Maximum principles for vectorial approximate minimizers of nonconvex functionals (Q1942221)

The author considers the \(\alpha\)-minimizers of functionals like \[ E(u,\Omega) = \int_{\Omega} \mathcal{L} (x, u, Du) dx \] where \(u \in (W^{1,q}(\Omega))^N\) (with some boundary condition) and an \(\alpha\)\textit{-minimizer} (\(\alpha \geqslant 0\)) is a function \(u\) satisfying \[ E(u,\Omega) \leqslant E(u + \phi,\Omega) + \alpha \qquad \text{ for every } \phi \in (W^{1,q}_0(\Omega))^N \, . \] A kind of maximum principle is the so called \textit{convex hull property} which states that \[ u(\Omega) \subset \overline{\text{co}} (u (\partial \Omega)) \, . \] In this paper the author establishes appropriate versions of the convex hull property for approximate minimizers.