Strict sub-solutions and Mañé potential in discrete weak KAM theory (Q662439): Difference between revisions

From the text: In this paper, we explain some facts on the discrete case of weak KAM theory. In that setting, the Lagrangian is replaced by a cost \(c:X\times X\to\mathbb{R}\), on a ``reasonable'' space \(X\). In a first part, we introduce the Lax-Oleinik semi-groups \(T^-_c\) and \(T^+_c\) and study their sub-solutions. We start with a (continuous) cost \(c:X^2\to\mathbb{R}\) which verifies a definition of uniform super-linearity and a definition of uniform boundedness. A function \(u\) is an \(\alpha\)-sub-solution for \(c\) if \[ \forall(x,y)\in X^2,\;u(y)-u(x)\leq c(x,y)+\alpha. \] The critical constant \(\alpha[0]\) is the smallest constant \(\alpha\) such that there are \(\alpha\)-sub-solutions. In the first part, we prove the existence of critical sub-solutions which are strict on a maximal set. The proof is done using the Lax-Oleinik semi-groups and the notion of Aubry set. The second part is devoted to the study of continuity of sub-solutions and of an analogue of Mañé potential. In the discrete case, we will see that sub-solutions are not necessarily continuous. However, analogously to the continuous case, the projected Aubry set is the set of points where all sub-solutions are continuous.