Multisymplectic observable reduction using constraint triples (Q6932065)

This paper deals with a higher analogue of symplectic reduction. Classical Marsden-Weinstein-Meyer reduction takes a symplectic manifold with a group action and an equivariant momentum map, considers a level set by fixing a value of the momentum map, and then quotients by the group to get a lower-dimensional symplectic manifold. This procedure requires good behavior (free and proper group action, regular value assumptions), which often fails and one has to deal with singular situations. Due to its importance, it is natural to ask whether one can perform reduction in the multisymplectic context as well. The motivation to approach the multisymplectic context lies in the limitations of applying the symplectic formalism to field theory. In particular, one works with a multisymplectic manifold naturally associated to the configuration bundle. In multisymplectic geometry, there are recent geometric and algebraic reduction procedures. It is important to remark that the algebraic perspective is very powerful because it works even in singular situations. The authors show how to reinterpret those constructions via constraint triples, a formalism that has been recently introduced by \textit{M. Dippell} et al. [Doc. Math. 24, 1811--1853 (2019; Zbl 1427.53105)], and they prove that one can canonically build an \(L_\infty\)-algebra of observables and reduce it. More precisely, as a first step the authors show that, when a certain cochain is chosen, any BV-module gives rise to an associated \(L_\infty\)-algebra of observables. As a second step, they develop the constraint algebra framework, introducing the notion of constraint BV-modules, showing how noncommutative differential calculus framework naturally integrates with the algebraic theory of constraint triple. Finally, they show how to construct a constraint \(L_\infty\)-algebra of observables from any constraint BV-module.