn+i+1,n+i+2, i = 1, n−1, E−αn = X1,1+n−X2n+1,1.
The highest root corresponds to Eθ = X32+n −X23+n, its negative counterpart corresponds to
E−θ = X2+n3 −X3+n2, the Coxeter number is h = 2n.
Example 3. Let g = sp(n), where sp(n) = {X ∈ gl(2n)|X = wX>w} and w =
(
0 1n
−1n 0
)
. In
such realization the Cartan subalgebra has a basisHi = Xi,i−Xi+n,i+n where i = 1, n, generators
of algebra that correspond to the simple roots are Eαi = Xi,i+1 − Xn+i+1,n+i, i = 1, n − 1,
Eαn = Xn,2n, their negative counterparts are E−αi = Xi+1,i −Xn+i,n+i+1, i = 1, n− 1, E−αn =
X2n,n. The highest root corresponds to Eθ = X11+n, its negative counterpart corresponds to
E−θ = X1+n1, the Coxeter number is h = 2n.
Example 4. Let g = so(2n), where so(2n) = {X ∈ gl(2n)|X = −sX>s} and s ≡ s2n, s2n =(
0 1n
1n 0
)
. In such realization the Cartan subalgebra has a basis Hi = Xi,i − Xi+n,i+n where
i = 1, n, generators of algebra that correspond to the simple roots are Eαi = Xi,i+1−Xn+i+1,n+i,
i = 1, n−1, Eαn = Xn,2n−1−Xn−1,2n, their negative counterparts are E−αi = Xi+1,i−Xn+i,n+i+1,
i = 1, n− 1, E−αn = X2n−1,n −X2n,n−1. The highest root corresponds to Eθ = X21+n −X12+n,
its negative counterpart corresponds to E−θ = X1+n2−X2+n1, the Coxeter number is h = 2n−2.
3 “Principal” quasigraded Lie algebras
3.1 General case
It is well-known [20] that having the “principal” grading of g it is possible to define corresponding
grading of loop space. Let g =
h−1∑
k=0
gk be the Z/hZ grading of g. Let us consider the subspace
4 T.V. Skrypnyk
g̃pr ⊂ g̃, where g̃ ≡ g⊗ P (λ, λ−1) of the following type:
g̃pr =
⊕
j∈Z
gj ⊗ λj .
Here j denotes the class of equivalence of the elements j ∈ Z mod hZ. From the fact that
[gi, gj ] ⊂ gi+j it follows that g̃pr is a closed Lie algebra with respect to the ordinary Lie bracket
on the tensor product:
[X ⊗ p(λ), Y ⊗ q(λ)] = [X,Y ]⊗ p(λ)q(λ),
where X ⊗ p(λ), Y ⊗ q(λ) ∈ g̃pr. It is evident from the definition itself that g̃pr is the graded Lie
algebra with the grading being defined by the degrees of the spectral parameter λ.
Let us introduce the structure of the quasigraded Lie algebra into the loop space g̃. In
order to do this we will deform Lie algebraic structure in loop algebras g̃ in the following way
[11, 12, 13, 14, 15, 16, 17]:
[X ⊗ p(λ), Y ⊗ q(λ)]F = [X,Y ]⊗ p(λ)q(λ)− F (X,Y )⊗ λp(λ)q(λ), (2)
where X ⊗ p(λ), Y ⊗ q(λ) ∈ g̃ and the map F : g× g → g is skew and satisfies the following two
requirements which are equivalent to the Jacobi identities:
(J1)
∑
c.p. {i,j,k}
(F ([Xi, Xj ], Xk) + [F (Xi, Xj), Xk]) = 0,
(J2)
∑
c.p. {i,j,k}
F (F (Xi, Xj), Xk) = 0.
Now we are interested in a possibility of defining the structure of the quasigraded algebra on
the space g̃pr. For this purpose we want bracket (2) to be correctly restricted to the space g̃pr.
By the direct verification one can prove the following proposition:
Proposition 1. The subspace g̃pr ⊂ g̃ is the closed Lie algebra if and only if:
F (gi, gj) ⊂ gi+j+1. (3)
In the next subsection we will present examples of the cocycles F on the finite-dimensional
Lie algebras g that satisfy conditions (J1), (J2) and (3).
3.2 Case of classical matrix Lie algebras
Let us now consider the classical matrix Lie algebras g of the type gl(n), so(n) and sp(n) over
the field K of the complex or real numbers. As in the above examples we will realize the
algebra so(n) as algebra of skew-symmetric matrices: so(n) = {X ∈ gl(n)|X = −sX>s} and
the algebra sp(n) as the following matrix Lie algebra: sp(n) = {X ∈ gl(2n)|X = wX>w}, where
s ∈ symm (n), s2 = 1, w ∈ so(2n) and w2 = −1.
