From a Reeb orbit trap to a Hamiltonian plug (Q330085): Difference between revisions

This is a kind of ``parergon'' to the authors' path breaking article [Invent. Math. 198, No. 1, 211--217 (2014; Zbl 1304.37019)]. In that article, the authors defined a trap for Reeb orbits. Here, they construct a plug for a smooth Hamiltonian flow on hypersurfaces of dimension at least five. A plug means a smooth non-singular vector field \(X\) on \(D\times I\) where \(D\) is an \((m-1)\)-dimensional disc of radius \(\delta\) and \(I=[-\epsilon,\epsilon]\) with coordinate \(z\) such that \(X=\partial_z\) near \(\partial(D\times I)\) and such that there is a trajectory of \(X\) that enters \(D\times I\) at \(D\times\{-\epsilon\}\) and never leaves \(D\times I\). If the flow of \(X\) on \(D\times I\) has no closed orbits then we have a half-plug. For a plug, a matching condition is required: any orbit that transverses \(D\times I\) enters and exits this set at a pair of points \((x,\pm\epsilon)\). In their previous article [loc. cit.] the authors constructed a half-plug for Reeb flows. Here, they show how by using this construction one may find a half-plug for Hamiltonian flows. Then one puts one half-plug on top of the other in order to satisfy the matching condition. The reason why one wants to have a plug is that it allows one to destroy a single periodic orbit without introducing new ones.