Limits of nonlinear Dirichlet problems in varying domains (Q1107812)

The authors consider a sequence of nonlinear variational problems in varying domains with Dirichlet boundary conditions: \[ m_ h(g):=\min_{u\in H_ 0^{1,p}(\Omega \setminus E_ h)}\{\int_{\Omega \setminus E_ h}f(x,Du) \quad dx+\int_{\omega \setminus E_ h}gu dx\}, \] where \(\Omega\) is a bounded open subset of R n (n\(\geq 2)\), \(\{E_ n\}\) is a sequence of closed subsets of \(\Omega\), f is a fixed nonnegative function satisfying some conditions and \(g\in L\) q(\(\Omega)\), 1/p \(+ 1/q=1\). They look for the general form of the limit, in the sense of \(\Gamma\)-convergence, of such problems. They state a compactness result and find the form of the limit problem. Namely, they prove that there exist a subsequence \(\{E_{\sigma (h)}\}\) and a nonnegative Borel measure \(\mu\), vanishing on subsets of \(\Omega\) with p-capacity zero, such that \[ \lim_{h\to +\infty}m_{\sigma (h)}(g)=m(\mu,g)\quad for\quad every\quad g\in L\quad q(\Omega), \] \[ where\quad m(\mu,g)=\min_{u\in H_ o^{1,p}(\Omega)}\int_{\Omega}f(x,Du)dx+\int_{\Omega}| u| \quad p d\mu +\int_{\Omega}gu dx. \] Moreover, if we denote by \(M_ h(g)\) and M(\(\mu\),g) the sets of minimizers for the functionals which appear in the definitions of \(m_ h(g)\) and m(\(\mu\),g), respectively, then for every neighborhood U of M(\(\mu\),g) in L p(\(\Omega)\) there exists k such that \(M_{\sigma (h)}(g)\subset U\) for \(h\geq K\).