Minimizers of Dirichlet functionals on the \(n\)-torus and the weak KAM theory (Q1012359)

This paper deals with the minimization of a Dirichlet-type functional \(F\) on the \(n\)-torus \({\mathbb T}^n\). It is first argued that the supremum of this functional is not attained in \(L^1({\mathbb T}^n)\), but in the set \(\overline{\mathcal M}\) of Borel probability measures on \({\mathbb T}^n\). This motivates the author to extend the domain of \(F\) from the set of nonnegative densities in \(L^1({\mathbb T}^n)\) to \(\overline{\mathcal M}\). First objective of the present paper is to define a generalized minimizer of \(F\). This is mainly done by means of the effective Hamiltonian, in the framework of the weak KAM theory. A second objective of this paper is to relate the generalized minimizer of \(F\) to the minimal Mather measure. Essentially, the minimal Mather measures of a given Lagrangian is connected with the measure which minimizes a certain optimal transportation plan. That is why the third main objective of the present paper is to approximate \(F\) by an optimal transportation function and to establish a combinatorial search algorithm.