No label defined (Q4469376)

The authors study the behaviour as \(p\rightarrow 1\) of minimizers \(u_p\) in \(W^{1,p}_0(\Omega)\) of the functional \[ J_p(u) = {1\over p} \int_\Omega | \nabla u| ^p\,dx - \int_\Omega fu\,dx. \] Here \(\Omega\) is a bounded domain in \({\mathbb R}^N\), \(N\geq 2\), and \(f\in L^N(\Omega)\). The main result is that if the \(L^N\) norm (or the Marcinkiewicz norm \(L^{N,\infty}\)) of \(f\) is small enough, then \(u_p\) converges to a minimizer \(u\) of the functional \[ J(u) = | Du| (\Omega) + \int_{\partial\Omega} | u| \,d{\mathcal H}^{N-1} - \int_\Omega fu\,dx, \] where \(| u| \) denotes the internal trace of \(u\) on \(\partial\Omega\), \({\mathcal H}^{N-1}\) is the \(N-1\) dimensional Hausdorff measure, and \(| Du| \) is the total variation of the distributional gradient \(Du\). Results of this type for the case \(f=1\) were obtained by \textit{B.~Kawohl} [J. Reine Angew. Math. 410, 1--22 (1990; Zbl 0701.35015)], who also found examples showing that if the norm of \(f\) is not sufficiently small, then the functional \(J\) is unbounded from below and \(\lim_{p\rightarrow 1}u_p(x)=+\infty\) in \(\Omega\).