Computation of the lower semicontinuous envelope of integral functionals and non-homogeneous microstructures (Q2475160): Difference between revisions

The author proposes the computation of the lower semicontinuous envelope of integral functional with non-quasiconvex integrands. The problem of minimizing an integral functional: \(F(u, \Omega)= \int_\Omega f(x,u(x), \nabla u(x))\,dx\) with non-quasiconvex integrand \(f: \Omega \times\mathbb R^m \times\mathbb R^{mn} \rightarrow\mathbb R^1\) in a set of admissible functions \(U(u_0, \Omega) = \{u \in W^{1,p}(\Omega,\mathbb R^m):u=u_0 \text{ on } \partial \Omega \}\), where \(\Omega \subset\mathbb R^n\) is a bounded open set with Lipschitz continuous boundary \(\partial \Omega\) and \(1 < p < \infty\), is considered. A numerical method for evaluating \(\Gamma^-\)-\(\lim F(u, \Omega)\) for a given \(u \in U(u_0, \Omega)\) under certain hypotheses (here the \(\Gamma^-\)-limit plays an important role, since it is the sequentially weakly lower semicontinuous envelope of \(F( \cdot , \Omega) \text{ in } U(u_0, \Omega))\) is established. The computing of non-homogeneous and curved laminated microstructures, a numerical method for the evaluation of \(\Gamma^-\)-\(\lim F(u, \Omega)\), is given. Notice that the basic idea is to approximate the value by the solutions to a set of subproblems defined on a finite element subdivision of \(\Omega\). Here the subproblems, each of which is concerned with a homogeneous integrand and boundary data, are then solved by any available numerical method (the rational transformation method with the incremental crystallization method is applied). Main result: The finite element solutions to the subproblems are used to construct a minimizing sequence of the original problem \(\Gamma^-\)-\(\lim F(u, \Omega)= \min \{\lim_{\alpha \rightarrow 0} \inf F(u_\alpha, \Omega):u_\alpha \in U(u_0, \Omega),u_\alpha \rightarrow u \text{ weakly in } W^{1,p}(\Omega,\mathbb R^m)\}\) The convergence of the method is proved and it is shown that the method is capable for capturing curved and non-homogeneous microstructures. Finally, numerical examples are presented, too. They show the effectiveness of the author's method.