Locally conformally symplectic reduction of the cotangent bundle (Q2116425)

Although there exist other characterizations, locally conformally symplectic (LCS for short) manifolds can be defined as manifolds that are locally conformal to symplectic manifolds. Part of their interest relies on the fact that they become natural phase spaces of Hamiltonian dynamical systems (see [\textit{C.-M. Marle}, J. Geom. Mech. 4, No. 2, 181--206 (2012; Zbl 1417.37187)] in relation to the Kepler problem); from a mathematical perspective, Eliashberg and Murphy remarkably proved that LCS manifolds are widespread, and hence these manifolds are much numerous than symplectic ones. For instance, \textit{S. Haller} and \textit{T. Rybicki} [Ann. Global Anal. Geom. 17, No. 5, 475--502 (1999; Zbl 0940.53044)] pointed out that in addition to the canonical symplectic structure, the cotangent bundle of a given manifold has various LCS structures. Hence, one can try to study to which extent properties of symplectic manifolds and techniques in classical symplectic geometry generalize to the LCS realm. As it is accounted in the beautiful survey [\textit{G. Bazzoni}, EMS Surv. Math. Sci. 5, No. 1--2, 129--154 (2018; Zbl 1411.53001)] (see references therein), this line of research is vigorous, since that a Moser trick, blow-up of a LCS manifold, the notion of Lagrangian submanifolds and their displacement, as well as neighbourhood theorems for them have been incorporated to the LCS realm. At this point, it is natural to ask whether the classical reduction theorems for symplectic manifolds can be extended to LCS manifolds. Haller and Rybicki proved that, in regular cases, the symplectic reduction is possible for LCS manifolds, preserving the conformal equivalence class of an LCS structure. The problem is that since they retain the idea of factoring by the group action itself, in the general case, the level sets of the momentum mapping no longer satisfy the hypotheses that they impose. In his previous article [\textit{M. Stanciu}, Ann. Global Anal. Geom. 56, No. 2, 245--275 (2019; Zbl 1422.53070)], the author did the opposite to successfully adapt the Marsden-Weinstein-Meyer symplectic reduction to LCS manifolds: he factors the level sets of the momentum mapping, as in the symplectic case, but not through the group action, but along a foliation derived by ``twisting'' the group action. Moreover, in the final section, he studied three concrete classes of examples on which the author applies his main theorem, including the reduction of the cotangent bundle in \S6.3. Note that there exist other approaches in the literature to the reduction procedure for LCS manifolds. \textit{T. Noda} [Tsukuba J. Math. 28, No. 1, 127--136 (2004; Zbl 1077.53067)] proved the classical symplectic reduction procedure for such manifolds with some additional assumptions (see the footnote in page 128). Also, interestingly, \textit{J. C. Marrero} et al. [Ann. Global Anal. Geom. 40, No. 3, 311--337 (2011; Zbl 1228.53093)] obtained universal models for several types of LCS manifolds via reduction. In this clear and self-contained article, the author deepens in the reduction procedure for the cotangent bundle in the LCS setting. The key point consists of the fact that the universality property with respect to reduction (as stated in Theorem 2.2.1 of [\textit{H. Cendra} et al., Lagrangian reduction by stages. Providence, RI: American Mathematical Society (AMS) (2001; Zbl 1193.37072)] and recorded as Theorem 3.3 in the article under review) holds, in the same conditions, for LCS structures of a cotangent bundle reduced at any regular value. This is Theorem 3.5, which is the main result of the article. To prove it, the author applies Theorem 2.15 (which is the main result of his above-mentioned previous article), making sure that the quotient by the foliation is smooth in order to be able to apply LCS reduction. As in the usual symplectic case, the theorem requires the existence of a 1-form \(\alpha_\xi\) on the manifold endowed with a free and proper Lie group action. However, whereas in the symplectic case the 1-form with the required properties can be constructed from certain connections (see Proposition 3.4), the conditions imposed by Theorem 3.5 for the existence of \(\alpha_\xi\) may not be satisfied. In fact, in the last section the author shows some manifolds on which the 1-form \(\alpha_\xi\) with the required properties cannot exist. In this way, he is able to construct a class of examples of cotangent bundles whose reductions at nonzero regular values are not cotangent bundles.