Let us consider the cochain F : g× g → g of the following form:
FA(X,Y ) = [X,Y ]A ≡ XAY − Y AX.
From the theory of consistent Poisson brackets it is known [25] to satisfy conditions (J1), (J2).
The following Proposition holds true:
Quasigraded Lie Algebras and Modified Toda Field Equations 5
Proposition 2. The cochain FA satisfies condition (3) if and only if the matrix A has the form:
1) A =
n−1∑
i=1
aiXii+1 + anXn1 if g = gl(n);
2) A =
n−1∑
i=1
ai(Xi+1,i+2 +Xn+i+2,n+i+1) + an(X1+n,1 +X1,2n+1) + an+1(X2+n3 +X3+n2)
if g = so(2n+ 1);
3) A =
n−1∑
i=1
ai(Xi,i+1 +Xn+i+1,n+i) if g = sp(n);
4) A =
n−1∑
i=1
ai(Xi,i+1 +Xn+i+1,n+i) + an(Xn,2n−1 +Xn−1,2n) + an+1(X1+n2 +X2+n1)
if g = so(2n),
where Xij is the standard matrix basis of gl(n), (Xij)αβ = δiαδγβ.
Proof. For the bracket constructed with the help of the cocycle FA to be correctly defined
on g̃pr we have to require the linear space g to be closed with respect to the bracket [·, ·]A
and linear space g̃pr as a space of matrix-valued function of λ to be closed with respect to the
bracket (4). These conditions are equivalent to the following requirement: [X,Y ]A ∈ gi+j+1
∀X ∈ gi, Y ∈ gj . The straightforward verification shows in each case that this requirement is
satisfied if and only if matrix A has the form described in the proposition. �
Hence in the case of the matrix Lie algebras and the matrices A defined in the above Propo-
sition we may introduce into the space g̃pr the new Lie bracket of the form:
[X ⊗ p(λ), Y ⊗ q(λ)] = [X,Y ]⊗ p(λ)q(λ)− [X,Y ]A ⊗ λp(λ)q(λ), (4)
where X⊗p(λ), Y ⊗ q(λ) ∈ g̃pr, [X,Y ] ≡ XY −Y X in the righthand side of this identity denote
ordinary Lie bracket in g and [X,Y ]A ≡ XAY − Y AX.
Definition. We will denote the linear space g̃pr with the bracket given by (4) by g̃
pr
A .
From the very definition of g̃
pr
A it follows that the algebra g̃
pr
A is Z-quasigraded and g̃
pr
A admits
the direct sum decomposition g̃
pr
A = g̃
pr+
A + g̃
pr−
A , where
g̃
pr+
A =
⊕
j≥0
gj ⊗ λj , g̃
pr−
A =
⊕
j<0
gj ⊗ λj .
In the subsequent exposition we will also use the following statement:
Proposition 3. For all the above classical matrix Lie algebras g and matrices A defined in the
Proposition 2 it is possible to introduce the following “pseudo-roots” βAi :
[X,Hi]A = βAi (X)Hi, (5)
where X ∈ g−1 and Hi are basic elements in Cartan subalgebra defined in the Section 2.
Remark 1. We call linear forms βAi to be “pseudo-roots” because gA = (g, [·, ·]) in the general
case is not isomorphic to g and g−1 ≡ (gA)0 is not commutative with respect to the [·, ·]A-bracket.
6 T.V. Skrypnyk
4 Dual space and Lie–Poisson bracket
In order to describe applications of the Lie algebras g̃
pr
A to the theory of finite-dimensional
integrable Hamiltonian systems it is necessary to define linear space (g̃pr
A )∗, the corresponding
Lie–Poisson bracket and the Casimir functions of g̃
pr
A .
4.1 Coadjoint representation and invariant functions of g̃
pr
A
In this subsection we will construct the dual space, coadjoint representation and its invariants
for the case of the ”principal” quasigraded Lie algebras g̃
pr
A . If h is the Coxeter number then from
properties of invariant form on simple Lie algebras it follows [20] that (gi, gj) = 0 if i+ j 6= 0
mod h. Hence we can define pairing between g̃
pr
A and (g̃pr
A )∗ in the following way:
〈X(λ), L(λ)〉 = resλ=0λ
−1 Tr (X(λ)L(λ)). (6)
From this definition it follows that the generic element L(λ) ∈ (g̃pr
A )∗ has the form:
L(λ) =
∑
j∈Z
dim gj∑
α=1
l(j)α X−j
α λ−j ,
where X−j
α is an element of basis of subspace g−j , l
(j)
α is a coordinate function on (g̃pr
A )∗.
Remark 2. We use the notion of infinite-dimensional Lie algebras g̃A and g̃
pr
A in the algebraic
sense of [20], i.e. we require that each of its elements consists of finite linear combination of
an infinite set of the natural elements of basis {Xj
αλj}. On the contrary, we let the dual
space g̃∗ to be wider, i.e. to contain also infinite linear combination of the elements of its basis
(Xj
αλj)∗ = X−j
α λ−j . Under such agreement all subsequent consideration will be correct and
consistent (see [18]).
