Erwin Schrödinger International Institute of Mathematical Physics, Wien, Austria. Sept. 01, 1999

Similar documents
Completely integrable systems: a generalization

The Erwin Schrodinger International Boltzmanngasse 9. Institute for Mathematical Physics A-1090 Wien, Austria

ISOMORPHISMS OF POISSON AND JACOBI BRACKETS

Deformation of the `embedding'

Institut für Mathematik, Universität Wien, Strudlhofgasse 4, A-1090 Wien, Austria.

The Erwin Schrödinger International Boltzmanngasse 9 Institute for Mathematical Physics A-1090 Wien, Austria

A global version of the quantum duality principle

Poisson Manifolds Bihamiltonian Manifolds Bihamiltonian systems as Integrable systems Bihamiltonian structure as tool to find solutions

The Erwin Schrodinger International Boltzmanngasse 9. Institute for Mathematical Physics A-1090 Wien, Austria

Cartan A n series as Drinfeld doubles

1 Differentiable manifolds and smooth maps

THEODORE VORONOV DIFFERENTIABLE MANIFOLDS. Fall Last updated: November 26, (Under construction.)

arxiv: v2 [math-ph] 11 Mar 2017

Quantum Algebras and Quantum Physics

GEOMETRIC QUANTIZATION

Combinatorial Dyson-Schwinger equations and systems I Feynma. Feynman graphs, rooted trees and combinatorial Dyson-Schwinger equations

The classical basis for κ-deformed Poincaré (super)algebra and the second κ-deformed supersymmetric Casimir.

Basic Concepts of Group Theory

THE MODULAR CLASS OF A LIE ALGEBROID COMORPHISM

Definition 5.1. A vector field v on a manifold M is map M T M such that for all x M, v(x) T x M.

The Toda Lattice. Chris Elliott. April 9 th, 2014

JANUSZ GRABOWSKI, GIUSEPPE MARMO, PETER W. MICHOR

MAPPING CLASS ACTIONS ON MODULI SPACES. Int. J. Pure Appl. Math 9 (2003),

THE CLASSICAL HOM-YANG-BAXTER EQUATION AND HOM-LIE BIALGEBRAS. Donald Yau

SYMPLECTIC LEAVES AND DEFORMATION QUANTIZATION

Vector Bundles on Algebraic Varieties

Representations of Sp(6,R) and SU(3) carried by homogeneous polynomials

LECTURE 25-26: CARTAN S THEOREM OF MAXIMAL TORI. 1. Maximal Tori

arxiv:solv-int/ v1 18 Oct 1999

Geometrical mechanics on algebroids

MANIN PAIRS AND MOMENT MAPS

CLASSIFICATION OF NON-ABELIAN CHERN-SIMONS VORTICES

Modern Geometric Structures and Fields

MULTIPLIER HOPF ALGEBRAS AND DUALITY

ON THE MATRIX EQUATION XA AX = X P

CALCULUS ON MANIFOLDS

A Note on n-ary Poisson Brackets

BRST and Dirac Cohomology

J.F. Cari~nena structures and that of groupoid. So we will nd topological groupoids, Lie groupoids, symplectic groupoids, Poisson groupoids and so on.

EXERCISES IN POISSON GEOMETRY

The Erwin Schrodinger International Pasteurgasse 6/7. Institute for Mathematical Physics A-1090 Wien, Austria

arxiv:math/ v1 [math.qa] 20 Feb 2001

MAT265 Mathematical Quantum Mechanics Brief Review of the Representations of SU(2)

A MARSDEN WEINSTEIN REDUCTION THEOREM FOR PRESYMPLECTIC MANIFOLDS

NAMBU-LIE GROUP ACTIONS. 1. Introduction

Chap. 1. Some Differential Geometric Tools

ALGEBRAIC GROUPS J. WARNER

Superintegrable 3D systems in a magnetic field and Cartesian separation of variables

Atiyah classes and homotopy algebras

Harmonic Oscillator Lie Bialgebras and their Quantization

9. The Lie group Lie algebra correspondence

Lecture 4 Super Lie groups

COULOMB SYSTEMS WITH CALOGERO INTERACTION

arxiv:math-ph/ v1 29 Dec 1999

arxiv:math-ph/ v1 25 Feb 2002

Symplectic and Poisson Manifolds

ne varieties (continued)

