When are the invariant submanifolds of symplectic dynamics Lagrangian? (Q379844): Difference between revisions

Lagrangian submanifolds are relevant tools for a better understanding of symplectic geometry as well as Hamiltonian dynamics. Indeed, \textit{W. M. Tulczyjew} [C. R. Acad. Sci., Paris, Sér. A 283, 15--18 (1976; Zbl 0327.58007); C. R. Acad. Sci., Paris, Sér. A 283, 675--678 (1976; Zbl 0334.58008)] has proved that the equations of motion in the Lagrangian and Hamiltonian descriptions of mechanics can be described by Lagrangian submanifolds of convenient symplectic manifolds (these results have been exploited in some recent papers, see [\textit{M. Barbero-Liñán} et al., Monatsh. Math. 171, No. 3--4, 269--290 (2013; Zbl 1312.70013)] and the references therein). In addition, Lagrangian submanifolds are just the geometric structure behind the Hamilton-Jacobi theory. The reasons behind the paper under review come from the study of symplectic dynamical systems, where the existence of invariant submanifolds have many interesting consequences. A paradigmatic case is just the so-called KAM theory (from Kolmogorov, Arnold and Moser); in fact, the KAM theorem gives the existence of a lot of such invariant tori of symplectic dynamics. But there may exist other invariants that are not KAM tori, and this is the subject of the present paper. More precisely, let \({\mathcal L}\) be a \(D\)-dimensional submanifold of a \(2D\)-dimensio\-nal exact symplectic manifold \((M, \omega)\) (remember that exact means that \(\omega = d \alpha\)), and let \(f : M \longrightarrow M\) be a symplectic diffeomophism. Then, the goal is to link the dynamics \(f_|{_{\mathcal L}}\) restricted to \({\mathcal L}\) and the geometry of \({\mathcal L}\). The author proves a series of very interesting results, in particular some related to the so-called Tonelli Hamiltonian functions.