The following proposition holds true:
Proposition 4. Let L(λ) ∈ (g̃pr
A )∗ be the generic element of the dual space. Then the functions
Imk (L(λ)) =
1
m
resλ=0λ
−k−1 Tr
(
L(λ)A(λ)−1
)m
,
are invariants of the coadjoint representation of the Lie algebra g̃
pr
A
Proof. It follows from the explicit form of the coadjoint action which has the following form:
ad∗X(λ) ◦ L(λ) = A(λ)X(λ)L(λ)− L(λ)X(λ)A(λ),
where A(λ) = (1− λA), X(λ), Y (λ) ∈ g̃
pr
A , L(λ) ∈ (g̃pr
A )∗. �
Remark 3. The matrix A(λ)−1 ≡ (1 − λA)−1 in the above Proposition has to be understood
as a power series in λ in the neighborhood of 0 or ∞: A(λ)−1 = (1 + Aλ + A2λ2 + · · · ) or
A(λ)−1 = −(A−1λ−1 + A−2λ−2 + · · · ). The corresponding decomposition should be chosen in
such a way that the restriction of the invariant function Imk (L(λ)) onto the dual space (g̃pr±
A )∗
are finite polynomials (see [18]). More precisely, when restricting invariant functions on (g̃pr−
A )∗
one has to choose the decomposition of A(λ)−1 in the neighborhood of 0, and when restricting
invariant function on (g̃pr+
A )∗ one has to chose the decomposition of A(λ)−1 in the neighborhood
of ∞.
Quasigraded Lie Algebras and Modified Toda Field Equations 7
4.2 Lie–Poisson bracket
Let us define Poisson structures in the space (g̃pr
A )∗. Using pairing (6) (described in the previous
section) we can define Lie–Poisson bracket on P ((g̃pr
A )∗) in the standard way:
{F1(L(λ)), F2(L(λ))} = 〈L(λ), [∇F1(L(λ)),∇F2(L(λ))]A(λ)〉, (7)
here ∇Fi(L(λ)) =
∑
j∈Z
dim gj∑
α=1
∂Fi
∂l
(j)
α
Xj
αλj and Xj
α is an element of basis of subspace gj .
From the Proposition 4 and standard arguments follows the next corollary:
Corollary 1. The functions Imk (L(λ)) are central (Casimir) functions for the Lie–Poisson
bracket (7).
Let us calculate Poisson bracket (7) explicitly. It is easy to show that for the coordinate
functions l(i)α , l(j)β , where l(i)α ∈ (gi)
∗, l(j)β ∈ (gj)
∗, this bracket will have the following form:
{l(i)α , l
(j)
β } =
∑
γ
Cγα,βl
(i+j)
γ −
∑
δ
Cδα,β(A)l(i+j+1)
δ ,
where lγ and lδ are the coordinate functions on (gi+j)
∗ and (gi+j+1)
∗. This bracket determines
the Lie algebra structure isomorphic to g̃
pr
A in the space of linear functions {liα} and, hence,
subspaces ((g̃pr
A )±)∗ are Poisson.
4.3 Infinite-component Hamiltonian systems via g̃
pr
A
In this subsection we construct Hamiltonian systems on the infinite-dimensional space (g̃pr
A )∗
possessing infinite number of independent, mutually commuting integrals of motion. Let
L∓(λ) =
∑
j∈Z±
dim gj∑
α=1
Lα(λ)X−j
α =
∑
j∈Z±
dim gj∑
α=1
l(j)α λ−jX−j
α
be the generic elements of the spaces (g̃pr±
A )∗. Let us consider the functions on (g̃pr
A )∗ of the form
Im∓k (L(λ)) ≡ Imk (L∓(λ)),
where {Imk (L(λ))} are Casimir functions of g̃
pr
A . Due to the Remark 3 functions Im∓k (L(λ))
are finite polynomials on (g̃pr±
A )∗. The Hamiltonian flows corresponding to the Hamiltonians
Im∓k (L(λ)) are written in a standard way:
∂Lα(λ)
∂tm∓k
=
{
Lα(λ), Im∓k (L(λ))
}
. (8)
The following theorem is true:
Theorem 1. (i) The time flows given by the equations (8) are correctly defined and commute
for all times tm+
k , tn−l .
(ii) Euler–Arnold equations (8) are written in the “deformed” Lax form:
∂L(λ)
∂tm∓k
= A(λ)Mm±
k (λ)L(λ)− L(λ)Mm±
k (λ)A(λ). (9)
where Mm±
k (λ) = ∇Imk (L∓(λ)) =
∑
j∈Z±
dim gj∑
α=1
∂Imk
∂l
(j)
α
Xj
aλj.
