Cohomology of toric origami manifolds with acyclic proper faces (Q1689660): Difference between revisions

A celebrated result of \textit{T. Delzant} [Bull. Soc. Math. Fr. 116, No. 3, 315--339 (1988; Zbl 0676.58029)] states that compact symplectic toric manifolds \((M,\omega,T,\mu)\) are in 1-1 bijection with a certain kind of polytopes, which are known as \textit{Delzant} polytopes. The polytope \(\mu(M)\subset\mathfrak{t}^*\) is, as a topological space, the orbit space \(M/T\). From the knowledge of the polytope one can recover topological information on the symplectic manifold, for instance its Betti numbers. There exist two main generalizations of symplectic manifolds, which allow some \textit{degenerations} for the symplectic form: in the first case the symplectic form is allowed to vanish transversally along a codimension-1 submanifold, in the second one, it is allowed to explode along it. The first case is \textit{folded} symplectic geometry, the second one is \textit{b-} or \textit{log-}symplectic geometry. \textit{Origami} manifolds are a special class of folded symplectic manifolds, those for which the vanishing locus of the symplectic form is a principal \(S^1\)-bundle over a compact base. A result similar to Delzant's for toric origami manifolds was obtained in [\textit{A. Cannas} et al., Int. Math. Res. Not. 2011, No. 18, 4252--4293 (2011; Zbl 1236.53062)]. There the authors prove that they are in 1-1 bijection with \textit{origami templates}, a certain kind of graphs explaining how different moment polytopes are spliced together. The paper under review is concerned with determining the topological properties of a compact toric origami manifold starting from the origami template. If the origami template is a tree, one can use methods of [\textit{M. Masuda} et al., Osaka J. Math. 43, No. 3, 711--746 (2006; Zbl 1111.57019)]. This paper deals with an \textit{orientable} compact toric origami manifold \(M\) for which the orbit space \(M/T\) is arbitrary, but every proper space is acyclic (this assumption is non-trivial). Emphasis is put on Betti numbers and equivariant cohomology \(H^*_T(M)\), in particular the study of the restriction homomorphism from the equivariant cohomology to the ordinary cohomology; further applications are to the 4-dimensional case. The last section deals with a possible generalization to the case of non-acyclic faces.