Morse families and Dirac systems (Q2303565)

Dirac structures were introduced by \textit{T. Courant} and \textit{A. Weinstein} as a unified approach to both pre-symplectic and Poisson geometries [\textit{T. J. Courant}, Trans. Am. Math. Soc. 319, No. 2, 631--661 (1990; Zbl 0850.70212); \textit{T. Courant} and \textit{A. Weinstein}, Trav. Cours 27, 39--49 (1988; Zbl 0698.58020)]. One of the motivations behind the definition of Dirac structures was the study of constrained systems, including the celebrated case of the constrained bracket induced by a degenerate Lagrangian function first studied by \textit{P. A. M. Dirac} [Can. J. Math. 2, 129--148 (1950; Zbl 0036.14104)]. The infinite-dimensional analogue was introduced by \textit{I. Y. Dorfman} [Phys. Lett. A, 125, 240--246 (1987)] in the context of integrable evolution equations. A unified approach to Dirac systems, based on the more general notion of Dirac algebroids, was provided by \textit{K. Grabowska} and \textit{J. Grabowski} [J. Geom. Phys. 61, No. 11, 2233--2253 (2011; Zbl 1223.37064)]. In the present paper the authors take an alternative point of view and generalize Dirac systems in such a way that the dynamics is defined by means of a Lagrangian submanifold of the phase space. They use Morse families to generate those Lagrangian submanifolds, following their approach for optimal control problems [\textit{M. Barbero-Liñán} et al., SIAM J. Control Optim. 53, No. 1, 414--433 (2015; Zbl 1319.49028)]. With this notion of generalized Dirac system they are able to recover many examples in the literature in a unified intrinsic formalism; these include nonholonomic and vakonomic mechanics, optimal control problems and constrained problems on linear almost Poisson manifolds.