(iii) The functions Ipq (L±) are constant along all times tm±k and tn∓l .
8 T.V. Skrypnyk
The proof of this theorem repeats the proof of the analogous theorem for the case of ordinary
loop algebras (see [3] and references therein).
Remark 4. Note that by virtue of the fact that functions Imk (L∓(λ)) are finite polynomials
M -operators Mm±
k (λ) = ∇Imk (L∓(λ)) belong to g̃
pr±
A , i.e. are finite linear combinations of the
basic elements Xj
aλj . This fact provides the correctness of the definition of equation (9).
5 “Modified” Toda field equations
5.1 Zero-curvature condition with the values in g̃
pr
A
In this subsection we will obtain zero curvature-type equations with the values in the Lie algeb-
ras g̃
pr
A . The following theorem holds true:
Theorem 2. Let the infinite-dimensional Lie algebras g̃
pr
A , g̃
pr±
A , their dual spaces and polyno-
mial Hamiltonians Imk (L±(λ)), Ins (L±(λ)) on them be defined as in previous sections. Then the
algebra-valued gradients of these functions satisfy the “deformed” zero-curvature equations:
∂∇Imk (L±(λ))
∂tn±l
− ∂∇Ins (L±(λ))
∂tm±k
+ [∇Imk (L±(λ)),∇Ins (L±(λ))]A(λ) = 0, (10)
∂∇Imk (L±(λ))
∂tn∓l
− ∂∇Ins (L∓(λ))
∂tm±k
+ [∇Imk (L±(λ)),∇Ins (L∓(λ))]A(λ) = 0. (11)
Idea of the Proof. The statement of the theorem, i.e. validity of the equations (10), (11),
follows from the commutativity of the “deformed” Lax flows constructed in the previous section.
Remark 5. Using the mentioned above realizations of g̃
pr
A the “deformed” zero-curvature and
Lax equations can be rewritten in the standard form, but in this case the corresponding U–V
and L–M pairs will be more complicated and forcing us to work with the zero-curvature and
Lax equations in the “deformed” form (10), (11).
Theorem 2 provides us with an infinite number of g̃
pr
A -valued U–V pairs that satisfy zero
curvature-type equations. The latter are non-linear equations in partial derivatives in the dy-
namical variables — matrix elements of the matrix L(λ). In the terminology of [2] equations
generated by the infinite set of U–V pairs are called “integrable in the kinematic sense”. In the
next subsections we will consider the simplest examples of such integrable equations and their
hierarchies.
5.2 Modified Toda-field equations: general case
In this subsection we obtain concrete examples of integrable equations satisfying the “deformed”
zero-curvature representations constructed in the previous subsection. The most interesting of
them will be a “modified” Toda field equations. We will construct these equations for all series
of classical simple (reductive) lie algebras.
Let us at first consider a general situation. Let the “principal quasigraded” Lie algebra g̃
pr
A
and its decomposition g̃
pr
A = g̃
pr+
A + g̃
pr−
A be defined as in the previous sections. Generic elements
L±(λ) ∈ (g̃pr∓
A )∗ of the dual spaces have the following form:
L+(λ) = λL(−1) + λ2L(−2) + · · · , L−(λ) = L(0) + λ−1L(1) + λ−2L(2) + · · · ,
where L(k) ∈ g−k. In particular L(0) =
n∑
i=1
l
(0)
i Hi, L(−1) =
∑
αi∈Π
⋃
−Θ
l
(−1)
i Eαi .
Now we can formulate the following theorem:
Quasigraded Lie Algebras and Modified Toda Field Equations 9
Theorem 3. Let g be one of the classical matrix Lie algebras gl(n), so(2n+1), sp(n) or so(2n).
Let the “deformation” matrix A corresponding to the “principal” grading of g be defined as in
the Proposition 2. Then:
(i) among the corresponding equations (10) there is the equation equivalent to the following
“Volterra coupled system”:
∂tui = uiβ
A
i (v), ∂xvi = viαi(u), (12)
where βAi are “pseudo-roots” on g−1 defined by (5) and depending on the “deformation” matrix A
and Lie algebra g, αi ∈ Π
⋃
{−Θ} are linear forms (roots) on h, Hi is the basis in h, E−αi, EΘ is
the basis in g−1, and
U = u =
dim h∑
i=1
uiHi, V = λ−1v = λ−1
∑
αi∈Π
⋃
−Θ
viE−αi
is the corresponding U–V pair.
(ii) Volterra coupled system (12) is written in the form of the “modified” Toda field equations:
∂2
xtψi = ∂xψi
∑
αj∈Π
⋃
{−Θ}
βAi,je
αj(ψ)
, (13)
where ψ =
n∑
i=1
ψiHi and ψi satisfy the following differential constraint ∂xψ1∂xψ2 · · · ∂xψn = c.
