Integrable systems, toric degenerations and Okounkov bodies (Q896254)

The problem of integrability, though dating back to the middle of the XIX century, is still an active subject with new methods being developed. The original problem concerned dynamical systems (often having mechanical origin), where the key observation was that providing some conservation laws and symmetries permits to solve the equations in a rather explicit way, or at least claim that the system is more regular than a generic one. A good language to formalize the notion is related to Poisson or symplectic manifolds. The ``regularity'' result then says that a sufficient number of independent Poisson-commuting conserved quantities (first integrals) guarantees that the phase space of the system can be foliated by invariant tori, the motion on any of them being quasi periodic. Similar philosophy applies also to partial differential equations and even to some field theories. But even for ordinary differential equations there is no general method to decide if a given system is integrable. A natural idea would be then to restrict the class of considered first integrals to, say, polynomial, real analytic or meromorphic functions. And in most of the cases differential and algebraic geometry is still a good language to discuss their properties. There are thus two natural questions people study: spell out obstructions to integrability and provide natural non-trivial examples of integrable systems, the current paper by M.~Harada and K.~Kaveh concerns rather the latter one. The authors consider a generalization of the classical construction to complex projective varieties. They give a definition and provide sufficient conditions for complete integrability in the setting, as well as give natural examples. More precisely, the first major part of the paper is devoted to thorough transfer of the Hamiltonian formalism and some concepts from differential equations to complex varieties, carefully taking into account smoothness and critical points. As for the sufficient conditions, the main tool is toric degenerations, which under some (natural) technical assumptions provide integrable systems. The second part is devoted to toric degenerations that appear in the theory of Newton--Okounkov bodies: the authors basically go through the construction of the first part and show that those satisfy all the hypothesis and thus provide examples of integrable systems. The paper is essentially self-contained and well written. The authors even claim that since they combine techniques from various fields of contemporary mathematics, they wanted to make it ``accessible to broad audience'' -- and this effort is clearly visible. The paper is thus rather long and detailed, so a specialist might want to skip some definitions and known facts, or even just read the introduction which provides a good flavour of results. For a non-specialist some knowledge or reading on algebraic varieties, schemes and maybe graded algebras and valuations is advised.