Topological mixing of random substitutions (Q6165163)

In dynamical systems a relevant problem is that of describing and measuring, from the mathematical point of view, disordered behavior. Such measures can be obtained using several different notions such as Lyapunov exponents, values of entropy, mixing properties and others. It can happen that using one of them one could conclude that the a given system is disordered even though another measure would confirm the opposite. In the paper the authors deal with this problem and its possible ambiguity for some type of \textit{subshifts} associated to random substitutions on several letters and having some additional properties, as for example on the spectrum of the associated matrix of the subshift. The authors use the notion of topological mixing to evaluate the measure of order or disorder of the subshift associated to the random substitutions. Let us recall that random substitutions are certain maps transforming letters from a finite alphabet of symbols to a finite family of words constructed over the same alphabet. The procedure consists in introducing some computable criteria which are equivalent to the topological mixing properties. Several examples of random substitutions are considered. One is the case of irreducible Pisot random substitutions. The authors prove that the associated subshift is not topologically mixing. Furthermore some methods are developed to show that the Fibonacci substitution subshift is not topologically mixing. In fact such subshift it is seen to be locally disordered but globally well-ordered. It was obtained in the setting of mathematical quasicrystals by \textit{C. Godrèche} and \textit{J. M. Luck} [J. Stat. Phys. 55, No. 1--2, 1--28 (1989; Zbl 0717.05025)]. The main goal of the paper is to study the topological mixing of compatible random substitutions. The starting point is recalling primitive random substitutions on two letters whose second eigenvalue of the associated matrix is greater than one in modulus. A computable criterion to detect that the associated subshift is topologically mixing is given. Such procedure extends previous results on deterministic substitutions. If it is applied to irreducible Pisot random substitutions, it shows that the associated subshift is not topologically mixing. Other criteria of ``recognisability'' for excluding mixing are considered too, see [\textit{D. Rust}, Monatsh. Math. 193, No. 3, 683--704 (2020; Zbl 1452.37058)]. In particular, one of them is used to show that the random Fibonacci substitution subshift is not topologically mixing. In the paper the authors prove that if a primitive compatible random substitution is irreducible in the sense of Pisot, then its subshift is \(C\)-balanced (see the definition in the paper) and this is used to prove that the random Pisot substitutions are not topologically mixing. Some other examples of such situations are given. Similar theorems can be proved in the setting of \(\mathbb{R}\)-actions on random substitution tiling spaces. A subshift \((X, \sigma)\) is topological mixing if for any two words \(u, v\in \mathcal{L}(X)\), there is a natural number \(N\) such that for all natural numbers \(n\geq N\) there is a word \(w\) of lenghth \(n\) such that the concatenation of the three words \(uvw\) belongs to \(\mathcal{L}(X)\). On the basis of the previous definition the following statement is proved. Theorem. Let \(\theta\) be a primitive aperiodic deterministic substitution on an alphabet \(\mathcal{L}\). If the subshift \(X_\theta\) is topologically mixing, then \[ \mathrm{gdc}|\theta^n (a)| : a\in \mathcal{L} = 1 \] for all \(n\geq 1\). When \(\theta\) is defined on a two letters alphabet and the second eigenvalue of the corresponding matrix is greater than \(1\) in modulus, then the two conditions are equivalent. Note that the above equivalence is not valid for more general subshifts. One of the main aims of the paper is to extend the result of the above theorem to random cases. The rest of the paper is devoted to some extensions of the previous results. Overall, this is a well-written and detailed article, also understandable by non-specialists in the subject.