Decision problems in the space of Dehn fillings (Q1868042): Difference between revisions

There is a vast literature on the space of Dehn fillings of a knot manifold, i.e. a compact, connected, orientable 3-manifold with connected boundary a torus, and particular attention has been devoted to the case of exceptional fillings of knot manifolds, for which many finiteness results and global bounds are known (see, for instance, the seminal paper by \textit{M. Culler, C. McA. Gordon, J. Luecke} and \textit{P. B. Shalen} [Ann. Math. 125, 237-300 (1987; Zbl 0633.57006)] or the survey by \textit{C. McA. Gordon} [Proc. Int. Congr. Math., Kyoto/Japan 1990, Vol. I, 631-642 (1991; Zbl 0743.57008)] and \textit{J. Luecke} [Proceedings of the International Congress of Mathematicians, ICM '94, Zürich, Birkhäuser (1995; Zbl 0855.57005)]). In this very rich paper, the authors address the problem of understanding the space of Dehn fillings from a constructive point of view, namely, they discuss methods to compute all the slopes giving ``special'' fillings. More precisely they determine the slopes for which the resulting manifold -- is reducible; -- is toroidal; -- contains an embedded, incompressible and two-sided surface; -- is Haken; -- fibres over \(S^1\); -- is homeomorphic with \(S^3\); -- is a lens space. The main tools to prove the results are the existence of special triangulations for knot manifolds and normal surface theory. The triangulations of the knot manifolds are chosen in such a way that they induce one-vertex triangulations of the boundary and can be extended to triangulations of the Dehn fillings, which, on their turn, restrict to ``layered'' triangulations of the added solid torus. Recall that every essential surface and every strongly irreducible Heegaard surface of an irreducible manifold has an almost normal representative in any triangulation of the manifold. Moreover, most ``interesting'' surfaces are to be found at the vertices of the projective solution space. These issues are discussed in detail in Sections 2, 3 and 4. Two types of algorithms are provided: -- Algorithms establishing particular properties (Section 5): First of all, a finite set of special slopes is determined. This set contains all the slopes whose associated Dehn fillings might (respectively might not) contain essential surfaces if the knot manifold does not (respectively does) contain essential surfaces. More precisely, essential spheres will be considered to detect reducible fillings, while essential surfaces of higher genus will give Haken manifolds. In this part various auxiliary algorithms are employed: to determine irreducibility of the manifolds (Rubinstein, Thompson), to determine incompressibility of a surface (Haken) and to determine whether a surface is a fibre of a fibration over \(S^1\) (Jaco, Tollefson). -- Algorithms of recognition (Section 6): in this case for each Dehn filling which might be a lens space, a strongly irreducible Heegaard surface is found, exploiting almost normal surface theory. Previously known algorithms are then applied to decide whether the manifolds are the \(3\)-sphere (Rubinstein, Thompson) or a lens space (Rubinstein). As a corollary, one can determine when a knot manifold is the complement of a knot in the \(3\)-sphere. Notice that these algorithms give an alternative proof of some of the finiteness results mentioned above. It is worth pointing out that the methods of the paper confirm the fact that, if a knot manifold enjoys certain properties, then the same properties are ``generically'' carried over to its Dehn fillings. In this spirit, it is also proved that it is sufficient to study the vertex solutions of the knot manifold to completely determine the ``interesting'' surfaces of all its Dehn fillings.