Towards a modular functor from quantum higher Teichmüller theory

The Erwin Schrodinger International Pasteurgasse 6/7. Institute for Mathematical Physics A-1090 Wien, Austria

THE QUANTUM DOUBLE AS A HOPF ALGEBRA

PROBLEMS, MATH 214A. Affine and quasi-affine varieties

= 0. = q i., q i = E

LECTURE 1: LINEAR SYMPLECTIC GEOMETRY

Hopf subalgebras of the Hopf algebra of rooted trees coming from Dyson-Schwinger equations and Lie algebras of Fa di Bruno type

PLEASE SCROLL DOWN FOR ARTICLE

arxiv:math/ v1 [math.qa] 4 Jul 2001

Quasi-classical analysis of nonlinear integrable systems

Ruijsenaars type deformation of hyperbolic BC n Sutherland m

Recursion Systems and Recursion Operators for the Soliton Equations Related to Rational Linear Problem with Reductions

arxiv:dg-ga/ v3 21 Oct 1999

R_ -MATRICES, TRIANGULAR L_ -BIALGEBRAS, AND QUANTUM_ GROUPS. Denis Bashkirov and Alexander A. Voronov. IMA Preprint Series #2444.

Hamiltonian Dynamics from Lie Poisson Brackets

Some half-bps solutions of M-theory

Quantum Groups and Link Invariants

Representations of algebraic groups and their Lie algebras Jens Carsten Jantzen Lecture III

A 2-COCYCLE ON A GROUP OF SYMPLECTOMORPHISMS

LIE ALGEBROIDS AND POISSON GEOMETRY, OLIVETTI SEMINAR NOTES

Notes on nilpotent orbits Computational Theory of Real Reductive Groups Workshop. Eric Sommers

AMADEU DELSHAMS AND RAFAEL RAMíREZ-ROS

On a Hamiltonian version of a 3D Lotka-Volterra system

Lecture I: Constrained Hamiltonian systems

arxiv:math/ v2 [math.dg] 14 Mar 2003

The Erwin Schrodinger International Boltzmanngasse 9. Institute for Mathematical Physics A-1090 Wien, Austria

Clifford Algebras and Spin Groups

THE INTRINSIC FORMALITY OF E n -OPERADS

Quasiperiodic Motion for the Pentagram Map

Entanglement entropy and gauge fields

Thermo Field Dynamics and quantum algebras

MATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA

Manifolds and tangent bundles. Vector fields and flows. 1 Differential manifolds, smooth maps, submanifolds

SYMPLECTIC MANIFOLDS, GEOMETRIC QUANTIZATION, AND UNITARY REPRESENTATIONS OF LIE GROUPS. 1. Introduction

Chern characters via connections up to homotopy. Marius Crainic. Department of Mathematics, Utrecht University, The Netherlands

Categorical quantum mechanics

The Dirac operator on quantum SU(2)

QUANTIZATION VIA DIFFERENTIAL OPERATORS ON STACKS

NO SLICES ON THE SPACE OF GENERALIZED CONNECTIONS. Peter W. Michor Hermann Schichl

arxiv:hep-th/ v1 2 Feb 1996

Chapter 19 Clifford and the Number of Holes

(1) is an invertible sheaf on X, which is generated by the global sections

Holonomy groups. Thomas Leistner. School of Mathematical Sciences Colloquium University of Adelaide, May 7, /15

Transcription:

Modern Physics Letters A 14, 30 1999) 109-118 CONSTRUCTION OF COMPLETELY INTEGRABLE SYSTEMS BY POISSON MAPPINGS J. Grabowsi, G. Marmo, P. W. Michor Erwin Schrödinger International Institute of Mathematical Physics, Wien, Austria Sept. 01, 1999 Abstract. Pulling bac sets of functions in involution by Poisson mappings and adding Casimir functions during the process allows to construct completely integrable systems. Some examples are investigated in detail. 1. Introduction The standard notion of complete integrability is the so called Liouville-Arnold integrability: a Hamiltonian system on a n-dimensional symplectic manifold M is said to be completely integrable if it has n first integrals in involution which are functionally independent on some open and dense subset of M. It is natural to extend the notion of complete integrability to sytems defined on Poisson manifolds N, Λ) by requiring that on each symplectic leaf such system defines a completely integrable system in the usual sense. This generalization implies that an integrable system is associated to a maximal abelian Poisson subalgebra of C N), {, } Λ ). The dynamical system Γ associated to a 1-form α on N via Γ = i α Λ will define a Hamiltonian system on a symplectic leaf S with embedding ε S : S N if we have ε sα = dh S. If this is the case for any symplectic leaf we may write α = f dg where the f are Casimir functions for Λ. When all the g s belong to a sufficiently large set of functions in involution which are functionally independent on each leaf, the dynamical system is completely integrable. For some cases, even if ε Sdα 0, we get an integrable system in the generalized sense of [1]. Of course, integrable systems are not easy to find. Recently, in the paper [3] we came accross a beautiful idea to construct completely integrable systems by using coproducts in Poisson-Hopf algebras. In this paper we put this construction into a geometric perspective in order to understand better which are the essential ideas 1991 Mathematics Subject Classification. 58F07. Key words and phrases. Completely integrable systems, Poisson mappings. J. Grabowsi was supported by KBN, grant Nr P03A 031 17. 1 Typeset by AMS-TEX

GRABOWSKI, MARMO, MICHOR that mae the construction possible. In addition to this, we construct a full family of Poisson-Hopf algebras associated with a parametrized family of Poisson-Lie structures on the group SB, C). The standard Lie-Poisson structures on SB, C) with SU) and SL, R) as dual groups are included in this scheme. This Poisson- Hopf algebras can be viewed as geometrical version of the corresponding quantum groups deformations of the universal enveloping algebra of the associated Lie algebras. We also present symplectic realizations of the corresponding commutation rules in the deformed algebras.. Constructing integrable systems by Poisson maps.1. Complete integrability on Poisson manifolds. If M, Λ) is a Poisson manifold, a Hamiltonian system H C M) is called completely integrable if it admits a complete set of first integrals in involution: There are f 1,..., f C M) with {f i, f } Λ = 0 and {f i, H} Λ = 0 such that on each symplectic leaf or an open dense set of symplectic leaves) H together with a suitable subset of f 1,..., f is Liouville-Arnold integrable... Constructing families of functions in involution by Poisson maps. Let Φ i : M i+1, Λ i+1 ) M i, Λ i ) be Poisson maps between manifolds, so that {f, g} i Φ i = {f Φ i, g Φ i } i+1. If we have a family of functions F 1 C M 1 ) in involution on M 1, Λ 1 ), we may consider the family F = F 1 Φ 1 ) C C M ) where C is a complete set of Casimir functions on M, Λ ), and so on: M 1, Λ 1 ) F 1, in involution Φ 1 M, Λ ) F = F 1 Φ 1 ) C Φ M 3, Λ 3 ) F 3 = F Φ ) C 3 Φ 3...3. Poisson actions and multiplications. We shall apply the procedure of. mainly in the following situation: Consider M 1 M, Λ 1 Λ ). Then for the algebras of smooth functions we have C M 1 M ) = C M 1 ) C M ) for some suitable completed tensor product, where f 1 f )x, y) = f 1 x)f y). Then {f 1 f, g 1 g } Λ1 Λ = {f 1, g 1 } Λ1 g 1 g + f 1 f {g 1, g } Λ1. So if c 1 is a Casimir function of M 1, Λ 1 ), then c 1 1 is a Casimir function of M 1 M, Λ 1 Λ ). In this sense The Casimir functions of M 1, Λ 1 ) and those of M, Λ ) extend both to Casimir functions on M 1 M, Λ 1 Λ ). If Φ : M M, Λ Λ) M, Λ) is a Poisson map for example the multiplication of a Lie Poisson group) we may use it for the procedure of.. If Φ is associative then Φ : f f Φ is coassociative. But this is not essential for applying the