Proof. To prove item (i) of the theorem it is enough to show that among the integrals Imk ,
Inl are such integrals Im0
k0
, In0
l0
that Im0
k0
(L+(λ)) ≡ Im0
k0
(L(−1)), In0
l0
(L−(λ)) ≡ In0
l0
(L(0)). Let us
explicitly find these functions. Let us consider the generating series Im0
k0
(L+(λ)):
Im0
k0
(L+(λ)) = Im0(A−1(λ)L+(λ)) = Im0((1 +Aλ+A2λ2 + · · · )(λL(−1) + λ2L(−2) + · · · )).
Taking into account that Im0 is a homogeneous polynomial on g∗ we easily obtain that Im0
m0
(L+(λ))
= Im0
m0
(L(−1)). Now, for the role of Im0 we have to take any Casimir function on g that has a
non-trivial restriction onto g1. Such Casimir function always exists, because the generic element
of the space g1 is never nilpotent.
Let us now consider the generating series In0(L−(λ)). If we chose the generating Casimir
function as follows: In0(L−(λ)) = detA(λ) Det (L−(λ)A(λ)−1) ≡ DetL−(λ), we obtain:
In0
0 (L−(λ)) = Det (L(0)).
For such the Hamiltonians their matrix gradients have the following form: ∇Im0
m0
(L+(λ)) ≡
u =
n∑
i=1
uiHi, ∇In0
0 (L−(λ)) ≡ λ−1v = λ−1(
∑
αi∈Π
viE−αi + v0EΘ), where ui ≡
∂I
m0
m0
∂l
(0)
i
, vi ≡
∂I
n0
0
∂l
(−1)
i
.
Substituting this into the corresponding zero-curvature condition (11) we obtain item (i) of the
theorem.
Making the substitution of variables: ui = ∂xψi, vi = eαi(ψ) we obtain equation (13). Finally,
taking into account the constancy of the Hamiltonian In0
l0
(L(0)) along all time flows we obtain
that
n∏
i=1
ui = const. That proves item (ii). �
Remark 6. In [22, 23] gl(n) coupled Volterra systems were obtained using the Bäcklund trans-
formation for the Toda integrable hierarchy. Such connection means that starting from the
known integrable hierarchy and Bäcklund transformation one can generate a new integrable
hierarchy (see [27]). The investigation of the possibility of such connection for the Volterra
coupled systems, associated with the other series of classical matrix Lie algebras, is a separate
problem and we will return to it in our subsequent publications.
10 T.V. Skrypnyk
5.3 Example: gl(n)-modified Toda field equation
Let us consider the Lie algebra g̃l(n)pr
A defined in the Example 1 and its natural decomposition
into the direct sum g̃l(n)pr
A = ˜gl(n)pr+
A + ˜gl(n)pr−
A where ˜gl(n)pr+
A =
∑
j≥0
gl(n)jλ
j , ˜gl(n)pr−
A =∑
j<0
gl(n)jλ
j and j = j mod n. The generic elements L±(λ) ∈ ( ˜gl(n)pr∓
A )∗ of the dual spaces
have the following form:
L+(λ) = λL(−1) + λ2L(−2) + · · · , L−(λ) = L(0) + λ−1L(1) + λ−2L(2) + · · · ,
where L(k) ∈ g−k: L
(0) =
n∑
i=1
l
(0)
i Xii, L(−1) =
n−1∑
i=1
l
(−1)
i Xi,i+1 + l
(−1)
n Xn1. The simplest integrals
are:
Dn−
0 (L(λ)) = l
(0)
1 l
(0)
2 · · · l(0)
n , Dn+
n (L(λ)) = l
(−1)
1 l
(−1)
2 · · · l(−1)
n−1 l
(−1)
n .
Their matrix gradients yield the following U–V pair:
U ≡ ∇Dn−
0 =
n∑
i=1
uiXii, V ≡ ∇Dn+
n = λ−1
(
n−1∑
i=1
viXi+1i + vnX1n
)
.
where ui ≡ Dn−
0 /l
(0)
i , vi ≡ Dn+
n /l
(−1)
i .
The corresponding “deformed” zero-curvature condition:
∂∇Dn−
0 (L(λ))
∂t
− ∂∇Dn+
n (L(λ))
∂x
+ [∇Dn−
0 (L(λ)),∇Dn+
n (L(λ))]A(λ) = 0,
yields the following equations2:
∂tu1 = u1(a1v1 − anvn), ∂xv1 = v1(u2 − u1),
∂tu2 = u2(a2v2 − a1v1), ∂xv2 = v2(u3 − u2),
· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
∂tun = un(anvn − an−1vn−1), ∂xvn = vn(u1 − un).
This system of the hyperbolic equations is the so-called “Volterra coupled systems”. In the case
of infinite n and ai ≡ 1 it was first considered in the paper [22]. In the non-periodic case it was
also considered in the book [23].
