Rank-one convexity: Geometry of matrix space and applications to PDE

Laszlo Szekelyhidi
Institute for Advanced Study

A function f:Rmxn->R is said to be rank-one convex if it is convex along rank 1 directions. Rank-one convexity gives rise to a variety of interesting issues such as how to characterize probability measures that satisfy Jensen's inequality for all rank-one convex functions, and how to tell when the rank-one convex hull of a given set of matrices is trivial. These issues in turn can be of great help in constructing (using convex integration) weak solutions to elliptic PDEs. In the talk I will present this on two examples: critical points to polyconvex functionals that are Lipschitz but nowhere C^1, and solutions to isotropic equations that are exactly at the threshold of higher integrability.