In this talk I will discuss an example of the use of fully nonlinear parabolic flows to prove geometric results. I will emphasise the fact that there is a wide variety of geometric parabolic equations to choose from, and to get the best results it can be very important to choose the best flow. I will illustrate this in the setting of surfaces in a three-dimensional sphere. There are quite a few relevant results for surfaces in the sphere satisfying various kinds of curvature equations, including totally umbillic surfaces, minimal surfaces and constant mean curvature surfaces, and intrinsically flat surfaces. Parabolic flows can strengthen such results by allowing classes of surfaces satisfying curvature inequalities rather than equalities: This was first done by Huisken, who used mean curvature flow to deform certain classes of surfaces to totally umbillic surfaces. This motivates the question "What is the optimal result of this kind?" - that is, what is the weakest pointwise curvature condition which defines a class of surfaces which retracts to the space of great spheres? The answer to this question can be guessed in view of the examples. To prove it requires a surprising choice of evolution equation, forced by the requirement that the pointwise curvature condition be preserved. I will conclude by mentioning some other geometric situations in which strong results can be proved by choosing the best possible evolution equation.
展开▼
