Equivariant geometry in Euclidean \(G\)-spheres of cohomogeneity two. I (Q1336241): Difference between revisions

The techniques of equivariant geometry are applied to a specific problem in Riemannian geometry, namely the study of constant mean curvature (and minimal) hypersurfaces of the Euclidean sphere \(S^ n\), subject to the constraint of being invariant with respect to a suitable connected symmetry group \(G \subset \text{Iso} (S^ n)\), \(\dim S^ n/G = 2\). The surfaces are studied by deriving and solving the reduced differential equation, which is an ODE with singularities on the boundary of the orbit space \(S^ n /G\). For \(G \neq U(1)\) the orbit space is (topologically) a 2-disk with 0, 1, 2 or 3 vertices. The case of no vertex was settled around 1970 [\textit{W.-Y. Hsiang} and \textit{H. B. Lawson jun.}, J. Differ. Geom. 5, 1-38 (1971; Zbl 0219.53045)] and the case of two fixed points as vertices was analyzed in the 1980's by Hsiang and his collaborators in their study of the spherical Bernstein problem. The case of one vertex is settled in this paper. The main problem is the determination of the volume function and the orbital distance metric, leading to the associated ODE. For convenience, the authors also employ numerical simulation and estimation to describe the qualitative behavior of solution curves, but this part of the proofs can be avoided by a more careful theoretical analysis.