Minimization with integrands composed of minimum of convex functions (Q5946001)

Let \(\Omega\) be a smooth bounded domain of \({\mathbb R}^n\). Then the paper is concerned with minimization problems of the form \[ \alpha=\inf\left\{\int_\Omega\min\{f(v,Dv),g(v,Dv)\}dx : v\in H^1(\Omega;{\mathbb R}^m)\right\}, \] under the assumption that the two integral functionals \[ v\to\int_\Omega f(v,Dv)dx\;\text{ and } v\to\int_\Omega g(v,Dv)dx \] are sequentially weakly lower semicontinuous and coercive in \(H^1(\Omega;{\mathbb R}^m)\). The proposed approach is to search not only for the minimizer \(u\) but also for functions \(\chi_f\) and \(\chi_g\colon\Omega\to[0,1]\) with \(\chi_f+\chi_g\equiv 1\) together with \({\mathcal R}={\mathcal R}(u,\chi_f,\chi_g)\) such that \[ \int_\Omega[\chi_ff(u,Du)+\chi_gg(u,Du)]dx-{\mathcal R}=\alpha. \] It is clear that in such approach explicit representation formulas for \({\mathcal R}\) play a central role. In this setting, an approximation result of \((u,\chi_f,\chi_g)\) is proposed, together with some explicit formulas for \(\mathcal R\). Exemplifications to one- and three-dimensional problems related to well potentials are also given.