Absolutely minimising generalised solutions to the equations of vectorial calculus of variations in \(L^\infty \) (Q522460)

The author considers the supremal functional \[ E_\infty(u,A)=\|\mathcal{L}(\cdot,u,Du)\|_{L^\infty(A)},\quad A\subseteq \Omega, \] applied to \(W^{1,\infty}\) maps \(u: \Omega\subseteq \mathbb R\to\mathbb R^N\), \(N\geq 1\) and \(\mathcal{L}\in C^2(\Omega\times \mathbb R^N\times\mathbb R^N)\). Under certain assumptions on \(\mathcal{L}\) it is proved that, for any given boundary data, there exists a map which is: \begin{itemize} \item[(i)] a vectorial absolute minimiser of \(E_\infty\) in the sense of Aronsson, \item[(ii)] a generalised solution to the associated ODE system \(\mathcal{F}(\cdot,u,Du.D^2u)=0\) as an analogue of the Euler-Lagrange equations, \item[(iii)] a limit of minimisers of the respective \(L^p\) functionals as \(p\to\infty\) for any \(q\geq 1\) in the strong \(W^{1,q}\) topology, \item[(iv)] partially \(C^2\) on \(\Omega\) off an exceptional compact nowhere dense set. \end{itemize} The method is based on \(L^p\) approximations and stable a priori partial regularity estimates.