On the dynamics of nonreducible cylindrical vortices (Q2890312)

The paper under review deals with the dynamics of cylindrical vortices. Given a compact metric space \(X\), a minimal homeomorphism \(T:X\rightarrow X\) and two continuous functions \(\Psi:X\rightarrow \mathrm{O}(\mathbb{R}^l)\), \(\rho:X\rightarrow \mathbb{R}^l\), a cylindrical vortex is defined as the transformation \((x,v)\mapsto (T(x),\Phi(x)v+\rho(x))\) from \(X\times \mathbb{R}^l\) into itself. In the particular case of \(l=1\) and \(\Psi=\mathrm{Identity},\) \((x,v)\mapsto (T(x),v+\rho(x))\) is called a cylindrical cascade.NEWLINENEWLINEThe main result in the first section of the paper establishes that no cylindrical vortex over a locally homogeneous space is minimal. To do it, the authors distinguish two cases: \(l=1\) and \(l\geq 2\). In the one-dimensional case, the above main result was proved by \textit{ {A. S. Besicovitch}} in [Proc. Cambridge Philos. Soc. 47, 38--45 (1951; Zbl 0054.07004)] for cylindrical cascades. Here the authors extend it to the general setting of cylindrical vortices by modifying in a suitable way Besicovitch's arguments. For higher dimensions, using a geometrical approach they prove the existence of invariant compact sets in the Alexandrov compactification in the case of affine Euclidean isometries, that is, \(\rho(x)\) equals to a constant vector \(\rho\), and use this fact to obtain finally the main result through the uniform approximation of \(T\) by homeomorphisms having periodic orbits. At the end of the section they propose the interesting question whether there exists a minimal cylindrical vortex with infinite-dimensional fiber.NEWLINENEWLINEIn the second part of the paper, the authors pay attention to the analysis of almost-reducibility of cylindrical vortices. They present a characterization of almost-reducibility for cylindrical cascades and they show the obstacles in order to extend it to the general cases of cylindrical vortices. The main result of this study asserts that if \(F\) is an almost-reducible, one-dimensional cylindrical vortex over an irrational rotation of the circle, then \(F\) admits no nonempty, proper, minimal invariant closed set whenever the function \(\rho\) has finite total variation. This is a slight extension of a result in [\textit{S. Matsumoto} and \textit{M. Shishikura}, Hiroshima Math. J. 32, 207--215 (2002; Zbl 1014.37031)], given for cylindrical cascades. Note that if \(F\) is a general one-dimensional cylindrical vortex, then \(F^2\) is a cylindrical cascade.NEWLINENEWLINEFinally, they concentrate on the map \(F(x,z)=(x+\alpha, e^{2\pi i \beta}z+\rho(x))\), defined as a self-map of \(\mathbb{S}^{1}\times \mathbb{C}\), so \(T\) is an irrational rotation, \(\rho:\mathbb{S}^1\rightarrow \mathbb{C}\) and \(\Psi\) is the rotation of angle \(\beta\), with \(\alpha,\beta\) rationally independent. They note that \(F\) is a factor of the cylindrical cascade \(G((x,y),z)=((x+\alpha,y-\beta), z+e^{2\pi i (y-\beta)}\rho(x))\) on \(\mathbb{T}^2\times \mathbb{C}\), so the dynamics of \(F\) is strongly related to that of \(G\); for instance, \(F\) is topologically transitive if and only if \(G\) is topologically transitive. The authors set out the question whether there exist cylindrical vortices \(F\) of the above type that are neither reducible (that is, the cohomological equation \(\rho(x)=\varphi(T(x))-\varphi(x)\) has no continuous solution \(\varphi:X\rightarrow \mathbb{C}\)) nor topologically transitive. Moreover, they give a very nice and ingenious example of a topologically transitive cylindrical vortex \(F\) of the above form. The main idea is to perform a sequence of approximations \(F_k\) inspired by the Anosov-Katok method; see [\textit{D. V. Anosov} and \textit{A. B. Katok}, Trans. Moscow Math. Soc. 23, 1--35 (1970; Zbl 0255.58007)]. Lastly, when \(\alpha,\beta\) verify a certain Diophantine condition, they prove that \(F\) is reducible if \(\rho\) is a \(C^{\infty}\) function, which implies that \(\beta\) cannot be approximated in a fast way by multiples of \(\alpha\).