Integrable systems in the realm of algebraic geometry (Q2563838): Difference between revisions

Today we observe the revival of interest in integrable systems motivated by a discovery of integrable nonlinear evolution equations of soliton type which have a lot of applications in modern physics. New deep understanding of the integrability includes now results of Lie theory, representation theory, differential and algebraic geometry [see e.g. \textit{A. M. Perelomov}, ``Integrable systems of classical mechanics and Lie algebras'', Moscow: Nauka (1990; Zbl 0717.70003)]. In many cases, the constructions that constitute the integrable systems are algebraic: the phase space has the structure of an affine algebraic variety, the Poisson brackets are regular functions, the integrals are polynomials etc. It means that for such a kind of integrable systems the algebraic-geometric tools are of prime importance in studying and solving them. Recently even the concept of completely integrable system was coined in mathematical physics in order to distinguish this type of systems. The author's book aims to present in a coherent way a theory of algebraic completely integrable systems of classical mechanics in the language of algebraic geometry. In the Introduction, the author gives an exposition of main ideas of the book, a history of their development, and explains the structure of the book. In Chapter 2, the author introduces the concept of an affine Poisson variety, which is just an affine variety endowed with a Poisson structure on its algebra of regular functions. Each function is associated with an appropriate Hamiltonian vector field. A function whose associated vector field is zero is called a Casimir. The Casimirs form a subalgebra of the algebra of regular functions and its spectrum forms an affine algebraic variety. Restricting the Poisson structure to a level set of the Casimirs one produces a new Poisson variety. Studies of different types of morphisms between affine Poisson varieties are carried out in detail. The author introduces also the important notion of the algebra of integrals, i.e., a set of regular functions in involution which has maximal dimension and is complete in the sense that every function in involution with all elements of the algebra is contained in it. Then an integrable Hamiltonian system is defined as an affine Poisson variety with a maximal algebra of Casimirs and a complete algebra of integrals. The author investigates carefully morphisms between integrable Hamiltonian systems and describes methods to construct new integrable Hamiltonian systems from given ones. In Chapter 3, the author generalizes a construction due to Mumford and builds a large family of integrable Hamiltonian systems on \(\mathbb{C}^{2d}\) with a Poisson structure of maximal rank \(2d\) corresponding to an arbitrary polynomial \(\phi(x,y)\) in two variables and with a complete algebra of integrals corresponding to another arbitrary polynomial \(F(x,y)\). The construction is described very clearly by explicit formulae. In the special case when the equation \(F(x,y)=0\) defines a hyperelliptic curve, a lot of simplifications occur and the vector fields of the integrable system may be presented in a Lax form. The fibers of the moment map are shown to be isomorphic to an affine part of the \(d\)-fold symmetric product of the curve \(\Gamma_F\). The author describes neatly the topology of the general fibers of the moment map which have in general nothing to do with Abelian varieties. For different choices of \(\Phi(x,y)\), all corresponding vector fields are tangent to the fibers of the moment map and are related to each other in a quite simple way. For the convenience of the reader, Chapter 4 is added in order to present the essentials of the algebraic geometry and Abelian varieties and thus to make the text more self-contained. The powerful machinery of algebraic geometry and Abelian varieties is used in Chapter 5 in order to analyse the notion of algebraic integrability. Loosely speaking, the Hamiltonian system is called algebraic completely integrable if the general level sets of the moment map are isomorphic to affine parts of Abelian varieties and the flow of the integrable vector fields is linearized by this isomorphism. Here Painlevé analysis is used as a tool to determine the nature of the general level sets and to embed the completed level sets explicitly in a projective space. Using results by Semenov-Tian-Shansky, the author discusses shortly the classical Yang-Baxter equations and the Lax equations which exhibit many aspects of integrable system in a unifying way. Further, in Chapter 6 the author studies the algebraic integrable subclasses of the hyperelliptic integrable Hamiltonian systems introduced in Chapter 3: 1) the so-called odd master systems constructed by Mumford and 2) the even ones introduced by the author. Painlevé analysis of these systems is shown to lead to a stratification of the Jacobians which appear as the fibers of the moment map. It is interesting that this stratification is also obtained from the natural stratification of the Sato Grassmannian via the Krichever map. Then the author considers some examples of the two-dimensional even master system which appears to be isomorphic to the so-called Bechlivanidis-van Moerbeke system. In particular he worked out the two-dimensional even master system which corresponds to an algebraic curve of genus two with an automorphism of the order three. This automorphism is extended to the Jacobian and leads to a singular quotient similar to the classical Kummer surface. At last, in Chapter 7 the author presents some examples of applications of the theory of algebraic completely integrable systems described above: 1) the Garnier systems and Abelian surfaces of type (1,4), 2) the Hénon-Heiles hierarchy and 3) the Toda lattices. In conclusion, we now have a good book on algebraic aspects of modern understanding of the integrability which unifies classical results of the nineteenth century and new achievements of the last one.