CONSTRUCTION OF COMPLETELY INTEGRABLE SYSTEMS 3 procedure in which M n = n M and Φ n is a Cartesian product of Φ with identities. For example, M Φ M M Φ Id M M M M Id M Φ Id M M M M M... We start with a set of functions F 1 C M) in involution and with a basis C of all Casimirs. Then F n C n M) is given recursively by F n+1 = F n Φ n ) {C 1 1, 1 C 1 1,... } and furnishes a family of functions in involution on n+1 M. If Φ is associative so Φ is coassociative) then the result does not depend on the path chosen to define the Φ n s. Another possibility is to consider a Poisson mapping Φ : M N, Λ M Λ N ) N, Λ N ) for example a Lie-Poisson action on N of a Lie Poisson group M) and to apply the procedure as follows: N Φ M N Id M Φ M M N Id M M Φ M M M N....4. We may extend the procedure described in.3 as follows. We assume that we have furthermore Poisson manifolds e.g. symplectic ones) N 1,... N n and Poisson mappings ϕ i : N i M. The product map ϕ = ϕ 1... ϕ n : N 1... N n n M is a Poisson map. Let F n be the set of functions in involution on n M constructed in.3. Then F n ϕ is a set of functions in involution on i N i. Standard examples of Poisson maps ϕ i : N i M are the canonical embeddings of symplectic leaves N i of the Poisson manifold M. In this case, the Casimir functions 1 1 c 1 are constants on N 1... N n, but the coproducts Φ c, Φ Id M ) Φ c, etc., are usually no longer Casimirs and hence sometimes give rise to completely integrable systems on N 1... N n. See example 3.1. 3. Examples 3.1. Example. Let M = su) be the dual space of the Lie algebra su). It carries a Kostant-Kirillov-Souriau Poisson structure which is given in linear coordinates by Λ = z x y + x y z + y z x. Since Λ is linear, we have the obvious Poisson map Φ : M M M, Φx 1, y 1, z 1, x, y, z ) = x 1 + x, y 1 + y, z 1 + z ). A Casimir function for Λ is c = x + y + z. According to our procedure in.3 the functions c Φ = x 1 + x ) + y 1 + y ) + z 1 + z ), c 1 = x 1 + y 1 + z 1, 1 c = x + y + z, f Φ = fx 1 + x, y 1 + y, z 1 + z ),

4 GRABOWSKI, MARMO, MICHOR are functions in involution on M M, where f is an arbitrary function on M. If we tae for N the symplectic leaf N = c 1 1) which is a -dimensional sphere S, the Casimir functions c 1 and 1 c pull bac to constants on N N = S S. However, c Φ and f Φ are in involution and hence H = 1 c Φ) 1 = x 1x + y 1 y + z 1 z defines a completely integrable system on the symplectic manifold N N M M. The system defined by the Hamiltonian function H is, in fact, completely integrable on each symplectic leaf of M M, so that we get a completely integrable system on M M whose dynamics is given by the vector field Γ = z y 1 y z 1 ) x1 + z 1 y y 1 z ) x + x z 1 z x 1 ) y1 + x 1 z z 1 x ) y + y x 1 x y 1 ) z1 + y 1 x x 1 y ) z. This vector field is tangent to all products of spheres since c 1 = x 1 + y 1 + z 1 and 1 c = x + y + z are first integrals, and on N N it induces the motion which can be interpreted as associated with a spin-spin -interaction: J 1 = J 1 J, J = J J 1 ; The points on the spheres move in such a way that the velocity of each of them is the vector product of the two position vectors. Stationary solutions occupy the same or opposite points on the sphere. In sperical coordinates, the same system can be given a different interpretation: z i = sin β i, y i = cos β i sin α i, x i = cos β i cos α i. In canonical coordinates p i = sin β i, q i = α i we get the Hamiltonian function in the form H = p 1 p + 1 p 1 )1 p ) cosq 1 q ). Since H 1 = z ) c) is in involution with z ) we can consider the completely integrable system given by the Hamiltonian H 1 = p 1 + p 1 p 1 )1 p ) cosq 1 q ). Let us remar that our Hamiltonian H is a slight modification of the Hamiltonian H 0 = p 1 p p 1 p cosq 1 q ) obtained in [3]. We can inductively apply our procedure to get a completely integrable system on M with Hamiltonian H ) = x i x j + y i y j + z i z j ) i<j which reduces in canonical coordinates on N to H ) = i<j ) p i p j + 1 p i )1 p j ) cosq i q j ).