This system goes together with the two constraints:
u1u2 · · ·un = const1, v1v2 · · · vn−1vn = const2,
They follow from the constancy of the Hamiltonians Dn−
0 and Dn+
n with respect to the all time
flows.
In order for the “Volterra coupled systems” to acquire the form of the periodic “modified”
Toda system it is enough to make the following change of variables [22]:
ui = ∂xψi, i = 1, n, vi = eψi+1−ψi , i = 1, n− 1, vn = eψ1−ψn .
in the result we obtain the following system of equations:
∂2
xtψ1 = ∂xψ1(a1e
ψ2−ψ1 − ane
ψ1−ψn),
2In the case ai 6= 0 one may make ai = 1 rescaling the variables vi.
Quasigraded Lie Algebras and Modified Toda Field Equations 11
∂2
xtψ2 = ∂xψ2(a2ε
ψ3−ψ2 − a1e
ψ2−ψ1),
· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
∂2
xtψn = ∂xψn(aneψ1−ψn − an−1e
ψn−ψn−1)
with additional differential constraint:
∂xψ1∂xψ2 · · · ∂xψn = c.
The simplest case, when the corresponding constraint can be solved, is the case ∂xψi = 0 for
some fixed i. This is equivalent to the reduction ui = 0 in the corresponding Volterra coupled
system. Let us put, for example, ψn = const = 0. In this case we obtain the following “open”
(n− 1)-component modified Toda chain:
∂2
xtψ1 = ∂xψ1(a1e
ψ2−ψ1 − ane
ψ1),
∂2
xtψ2 = ∂xψ2(a2ε
ψ3−ψ2 − a1e
ψ2−ψ1),
· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
∂2
xtψn−1 = ∂xψn−1(an−1e
−ψn−1 − an−2e
ψn−1−ψn−2).
5.4 Example: g = so(2n)-modified Toda field equation
Let us consider the Lie algebra ˜so(2n)pr
A being defined in the Example 4 and its natural de-
composition into the direct sum ˜so(2n)pr
A = ˜so(2n)pr+
A + ˜so(2n)pr−
A . Generic elements L±(λ) ∈
( ˜so(2n)pr∓
A )∗ of the dual spaces have the following form:
L+(λ) = λL(−1) + λ2L(−2) + · · · , L−(λ) = L(0) + λ−1L(1) + λ−2L(2) + · · · ,
where L(k) ∈ so(2n)−k mod (2n−2).
L(0) =
n∑
i=1
l
(0)
i (Xii −Xi+ni+n),
L(−1) =
n−1∑
i=1
l
(−1)
i (Xii+1 −Xi+n+1i+n) + l(−1)
n (Xn2n−1 −Xn−12n) + l
(−1)
n+1 (Xn+12 −Xn+21).
It turned out that the form of the needed Casimir functions and Hamiltonians differs for the
cases n ≤ 3 and n > 3. We will consider here only the case n > 3. The simplest non-trivial
integrals are:
D2n−
0 (L(λ)) =
(
l
(0)
1 l
(0)
2 · · · l(0)
n
)2
,
T
(2n−2)+
2n−2 (L(λ)) = l
(−1)
1
(
l
(−1)
2
)2 · · · (l(−1)
n−2
)2
l
(−1)
n−1 l
(−1)
n l
(−1)
n+1 .
Their matrix gradients are:
U ≡ ∇D2n−
0 =
n∑
i=1
ui(Xii −Xi+ni+n),
V ≡ ∇T (2n−2)+
2n−2 = λ−1
(
n−1∑
i=1
vi(Xi+1i −Xi+ni+n+1) + vn(X2n−1n −X2nn−1)
+ vn+1(X2n+1 −X1n+2)
)
,
where ui ≡ D2n−
0 /l
(0)
i , vi ≡ T
(2n−2)+
2n−2 /l
(−1)
i .
12 T.V. Skrypnyk
The corresponding zero-curvature condition
∂∇D2n−
0 (L(λ))
∂t
−
∂∇T (2n−2)+
2n−2 (L(λ))
∂x
+
[
∇D2n−
0 (L(λ)), T (2n−2)+
2n−2 (L(λ))
]
A(λ)
= 0,
yields the following equations:
∂tu1 = u1(a1v1 − an+1vn+1), ∂xv1 = v1(u2 − u1),
∂tu2 = u2(a2v2 − a1v1 + an+1vn+1), ∂xv2 = v2(u3 − u2),
∂tu3 = u3(a3v3 − a2v2), ∂xv3 = v3(u4 − u3),
· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
∂tun−2 = un−2(an−2vn−2 − an−3vn−3), ∂xvn−1 = vn−1(un − un−1),
∂tun−1 = un−1(an−1vn−1 − an−2vn−2 − anvn), ∂xvn = −vn(un + un−1),
∂tun = un(anvn − an−1vn−1), ∂xvn+1 = vn+1(u1 + u2).
