Heegaard surfaces and measured laminations. I: the Waldhausen conjecture (Q863439)

\textit{K. Johannson} [Bull. Am. Math. Soc., New Ser. 23, No. 1, 91--98 (1990; Zbl 0715.57006), Topology and combinatorics of 3-manifolds, Lecture Notes in Mathematics. 1599. Berlin: Springer-Verlag (1995; Zbl 0820.57001)] proved that a Haken 3-manifold can have only finitely many Heegaard splittings in each genus, up to isotopy. F. Waldhausen conjectured that this holds for (orientable) atoroidal 3-manifolds in general, a conjecture which is proven in the paper under review. The main tools used in the proof are branched surfaces, normal surfaces, and 0-efficient triangulations, the latter developed by \textit{W. Jaco} and \textit{J. H. Rubinstein} [J. Differ. Geom. 65, No. 1, 61--168 (2003; Zbl 1068.57023)], who have also announced a proof of the Waldhausen conjecture using different methods. Using a result of \textit{M. Stocking} [Trans. Am. Math. Soc. 352, No. 1, 171--207 (2000; Zbl 0933.57016)], every strongly irreducible Heegaard splitting is isotopic to an almost normal surface. The author constructs a finite collection of branched surfaces such that every almost normal strongly irreducible Heegaard surface is fully carried by a branched surface in the collection. Using complicated topological arguments, each of these branched surfaces is split into a finite collection of branched surfaces such that each almost normal irreducible Heegaard surface is still carried by a branched surface in the new collection, and moreover no branched surface in the collection carries any normal torus. The Waldhausen conjecture follows relatively easily from the existence of such a collection. In a sequel to this paper [\textit{T. Li}, J. Am. Math. Soc. 19, No. 3, 625--657 (2006; Zbl 1108.57015)], the author proves the stronger result that a closed orientable non-Haken 3-manifold has only finitely many irreducible Heegaard splittings.