On the existence of minimizers of the variable exponent Dirichlet energy integral (Q2474755)

The \(p\)-Dirichlet boundary value problem is to find a solution to the \(p\)-Laplace equation \[ \text{div} (|\nabla u|^{p-2}\nabla u)=0,\;x\in\Omega, \] which satisfies \(u=w\) on \(\partial\Omega\). This solution is treated in the sense of the minimization of the energy integral, the \(p\)-Laplace equation being the Euler-Lagrange equation for the variational integral \(\int_\Omega|\nabla u(x)|dx\). The author studies the variable exponent case where \(p=p(x)\), which attracted a lot of attention over the last two decades in the papers of Zhikov, Aktukhov, Acerbi, Mingione, Fan, Hästö, Harjulehto, and others. In the present paper, the author complements recent results of the paper by \textit{P.\,Harjulehto, P.\,Hästö, M.\,Koskenoja} and \textit{S.\,Varonen} [Potential Anal.\ 25, No.\,3, 205--222 (2006; Zbl 1120.46016)]. The main result is Theorem 2 which states that the minimizer of the Dirichlet energy integral always exists when the boundary values are bounded. It is known that in the one-dimensional case the boundary values are necessarily bounded. The author also provides an example showing that if the ``jump condition'' introduced in the above cited paper is violated, then the minimizer may not exist.