This system goes together with the natural constraints:
u1u2 · · ·un = const1, v1(v2)2 · · · (vn−2)2vn−1vnvn+1 = const2.
which follows from the constancy of the Hamiltonians D2n−
0 and T
(2n−2)+
2n−2 with respect to the
all time flows. Making the following replacement of variables:
ui = ∂xψi, vi = eψi+1−ψi , where i = 1, n− 1,
vn = e−(ψn+1+ψn), vn+1 = e−(ψ2+ψ1)
we obtain the so(2n)-modified Toda chain:
∂2
xtψ1 = ∂xψ1(a1e
ψ2−ψ1 − an+1e
ψ2+ψ1),
∂2
xtψ2 = ∂xψ2(a2e
ψ3−ψ2 − a1e
ψ2−ψ1 + an+1e
ψ2+ψ1),
∂2
xtψ3 = ∂xψ3(a3e
ψ4−ψ3 − a2e
ψ3−ψ2),
· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
∂2
xtψn−2 = ∂xψn−2(an−2e
ψn−1−ψn−2 − an−3e
ψn−2−ψn−3),
∂2
xtψn−1 = ∂xψn−1(an−1e
ψn−ψn−1 − an−2e
ψn−1−ψn−2 − ane
−(ψn−1+ψn)),
∂2
xtψn = ∂xψn(ane−(ψn−1+ψn) − an−1e
ψn−ψn−1).
These equations goes also with additional differential constraint: ∂xψ1∂xψ2 · · · ∂xψn = c.
6 Conclusion and discussion
In the present paper we have constructed a family of quasigraded Lie algebras that coincide
with deformations of the “principal” subalgebras of the loop algebras and admit Kostant–Adler–
Symes scheme. Using the constructed algebras we have obtained new Volterra coupled systems
and modified Toda field equations for all series of classical matrix Lie algebras g.
Further, more detailed investigation of the constructed integrable hierarchies would be very
interesting. In particular, an interesting open problem is to find a possible connection of the
constructed new Volterra coupled systems and new modified Toda field equations with the
Bäcklund transformations and the usual Toda field equations [19, 21, 23]. It would be also
interesting to find explicit forms of the simplest equations of the “negative” and “positive”
subhierarchies of the constructed integrable hierarchies in the case of general g. They should
Quasigraded Lie Algebras and Modified Toda Field Equations 13
coincide with the “deformations” of the generalized mKdV equations [28] (see [29] where gl(2)
case was considered in details). It would be also very nice to find out whether there exists
an analog of the “Drienfield–Sokolov” reduction [21] and the corresponding “deformed” KdV
equations.
Another interesting development is to generalize the construction of the present paper onto
the case of the other graded subalgebras of the loop algebras (see [20, 28]). The work in this
direction is now in progress and the corresponding results will soon be published.
Acknowledgements
Author is grateful to P. Holod, S. Zykov and M. Pavlov for discussions. The research was sup-
ported by INTAS YS Fellowship Nr 03-55-2233 and by the French-Ukrainian project “Dnipro”.
[1] Zaharov V., Shabat A., Integration of nonlinear equations of mathematical physics using the method of
inverse scattering problem. II, Funktsional. Anal. i Prilozhen., 1979, V.13, N 3, 13–21 (in Russian).
[2] Takhtadzhyan L.A., Faddeev L.D., The Hamiltonian approach in soliton theory, Moscow, Nauka, 1986 (in
Russian).
[3] Newell A., Solitons in mathematics and physics, University of Arizona, Society for Industrial and Applied
Mathematics, 1985.
[4] Flaschka H., Newell A., Ratiu T., Kac–Moody Lie algebras and soliton equations. II. Lax equations associated
with A
(1)
1 , Phys. D, 1983, V.9, 303–323.
[5] Flaschka H., Newell A., Ratiu T., Kac–Moody Lie algebras and soliton equations. III. Stationary equations
associated with A
(1)
1 , Phys. D, 1983, V.9, 324–332.
[6] Holod P., Integrable Hamiltonian systems on the orbits of affine Lie groups and periodical problem for
mKdV equation, Preprint ITF-82-144R, Kyiv, Institute for Theoretical Physics, 1982 (in Russian).
[7] Reyman A., Semenov-Tian-Shansky M., Group theoretical methods in the theory of finite-dimensional
integrable systems, VINITI, Current Problems in Mathematics. Fundamental Directions, 1989, V.6, 145–
147 (in Russian).
[8] Holod P., Skrypnyk T., Anisotropic quasigraded Lie algebras on the algebraic curves and integrable Hamil-
tonian systems, Naukovi Zapysky NAUKMA, Ser. Phys.-Math. Sciences, 2000, V.18, 20–25 (in Ukrainian).
