Stability of regular shrinkers in the network flow
A network \(\Gamma \) in an open set \(\Omega \subset \mathbb{R}^{2}\) is a finite collection of smooth curves \(\gamma _{i}:[a_{i},b_{i}]\rightarrow \Omega \) such that the interior of each \(\gamma _{i}\) is embedded. Two or more curves may intersect at their endpoints. An endpoint of a curve either lies on \(\partial \Omega \) or coincides with the endpoint of another curve. In the latter case, the common endpoint is called a multi-junction; in the former, it is called a boundary point of \(\Gamma \). If \(\Omega =\mathbb{R} ^{2}\), all boundary points must lie at infinity, and the network is called an open network. A network \(\Gamma \) is regular if every multi-junction is the meeting point of exactly three curves, and the angles between any two of them are all equal to \(2\pi /3\) (Herring's condition). \N\NA one-parameter family of regular networks \(\{\Gamma _{t}\}_{t\geq 0}\) evolves by the network flow if, on each curve \(\gamma _{i}\) of \(\Gamma _{t}\), the velocity vector satisfies \((\frac{\partial x}{\partial t})^{\perp }=k_{i}N_{i}\), where \(k_{i}\) and \(N_{i}\) are the curvature and unit normal vector of \( \gamma _{i}\) at the point \(x\), respectively. A network \(\Gamma \subset \mathbb{R}^{2}\) is called a regular shrinker if it is regular and satisfies the equation \(k+\frac{\left\langle x,N\right\rangle }{2}=0\), on each curve \( \gamma _{i}\) of the network, where \(k\) is the curvature, \(N\) the unit normal vector, and \(x\) the position vector. The curvature is defined by the relation \(\frac{d^{2}x}{d\sigma ^{2}}=kN\), where \(\sigma \) denotes arc length. \N\NLet \(\Gamma \)\ be a possibly open network. For any point \(x_{0}\in \mathbb{R}^{2}\) and time \(t_{0}>0\), the Gaussian-weighted length functional (\(F\)-functional) is defined through \N\[\NF_{x_{0},t_{0}}(\Gamma )=\frac{1}{\sqrt{ 4\pi t_{0}}}\int_{\Gamma }e^{-\frac{\left\vert x-x_{0}\right\vert ^{2}}{ 4t_{0}}}d\sigma, \N\]\N where \(d\sigma \) denotes the arc-length element of \( \Gamma \). A regular shrinker is \(F\)-stable if for every compactly supported variation \(\Gamma _{s}\), there exist corresponding variations \(x_{s}\) of \( x_{0}\) and \(t_{s}\) of \(t_{0}\) with \((x_{0},t_{0})=(0,1)\) such that \N\[\N\frac{ d^{2}}{ds^{2}}F_{x_{s},t_{s}}(\Gamma _{s})\mid _{s=0}\geq 0.\N\] \NIf this condition fails, \(\Gamma \) is \(F\)-unstable or it exhibits \(F\)-instability. A regular shrinker is entropy-stable if it is a local minimizer of the entropy functional \N\[\N\lambda (\Gamma )=\sup_{x_{0}\in \mathbb{R}^{2},t_{0}\in \mathbb{ R}_{+}}F_{x_{0},t_{0}}(\Gamma ).\N\] \NThe main result of the study establishes the \(F\)-instability of some regular shrinkers: The 4-ray star, 5-ray star, fish, and rocket are \(F\)-unstable and entropy-unstable regular shrinkers. \N\NThe author presents figures of such networks. The author first considers the case of regular shrinkers with two or more enclosed regions, observing that the rate of area decrease of an enclosed region under the network flow depends solely on the number of its bounding edges. In the case of regular shrinkers that enclose a single region, the author observes that the \(F\) -functional can be represented in terms of the backward heat kernel \(\Phi : \mathbb{R}^{2}\times (-\infty ,0)\rightarrow \mathbb{R}\) defined as \(\Phi (x,t)=\frac{1}{\sqrt{-4\pi t}}e^{\frac{\left\vert x\right\vert ^{2}}{4t}}\) through \N\[\NF_{x_{0},t_{0}}(\Gamma )=\int_{\Gamma }\Phi _{(x_{0},t_{0})}(x,t)d\sigma. \N\] \NHe derives translation, scaling and monotonicity properties of \(F\). He computes the first variation of the \(F\)-functional and proves its connection to self-similarly shrinking solutions. He computes the second variation of the \(F\)-functional for a regular shrinker, assuming that all multi-junctions satisfy the Herring condition. He proves that the tangential component of a variation vector field does not contribute to the second derivative, thus focusing on functions representing the normal component. He finally analyzes the eigenfunction problem from the second variation formula. \N\NThe author also proves that, if \(\Gamma \) is a regular shrinker with two or more enclosed regions, then there exists a perturbation such that the \(t=-1\) slice of any tangent flow of the perturbed network is not \(\Gamma \).
- A stable manifold theorem for the curve shortening equation
- Asymptotic behavior for singularities of the mean curvature flow
- Curvature evolution of nonconvex lens-shaped domains
- Evolution of convex lens-shaped networks under the curve shortening flow
- Evolution of networks with multiple junctions
- Evolution of spoon-shaped networks
- Existence and regularity theorems of one-dimensional Brakke flows
- Generic mean curvature flow. I: Generic singularities
- scientific article; zbMATH DE number 1521306 (Why is no real title available?)
- Lectures on curvature flow of networks
- Motion by curvature of networks with two triple junctions
- Motion by curvature of planar networks
- Motion by curvature of planar networks. II.
- Networks self-similarly moving by curvature with two triple junctions
- Non–existence oftheta–shaped self–similarly shrinking networks moving by curvature
- On short time existence for the planar network flow
- On the mean curvature flow of grain boundaries
- On three-phase boundary motion and the singular limit of a vector-valued Ginzburg-Landau equation
- Self-similar solutions of a 2-D multiple-phase curvature flow
- The normalized curve shortening flow and homothetic solutions
- Uniqueness of regular shrinkers with two enclosed regions
This page was built for publication: Stability of regular shrinkers in the network flow
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6844339)
