An exhaustive criterion for the non-existence of invariant circles for area-preserving twist maps (Q1116514): Difference between revisions

Let \(M=T^ 1\times R\) be the cylinder or annulus. An area-preserving twist map f: \(M\to M\) is an area, orientation and end preserving \(C^ 1\) diffeomorphism which satisfies \(\partial \theta '/\partial y\geq K>0\), where \(f(\theta,y)=(\theta ',y'),\) (\(\theta\),y)\(\in M\) and K a constant. An invariant circle for an area-preserving twist map f is a homeomorphic image of \(T^ 1\) which is invariant under f. A rotational invariant circle is one which is homotopic to the generator of the fundamental group of \(M=T^ 1\times R\). It is known by Birkhoff that every rotational invariant circle \(\Gamma\) is the graph of a Lipschitz function \(\gamma\) : \(T^ 1\to R\). For a given area-preserving twist map f, the conefield \(K_ m\) \((m=0,1,2,3,...)\) are constructed. Using the conefield \(K_ m\) and Birkhoff's theorem, the cone-crossing criterion and the Killends criterion of MacKay and Percival are introduced. The notion of net flux or Calabi invariant of a twist map f is defined. Theorem: Given an area-preserving twist map with zero net flux, if x is a point such that for all \(n\in N\) there is a curve through x satisfying \(K_ n\), then there is a rotational invariant circle through x. This theorem implies the following conjecture made by MacKay and Percival. Finite Computation Conjecture; given a compact region through which no rotational invariant circles pass, then the algorithm will prove so using a finite amount of computation.