Reduction of multisymplectic manifolds (Q830418)

The main goal of this very well written paper is to extend to the multi-symplectic setting both the Marsden-Weinstein-Mayer reduction theorem and the analysis of the dependence of the reduced space on the reduction parameters. It is worth recalling that the above cited theorem plays a crucial role in symplectic geometry as a tool to construct new symplectic manifolds starting from old ones carrying a suitable Lie group action. The suitableness of the \(G\)-action is measured by the possibility to define an associated \textit{moment map}, which is used to cut the \textit{irrelevant} degrees of freedom of the original symplectic manifold to produce a new one. When \(G\) is a torus, a fundamental result of Duistermaat-Heckman states that the difference of de Rham classes of the reduced symplectic forms, obtained reducing the original one choosing two different regular values of the moment map, is controlled by the common Chern class of the corresponding fibrations. When one moves from classical mechanics to classical field theory, the symplectic formalism is traded with the multi-symplectic one. In this enhanced setting the phase space is a smooth manifold \(M\) endowed with a \(k\)-symplectic structure, i.e., a closed \((k+1)\)-form \(\omega\) such that the corresponding \(\flat\)-map, \(\omega^\flat(X)=i_X\omega\), is injective. The notions of locally and globally Hamiltonian vector fields can be naturally extended to this setting, and this extension yields a rich supply of (local) symmetries of the underlying multi-symplectic structure. In particular, \(\Omega^{k-1}_H(M)\), the vector space of the \textit{Hamiltonian} \((k-1)\)-forms, can be equipped with a remarkable bilinear form, playing the role of a Poisson bracket, see Definition 5 in Section 2. It is interesting to note that the definition of this bracket suffers a certain arbitrariness and that the choice made in the paper yields a bilinear form defining a structure of a Leibniz algebra on \(\Omega^{k-1}_H(M)\), see Lemma 1. In Section 3 the multi-symplectic analogue of a Hamiltonian \(G\)-space is introduced. There, first the relevant notion of comomentum map is introduced, then the relation of the chosen definition with the many others available in the literature is analyzed, see Item 3 in Remark 4. In Definition 8 the notion of split moment map is presented and then it is used as the main ingredient of the reduction theorem (Theorem 1) discussed in details in Section 4. As remarked there, the reduced multi-symplectic structure can be singular, in analogy to what happens in the reduction process of the so-called poly-symplectic structures, see the paragraph just above Remark 6 in this section. Theorem 2 deals with the multi-symplectic analogue of the Noether theorem, i.e., the reduction of the dynamics from the original to the reduced multi-symplectic manifold. Section 5 is devoted to the multi-symplectic version of the Duistermaat-Heckman theorem and Section 6 delves into the analysis of the relation between equivariant localization and multi-symplectic geometry. The last section is a nice collection of observations about possible future directions of investigation. I found the paper very readable, very well written and informative.