[9] Skrypnyk T., Lie algebras on algebraic curves and finite-dimensional integrable systems, nlin.SI-0010005
(talk on the XXIII International Colloquium on Group Theoretical Methods in Physics, July 31 – August 5,
2000, Dubna, Russia).
[10] Skrypnyk T., Quasigraded Lie algebras on hyperelliptic curves and classical integrable systems, J. Math.
Phys., 2001, V.42, 4570–4581.
[11] Skrypnyk T., Quasigraded deformations of Lie algebras and finite-dimensional integrable systems, Czechoslo-
vak J. Phys., 2002, V.52, 1283–1288.
[12] Skrypnyk T., Quasigraded Lie algebras and hierarchies of integrable equations, Czechoslovak J. Phys., 2003,
V.53, 1119–1124.
[13] Golubchik I., Sokolov V., Compatible Lie brackets and integrable equations of the principle chiral model
type, Funktsional. Anal. i Prilozhen., 2002, V.36, N 3, 9–19 (in Russian).
[14] Holod P., Hamiltonian systems connected with anisotropic affine Lie algebras and higher Landau–Lifshits
equations, Dokl. Akad. Nauk Ukrain. SSR Ser. A, 1984, N 5, 5–8 (in Russian).
[15] Holod P., The hidden symmetry of the Landau–Lifshits equation, the hierarchy of higher equations and
a dual equation for an asymmetric chiral field, Teoret. Mat. Fiz., 1987, V.70, N 1, 18–29 (in Russian).
[16] Skrypnyk T., Doubled generalized Landau–Lifshitz hierarchies and special quasigraded Lie algebras,
J. Phys. A: Math. Gen., 2004, V.37, 7755–7768, nlin.SI/0403046.
[17] Skrypnyk T., Deformations of loop algebras and integrable systems: hierarchies of integrable equations,
J. Math. Phys., 2004, V.45, 4578–4595.
[18] Skrypnyk T., Quasigraded Lie algebras, the Kostant–Adler scheme, and integrable hierarchies, Teoret. Mat.
Fiz., 2005, V.142, N 2, 275–288 (in Russian).
[19] Mikhailov A., The reduction in integrable systems. Groups of reduction, Pis’ma Zh. Eksper. Teoret. Fiz.,
1980, V.32, N 1, 187–192 (in Russian).
14 T.V. Skrypnyk
[20] Kac V., Infinite-dimentional Lie algebras, Moscow, Mir, 1993 (in Russian).
[21] Drinfel’d V., Sokolov V., Lie algebras and KdV-type equations, VINITI, Current Problems in Mathematics.
Fundamental Directions, 1984, V.24, 81–180 (in Russian).
[22] Shabat A., Yamilov R., To a transformation theory of two-dimensional integrable systems, Phys. Lett. A,
1997, V.277, 15–23.
[23] Leznov A., Saveliev M., Group methods for the integration of nonlinear dynamical systems, Moscow, Nauka,
1985 (in Russian).
[24] Krichiver I., Novikov S., Virasoro-type algebras, Riemannian surfaces and structures of the soliton theory,
Funktsional. Anal. i Prilozhen., 1987, V.21, N 2, 46–64 (in Russian).
[25] Cantor I., Persits D., Closed stacks of Poisson brackets, in Proceedinds of the IX USSR Conference in
Geometry, Kishinev, Shtinitsa, 1988, 141 (in Russian).
[26] Bolsinov A., Completeness of families of functions in involution that are connected with compatible Poisson
brackets, Trudy Sem. Vektor. Tenzor. Anal., 1988, N 23, 18–38 (in Russian).
[27] Borisov A., Zykov S., The chain of the dressing discrete symmetries and generation of the non-linear equa-
tions, Teoret. Mat. Fiz., 1998, V.115, 199–214 (in Russian).
[28] de Groot M., Hollowood T., Miramontes J., Generalized Drienfield–Sokolov hierarchies, Comm. Math. Phys.,
1992, V.145, 57–78.
[29] Skrypnyk T., Integrable deformations of the mKdV and SG hierarchies and quasigraded Lie algebras,
Phys. D, to appear.
1 Introduction
2 Principal grading of simple Lie algebras
3 ``Principal'' quasigraded Lie algebras
3.1 General case
3.2 Case of classical matrix Lie algebras
4 Dual space and Lie-Poisson bracket
4.1 Coadjoint representation and invariant functions of g"0365gprA
4.2 Lie-Poisson bracket
4.3 Infinite-component Hamiltonian systems via g"0365gprA
5 ``Modified'' Toda field equations
5.1 Zero-curvature condition with the values in g"0365gprA
5.2 Modified Toda-field equations: general case
5.3 Example: gl(n)-modified Toda field equation
5.4 Example: g=so(2n)-modified Toda field equation
6 Conclusion and discussion
|