Stabilizing and destabilizing Heegaard splittings of sufficiently complicated 3-manifolds (Q1938067)

The stabilization problem asks how many stabilizations are required for a pair of Heegaard splittings of a manifold to become equivalent. The upper bounds for this problem were established by Rubinstein and Scharlemann. For a large number of manifolds, it has been shown that the lower bound for the stabilization problem is just one. No example has been found for a long time which requires more that one stabilization. Based on these observations, the stabilization conjecture was established: Any pair of Heegaard splittings requires at most one stabilization to become equivalent. Recently, counter-examples to the stabilization conjecture were found, [\textit{J. Hass, A. Thompson} and \textit{W. Thurston}, Geom. Topol. 13, No. 4, 2029--2050 (2009; Zbl 1177.57018)]; \textit{J. Johnson}, [J. Topol. 3, No. 3, 668--690 (2010; Zbl 1246.57044)] and the paper under review. In the present paper, the author introduces a different way of constructing manifolds, i.e. gluing two manifolds together by complicated homeomorphism along some boundary component \(F\) of the manifolds, and \(F\) becoming a ``barrier'' to low genus topologically minimal surfaces. A similar idea was used by Lackenby when he studied the behavior of Heegaard genus under complicated gluings. The same idea leads to some other results, among them a resolution of the higher genus analogue of Gordon's conjecture (Gordon's original conjecture was proved by the author and Scharlemann and Qiu simultaneously) and a result about the uniqueness of expressions of Heegaard splittings as amalgamations.