CONSTRUCTION OF COMPLETELY INTEGRABLE SYSTEMS 5 3.. Example. We consider the following symplectic realization [8] of the Lie algebra su) in T R : x = 1 q 1q + p 1 p ), y = 1 p 1q q 1 p ), z = 1 4 p 1 + q 1 p q ). This defines a Poisson morphism ψ : T R su), q 1, q, p 1, p ) x, y, z), which is the momentum map of the corresponding Hamiltonian action of the group SU). As before, we consider the Casimir function c = x + y + z on su). This time, however, F = c ψ = 1 16 p 1 + p + q1 + q) is not a Casimir function on the symplectic manifold T R. The functions in involution on su) su) from example 3.1 give rise to functions in involution on T R T R = T R 4 as in.4, where q 1 etc. denote the functions on the second copy of T R : where F 1 = F q 1, q, p 1, p ), F = F q 1, q, p 1, p ), c) ψ = F 1 + F + H, H = 1 4 ) q 1 q + p 1 p ) q 1 q + p 1 p ) + p 1 q q 1 p ) p 1 q q 1 p ) and f) ψ = fg 1, G, G 3 ), where G 1 = 1 q 1q + p 1 p + q 1 q + p 1 p ), G = 1 p 1q q 1 p + p 1 q q 1 p ), + 1 16 p 1 + q 1 p q ) p 1 + q 1 p q ) G 3 = 1 4 p 1 + q 1 p q + p 1 + q 1 p q ). Hence, we have 4 independent functions in involution on T R 4 which define completely integrable systems. As Hamiltonian functions we can tae the pure interaction term H. The trajectories of the corresponding dynamics Γ H lie on the intersections of the level sets of F 1 and F which are, topologically, products of 3-dimensional spheres) and the level sets of H, and, say, G 1 which are, generically, 4-dimensional tori). Note that G and G 3 are additional constants of the motion. The whole set {F 1, F, H, G 1, G, G 3 } is, however, not independent, since G 1 + G + G 3 = F 1 + F + H. The dynamics Γ H on T R 4 = R 8 is described by a rather complicated vector field whose coefficients are polynomials of degree 3. The functions G 1, G, G 3 define the diagonal action of SU) on T R T R which preserves Γ H. The dynamics on S S from example 3.1 can be obtained via symplectic reduction with respect to this action.

6 GRABOWSKI, MARMO, MICHOR 3.3. Example. Let us now consider the Lie group M = SB, C) of all matrices of the form ) e z/ x + iy A = 0 e z/, where 0 is fixed and z, x, y R may be viewed as global coordinates on SB, C). The coproduct corresponding to the group multiplication Φ : SB, C) SB, C) SB, C) is given by z) = z 1 + 1 z, x) = x e z/ + e z/ x, y) = y e z/ + e z/ y. Φ is a Poisson map for each of the following Poisson structures parameterized by α, β, γ R): *) {z, x} = βy, {y, z} = αx, {x, y} = γ sinz). For α = β = γ = 1 we get the classsical realization of the quantum SU) group, and for α = β = γ = 1 we get the classsical realization of the quantum SL, R) of Drinfeld and Jimbo, [6]. For 0 we get the Lie algebras {z, x} = βy, {y, z} = αx, {x, y} = γz, with the standard cobracets u) = u 1 + 1 u, corresponding to the addition in g. A Casimir function for the Poisson bracet *) is c = αx +βy + 4γ sinh z/), which for 0 goes to c 0 = αx + βy + γz, see [7]. The generic symplectic leaves of the Poisson structure are -dimensional except for the trivial case α = β = γ = 0). As in example 3.1 the Hamiltonian H = 1 c) defines a completely integrable system on SB, C) SB, C). In the coordinates x, y, z the Hamiltonian H has the form H = 1 αx 1 + βy1 + 4γ sinh z1 )) e z + 1 αx + βy + 4γ sinh z )) e z1 + αx 1 x + βy 1 y + 4γ sinh z1 z ) sinh )) e z1 z)/. In the limit for 0 we get H 0 = c 0 1 + 1 c 0 + αx 1 x + βy 1 y + γz 1 z ) and for α = β = γ = 1 we are in the situation of example 3.1. In all generality, however, it is difficult to express the dynamics explicitly since we deal simultaneously with a parametrized family of structures for which even the topology of the symplectic leaves changes.

