On the Gromov width of polygon spaces (Q1742057): Difference between revisions

The authors study the space \(\mathcal M(r)\) of \(n\)-gons in \(\mathbb R^3\), where \(r=(r_1,\dots,r_n)\in \mathbb R^n_+\) corresponds to the lengths of edges. For a generic \(r\), the authors show that the Gromov width of the space of smooth \(5\)-gons and \(6\)-gons (under certain extra conditions on \(r\)) is given by \(2\pi \min_j (2r_j, (\sum_{i\neq j} r_i) - r_j)\). The same formula is a lower bound for all smooth spaces of \(6\)-gons. In addition, the authors show that the same formula for the Gromov width holds when \(\mathcal M(r)\) is symplectomorphic to \(\mathbb C P^{n-3}\) for \(n\geq 4\). The most important tool used by the authors is a toric action, called the bending action, defined on a dense open subset of \(\mathcal M(r)\). For many values of \(r\), this action can be defined on the whole \(\mathcal M(r)\). The flow of the bending action is used to construct symplectic embeddings of balls and thus obtain lower bounds for the Gromov width. In addition, the bending action plays a crucial role in determining upper bounds for the Gromov width, as certain tools are available for toric manifolds which are Fano, or blow-ups of Fano toric manifolds at toric fixed points. Using a Moser-type continuity argument the authors obtain upper bounds for the Gromov width of some non-toric spaces \(\mathcal M (r)\). The upper bounds coincide with the lower bounds the authors have determined by embedding techniques, and so the authors determine an explicit formula for the Gromov width of (some) polygon spaces.