The area preserving curve shortening flow with Neumann free boundary conditions (Q326953): Difference between revisions

Let \(M^n\) be an \(n\)-dimensional compact variety. By definition a map \(F\) from \(M^n\times [0,T)\) to \(\mathbb R^{n+1}\) is a solution to the volume preserving curvature flow with Neumann free boundary whenever the rank of the derivative linear map associated to \(F\) is equal to \(n+1\), and \(F\) is a solution of \[ (P)_n:\left\{\begin{aligned} &\frac{d F(p,t)}{dt}=\big(H(p,t)-h(t)\big)\nu(p,t)\text{ for }(p,t)\in M^n\times [0,T)\\ &F(p,0)=F_0(p)\text{ for }p\in M^n,\\ &F(p,t)\in\Sigma\text{ for }(p,t)\in \partial M^n\times [0,T),\\ &\nu(p,t)\perp^\Sigma\vec{\nu}\left(F(p,t)\right)\text{ for }(p,t)\in \partial M^n\times [0,T). \end{aligned}\right. \] where \(H,h\) stand for the mean curvature and its average, respectively, and \(\nu(p,t)\) the outer unit normal at the point \((p,t)\), \(^\Sigma {\vec{\nu}}\) is a \(\mathbb S^{n}\)-valued map on \(\Sigma\), a smooth embedded hypersurface in \(\mathbb R^{n+1}\), and the symbol \(\perp\) means orthogonal. The author studies solutions properties of \((P)_1\) in the particular case \(M^1=[a,b]\). Principally, the main result states that a curve \(c(\cdot,\cdot)\) solution of \((P_1\) has not a singularity and that for every no convergent sequence \((t_j)_{j\in \mathbb N}\subset [0,\infty)\), the succession of curves \(\left(c(\cdot,t_j)\right)_{j\in\mathbb N}\) subconverges smoothly to \(\gamma_\infty\), an embedded arc of a circle, such that the average of its curvature belongs to \([\pi,2\pi)\) besides \(\gamma_\infty\) encloses an area. These results are fulfilled whenever the initial curve \(c(\cdot,0)=c_0\) has a positive curvature in the interval \([a,b]\), has not a self-intersection, contained in the outer domain created by \(\Sigma\), and its length has a suitable upper bound (Theorem 1.1). The article is organized as follows. In the second section the author recalls a few properties of planar curves. In the third section, the author provides a uniform upper/lower bound of the average of the curvature of \(c(\cdot,\cdot)\) whenever \(c_0\) satisfies, almost, the hypothesis of Theorem 1.1 and which are mentioned above (Theorem 3.10). In the fourth section, the author defines a singularity of type I (resp. a singularity at \(T\)) of \(c(\cdot,\cdot)\), i.e., if there is a uniform constant \(C\) such that \[ \sup_{p\in [a,b]}k^2(p,t)\leq \frac{C}{2\big(T-t\big)} \] for all \[ t\in [0,T\big)\left(\text{resp. }\lim_{t\to T^-}\left[\max_{p\in[a,b]}|k(p,t)|\right]=\infty\right) \] where \(k(p,t)\) is the curvature at the point \((p,t)\) in \(c\big([a,b]\times [0,T)\big)\), then by assuming that \(c_0\) is convex, the absolute value of the average of \(k(\cdot,t)\) on \([a,b]\) is bounded, and the length of \(c(\cdot,t)\) is lower bounded by a positive constant. The author states that \(c(\cdot,\cdot)\) develops a singularity at \(T\) which cannot be of type I (Proposition 4.12). Germane to the sixth section, the author supposes that \(c_0\) satisfies the assumptions of Theorem 1.1, and shows that \(c(\cdot,\cdot)\) does not develop singularity at finite time, see the statement of Theorem 6.5. Finally, by gathering the previous results, the author provides the proof of the main result in the seventh section.