CONSTRUCTION OF COMPLETELY INTEGRABLE SYSTEMS 7 Even in the case α = β = γ = 1, where it is nown [] that SB, C), Λ) is equivalent, as a Poisson manifold, to su) with the Kostant-Kirillov-Souriau structure Λ 0 described in example 3.1, the dynamics described by H in the deformed case may differ from that of example 3.1. The reason is that SB, C), Λ) is not equivalent to su), Λ 0 ) as a Lie Poisson group, since SU, C) is not commutative. In particular, the deformed coproduct is not cocommutative and the interaction we obtain is not symmetric. In order to wor in canonical coordinates let us introduce a symplectic realization of the commutation rules *) with α = δ > 0 and β = 1: X = a 4γ sinh p ) sinδq), Y = δ a 4γ sinh p ) cosδq), Z = p, where a 0 and a > 0 if γ = 0. In particular, if γ > 0 we get the deformed SU), and if γ < 0 we get the deformed SL, R). In this realization the Casimir function is c = a and the Hamiltonian reads H = e p p1)/ a 4γ sinh p1 ) ) a 4γ sinh p ) ) cosδq 1 q ))+ + a cosh p1+p ) + 4γ sinh p1 ) sinh p ) ). This Hamiltonian is quite complicated. But if we put a = 0, γ = 1, and δ = 1 we get H 1 = 4e p p1)/ 1 sinh p1 p ) sinh )cosq 1 q ) 1) which is the Hamiltonian obtained in [3] for the deformed SL, R). 3.4. Example. A slight modification of the previous example which, at least formally, is not dealing with Lie-Poisson groups, is the the following. Let M be the space of all upper triangular matrices of the form A = a x + iy 0 b where a, b, x, y R. We use these as coordinates on M. We consider the Poisson structure Λ on M with Poisson bracet ), **) {x, a} = βya, {x, b} = βyb, {x, y} = γb a ), {y, a} = αxa, {y, b} = αxb, {a, b} = 0, where α, β, γ R paramerterize a family of Poisson bracets. Matrix multiplication on M leads to the coproduct: a) = a a, b) = b b, x) = a x + x b, y) = a y + y b. The matrix multiplication turns out to be a Poisson mapping with respect to all bracets **). But M is not a Lie-Poisson group since it contains elements which are not invertible. On the other hand, all Λ are tangent to SB, C) M and give

8 GRABOWSKI, MARMO, MICHOR there the bracets *) with slightly modified coeficients α, β, γ, if we parameterize a = e z/, b = e z/. The Poisson tensor has independent Casimirs: c 1 = ab and c = αx + βy + γa + b ). The symplectic leaves are, generically, -dimensional we assume that α + β + γ 0), and as before, H = 1 c ) = 1 a 1αx + βy + γa + b ))+ 1 b αx 1 + βy 1 + γa 1 + b 1))+ a 1 b αx 1 x + βy 1 y γa 1 b ) describes a completely integrable system on M M which, on SB, C) = c 1 1 1), coincides with a system from example 3.3. 3.5. Example. This is of different type. Let D = G.G = G.G be a complete Drinfeld double group. The decompositions give us two Poisson projections π 1, π : D G onto the Lie-Poisson group G. Let F, F be two families of functions in involution on G. It is nown that pull bacs by π 1 and by π commute with respect to the symplectic structure on D. So we can tae F = F 1 π 1 ) F π ) as a set of functions in involution on D. For example, SL, C) = SU).SB, C) = SB, C).SU), where π 1 denotes the projection onto the left factor SB, C), and π onto the the right one. Let F consist of the Casimir C and some function f. Then the functions c π 1, f π 1, g π are in involution on D, where f and g are arbitrary in C SB, C)), and they are generically independent, so we have: H = f π 1 is a completely integrable system on the symplectic SL, C) for any f C SB, C)). 3.6. Concluding remars. 1. We have shown that by using Poisson-compatible coproducts it is possible to generate interacting systems while preserving the complete integrability. We have given some examples. These can be extended to arbitrary Lie Poisson pairs.. The interaction we get is a -body interaction, one may wonder if it would not be possible to obtain non-factorizable n-body interactions by using n-ary bracets or n-ary operations [7], [8], [9]). 3. The composition procedure does not use the existence of an inverse for each element in the product, therefore the procedure may be extended to any algebra. Is it possible to obtain interacting systems with fermionic degrees of freedom by using graded algebras? 4. The procedure may clearly be extended to infinite dimensions. Can one use it to obtain interacting fields? We shall come bac to some of these questions in the near future.

