Normalization of ternary generalized pseudostandard words (Q2420613)

The paper deals with generalized pseudostandard words. Generalized pseudostandard words are defined as follows: Let \(\Delta=(\delta_n)\) be an infinite sequence of letters and let \(\Theta=(\theta_n)\) be an infinite sequence of involutory antimorphisms. The generalized pseudostandard word generated by the directive bi-sequence \((\Delta,\Theta)\) is the limit of the sequence of words \((w_n)\), where \(w_0\) is the empty word and \(w_{n+1} \) is the shortest \(\theta_{n+1}\)-palindrome starting with \(w_n\cdot \delta_{n+1}\). Generalized pseudostandard words are a generalization of standard Sturmian, standard episturmian and pseudostandard words. A generalized pseudostandard word may be generated by different directive bi-sequences. The directive bi-sequence \((\Delta,\Theta)\) is called normalized if the sequence \((w_n)\) contains all pseudopalindromic prefixes of its limit. Every directive bi-sequence can be normalized; the problem is therefore to give an algorithmic procedure to find the normalization. In this paper, the authors give such an algorithm in the case of an alphabet of size \(3\), thus generalizing the solution given for binary alphabets in [\textit{A. B. Massé} et al., J. Integer Seq. 16, No. 2, Article 13.2.11, 28 p. (2013; Zbl 1310.68175)]. Alongside the theoretical work, the authors implemented and tested the new normalization algorithm. The documented code and examples are publicly available at \url{https://github.com/velkater/tgpc}.