Motion of a set by the curvature of its boundary
journal contributionposted on 01.01.1990, 00:00 by H. Mete. Soner
Any type of content formally published in an academic journal, usually following a peer-review process.
Abstract: "The study of a crystal shrinking or growing in a melt gives rise to equations relating the normal velocity of the motion to the curvature of the crystal boundary. Often these equations are anisotropic, indicating the prefered [sic] directions of the crystal structure. In the isotropic case this equation is called the mean curvature flow or the curve shortening equation, and has been studied by differential geometric tools. In particular, it is known that there are no classical solutions to these equations. In this paper we develop a weak theory for the 'generalized mean curvature' equation using the newly developed theory of viscosity solutions. Our approach is closely related to that of Osher & Sethian, Chen, Giga & Goto, and Evans & Spruck, who view the boundary of the crystal as the level set of a solution to a nonlinear parabolic equation. Altough [sic] we use their results in an essential way, we do not require that the boundary is a level set. Our main results are the existence of a solution, large time asymptotics of this solution, and its connection to the level set solution of Osher & Sethian, Chen, Giga & Goto, and Evans & Spruck. In general there is no uniqueness, even for classical solutions, but we prove a uniqueness result under restrictive assumptions. We also construct a class of explicit solutions which are dialations of Wullf crystals."