Connections of Berry and Hannay type for moving Lagrangian submanifolds (Q749946)

The geometric phase was introduced in a seminal work by \textit{M. V. Berry} [Proc. R. Soc. Lond. A 392, 47-57 (1984)]. A classical analogue was obtained by \textit{J. Hannay}, in the case of integrable Hamiltonians subjected to adiabatic variations [J. Phys. A 18, 221-230 (1985)]. An important question, namely how to obtain the classical limit of Berry's phase in the general case, is solved in this paper. The author considers the space of Lagrangian submanifolds of a symplectic manifold P, equipped with smooth densities of total measure 1. These represent semiclassically the WKB amplitudes, while the Lagrangian submanifolds represent phase functions. Motivated by the physics of adiabatic systems, the motion of the Lagrangian submanifolds is restricted to isodrasts, defined by the constancy of action integrals. To have a suitable space in which the phases take their values, a prequantization of P is considered, i.e., a bundle with a connection whose curvature is the symplectic structure. From the introduction: ``If we have a prescription for assigning a weight to each Lagrangian submanifold, we get a connection over the Lagrangian submanifolds themselves, whose curvature depends on how we assigned the weights. Sometimes, this connection is flat, in which case we get a homotopy invariant for loops of Lagrangian submanifolds. This occurs, for instance, if we look at graphs of symplectomorphisms with the symplectic volume as the weighting [the author, Math. Z. 201, No.1, 75-82 (1989; Zbl 0644.57024)]. On the other hand, if the weightings are determined by a Kähler metric on P compatible with the symplectic structure, the Berry curvature vanishes at a Lagrangian submanifold L if and only if L is minimal.''