No label defined (Q4369310)

The paper considers integral functionals \[ I(u)= \int_\Omega [F(u(x)- a(x))+ \varphi(\nabla u(x))]dx,\quad u\in W^{1,\infty}(\Omega),\quad u(x)= a(x)\quad\text{on }\partial\Omega, \] where the function \(\varphi\) is such that \(\varphi(w_i)= 0\), \(i=1,\dots, k\), \(\varphi(w)>0\) if \(w\neq w_i\), \(i=1,\dots, k\). Here \(\Omega\subset\mathbb{R}^n\) is a convex domain and \(w_i\in\mathbb{R}^n\), \(i= 1,\dots, k\), are given. Under some assumptions on \(F\), \(\varphi\), \(\{w_i\}\) and \(a\) it is shown that the finite element method gives minimizing sequences \(\{u_h\}\) for which \(I(u_h)\to 0\) as \(h\to 0\). Here \(h\) is the triangulation's parameter. Some additional assumptions, among them \[ \nabla a(x)= \sum^k_{i= 1} \alpha_i(x) w_i,\quad 0\leq\alpha_i(x)\leq 1,\quad i= 1,\dots, k,\quad \alpha_1(x)+\cdots+ \alpha_k(x)= 1, \] ensure that the sequences \(\{\nabla u_h\}\) determine the Young measure \(\nu_x= \sum^k_{i= 1}\alpha_i(x)\delta_{w_i}\). For the entire collection see [Zbl 0877.00012].