Discrete weak KAM theory. An introduction through examples and its applications to twist maps (Q6955780)

The Kolmogorov-Arnold-Moser (KAM, for short) method is based on a theorem in the qualitative theory of ordinary differential equations and in the theory of dynamical systems, concerning quasiperiodic motions under small perturbations. The theorem partly resolves the small-divisor problem that arises in perturbation theory in physics, especially in mechanics.\N\NThe weak KAM theory was introduced by \textit{A. Fathi} [C. R. Acad. Sci., Paris, Sér. I 324, No. 9, 1043--1046 (1997; Zbl 0885.58022)] in the 1990s. It was designed to understand the mathematical models of Aubry-Mather theory for Tonelli Hamiltonian/Lagrangian systems through particular functions called {critical subsolutions} and {weak KAM solutions}. The term weak KAM was chosen in analogy with KAM tori. Below is the definition of viscosity solutions given by \textit{M. G. Crandall} and \textit{P.-L. Lions} [C. R. Acad. Sci., Paris, Sér. I 292, 183--186 (1981; Zbl 0469.49023); Trans. Am. Math. Soc. 277, 1--42 (1983; Zbl 0599.35024); J. Funct. Anal. 62, 379--396 (1985; Zbl 0627.49013)]. Fix bounded and uniformly continuous initial data \(u_0:\mathbb{R}^N\to\mathbb{R}\) and let \(\varepsilon>0\). Then the evolutionary Hamilton-Jacobi equation\N\[\N\partial_t U + H(x/\varepsilon, \partial_x U)=0,\qquad U(0,\cdot)=u_0,\N\]\Nadmits a unique viscosity solution \(U_{\varepsilon}:[0,+\infty)\times\mathbb{R}^N\to\mathbb{R}\). It is known that the functions \(U_{\varepsilon}\) converge locally uniformly to a function \(U_0\), characterized as the unique solution of\N\[\N\partial_t U + \overline{H}(\partial_x U)=0,\qquad U(0,\cdot)=u_0,\N\]\Nwhere \(\overline{H}\) is called the effective Hamiltonian. For \(P\in\mathbb{R}^N\), \(\overline{H}(P)\) is the unique constant such that the cell problem\N\[\NH(x, P+\partial_x u)=\overline{H}(P)\N\]\Nadmits a \(\mathbb{Z}^N\)-periodic viscosity solution \(u\). In homogenization theory, the state variable \(x\) takes values in the universal cover of the flat torus \(\mathbb{T}^N\). A fundamental domain of the covering map is then the cell \([0,1)^N\), in the sense that the knowledge of a \(\mathbb{Z}^N\)-periodic function on \(\mathbb{R}^N\) is equivalent to its restriction to this cell. The author then discusses some applications of the theory.\N\NNext the discrete Lax-Oleinik operator and its discounted counterpart are defined, and the discrete weak KAM Theorem is proved. Moreover, the fundamental Aubry set is introduced, together with its links to subsolutions and strict subsolutions. Here the author shows new connections between the classical and the discrete theories when the cost function is given by the action functional. The discrete setting and the Lax-Oleinik semigroup are applied. Graph properties, the twist condition, and the dynamics on the Aubry set are also considered.\N\NOne chapter is devoted to minimizing Mather measures and the discounted semigroups. The similarities and differences between discrete and continuous Mather measures in the classical Lagrangian setting are studied. The study of the positive counterpart of the discounted equations is, to the author's knowledge, new, both in the discrete and in the continuous setting.\N\NThe final chapter applies and develops discrete weak KAM theory in the context of conservative twist maps of the annulus. The author presents the special features of Aubry and Mather sets in the low-dimensional setting. Weak KAM solutions for twist maps are studied and characterized, and the shape of their derivatives is described. Finally, results concerning weaker or stronger integrability properties of twist maps are discussed.\N\NThe bibliography contains 208 articles and books concerning the topics under consideration.