Mean curvature flow of pinched submanifolds to spheres (Q621858): Difference between revisions

\textit{G. Huisken} proved in [J. Differ. Geom. 20, 237--266 (1984; Zbl 0556.53001)] that the mean curvature flow (MCF) deforms a compact convex hypersurface in \(\mathbb R^{n+1}\) to a point in finite time and the solution is asymptotic to a shrinking sphere \(S^n\). The current paper generalizes the result to the case of higher co-dimension. More specifically, the authors obtain the results for \(n\)-dimensional compact submanifold \(\Sigma\) in \(\mathbb R^{n+k}\) satisfying \(\|H\| \neq 0\) and \(\|A\|^2 \leq c \|H\|^2\) for an explicit constant \(c\) depending only on \(n\). The constant obtained is optimal for \(n\geq 4\). Here \(H\) and \(A\) are the mean curvature vector and the second fundamental form of the initial submanifold \(\Sigma\), respectively. Although the set of such pinching conditions is not convex and the maximum principle for system cannot apply directly, the authors still manage to show that the conditions are preserved under MCF. The rest of the proof is based on Huisken's approach in the above mentioned paper. However, the argument is much more involved and complicated in the high co-dimensional situation. Many efforts are required to prove a pinching estimate for the traceless second fundamental form, which is a key estimate to show the submanifold evolving to a ``round'' point. A detailed introduction to the machinery of connections on vector bundle is also provided in the paper. It is applied to the space-time image, which simplifies some computation in deriving the evolution equations for related quantities and also gives a natural interpretation of the ``Uhlenbeck trick''.