CONSTRUCTION OF COMPLETELY INTEGRABLE SYSTEMS 9 References [1] Aleseevsy, D.; Grabowsi, J.; Marmo, G.; Michor, P.W., Completely integrable systems: a generalization, Modern Physics Letters A 1 1997), 1637 1648. [] Aleseevsy, D.; Grabowsi, J.; Marmo, G.; Michor, P.W., Poisson structures on double Lie groups, J. Geom. Physics 6 1998), 340-379, math.dg/980108. [3] Ballesteros, A; Consetti, M.; Ragnisco, O., N-dimsional classical integrable systems from Hopf algebras, Czech. J. Phys. 46 1996), 1153. [4] Ballesteros, A.; Ragnisco, O., A systematic construction of completely integrable Hamiltonians from coalgebras, J. Phys. A: Math. Gen. 31 1998), 3791-3813. [5] Ballesteros, A.; Ragnisco, O., N = Hamiltonians with sl)-coalgebra symmetry and their integrable deformations, Roma TRE -prepint, February 1999. [6] Drinfeld, V.G., Quantum groups, Proc. Intern. Congress Math., Bereley 1986. [7] J. Grabowsi and G. Marmo, Generalized n-poisson Bracets on a Symplectic Manifold, ESI preprint 669, math.dg/99019,, Modern Phys. Lett. A 13 1998), 3185 319. [8] J. Grabowsi and G. Marmo, On Filippov algebroids and multiplicative Nambu-Poisson structures, ESI preprint 668, math.dg/99017, to appear in Diff. Geom. Appl.. [9] J. Grabowsi and G. Marmo, Remars on Nambu-Poisson and Nambu-Jacobi bracets, ESI preprint 670, math.dg/99018, J. Phys. A: Math. Gen. 3 1999), 439 447. [10] Grabowsi, J.; Marmo, G.; Perelomov, A., Poisson structures: towards a classification, Mod. Phys. Lett. A 8 1993), 1719 1733. [11] Man o, V.I; Marmo, G; Vitale, P; Zaccaria, F., A generalization of the Jordan-Schwinger map: the classical version and its q-deformations, Int. J. Mod. Phys. A 9 1994), 5541 5561. J. Grabowsi: Institute of Mathematics, University of Warsaw, ul. Banacha, PL 0-097 Warsaw, Poland; and Mathematical Institute, Polish Academy of Sciences, ul. Śniadecich 8, P.O. Box 137, PL 00-950 Warsaw, Poland E-mail address: jagrab@mimuw.edu.pl G. Marmo: Dipart. di Scienze Fisiche - Università di Napoli, Mostra d Oltremare, Pad.19, I-8015 Napoli, Italy. E-mail address: gimarmo@na.infn.it P. W. Michor: Institut für Mathemati, Universität Wien, Strudlhofgasse 4, A-1090 Wien, Austria; and Erwin Schrödinger International Institute of Mathematical Physics, Boltzmanngasse 9, A-1090 Wien, Austria E-mail address: peter.michor@univie.ac.at, peter.michor@esi.ac.at