Abstract:We consider the motion by curvature of a network of curves in the plane and we discuss existence, uniqueness, singularity formation and asymptotic behavior of the flow.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the motion of curve networks in a plane under curvature flow. Specifically, the authors explore the existence, uniqueness, singularity formation and asymptotic behavior of the motion of such networks. The main purpose of the paper is to provide a comprehensive understanding of the evolution process of networks composed of multiple curves under the action of curvature flow, which includes:
1. **Existence and Uniqueness**: Study whether there are solutions for the motion of curve networks under curvature flow under given initial conditions, and whether these solutions are unique. For example, for a tri - point network (called triod) composed of three curves, Bronlard and Reitich first considered the existence problem of the curvature flow of a regular tri - point network with fixed endpoints in [17], and proved the existence and uniqueness in a short time.
2. **Singularity Formation**: Analyze whether singularities (i.e., some geometric quantities such as curvature become unbounded) will be formed in the network during the curvature flow process. If singularities are formed, study the properties of these singularities and their influence on the overall structure of the network. For example, when the lengths of some curves in the network approach zero, it may lead to the collision of tri - points and form higher - order multi - points.
3. **Asymptotic Behavior**: Explore the final state or asymptotic behavior of the network during the network evolution process, especially on a long - time scale. For example, if the lengths of the curves in the network remain far from zero, the maximum curvature of the network must tend to infinity (see Corollary 4.15 and Theorem 5.7).
4. **Restarting the Evolution**: Study how to restart the evolution process of the network after singularities appear. For example, after the network forms a singularity at a certain moment, how to define a new initial condition so that the evolution can continue. Ilmanen, Neves and the fourth author independently proved a short - time existence result for a non - regular network in [58], which allows the network to evolve smoothly immediately after regularization.
5. **Stability**: Study the stability of the network under small perturbations. For example, if a network is close enough to a regular network \( S^* \) composed of straight line segments in the \( W^{2,2} \) norm, then its curvature flow exists all the time and converges smoothly to a regular network still composed of straight line segments and having the same length (see Proposition 12.6).
6. **Global Evolution of Special Networks**: Describe in detail the global evolution process of simple networks with at most two tri - points. Although such networks are simple, they already contain most of the relevant phenomena, such as the smooth evolution of tree - like networks (networks with only one tri - point and three fixed endpoints) under curvature flow and asymptotic convergence to a Steiner network (see Section 14).
Through these studies, the authors hope to provide a theoretical basis for more complex multi - phase system models, especially in materials science, such as the grain boundary growth model. In addition, the paper also discusses some open problems, pointing out the direction for further research.